Hi,
This patch-set takes advantage of the new statistics that are going to be available in the kernel thanks to the per-entity load-tracking: http://thread.gmane.org/gmane.linux.kernel/1348522. It packs the small tasks in as few as possible CPU/Cluster/Core. The main goal of packing small tasks is to reduce the power consumption by minimizing the number of power domain that are used. The packing is done in 2 steps:
The 1st step looks for the best place to pack tasks on a system according to its topology and it defines a pack buddy CPU for each CPU if there is one available. The policy for setting a pack buddy CPU is that we pack at all levels where the power line is not shared by groups of CPUs. For describing this capability, a new flag has been introduced SD_SHARE_POWERLINE that is used to describe where CPUs of a scheduling domain are sharing their power rails. This flag has been set in all sched_domain in order to keep unchanged the default behaviour of the scheduler.
In a 2nd step, the scheduler checks the load level of the task which wakes up and the business of the buddy CPU. Then, It can decide to migrate the task on the buddy.
The patch-set has been tested on ARM platforms: quad CA-9 SMP and TC2 HMP (dual CA-15 and 3xCA-7 cluster). For ARM platform, the results have demonstrated that it's worth packing small tasks at all topology levels.
The performance tests have been done on both platforms with sysbench. The results don't show any performance regressions. These results are aligned with the policy which uses the normal behavior with heavy use cases.
test: sysbench --test=cpu --num-threads=N --max-requests=R run
Results below is the average duration of 3 tests on the quad CA-9.
default is the current scheduler behavior (pack buddy CPU is -1)
pack is the scheduler with the pack mecanism
| default | pack |
-----------------------------------
N=8; R=200 | 3.1999 | 3.1921 |
N=8; R=2000 | 31.4939 | 31.4844 |
N=12; R=200 | 3.2043 | 3.2084 |
N=12; R=2000 | 31.4897 | 31.4831 |
N=16; R=200 | 3.1774 | 3.1824 |
N=16; R=2000 | 31.4899 | 31.4897 |
-----------------------------------
The power consumption tests have been done only on TC2 platform which has got accessible power lines and I have used cyclictest to simulate small tasks. The tests show some power consumption improvements.
test: cyclictest -t 8 -q -e 1000000 -D 20 & cyclictest -t 8 -q -e 1000000 -D 20
The measurements have been done during 16 seconds and the result has been normalized to 100
| CA15 | CA7 | total |
-------------------------------------
default | 100 | 40 | 140 |
pack | <1 | 45 | <46 |
-------------------------------------
The A15 cluster is less power efficient than the A7 cluster but if we assume that the tasks is well spread on both clusters, we can guest estimate that the power consumption on a dual cluster of CA7 would have been for a default kernel:
| CA7 | CA7 | total |
-------------------------------------
default | 40 | 40 | 80 |
-------------------------------------
Vincent Guittot (6):
Revert "sched: introduce temporary FAIR_GROUP_SCHED dependency for
load-tracking"
sched: add a new SD SHARE_POWERLINE flag for sched_domain
sched: pack small task at wakeup
sched: secure access to other CPU statistics
sched: pack the idle load balance
ARM: sched: clear SD_SHARE_POWERLINE
arch/arm/kernel/topology.c | 5 ++
arch/ia64/include/asm/topology.h | 1 +
arch/tile/include/asm/topology.h | 1 +
include/linux/sched.h | 9 +--
include/linux/topology.h | 3 +
kernel/sched/core.c | 13 ++--
kernel/sched/fair.c | 155 +++++++++++++++++++++++++++++++++++---
kernel/sched/sched.h | 10 +--
8 files changed, 165 insertions(+), 32 deletions(-)
--
1.7.9.5
This reverts commit f0494c69d387db9496b6f9aa71b3bc49f1dcf474.
Signed-off-by: Vincent Guittot <[email protected]>
---
include/linux/sched.h | 8 +-------
kernel/sched/core.c | 7 +------
kernel/sched/fair.c | 13 ++-----------
kernel/sched/sched.h | 9 +--------
4 files changed, 5 insertions(+), 32 deletions(-)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index f45da5f..4786b20 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1214,13 +1214,7 @@ struct sched_entity {
/* rq "owned" by this entity/group: */
struct cfs_rq *my_q;
#endif
-/*
- * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be
- * removed when useful for applications beyond shares distribution (e.g.
- * load-balance).
- */
-#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED)
- /* Per-entity load-tracking */
+#ifdef CONFIG_SMP
struct sched_avg avg;
#endif
};
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index b81915c..d50fbac 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -1715,12 +1715,7 @@ static void __sched_fork(struct task_struct *p)
p->se.vruntime = 0;
INIT_LIST_HEAD(&p->se.group_node);
-/*
- * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be
- * removed when useful for applications beyond shares distribution (e.g.
- * load-balance).
- */
-#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED)
+#ifdef CONFIG_SMP
p->se.avg.runnable_avg_period = 0;
p->se.avg.runnable_avg_sum = 0;
#endif
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 095d86c..4f4a4f6 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -877,8 +877,7 @@ static inline void update_cfs_shares(struct cfs_rq *cfs_rq)
}
#endif /* CONFIG_FAIR_GROUP_SCHED */
-/* Only depends on SMP, FAIR_GROUP_SCHED may be removed when useful in lb */
-#if defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED)
+#ifdef CONFIG_SMP
/*
* We choose a half-life close to 1 scheduling period.
* Note: The tables below are dependent on this value.
@@ -3204,12 +3203,6 @@ unlock:
}
/*
- * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be
- * removed when useful for applications beyond shares distribution (e.g.
- * load-balance).
- */
-#ifdef CONFIG_FAIR_GROUP_SCHED
-/*
* Called immediately before a task is migrated to a new cpu; task_cpu(p) and
* cfs_rq_of(p) references at time of call are still valid and identify the
* previous cpu. However, the caller only guarantees p->pi_lock is held; no
@@ -3232,7 +3225,6 @@ migrate_task_rq_fair(struct task_struct *p, int next_cpu)
atomic64_add(se->avg.load_avg_contrib, &cfs_rq->removed_load);
}
}
-#endif
#endif /* CONFIG_SMP */
static unsigned long
@@ -5812,9 +5804,8 @@ const struct sched_class fair_sched_class = {
#ifdef CONFIG_SMP
.select_task_rq = select_task_rq_fair,
-#ifdef CONFIG_FAIR_GROUP_SCHED
.migrate_task_rq = migrate_task_rq_fair,
-#endif
+
.rq_online = rq_online_fair,
.rq_offline = rq_offline_fair,
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 81135f9..a95d5c1 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -225,12 +225,6 @@ struct cfs_rq {
#endif
#ifdef CONFIG_SMP
-/*
- * Load-tracking only depends on SMP, FAIR_GROUP_SCHED dependency below may be
- * removed when useful for applications beyond shares distribution (e.g.
- * load-balance).
- */
-#ifdef CONFIG_FAIR_GROUP_SCHED
/*
* CFS Load tracking
* Under CFS, load is tracked on a per-entity basis and aggregated up.
@@ -240,8 +234,7 @@ struct cfs_rq {
u64 runnable_load_avg, blocked_load_avg;
atomic64_t decay_counter, removed_load;
u64 last_decay;
-#endif /* CONFIG_FAIR_GROUP_SCHED */
-/* These always depend on CONFIG_FAIR_GROUP_SCHED */
+
#ifdef CONFIG_FAIR_GROUP_SCHED
u32 tg_runnable_contrib, tg_usage_contrib;
u64 tg_load_contrib;
--
1.7.9.5
During sched_domain creation, we define a pack buddy CPU if available.
On a system that share the powerline at all level, the buddy is set to -1
On a dual clusters / dual cores system which can powergate each core and
cluster independantly, the buddy configuration will be :
| CPU0 | CPU1 | CPU2 | CPU3 |
-----------------------------------
buddy | CPU0 | CPU0 | CPU0 | CPU2 |
Small tasks tend to slip out of the periodic load balance.
The best place to choose to migrate them is at their wake up.
Signed-off-by: Vincent Guittot <[email protected]>
---
kernel/sched/core.c | 1 +
kernel/sched/fair.c | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++
kernel/sched/sched.h | 1 +
3 files changed, 111 insertions(+)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index dab7908..70cadbe 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
rcu_assign_pointer(rq->sd, sd);
destroy_sched_domains(tmp, cpu);
+ update_packing_domain(cpu);
update_top_cache_domain(cpu);
}
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 4f4a4f6..8c9d3ed 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -157,6 +157,63 @@ void sched_init_granularity(void)
update_sysctl();
}
+
+/*
+ * Save the id of the optimal CPU that should be used to pack small tasks
+ * The value -1 is used when no buddy has been found
+ */
+DEFINE_PER_CPU(int, sd_pack_buddy);
+
+/* Look for the best buddy CPU that can be used to pack small tasks
+ * We make the assumption that it doesn't wort to pack on CPU that share the
+ * same powerline. We looks for the 1st sched_domain without the
+ * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the lowest
+ * power per core based on the assumption that their power efficiency is
+ * better */
+void update_packing_domain(int cpu)
+{
+ struct sched_domain *sd;
+ int id = -1;
+
+ sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
+ if (!sd)
+ sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
+ else
+ sd = sd->parent;
+
+ while (sd) {
+ struct sched_group *sg = sd->groups;
+ struct sched_group *pack = sg;
+ struct sched_group *tmp = sg->next;
+
+ /* 1st CPU of the sched domain is a good candidate */
+ if (id == -1)
+ id = cpumask_first(sched_domain_span(sd));
+
+ /* loop the sched groups to find the best one */
+ while (tmp != sg) {
+ if (tmp->sgp->power * sg->group_weight <
+ sg->sgp->power * tmp->group_weight)
+ pack = tmp;
+ tmp = tmp->next;
+ }
+
+ /* we have found a better group */
+ if (pack != sg)
+ id = cpumask_first(sched_group_cpus(pack));
+
+ /* Look for another CPU than itself */
+ if ((id != cpu)
+ || ((sd->parent) && !(sd->parent->flags && SD_LOAD_BALANCE)))
+ break;
+
+ sd = sd->parent;
+ }
+
+ pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
+ per_cpu(sd_pack_buddy, cpu) = id;
+}
+
#if BITS_PER_LONG == 32
# define WMULT_CONST (~0UL)
#else
@@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct *p, int target)
return target;
}
+static inline bool is_buddy_busy(int cpu)
+{
+ struct rq *rq = cpu_rq(cpu);
+
+ /*
+ * A busy buddy is a CPU with a high load or a small load with a lot of
+ * running tasks.
+ */
+ return ((rq->avg.usage_avg_sum << rq->nr_running) >
+ rq->avg.runnable_avg_period);
+}
+
+static inline bool is_light_task(struct task_struct *p)
+{
+ /* A light task runs less than 25% in average */
+ return ((p->se.avg.usage_avg_sum << 2) < p->se.avg.runnable_avg_period);
+}
+
+static int check_pack_buddy(int cpu, struct task_struct *p)
+{
+ int buddy = per_cpu(sd_pack_buddy, cpu);
+
+ /* No pack buddy for this CPU */
+ if (buddy == -1)
+ return false;
+
+ /*
+ * If a task is waiting for running on the CPU which is its own buddy,
+ * let the default behavior to look for a better CPU if available
+ * The threshold has been set to 37.5%
+ */
+ if ((buddy == cpu)
+ && ((p->se.avg.usage_avg_sum << 3) < (p->se.avg.runnable_avg_sum * 5)))
+ return false;
+
+ /* buddy is not an allowed CPU */
+ if (!cpumask_test_cpu(buddy, tsk_cpus_allowed(p)))
+ return false;
+
+ /*
+ * If the task is a small one and the buddy is not overloaded,
+ * we use buddy cpu
+ */
+ if (!is_light_task(p) || is_buddy_busy(buddy))
+ return false;
+
+ return true;
+}
+
/*
* sched_balance_self: balance the current task (running on cpu) in domains
* that have the 'flag' flag set. In practice, this is SD_BALANCE_FORK and
@@ -3098,6 +3204,9 @@ select_task_rq_fair(struct task_struct *p, int sd_flag, int wake_flags)
if (p->nr_cpus_allowed == 1)
return prev_cpu;
+ if (check_pack_buddy(cpu, p))
+ return per_cpu(sd_pack_buddy, cpu);
+
if (sd_flag & SD_BALANCE_WAKE) {
if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p)))
want_affine = 1;
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index a95d5c1..086d8bf 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -875,6 +875,7 @@ static inline void idle_balance(int cpu, struct rq *rq)
extern void sysrq_sched_debug_show(void);
extern void sched_init_granularity(void);
+extern void update_packing_domain(int cpu);
extern void update_max_interval(void);
extern void update_group_power(struct sched_domain *sd, int cpu);
extern int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu);
--
1.7.9.5
The ARM platforms take advantage of packing small tasks on few cores.
This is true even when the cores of a cluster can't be powergated
independantly.
Signed-off-by: Vincent Guittot <[email protected]>
---
arch/arm/kernel/topology.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c
index 26c12c6..00511d0 100644
--- a/arch/arm/kernel/topology.c
+++ b/arch/arm/kernel/topology.c
@@ -226,6 +226,11 @@ static inline void update_cpu_power(unsigned int cpuid, unsigned int mpidr) {}
*/
struct cputopo_arm cpu_topology[NR_CPUS];
+int arch_sd_share_power_line(void)
+{
+ return 0*SD_SHARE_POWERLINE;
+}
+
const struct cpumask *cpu_coregroup_mask(int cpu)
{
return &cpu_topology[cpu].core_sibling;
--
1.7.9.5
The atomic update of runnable_avg_sum and runnable_avg_period are ensured
by their size and the toolchain. But we must ensure to not read an old value
for one field and a newly updated value for the other field. As we don't
want to lock other CPU while reading these fields, we read twice each fields
and check that no change have occured in the middle.
Signed-off-by: Vincent Guittot <[email protected]>
---
kernel/sched/fair.c | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 8c9d3ed..6df53b5 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3133,13 +3133,28 @@ static int select_idle_sibling(struct task_struct *p, int target)
static inline bool is_buddy_busy(int cpu)
{
struct rq *rq = cpu_rq(cpu);
+ volatile u32 *psum = &rq->avg.runnable_avg_sum;
+ volatile u32 *pperiod = &rq->avg.runnable_avg_period;
+ u32 sum, new_sum, period, new_period;
+ int timeout = 10;
+
+ while (timeout) {
+ sum = *psum;
+ period = *pperiod;
+ new_sum = *psum;
+ new_period = *pperiod;
+
+ if ((sum == new_sum) && (period == new_period))
+ break;
+
+ timeout--;
+ }
/*
* A busy buddy is a CPU with a high load or a small load with a lot of
* running tasks.
*/
- return ((rq->avg.usage_avg_sum << rq->nr_running) >
- rq->avg.runnable_avg_period);
+ return ((new_sum << rq->nr_running) > new_period);
}
static inline bool is_light_task(struct task_struct *p)
--
1.7.9.5
Look for an idle CPU close the pack buddy CPU whenever possible.
The goal is to prevent the wake up of a CPU which doesn't share the power
line of the pack CPU
Signed-off-by: Vincent Guittot <[email protected]>
---
kernel/sched/fair.c | 18 ++++++++++++++++++
1 file changed, 18 insertions(+)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 6df53b5..f76acdc 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -5158,7 +5158,25 @@ static struct {
static inline int find_new_ilb(int call_cpu)
{
+ struct sched_domain *sd;
int ilb = cpumask_first(nohz.idle_cpus_mask);
+ int buddy = per_cpu(sd_pack_buddy, call_cpu);
+
+ /*
+ * If we have a pack buddy CPU, we try to run load balance on a CPU
+ * that is close to the buddy.
+ */
+ if (buddy != -1)
+ for_each_domain(buddy, sd) {
+ if (sd->flags & SD_SHARE_CPUPOWER)
+ continue;
+
+ ilb = cpumask_first_and(sched_domain_span(sd),
+ nohz.idle_cpus_mask);
+
+ if (ilb < nr_cpu_ids)
+ break;
+ }
if (ilb < nr_cpu_ids && idle_cpu(ilb))
return ilb;
--
1.7.9.5
This new flag SD SHARE_POWERLINE reflects the sharing of the power rail
between the members of a domain. As this is the current assumption of the
scheduler, the flag is added to all sched_domain
Signed-off-by: Vincent Guittot <[email protected]>
---
arch/ia64/include/asm/topology.h | 1 +
arch/tile/include/asm/topology.h | 1 +
include/linux/sched.h | 1 +
include/linux/topology.h | 3 +++
kernel/sched/core.c | 5 +++++
5 files changed, 11 insertions(+)
diff --git a/arch/ia64/include/asm/topology.h b/arch/ia64/include/asm/topology.h
index a2496e4..065c720 100644
--- a/arch/ia64/include/asm/topology.h
+++ b/arch/ia64/include/asm/topology.h
@@ -65,6 +65,7 @@ void build_cpu_to_node_map(void);
| SD_BALANCE_EXEC \
| SD_BALANCE_FORK \
| SD_WAKE_AFFINE, \
+ | arch_sd_share_power_line() \
.last_balance = jiffies, \
.balance_interval = 1, \
.nr_balance_failed = 0, \
diff --git a/arch/tile/include/asm/topology.h b/arch/tile/include/asm/topology.h
index 7a7ce39..d39ed0b 100644
--- a/arch/tile/include/asm/topology.h
+++ b/arch/tile/include/asm/topology.h
@@ -72,6 +72,7 @@ static inline const struct cpumask *cpumask_of_node(int node)
| 0*SD_PREFER_LOCAL \
| 0*SD_SHARE_CPUPOWER \
| 0*SD_SHARE_PKG_RESOURCES \
+ | arch_sd_share_power_line() \
| 0*SD_SERIALIZE \
, \
.last_balance = jiffies, \
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 4786b20..74f2daf 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -862,6 +862,7 @@ enum cpu_idle_type {
#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
#define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
+#define SD_SHARE_POWERLINE 0x0100 /* Domain members share power domain */
#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
#define SD_ASYM_PACKING 0x0800 /* Place busy groups earlier in the domain */
diff --git a/include/linux/topology.h b/include/linux/topology.h
index fec12d6..20964ab 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -99,6 +99,7 @@ int arch_update_cpu_topology(void);
| 1*SD_WAKE_AFFINE \
| 1*SD_SHARE_CPUPOWER \
| 1*SD_SHARE_PKG_RESOURCES \
+ | arch_sd_share_power_line() \
| 0*SD_SERIALIZE \
| 0*SD_PREFER_SIBLING \
| arch_sd_sibling_asym_packing() \
@@ -132,6 +133,7 @@ int arch_update_cpu_topology(void);
| 0*SD_PREFER_LOCAL \
| 0*SD_SHARE_CPUPOWER \
| 1*SD_SHARE_PKG_RESOURCES \
+ | arch_sd_share_power_line() \
| 0*SD_SERIALIZE \
, \
.last_balance = jiffies, \
@@ -163,6 +165,7 @@ int arch_update_cpu_topology(void);
| 0*SD_PREFER_LOCAL \
| 0*SD_SHARE_CPUPOWER \
| 0*SD_SHARE_PKG_RESOURCES \
+ | arch_sd_share_power_line() \
| 0*SD_SERIALIZE \
| 1*SD_PREFER_SIBLING \
, \
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index d50fbac..dab7908 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -6407,6 +6407,11 @@ int __weak arch_sd_sibling_asym_packing(void)
return 0*SD_ASYM_PACKING;
}
+int __weak arch_sd_share_power_line(void)
+{
+ return 1*SD_SHARE_POWERLINE;
+}
+
/*
* Initializers for schedule domains
* Non-inlined to reduce accumulated stack pressure in build_sched_domains()
--
1.7.9.5
Vincent,
Few comments/questions.
On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
> This new flag SD SHARE_POWERLINE reflects the sharing of the power rail
> between the members of a domain. As this is the current assumption of the
> scheduler, the flag is added to all sched_domain
>
> Signed-off-by: Vincent Guittot <[email protected]>
> ---
> arch/ia64/include/asm/topology.h | 1 +
> arch/tile/include/asm/topology.h | 1 +
> include/linux/sched.h | 1 +
> include/linux/topology.h | 3 +++
> kernel/sched/core.c | 5 +++++
> 5 files changed, 11 insertions(+)
>
> diff --git a/arch/ia64/include/asm/topology.h b/arch/ia64/include/asm/topology.h
> index a2496e4..065c720 100644
> --- a/arch/ia64/include/asm/topology.h
> +++ b/arch/ia64/include/asm/topology.h
> @@ -65,6 +65,7 @@ void build_cpu_to_node_map(void);
> | SD_BALANCE_EXEC \
> | SD_BALANCE_FORK \
> | SD_WAKE_AFFINE, \
> + | arch_sd_share_power_line() \
> .last_balance = jiffies, \
> .balance_interval = 1, \
> .nr_balance_failed = 0, \
> diff --git a/arch/tile/include/asm/topology.h b/arch/tile/include/asm/topology.h
> index 7a7ce39..d39ed0b 100644
> --- a/arch/tile/include/asm/topology.h
> +++ b/arch/tile/include/asm/topology.h
> @@ -72,6 +72,7 @@ static inline const struct cpumask *cpumask_of_node(int node)
> | 0*SD_PREFER_LOCAL \
> | 0*SD_SHARE_CPUPOWER \
> | 0*SD_SHARE_PKG_RESOURCES \
> + | arch_sd_share_power_line() \
> | 0*SD_SERIALIZE \
> , \
> .last_balance = jiffies, \
> diff --git a/include/linux/sched.h b/include/linux/sched.h
> index 4786b20..74f2daf 100644
> --- a/include/linux/sched.h
> +++ b/include/linux/sched.h
> @@ -862,6 +862,7 @@ enum cpu_idle_type {
> #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
> #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks local to this domain */
> #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
> +#define SD_SHARE_POWERLINE 0x0100 /* Domain members share power domain */
If you ignore the current use of SD_SHARE_CPUPOWER, isn't the meaning of
CPUPOWER and POWERLINE is same here. Just trying to understand the clear
meaning of this new flag. Have you not considered SD_SHARE_CPUPOWER
because it is being used for cpu_power and needs at least minimum two
domains ? SD_PACKING would have been probably more appropriate based
on the way it is being used in further series.
Regards
Santosh
Vincent,
Few comments/questions.
On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
> During sched_domain creation, we define a pack buddy CPU if available.
>
> On a system that share the powerline at all level, the buddy is set to -1
>
> On a dual clusters / dual cores system which can powergate each core and
> cluster independantly, the buddy configuration will be :
> | CPU0 | CPU1 | CPU2 | CPU3 |
> -----------------------------------
> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
^
Is that a typo ? Should it be CPU2 instead of
CPU0 ?
> Small tasks tend to slip out of the periodic load balance.
> The best place to choose to migrate them is at their wake up.
>
I have tried this series since I was looking at some of these packing
bits. On Mobile workloads like OSIdle with Screen ON, MP3, gallary,
I did see some additional filtering of threads with this series
but its not making much difference in power. More on this below.
> Signed-off-by: Vincent Guittot <[email protected]>
> ---
> kernel/sched/core.c | 1 +
> kernel/sched/fair.c | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++
> kernel/sched/sched.h | 1 +
> 3 files changed, 111 insertions(+)
>
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index dab7908..70cadbe 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
> rcu_assign_pointer(rq->sd, sd);
> destroy_sched_domains(tmp, cpu);
>
> + update_packing_domain(cpu);
> update_top_cache_domain(cpu);
> }
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 4f4a4f6..8c9d3ed 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
> update_sysctl();
> }
>
> +
> +/*
> + * Save the id of the optimal CPU that should be used to pack small tasks
> + * The value -1 is used when no buddy has been found
> + */
> +DEFINE_PER_CPU(int, sd_pack_buddy);
> +
> +/* Look for the best buddy CPU that can be used to pack small tasks
> + * We make the assumption that it doesn't wort to pack on CPU that share the
s/wort/worth
> + * same powerline. We looks for the 1st sched_domain without the
> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the lowest
> + * power per core based on the assumption that their power efficiency is
> + * better */
Commenting style..
/*
*
*/
Can you please expand the why the assumption is right ?
"it doesn't wort to pack on CPU that share the same powerline"
Think about a scenario where you have quad core, ducal cluster system
|Cluster1| |cluster 2|
| CPU0 | CPU1 | CPU2 | CPU3 | | CPU0 | CPU1 | CPU2 | CPU3 |
Both clusters run from same voltage rail and have same PLL
clocking them. But the cluster have their own power domain
and all CPU's can power gate them-self to low power states.
Clusters also have their own level2 caches.
In this case, you will still save power if you try to pack
load on one cluster. No ?
> +void update_packing_domain(int cpu)
> +{
> + struct sched_domain *sd;
> + int id = -1;
> +
> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
> + if (!sd)
> + sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
> + else
> + sd = sd->parent;
> +
> + while (sd) {
> + struct sched_group *sg = sd->groups;
> + struct sched_group *pack = sg;
> + struct sched_group *tmp = sg->next;
> +
> + /* 1st CPU of the sched domain is a good candidate */
> + if (id == -1)
> + id = cpumask_first(sched_domain_span(sd));
> +
> + /* loop the sched groups to find the best one */
> + while (tmp != sg) {
> + if (tmp->sgp->power * sg->group_weight <
> + sg->sgp->power * tmp->group_weight)
> + pack = tmp;
> + tmp = tmp->next;
> + }
> +
> + /* we have found a better group */
> + if (pack != sg)
> + id = cpumask_first(sched_group_cpus(pack));
> +
> + /* Look for another CPU than itself */
> + if ((id != cpu)
> + || ((sd->parent) && !(sd->parent->flags && SD_LOAD_BALANCE)))
Is the condition "!(sd->parent->flags && SD_LOAD_BALANCE)" for
big.LITTLE kind of system where SD_LOAD_BALANCE may not be used ?
> + break;
> +
> + sd = sd->parent;
> + }
> +
> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
> + per_cpu(sd_pack_buddy, cpu) = id;
> +}
> +
> #if BITS_PER_LONG == 32
> # define WMULT_CONST (~0UL)
> #else
> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct *p, int target)
> return target;
> }
>
> +static inline bool is_buddy_busy(int cpu)
> +{
> + struct rq *rq = cpu_rq(cpu);
> +
> + /*
> + * A busy buddy is a CPU with a high load or a small load with a lot of
> + * running tasks.
> + */
> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
> + rq->avg.runnable_avg_period);
I agree busy CPU is the one with high load, but many small threads may
not make CPU fully busy, right ? Should we just stick to the load
parameter alone here ?
> +}
> +
> +static inline bool is_light_task(struct task_struct *p)
> +{
> + /* A light task runs less than 25% in average */
> + return ((p->se.avg.usage_avg_sum << 2) < p->se.avg.runnable_avg_period);
> +}
Since the whole packing logic relies on the light threads only, the
overall effectiveness is not significant. Infact with multiple tries on
Dual core system, I didn't see any major improvement in power. I think
we need to be more aggressive here. From the cover letter, I noticed
that, you were concerned about any performance drop due to packing and
may be that is the reason you chose the conservative threshold. But the
fact is, if we want to save meaningful power, there will be slight
performance drop which is expected.
> +static int check_pack_buddy(int cpu, struct task_struct *p)
> +{
> + int buddy = per_cpu(sd_pack_buddy, cpu);
> +
> + /* No pack buddy for this CPU */
> + if (buddy == -1)
> + return false;
> +
> + /*
> + * If a task is waiting for running on the CPU which is its own buddy,
> + * let the default behavior to look for a better CPU if available
> + * The threshold has been set to 37.5%
> + */
> + if ((buddy == cpu)
> + && ((p->se.avg.usage_avg_sum << 3) < (p->se.avg.runnable_avg_sum * 5)))
> + return false;
I lost you here on better CPU , 37.5 % and last two conditions.
Isn't the first condition 'buddy==cpu' enough to return since nothing
really needs to be done in that case. Can you please expand this a bit?
> +
> + /* buddy is not an allowed CPU */
> + if (!cpumask_test_cpu(buddy, tsk_cpus_allowed(p)))
> + return false;
> +
> + /*
> + * If the task is a small one and the buddy is not overloaded,
> + * we use buddy cpu
> + */
> + if (!is_light_task(p) || is_buddy_busy(buddy))
> + return false;
This is right but both the evaluation needs update to be effective.
Regards
Santosh
$subject is bit confusing here.
On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
> The atomic update of runnable_avg_sum and runnable_avg_period are ensured
> by their size and the toolchain. But we must ensure to not read an old value
> for one field and a newly updated value for the other field. As we don't
> want to lock other CPU while reading these fields, we read twice each fields
> and check that no change have occured in the middle.
>
> Signed-off-by: Vincent Guittot <[email protected]>
> ---
> kernel/sched/fair.c | 19 +++++++++++++++++--
> 1 file changed, 17 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 8c9d3ed..6df53b5 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -3133,13 +3133,28 @@ static int select_idle_sibling(struct task_struct *p, int target)
> static inline bool is_buddy_busy(int cpu)
> {
> struct rq *rq = cpu_rq(cpu);
> + volatile u32 *psum = &rq->avg.runnable_avg_sum;
> + volatile u32 *pperiod = &rq->avg.runnable_avg_period;
> + u32 sum, new_sum, period, new_period;
> + int timeout = 10;
So it can be 2 times read or more as well.
> +
> + while (timeout) {
> + sum = *psum;
> + period = *pperiod;
> + new_sum = *psum;
> + new_period = *pperiod;
> +
> + if ((sum == new_sum) && (period == new_period))
> + break;
> +
> + timeout--;
> + }
>
Seems like you did notice incorrect pair getting read
for rq runnable_avg_sum and runnable_avg_period. Seems
like the fix is to update them together under some lock
to avoid such issues.
Regards
Santosh
On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
> Look for an idle CPU close the pack buddy CPU whenever possible.
s/close/close to
> The goal is to prevent the wake up of a CPU which doesn't share the power
> line of the pack CPU
>
> Signed-off-by: Vincent Guittot <[email protected]>
> ---
> kernel/sched/fair.c | 18 ++++++++++++++++++
> 1 file changed, 18 insertions(+)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 6df53b5..f76acdc 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -5158,7 +5158,25 @@ static struct {
>
> static inline int find_new_ilb(int call_cpu)
> {
> + struct sched_domain *sd;
> int ilb = cpumask_first(nohz.idle_cpus_mask);
> + int buddy = per_cpu(sd_pack_buddy, call_cpu);
> +
> + /*
> + * If we have a pack buddy CPU, we try to run load balance on a CPU
> + * that is close to the buddy.
> + */
> + if (buddy != -1)
> + for_each_domain(buddy, sd) {
> + if (sd->flags & SD_SHARE_CPUPOWER)
> + continue;
Do you mean SD_SHARE_POWERLINE here ?
> +
> + ilb = cpumask_first_and(sched_domain_span(sd),
> + nohz.idle_cpus_mask);
> +
> + if (ilb < nr_cpu_ids)
> + break;
> + }
>
> if (ilb < nr_cpu_ids && idle_cpu(ilb))
> return ilb;
>
Can you please expand "idle CPU _close_ the pack buddy CPU" ?
Regards
santosh
On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
> The ARM platforms take advantage of packing small tasks on few cores.
> This is true even when the cores of a cluster can't be powergated
> independently.
>
> Signed-off-by: Vincent Guittot <[email protected]>
> ---
> arch/arm/kernel/topology.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c
> index 26c12c6..00511d0 100644
> --- a/arch/arm/kernel/topology.c
> +++ b/arch/arm/kernel/topology.c
> @@ -226,6 +226,11 @@ static inline void update_cpu_power(unsigned int cpuid, unsigned int mpidr) {}
> */
> struct cputopo_arm cpu_topology[NR_CPUS];
>
> +int arch_sd_share_power_line(void)
> +{
> + return 0*SD_SHARE_POWERLINE;
> +}
Making this selection of policy based on sched domain will better. Just
gives the flexibility to choose a separate scheme for big and little
systems which will be very convenient.
Regards
Santosh
On 24 October 2012 17:17, Santosh Shilimkar <[email protected]> wrote:
> Vincent,
>
> Few comments/questions.
>
>
> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>
>> This new flag SD SHARE_POWERLINE reflects the sharing of the power rail
>> between the members of a domain. As this is the current assumption of the
>> scheduler, the flag is added to all sched_domain
>>
>> Signed-off-by: Vincent Guittot <[email protected]>
>> ---
>> arch/ia64/include/asm/topology.h | 1 +
>> arch/tile/include/asm/topology.h | 1 +
>> include/linux/sched.h | 1 +
>> include/linux/topology.h | 3 +++
>> kernel/sched/core.c | 5 +++++
>> 5 files changed, 11 insertions(+)
>>
>> diff --git a/arch/ia64/include/asm/topology.h
>> b/arch/ia64/include/asm/topology.h
>> index a2496e4..065c720 100644
>> --- a/arch/ia64/include/asm/topology.h
>> +++ b/arch/ia64/include/asm/topology.h
>> @@ -65,6 +65,7 @@ void build_cpu_to_node_map(void);
>> | SD_BALANCE_EXEC \
>> | SD_BALANCE_FORK \
>> | SD_WAKE_AFFINE, \
>> + | arch_sd_share_power_line() \
>> .last_balance = jiffies, \
>> .balance_interval = 1, \
>> .nr_balance_failed = 0, \
>> diff --git a/arch/tile/include/asm/topology.h
>> b/arch/tile/include/asm/topology.h
>> index 7a7ce39..d39ed0b 100644
>> --- a/arch/tile/include/asm/topology.h
>> +++ b/arch/tile/include/asm/topology.h
>> @@ -72,6 +72,7 @@ static inline const struct cpumask *cpumask_of_node(int
>> node)
>> | 0*SD_PREFER_LOCAL \
>> | 0*SD_SHARE_CPUPOWER \
>> | 0*SD_SHARE_PKG_RESOURCES \
>> + | arch_sd_share_power_line() \
>> | 0*SD_SERIALIZE \
>> , \
>> .last_balance = jiffies, \
>> diff --git a/include/linux/sched.h b/include/linux/sched.h
>> index 4786b20..74f2daf 100644
>> --- a/include/linux/sched.h
>> +++ b/include/linux/sched.h
>> @@ -862,6 +862,7 @@ enum cpu_idle_type {
>> #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU
>> */
>> #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks
>> local to this domain */
>> #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power
>> */
>> +#define SD_SHARE_POWERLINE 0x0100 /* Domain members share power
>> domain */
>
> If you ignore the current use of SD_SHARE_CPUPOWER, isn't the meaning of
> CPUPOWER and POWERLINE is same here. Just trying to understand the clear
> meaning of this new flag. Have you not considered SD_SHARE_CPUPOWER
> because it is being used for cpu_power and needs at least minimum two
> domains ? SD_PACKING would have been probably more appropriate based
> on the way it is being used in further series.
CPUPOWER reflects the share of hw ressources between cores like for
hyper threading. POWERLINE describes the fact that cores are sharing
the same power line amore precisely the powergate.
>
> Regards
> Santosh
>
It looks like i need to describe more what
On 29 October 2012 10:40, Vincent Guittot <[email protected]> wrote:
> On 24 October 2012 17:17, Santosh Shilimkar <[email protected]> wrote:
>> Vincent,
>>
>> Few comments/questions.
>>
>>
>> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>>
>>> This new flag SD SHARE_POWERLINE reflects the sharing of the power rail
>>> between the members of a domain. As this is the current assumption of the
>>> scheduler, the flag is added to all sched_domain
>>>
>>> Signed-off-by: Vincent Guittot <[email protected]>
>>> ---
>>> arch/ia64/include/asm/topology.h | 1 +
>>> arch/tile/include/asm/topology.h | 1 +
>>> include/linux/sched.h | 1 +
>>> include/linux/topology.h | 3 +++
>>> kernel/sched/core.c | 5 +++++
>>> 5 files changed, 11 insertions(+)
>>>
>>> diff --git a/arch/ia64/include/asm/topology.h
>>> b/arch/ia64/include/asm/topology.h
>>> index a2496e4..065c720 100644
>>> --- a/arch/ia64/include/asm/topology.h
>>> +++ b/arch/ia64/include/asm/topology.h
>>> @@ -65,6 +65,7 @@ void build_cpu_to_node_map(void);
>>> | SD_BALANCE_EXEC \
>>> | SD_BALANCE_FORK \
>>> | SD_WAKE_AFFINE, \
>>> + | arch_sd_share_power_line() \
>>> .last_balance = jiffies, \
>>> .balance_interval = 1, \
>>> .nr_balance_failed = 0, \
>>> diff --git a/arch/tile/include/asm/topology.h
>>> b/arch/tile/include/asm/topology.h
>>> index 7a7ce39..d39ed0b 100644
>>> --- a/arch/tile/include/asm/topology.h
>>> +++ b/arch/tile/include/asm/topology.h
>>> @@ -72,6 +72,7 @@ static inline const struct cpumask *cpumask_of_node(int
>>> node)
>>> | 0*SD_PREFER_LOCAL \
>>> | 0*SD_SHARE_CPUPOWER \
>>> | 0*SD_SHARE_PKG_RESOURCES \
>>> + | arch_sd_share_power_line() \
>>> | 0*SD_SERIALIZE \
>>> , \
>>> .last_balance = jiffies, \
>>> diff --git a/include/linux/sched.h b/include/linux/sched.h
>>> index 4786b20..74f2daf 100644
>>> --- a/include/linux/sched.h
>>> +++ b/include/linux/sched.h
>>> @@ -862,6 +862,7 @@ enum cpu_idle_type {
>>> #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU
>>> */
>>> #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks
>>> local to this domain */
>>> #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power
>>> */
>>> +#define SD_SHARE_POWERLINE 0x0100 /* Domain members share power
>>> domain */
>>
>> If you ignore the current use of SD_SHARE_CPUPOWER, isn't the meaning of
>> CPUPOWER and POWERLINE is same here. Just trying to understand the clear
>> meaning of this new flag. Have you not considered SD_SHARE_CPUPOWER
>> because it is being used for cpu_power and needs at least minimum two
>> domains ? SD_PACKING would have been probably more appropriate based
>> on the way it is being used in further series.
>
> CPUPOWER reflects the share of hw ressources between cores like for
> hyper threading. POWERLINE describes the fact that cores are sharing
> the same power line amore precisely the powergate.
Sorry, the mail has been sent too early while I was writing it
CPUPOWER reflects the share of hw ressources between cores like for
hyper threading. POWERLINE describes the fact that cores are sharing
the same power line and more precisely the same power gating. It looks
like I need to describe more precisely what i would mean with
SHARE_POWERLINE.
I don't want to use PACKING because it's more a behavior than a
feature. If cores can power gate independently (!SD_SHARE_POWERLINE),
packing small tasks is one interesting behavior but it may be not the
only one. I want to make a difference between the HW configuration and
the behavior we want to have above it
Vincent
>>
>> Regards
>> Santosh
>>
On 24 October 2012 17:20, Santosh Shilimkar <[email protected]> wrote:
> Vincent,
>
> Few comments/questions.
>
>
> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>
>> During sched_domain creation, we define a pack buddy CPU if available.
>>
>> On a system that share the powerline at all level, the buddy is set to -1
>>
>> On a dual clusters / dual cores system which can powergate each core and
>> cluster independantly, the buddy configuration will be :
>> | CPU0 | CPU1 | CPU2 | CPU3 |
>> -----------------------------------
>> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
>
> ^
> Is that a typo ? Should it be CPU2 instead of
> CPU0 ?
No it's not a typo.
The system packs at each scheduling level. It starts to pack in
cluster because each core can power gate independently so CPU1 tries
to pack its tasks in CPU0 and CPU3 in CPU2. Then, it packs at CPU
level so CPU2 tries to pack in the cluster of CPU0 and CPU0 packs in
itself
>
>
>> Small tasks tend to slip out of the periodic load balance.
>> The best place to choose to migrate them is at their wake up.
>>
> I have tried this series since I was looking at some of these packing
> bits. On Mobile workloads like OSIdle with Screen ON, MP3, gallary,
> I did see some additional filtering of threads with this series
> but its not making much difference in power. More on this below.
Can I ask you which configuration you have used ? how many cores and
cluster ? Can they be power gated independently ?
>
>
>> Signed-off-by: Vincent Guittot <[email protected]>
>> ---
>> kernel/sched/core.c | 1 +
>> kernel/sched/fair.c | 109
>> ++++++++++++++++++++++++++++++++++++++++++++++++++
>> kernel/sched/sched.h | 1 +
>> 3 files changed, 111 insertions(+)
>>
>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> index dab7908..70cadbe 100644
>> --- a/kernel/sched/core.c
>> +++ b/kernel/sched/core.c
>> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct
>> root_domain *rd, int cpu)
>> rcu_assign_pointer(rq->sd, sd);
>> destroy_sched_domains(tmp, cpu);
>>
>> + update_packing_domain(cpu);
>> update_top_cache_domain(cpu);
>> }
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index 4f4a4f6..8c9d3ed 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
>> update_sysctl();
>> }
>>
>> +
>> +/*
>> + * Save the id of the optimal CPU that should be used to pack small tasks
>> + * The value -1 is used when no buddy has been found
>> + */
>> +DEFINE_PER_CPU(int, sd_pack_buddy);
>> +
>> +/* Look for the best buddy CPU that can be used to pack small tasks
>> + * We make the assumption that it doesn't wort to pack on CPU that share
>> the
>
> s/wort/worth
yes
>
>> + * same powerline. We looks for the 1st sched_domain without the
>> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the
>> lowest
>> + * power per core based on the assumption that their power efficiency is
>> + * better */
>
> Commenting style..
> /*
> *
> */
>
yes
> Can you please expand the why the assumption is right ?
> "it doesn't wort to pack on CPU that share the same powerline"
By "share the same power-line", I mean that the CPUs can't power off
independently. So if some CPUs can't power off independently, it's
worth to try to use most of them to race to idle.
>
> Think about a scenario where you have quad core, ducal cluster system
>
> |Cluster1| |cluster 2|
> | CPU0 | CPU1 | CPU2 | CPU3 | | CPU0 | CPU1 | CPU2 | CPU3 |
>
>
> Both clusters run from same voltage rail and have same PLL
> clocking them. But the cluster have their own power domain
> and all CPU's can power gate them-self to low power states.
> Clusters also have their own level2 caches.
>
> In this case, you will still save power if you try to pack
> load on one cluster. No ?
yes, I need to update the description of SD_SHARE_POWERLINE because
I'm afraid I was not clear enough. SD_SHARE_POWERLINE includes the
power gating capacity of each core. For your example above, the
SD_SHARE_POWERLINE shoud be cleared at both MC and CPU level.
>
>
>> +void update_packing_domain(int cpu)
>> +{
>> + struct sched_domain *sd;
>> + int id = -1;
>> +
>> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
>> + if (!sd)
>> + sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
>> + else
>> + sd = sd->parent;
>> +
>> + while (sd) {
>> + struct sched_group *sg = sd->groups;
>> + struct sched_group *pack = sg;
>> + struct sched_group *tmp = sg->next;
>> +
>> + /* 1st CPU of the sched domain is a good candidate */
>> + if (id == -1)
>> + id = cpumask_first(sched_domain_span(sd));
>> +
>> + /* loop the sched groups to find the best one */
>> + while (tmp != sg) {
>> + if (tmp->sgp->power * sg->group_weight <
>> + sg->sgp->power *
>> tmp->group_weight)
>> + pack = tmp;
>> + tmp = tmp->next;
>> + }
>> +
>> + /* we have found a better group */
>> + if (pack != sg)
>> + id = cpumask_first(sched_group_cpus(pack));
>> +
>> + /* Look for another CPU than itself */
>> + if ((id != cpu)
>> + || ((sd->parent) && !(sd->parent->flags &&
>> SD_LOAD_BALANCE)))
>
> Is the condition "!(sd->parent->flags && SD_LOAD_BALANCE)" for
> big.LITTLE kind of system where SD_LOAD_BALANCE may not be used ?
No, packing small tasks is part of the load balance so if the
LOAD_BALANCE flag is cleared, we will not try to pack which is a kind
of load balance. There is no link with big.LITTLE
>
>
>> + break;
>> +
>> + sd = sd->parent;
>> + }
>> +
>> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
>> + per_cpu(sd_pack_buddy, cpu) = id;
>> +}
>> +
>> #if BITS_PER_LONG == 32
>> # define WMULT_CONST (~0UL)
>> #else
>> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct
>> *p, int target)
>> return target;
>> }
>>
>> +static inline bool is_buddy_busy(int cpu)
>> +{
>> + struct rq *rq = cpu_rq(cpu);
>> +
>> + /*
>> + * A busy buddy is a CPU with a high load or a small load with a
>> lot of
>> + * running tasks.
>> + */
>> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
>> + rq->avg.runnable_avg_period);
>
> I agree busy CPU is the one with high load, but many small threads may
> not make CPU fully busy, right ? Should we just stick to the load
> parameter alone here ?
IMO, the busy state of a CPU isn't only the load but also how many
threads are waiting for running on it. This formula tries to take into
account both inputs. If you have dozen of small tasks on a CPU, the
latency can be large even if the tasks are small.
>
>
>> +}
>> +
>> +static inline bool is_light_task(struct task_struct *p)
>> +{
>> + /* A light task runs less than 25% in average */
>> + return ((p->se.avg.usage_avg_sum << 2) <
>> p->se.avg.runnable_avg_period);
>> +}
>
> Since the whole packing logic relies on the light threads only, the
> overall effectiveness is not significant. Infact with multiple tries on
> Dual core system, I didn't see any major improvement in power. I think
> we need to be more aggressive here. From the cover letter, I noticed
> that, you were concerned about any performance drop due to packing and
> may be that is the reason you chose the conservative threshold. But the
> fact is, if we want to save meaningful power, there will be slight
> performance drop which is expected.
I think that everybody agrees that packing small tasks will save power
whereas it seems to be not so obvious for heavy task. But I may have
set the threshold a bit too low
Up to which load, you would like to pack on 1 core of your dual core system ?
Can you provide more details of your load ? Have you got a trace that
you can share ?
>
>
>> +static int check_pack_buddy(int cpu, struct task_struct *p)
>> +{
>> + int buddy = per_cpu(sd_pack_buddy, cpu);
>> +
>> + /* No pack buddy for this CPU */
>> + if (buddy == -1)
>> + return false;
>> +
>> + /*
>> + * If a task is waiting for running on the CPU which is its own
>> buddy,
>> + * let the default behavior to look for a better CPU if available
>> + * The threshold has been set to 37.5%
>> + */
>> + if ((buddy == cpu)
>> + && ((p->se.avg.usage_avg_sum << 3) < (p->se.avg.runnable_avg_sum
>> * 5)))
>> + return false;
>
> I lost you here on better CPU , 37.5 % and last two conditions.
> Isn't the first condition 'buddy==cpu' enough to return since nothing really
> needs to be done in that case. Can you please expand this a bit?
If you have a lot of small tasks waking up and running simultaneously,
Some tasks will wait for runnning and we could short the running time
by parallelizing tasks if possible (at MC level for example)
>
>
>> +
>> + /* buddy is not an allowed CPU */
>> + if (!cpumask_test_cpu(buddy, tsk_cpus_allowed(p)))
>> + return false;
>> +
>> + /*
>> + * If the task is a small one and the buddy is not overloaded,
>> + * we use buddy cpu
>> + */
>> + if (!is_light_task(p) || is_buddy_busy(buddy))
>> + return false;
>
> This is right but both the evaluation needs update to be effective.
>
> Regards
> Santosh
On 24 October 2012 17:21, Santosh Shilimkar <[email protected]> wrote:
> $subject is bit confusing here.
>
>
> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>
>> The atomic update of runnable_avg_sum and runnable_avg_period are ensured
>> by their size and the toolchain. But we must ensure to not read an old
>> value
>> for one field and a newly updated value for the other field. As we don't
>> want to lock other CPU while reading these fields, we read twice each
>> fields
>> and check that no change have occured in the middle.
>>
>> Signed-off-by: Vincent Guittot <[email protected]>
>> ---
>> kernel/sched/fair.c | 19 +++++++++++++++++--
>> 1 file changed, 17 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index 8c9d3ed..6df53b5 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -3133,13 +3133,28 @@ static int select_idle_sibling(struct task_struct
>> *p, int target)
>> static inline bool is_buddy_busy(int cpu)
>> {
>> struct rq *rq = cpu_rq(cpu);
>> + volatile u32 *psum = &rq->avg.runnable_avg_sum;
>> + volatile u32 *pperiod = &rq->avg.runnable_avg_period;
>> + u32 sum, new_sum, period, new_period;
>> + int timeout = 10;
>
> So it can be 2 times read or more as well.
>
>> +
>> + while (timeout) {
>> + sum = *psum;
>> + period = *pperiod;
>> + new_sum = *psum;
>> + new_period = *pperiod;
>> +
>> + if ((sum == new_sum) && (period == new_period))
>> + break;
>> +
>> + timeout--;
>> + }
>>
> Seems like you did notice incorrect pair getting read
> for rq runnable_avg_sum and runnable_avg_period. Seems
> like the fix is to update them together under some lock
> to avoid such issues.
My goal is to have a lock free mechanism because I don't want to lock
another CPU while reading its statistic
>
> Regards
> Santosh
>
On 24 October 2012 17:21, Santosh Shilimkar <[email protected]> wrote:
> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>
>> Look for an idle CPU close the pack buddy CPU whenever possible.
>
> s/close/close to
yes
>
>> The goal is to prevent the wake up of a CPU which doesn't share the power
>> line of the pack CPU
>>
>> Signed-off-by: Vincent Guittot <[email protected]>
>> ---
>> kernel/sched/fair.c | 18 ++++++++++++++++++
>> 1 file changed, 18 insertions(+)
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index 6df53b5..f76acdc 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -5158,7 +5158,25 @@ static struct {
>>
>> static inline int find_new_ilb(int call_cpu)
>> {
>> + struct sched_domain *sd;
>> int ilb = cpumask_first(nohz.idle_cpus_mask);
>> + int buddy = per_cpu(sd_pack_buddy, call_cpu);
>> +
>> + /*
>> + * If we have a pack buddy CPU, we try to run load balance on a
>> CPU
>> + * that is close to the buddy.
>> + */
>> + if (buddy != -1)
>> + for_each_domain(buddy, sd) {
>> + if (sd->flags & SD_SHARE_CPUPOWER)
>> + continue;
>
> Do you mean SD_SHARE_POWERLINE here ?
No, I just don't want to take hyperthread level for ILB
>
>> +
>> + ilb = cpumask_first_and(sched_domain_span(sd),
>> + nohz.idle_cpus_mask);
>> +
>> + if (ilb < nr_cpu_ids)
>> + break;
>> + }
>>
>> if (ilb < nr_cpu_ids && idle_cpu(ilb))
>> return ilb;
>>
> Can you please expand "idle CPU _close_ the pack buddy CPU" ?
The goal is to packed the tasks on the pack buddy CPU so when the
scheduler needs to start ILB, I try to wake up a CPU that is close to
the buddy and preferably in the same cluster
>
> Regards
> santosh
On 24 October 2012 17:21, Santosh Shilimkar <[email protected]> wrote:
> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>
>> The ARM platforms take advantage of packing small tasks on few cores.
>> This is true even when the cores of a cluster can't be powergated
>> independently.
>>
>>
>> Signed-off-by: Vincent Guittot <[email protected]>
>> ---
>> arch/arm/kernel/topology.c | 5 +++++
>> 1 file changed, 5 insertions(+)
>>
>> diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c
>> index 26c12c6..00511d0 100644
>> --- a/arch/arm/kernel/topology.c
>> +++ b/arch/arm/kernel/topology.c
>> @@ -226,6 +226,11 @@ static inline void update_cpu_power(unsigned int
>> cpuid, unsigned int mpidr) {}
>> */
>> struct cputopo_arm cpu_topology[NR_CPUS];
>>
>> +int arch_sd_share_power_line(void)
>> +{
>> + return 0*SD_SHARE_POWERLINE;
>> +}
>
>
> Making this selection of policy based on sched domain will better. Just
> gives the flexibility to choose a separate scheme for big and little
> systems which will be very convenient.
I agree that it would be more flexible to be able to set it for each level
>
> Regards
> Santosh
>
>
>
>
>
On Monday 29 October 2012 03:20 PM, Vincent Guittot wrote:
> It looks like i need to describe more what
>
> On 29 October 2012 10:40, Vincent Guittot <[email protected]> wrote:
>> On 24 October 2012 17:17, Santosh Shilimkar <[email protected]> wrote:
>>> Vincent,
>>>
>>> Few comments/questions.
>>>
>>>
>>> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>>>
>>>> This new flag SD SHARE_POWERLINE reflects the sharing of the power rail
>>>> between the members of a domain. As this is the current assumption of the
>>>> scheduler, the flag is added to all sched_domain
>>>>
>>>> Signed-off-by: Vincent Guittot <[email protected]>
>>>> ---
>>>> arch/ia64/include/asm/topology.h | 1 +
>>>> arch/tile/include/asm/topology.h | 1 +
>>>> include/linux/sched.h | 1 +
>>>> include/linux/topology.h | 3 +++
>>>> kernel/sched/core.c | 5 +++++
>>>> 5 files changed, 11 insertions(+)
>>>>
>>>> diff --git a/arch/ia64/include/asm/topology.h
>>>> b/arch/ia64/include/asm/topology.h
>>>> index a2496e4..065c720 100644
>>>> --- a/arch/ia64/include/asm/topology.h
>>>> +++ b/arch/ia64/include/asm/topology.h
>>>> @@ -65,6 +65,7 @@ void build_cpu_to_node_map(void);
>>>> | SD_BALANCE_EXEC \
>>>> | SD_BALANCE_FORK \
>>>> | SD_WAKE_AFFINE, \
>>>> + | arch_sd_share_power_line() \
>>>> .last_balance = jiffies, \
>>>> .balance_interval = 1, \
>>>> .nr_balance_failed = 0, \
>>>> diff --git a/arch/tile/include/asm/topology.h
>>>> b/arch/tile/include/asm/topology.h
>>>> index 7a7ce39..d39ed0b 100644
>>>> --- a/arch/tile/include/asm/topology.h
>>>> +++ b/arch/tile/include/asm/topology.h
>>>> @@ -72,6 +72,7 @@ static inline const struct cpumask *cpumask_of_node(int
>>>> node)
>>>> | 0*SD_PREFER_LOCAL \
>>>> | 0*SD_SHARE_CPUPOWER \
>>>> | 0*SD_SHARE_PKG_RESOURCES \
>>>> + | arch_sd_share_power_line() \
>>>> | 0*SD_SERIALIZE \
>>>> , \
>>>> .last_balance = jiffies, \
>>>> diff --git a/include/linux/sched.h b/include/linux/sched.h
>>>> index 4786b20..74f2daf 100644
>>>> --- a/include/linux/sched.h
>>>> +++ b/include/linux/sched.h
>>>> @@ -862,6 +862,7 @@ enum cpu_idle_type {
>>>> #define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU
>>>> */
>>>> #define SD_PREFER_LOCAL 0x0040 /* Prefer to keep tasks
>>>> local to this domain */
>>>> #define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power
>>>> */
>>>> +#define SD_SHARE_POWERLINE 0x0100 /* Domain members share power
>>>> domain */
>>>
>>> If you ignore the current use of SD_SHARE_CPUPOWER, isn't the meaning of
>>> CPUPOWER and POWERLINE is same here. Just trying to understand the clear
>>> meaning of this new flag. Have you not considered SD_SHARE_CPUPOWER
>>> because it is being used for cpu_power and needs at least minimum two
>>> domains ? SD_PACKING would have been probably more appropriate based
>>> on the way it is being used in further series.
>>
>> CPUPOWER reflects the share of hw ressources between cores like for
>> hyper threading. POWERLINE describes the fact that cores are sharing
>> the same power line amore precisely the powergate.
>
> Sorry, the mail has been sent too early while I was writing it
>
> CPUPOWER reflects the share of hw ressources between cores like for
> hyper threading. POWERLINE describes the fact that cores are sharing
> the same power line and more precisely the same power gating. It looks
> like I need to describe more precisely what i would mean with
> SHARE_POWERLINE.
>
Yes. More description will help. I see bit of overlap POWERLINE
flag with SD_SHARE_CPUPOWER and SD_SHARE_PKG_RESOURCES and hence
the questions.
> I don't want to use PACKING because it's more a behavior than a
> feature. If cores can power gate independently (!SD_SHARE_POWERLINE),
> packing small tasks is one interesting behavior but it may be not the
> only one. I want to make a difference between the HW configuration and
> the behavior we want to have above it
>
Fair enough. Thanks for clarification.
Regards,
Santosh
On Monday 29 October 2012 06:42 PM, Vincent Guittot wrote:
> On 24 October 2012 17:20, Santosh Shilimkar <[email protected]> wrote:
>> Vincent,
>>
>> Few comments/questions.
>>
>>
>> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>>
>>> During sched_domain creation, we define a pack buddy CPU if available.
>>>
>>> On a system that share the powerline at all level, the buddy is set to -1
>>>
>>> On a dual clusters / dual cores system which can powergate each core and
>>> cluster independantly, the buddy configuration will be :
>>> | CPU0 | CPU1 | CPU2 | CPU3 |
>>> -----------------------------------
>>> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
>>
>> ^
>> Is that a typo ? Should it be CPU2 instead of
>> CPU0 ?
>
> No it's not a typo.
> The system packs at each scheduling level. It starts to pack in
> cluster because each core can power gate independently so CPU1 tries
> to pack its tasks in CPU0 and CPU3 in CPU2. Then, it packs at CPU
> level so CPU2 tries to pack in the cluster of CPU0 and CPU0 packs in
> itself
>
I get it. Though in above example a task may migrate from say
CPU3->CPU2->CPU0 as part of packing. I was just thinking whether
moving such task from say CPU3 to CPU0 might be best instead.
>>
>>> Small tasks tend to slip out of the periodic load balance.
>>> The best place to choose to migrate them is at their wake up.
>>>
>> I have tried this series since I was looking at some of these packing
>> bits. On Mobile workloads like OSIdle with Screen ON, MP3, gallary,
>> I did see some additional filtering of threads with this series
>> but its not making much difference in power. More on this below.
>
> Can I ask you which configuration you have used ? how many cores and
> cluster ? Can they be power gated independently ?
>
I have been trying with couple of setups. Dual Core ARM machine and
Quad core X86 box with single package thought most of the mobile
workload analysis I was doing on ARM machine. On both setups
CPUs can be gated independently.
>>
>>
>>> Signed-off-by: Vincent Guittot <[email protected]>
>>> ---
>>> kernel/sched/core.c | 1 +
>>> kernel/sched/fair.c | 109
>>> ++++++++++++++++++++++++++++++++++++++++++++++++++
>>> kernel/sched/sched.h | 1 +
>>> 3 files changed, 111 insertions(+)
>>>
>>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>>> index dab7908..70cadbe 100644
>>> --- a/kernel/sched/core.c
>>> +++ b/kernel/sched/core.c
>>> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct
>>> root_domain *rd, int cpu)
>>> rcu_assign_pointer(rq->sd, sd);
>>> destroy_sched_domains(tmp, cpu);
>>>
>>> + update_packing_domain(cpu);
>>> update_top_cache_domain(cpu);
>>> }
>>>
>>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>>> index 4f4a4f6..8c9d3ed 100644
>>> --- a/kernel/sched/fair.c
>>> +++ b/kernel/sched/fair.c
>>> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
>>> update_sysctl();
>>> }
>>>
>>> +
>>> +/*
>>> + * Save the id of the optimal CPU that should be used to pack small tasks
>>> + * The value -1 is used when no buddy has been found
>>> + */
>>> +DEFINE_PER_CPU(int, sd_pack_buddy);
>>> +
>>> +/* Look for the best buddy CPU that can be used to pack small tasks
>>> + * We make the assumption that it doesn't wort to pack on CPU that share
>>> the
>>
>> s/wort/worth
>
> yes
>
>>
>>> + * same powerline. We looks for the 1st sched_domain without the
>>> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the
>>> lowest
>>> + * power per core based on the assumption that their power efficiency is
>>> + * better */
>>
>> Commenting style..
>> /*
>> *
>> */
>>
>
> yes
>
>> Can you please expand the why the assumption is right ?
>> "it doesn't wort to pack on CPU that share the same powerline"
>
> By "share the same power-line", I mean that the CPUs can't power off
> independently. So if some CPUs can't power off independently, it's
> worth to try to use most of them to race to idle.
>
In that case I suggest we use different word here. Power line can be
treated as voltage line, power domain.
May be SD_SHARE_CPU_POWERDOMAIN ?
>>
>> Think about a scenario where you have quad core, ducal cluster system
>>
>> |Cluster1| |cluster 2|
>> | CPU0 | CPU1 | CPU2 | CPU3 | | CPU0 | CPU1 | CPU2 | CPU3 |
>>
>>
>> Both clusters run from same voltage rail and have same PLL
>> clocking them. But the cluster have their own power domain
>> and all CPU's can power gate them-self to low power states.
>> Clusters also have their own level2 caches.
>>
>> In this case, you will still save power if you try to pack
>> load on one cluster. No ?
>
> yes, I need to update the description of SD_SHARE_POWERLINE because
> I'm afraid I was not clear enough. SD_SHARE_POWERLINE includes the
> power gating capacity of each core. For your example above, the
> SD_SHARE_POWERLINE shoud be cleared at both MC and CPU level.
>
Thanks for clarification.
>>
>>
>>> +void update_packing_domain(int cpu)
>>> +{
>>> + struct sched_domain *sd;
>>> + int id = -1;
>>> +
>>> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
>>> + if (!sd)
>>> + sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
>>> + else
>>> + sd = sd->parent;
>>> +
>>> + while (sd) {
>>> + struct sched_group *sg = sd->groups;
>>> + struct sched_group *pack = sg;
>>> + struct sched_group *tmp = sg->next;
>>> +
>>> + /* 1st CPU of the sched domain is a good candidate */
>>> + if (id == -1)
>>> + id = cpumask_first(sched_domain_span(sd));
>>> +
>>> + /* loop the sched groups to find the best one */
>>> + while (tmp != sg) {
>>> + if (tmp->sgp->power * sg->group_weight <
>>> + sg->sgp->power *
>>> tmp->group_weight)
>>> + pack = tmp;
>>> + tmp = tmp->next;
>>> + }
>>> +
>>> + /* we have found a better group */
>>> + if (pack != sg)
>>> + id = cpumask_first(sched_group_cpus(pack));
>>> +
>>> + /* Look for another CPU than itself */
>>> + if ((id != cpu)
>>> + || ((sd->parent) && !(sd->parent->flags &&
>>> SD_LOAD_BALANCE)))
>>
>> Is the condition "!(sd->parent->flags && SD_LOAD_BALANCE)" for
>> big.LITTLE kind of system where SD_LOAD_BALANCE may not be used ?
>
> No, packing small tasks is part of the load balance so if the
> LOAD_BALANCE flag is cleared, we will not try to pack which is a kind
> of load balance. There is no link with big.LITTLE
>
Now it make sense to me.
>>
>>
>>> + break;
>>> +
>>> + sd = sd->parent;
>>> + }
>>> +
>>> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
>>> + per_cpu(sd_pack_buddy, cpu) = id;
>>> +}
>>> +
>>> #if BITS_PER_LONG == 32
>>> # define WMULT_CONST (~0UL)
>>> #else
>>> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct
>>> *p, int target)
>>> return target;
>>> }
>>>
>>> +static inline bool is_buddy_busy(int cpu)
>>> +{
>>> + struct rq *rq = cpu_rq(cpu);
>>> +
>>> + /*
>>> + * A busy buddy is a CPU with a high load or a small load with a
>>> lot of
>>> + * running tasks.
>>> + */
>>> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
>>> + rq->avg.runnable_avg_period);
>>
>> I agree busy CPU is the one with high load, but many small threads may
>> not make CPU fully busy, right ? Should we just stick to the load
>> parameter alone here ?
>
> IMO, the busy state of a CPU isn't only the load but also how many
> threads are waiting for running on it. This formula tries to take into
> account both inputs. If you have dozen of small tasks on a CPU, the
> latency can be large even if the tasks are small.
>
Sure. Your point is to avoid throttling and probably use race for
idle.
>>
>>
>>> +}
>>> +
>>> +static inline bool is_light_task(struct task_struct *p)
>>> +{
>>> + /* A light task runs less than 25% in average */
>>> + return ((p->se.avg.usage_avg_sum << 2) <
>>> p->se.avg.runnable_avg_period);
>>> +}
>>
>> Since the whole packing logic relies on the light threads only, the
>> overall effectiveness is not significant. Infact with multiple tries on
>> Dual core system, I didn't see any major improvement in power. I think
>> we need to be more aggressive here. From the cover letter, I noticed
>> that, you were concerned about any performance drop due to packing and
>> may be that is the reason you chose the conservative threshold. But the
>> fact is, if we want to save meaningful power, there will be slight
>> performance drop which is expected.
>
> I think that everybody agrees that packing small tasks will save power
> whereas it seems to be not so obvious for heavy task. But I may have
> set the threshold a bit too low
>
I agree on packing saves power part for sure.
> Up to which load, you would like to pack on 1 core of your dual core system ?
> Can you provide more details of your load ? Have you got a trace that
> you can share ?
>
More than how much load to pack, I was more looking from the power
savings delta we can achieve by doing it. Some of the usecases like
osidle, mp3, gallary are already very low power and that might be
the reason I didn't notice major mA delta. Though the perf
traces did show some filtering even at 25 % load. i tried upto
50 % threshold to see the effectiveness and there was more
improvement and hence the suggestion about aggressiveness.
May be you can try some of these use-cases on your setup instead of
synthetic workload and see the results.
Regards
Santosh
On Monday 29 October 2012 06:57 PM, Vincent Guittot wrote:
> On 24 October 2012 17:21, Santosh Shilimkar <[email protected]> wrote:
>> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>>
>>> Look for an idle CPU close the pack buddy CPU whenever possible.
>>
>> s/close/close to
>
> yes
>
>>
>>> The goal is to prevent the wake up of a CPU which doesn't share the power
>>> line of the pack CPU
>>>
>>> Signed-off-by: Vincent Guittot <[email protected]>
>>> ---
>>> kernel/sched/fair.c | 18 ++++++++++++++++++
>>> 1 file changed, 18 insertions(+)
>>>
>>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>>> index 6df53b5..f76acdc 100644
>>> --- a/kernel/sched/fair.c
>>> +++ b/kernel/sched/fair.c
>>> @@ -5158,7 +5158,25 @@ static struct {
>>>
>>> static inline int find_new_ilb(int call_cpu)
>>> {
>>> + struct sched_domain *sd;
>>> int ilb = cpumask_first(nohz.idle_cpus_mask);
>>> + int buddy = per_cpu(sd_pack_buddy, call_cpu);
>>> +
>>> + /*
>>> + * If we have a pack buddy CPU, we try to run load balance on a
>>> CPU
>>> + * that is close to the buddy.
>>> + */
>>> + if (buddy != -1)
>>> + for_each_domain(buddy, sd) {
>>> + if (sd->flags & SD_SHARE_CPUPOWER)
>>> + continue;
>>
>> Do you mean SD_SHARE_POWERLINE here ?
>
> No, I just don't want to take hyperthread level for ILB
>
>>
>>> +
>>> + ilb = cpumask_first_and(sched_domain_span(sd),
>>> + nohz.idle_cpus_mask);
>>> +
>>> + if (ilb < nr_cpu_ids)
>>> + break;
>>> + }
>>>
>>> if (ilb < nr_cpu_ids && idle_cpu(ilb))
>>> return ilb;
>>>
>> Can you please expand "idle CPU _close_ the pack buddy CPU" ?
>
> The goal is to packed the tasks on the pack buddy CPU so when the
> scheduler needs to start ILB, I try to wake up a CPU that is close to
> the buddy and preferably in the same cluster
>
I see your point now. Thanks for clarification.
Regards
santosh
On Monday 29 October 2012 06:58 PM, Vincent Guittot wrote:
> On 24 October 2012 17:21, Santosh Shilimkar <[email protected]> wrote:
>> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>>
>>> The ARM platforms take advantage of packing small tasks on few cores.
>>> This is true even when the cores of a cluster can't be powergated
>>> independently.
>>>
>>>
>>> Signed-off-by: Vincent Guittot <[email protected]>
>>> ---
>>> arch/arm/kernel/topology.c | 5 +++++
>>> 1 file changed, 5 insertions(+)
>>>
>>> diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c
>>> index 26c12c6..00511d0 100644
>>> --- a/arch/arm/kernel/topology.c
>>> +++ b/arch/arm/kernel/topology.c
>>> @@ -226,6 +226,11 @@ static inline void update_cpu_power(unsigned int
>>> cpuid, unsigned int mpidr) {}
>>> */
>>> struct cputopo_arm cpu_topology[NR_CPUS];
>>>
>>> +int arch_sd_share_power_line(void)
>>> +{
>>> + return 0*SD_SHARE_POWERLINE;
>>> +}
>>
>>
>> Making this selection of policy based on sched domain will better. Just
>> gives the flexibility to choose a separate scheme for big and little
>> systems which will be very convenient.
>
> I agree that it would be more flexible to be able to set it for each level
>
Will you be addressing that in next version then ?
Regards
santosh
On Fri, Nov 02, 2012 at 10:53:47AM +0000, Santosh Shilimkar wrote:
> On Monday 29 October 2012 06:42 PM, Vincent Guittot wrote:
> > On 24 October 2012 17:20, Santosh Shilimkar <[email protected]> wrote:
> >> Vincent,
> >>
> >> Few comments/questions.
> >>
> >>
> >> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
> >>>
> >>> During sched_domain creation, we define a pack buddy CPU if available.
> >>>
> >>> On a system that share the powerline at all level, the buddy is set to -1
> >>>
> >>> On a dual clusters / dual cores system which can powergate each core and
> >>> cluster independantly, the buddy configuration will be :
> >>> | CPU0 | CPU1 | CPU2 | CPU3 |
> >>> -----------------------------------
> >>> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
> >>
> >> ^
> >> Is that a typo ? Should it be CPU2 instead of
> >> CPU0 ?
> >
> > No it's not a typo.
> > The system packs at each scheduling level. It starts to pack in
> > cluster because each core can power gate independently so CPU1 tries
> > to pack its tasks in CPU0 and CPU3 in CPU2. Then, it packs at CPU
> > level so CPU2 tries to pack in the cluster of CPU0 and CPU0 packs in
> > itself
> >
> I get it. Though in above example a task may migrate from say
> CPU3->CPU2->CPU0 as part of packing. I was just thinking whether
> moving such task from say CPU3 to CPU0 might be best instead.
To me it seems suboptimal to pack the task twice, but the alternative is
not good either. If you try to move the task directly to CPU0 you may
miss packing opportunities if CPU0 is already busy, while CPU2 might
have enough capacity to take it. It would probably be better to check
the business of CPU0 and then back off and try CPU2 if CP0 is busy. This
would require a buddy list for each CPU rather just a single buddy and
thus might become expensive.
>
> >>
> >>> Small tasks tend to slip out of the periodic load balance.
> >>> The best place to choose to migrate them is at their wake up.
> >>>
> >> I have tried this series since I was looking at some of these packing
> >> bits. On Mobile workloads like OSIdle with Screen ON, MP3, gallary,
> >> I did see some additional filtering of threads with this series
> >> but its not making much difference in power. More on this below.
> >
> > Can I ask you which configuration you have used ? how many cores and
> > cluster ? Can they be power gated independently ?
> >
> I have been trying with couple of setups. Dual Core ARM machine and
> Quad core X86 box with single package thought most of the mobile
> workload analysis I was doing on ARM machine. On both setups
> CPUs can be gated independently.
>
> >>
> >>
> >>> Signed-off-by: Vincent Guittot <[email protected]>
> >>> ---
> >>> kernel/sched/core.c | 1 +
> >>> kernel/sched/fair.c | 109
> >>> ++++++++++++++++++++++++++++++++++++++++++++++++++
> >>> kernel/sched/sched.h | 1 +
> >>> 3 files changed, 111 insertions(+)
> >>>
> >>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> >>> index dab7908..70cadbe 100644
> >>> --- a/kernel/sched/core.c
> >>> +++ b/kernel/sched/core.c
> >>> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct
> >>> root_domain *rd, int cpu)
> >>> rcu_assign_pointer(rq->sd, sd);
> >>> destroy_sched_domains(tmp, cpu);
> >>>
> >>> + update_packing_domain(cpu);
> >>> update_top_cache_domain(cpu);
> >>> }
> >>>
> >>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> >>> index 4f4a4f6..8c9d3ed 100644
> >>> --- a/kernel/sched/fair.c
> >>> +++ b/kernel/sched/fair.c
> >>> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
> >>> update_sysctl();
> >>> }
> >>>
> >>> +
> >>> +/*
> >>> + * Save the id of the optimal CPU that should be used to pack small tasks
> >>> + * The value -1 is used when no buddy has been found
> >>> + */
> >>> +DEFINE_PER_CPU(int, sd_pack_buddy);
> >>> +
> >>> +/* Look for the best buddy CPU that can be used to pack small tasks
> >>> + * We make the assumption that it doesn't wort to pack on CPU that share
> >>> the
> >>
> >> s/wort/worth
> >
> > yes
> >
> >>
> >>> + * same powerline. We looks for the 1st sched_domain without the
> >>> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the
> >>> lowest
> >>> + * power per core based on the assumption that their power efficiency is
> >>> + * better */
> >>
> >> Commenting style..
> >> /*
> >> *
> >> */
> >>
> >
> > yes
> >
> >> Can you please expand the why the assumption is right ?
> >> "it doesn't wort to pack on CPU that share the same powerline"
> >
> > By "share the same power-line", I mean that the CPUs can't power off
> > independently. So if some CPUs can't power off independently, it's
> > worth to try to use most of them to race to idle.
> >
> In that case I suggest we use different word here. Power line can be
> treated as voltage line, power domain.
> May be SD_SHARE_CPU_POWERDOMAIN ?
>
How about just SD_SHARE_POWERDOMAIN ?
> >>
> >> Think about a scenario where you have quad core, ducal cluster system
> >>
> >> |Cluster1| |cluster 2|
> >> | CPU0 | CPU1 | CPU2 | CPU3 | | CPU0 | CPU1 | CPU2 | CPU3 |
> >>
> >>
> >> Both clusters run from same voltage rail and have same PLL
> >> clocking them. But the cluster have their own power domain
> >> and all CPU's can power gate them-self to low power states.
> >> Clusters also have their own level2 caches.
> >>
> >> In this case, you will still save power if you try to pack
> >> load on one cluster. No ?
> >
> > yes, I need to update the description of SD_SHARE_POWERLINE because
> > I'm afraid I was not clear enough. SD_SHARE_POWERLINE includes the
> > power gating capacity of each core. For your example above, the
> > SD_SHARE_POWERLINE shoud be cleared at both MC and CPU level.
> >
> Thanks for clarification.
>
> >>
> >>
> >>> +void update_packing_domain(int cpu)
> >>> +{
> >>> + struct sched_domain *sd;
> >>> + int id = -1;
> >>> +
> >>> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
> >>> + if (!sd)
> >>> + sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
> >>> + else
> >>> + sd = sd->parent;
> >>> +
> >>> + while (sd) {
> >>> + struct sched_group *sg = sd->groups;
> >>> + struct sched_group *pack = sg;
> >>> + struct sched_group *tmp = sg->next;
> >>> +
> >>> + /* 1st CPU of the sched domain is a good candidate */
> >>> + if (id == -1)
> >>> + id = cpumask_first(sched_domain_span(sd));
> >>> +
> >>> + /* loop the sched groups to find the best one */
> >>> + while (tmp != sg) {
> >>> + if (tmp->sgp->power * sg->group_weight <
> >>> + sg->sgp->power *
> >>> tmp->group_weight)
> >>> + pack = tmp;
> >>> + tmp = tmp->next;
> >>> + }
> >>> +
> >>> + /* we have found a better group */
> >>> + if (pack != sg)
> >>> + id = cpumask_first(sched_group_cpus(pack));
> >>> +
> >>> + /* Look for another CPU than itself */
> >>> + if ((id != cpu)
> >>> + || ((sd->parent) && !(sd->parent->flags &&
> >>> SD_LOAD_BALANCE)))
> >>
> >> Is the condition "!(sd->parent->flags && SD_LOAD_BALANCE)" for
> >> big.LITTLE kind of system where SD_LOAD_BALANCE may not be used ?
> >
> > No, packing small tasks is part of the load balance so if the
> > LOAD_BALANCE flag is cleared, we will not try to pack which is a kind
> > of load balance. There is no link with big.LITTLE
> >
> Now it make sense to me.
>
> >>
> >>
> >>> + break;
> >>> +
> >>> + sd = sd->parent;
> >>> + }
> >>> +
> >>> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
> >>> + per_cpu(sd_pack_buddy, cpu) = id;
> >>> +}
> >>> +
> >>> #if BITS_PER_LONG == 32
> >>> # define WMULT_CONST (~0UL)
> >>> #else
> >>> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct
> >>> *p, int target)
> >>> return target;
> >>> }
> >>>
> >>> +static inline bool is_buddy_busy(int cpu)
> >>> +{
> >>> + struct rq *rq = cpu_rq(cpu);
> >>> +
> >>> + /*
> >>> + * A busy buddy is a CPU with a high load or a small load with a
> >>> lot of
> >>> + * running tasks.
> >>> + */
> >>> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
> >>> + rq->avg.runnable_avg_period);
> >>
> >> I agree busy CPU is the one with high load, but many small threads may
> >> not make CPU fully busy, right ? Should we just stick to the load
> >> parameter alone here ?
> >
> > IMO, the busy state of a CPU isn't only the load but also how many
> > threads are waiting for running on it. This formula tries to take into
> > account both inputs. If you have dozen of small tasks on a CPU, the
> > latency can be large even if the tasks are small.
> >
> Sure. Your point is to avoid throttling and probably use race for
> idle.
>
> >>
> >>
> >>> +}
> >>> +
> >>> +static inline bool is_light_task(struct task_struct *p)
> >>> +{
> >>> + /* A light task runs less than 25% in average */
> >>> + return ((p->se.avg.usage_avg_sum << 2) <
> >>> p->se.avg.runnable_avg_period);
> >>> +}
> >>
> >> Since the whole packing logic relies on the light threads only, the
> >> overall effectiveness is not significant. Infact with multiple tries on
> >> Dual core system, I didn't see any major improvement in power. I think
> >> we need to be more aggressive here. From the cover letter, I noticed
> >> that, you were concerned about any performance drop due to packing and
> >> may be that is the reason you chose the conservative threshold. But the
> >> fact is, if we want to save meaningful power, there will be slight
> >> performance drop which is expected.
> >
> > I think that everybody agrees that packing small tasks will save power
> > whereas it seems to be not so obvious for heavy task. But I may have
> > set the threshold a bit too low
> >
> I agree on packing saves power part for sure.
>
I'm not fully convinced that packing always saves power. For systems
with multiple cpu clusters where each cluster is a power domain and the
cpus have no individual power saving states it would probably be more
power efficient to spread the tasks and hope for more opportunities for
hitting cluster shutdown. If all tasks are packed on one cpu it will
keep the whole cluster up, while the remaining cpus are idling without
possibility for entering efficient power states.
Regards,
Morten
> > Up to which load, you would like to pack on 1 core of your dual core system ?
> > Can you provide more details of your load ? Have you got a trace that
> > you can share ?
> >
> More than how much load to pack, I was more looking from the power
> savings delta we can achieve by doing it. Some of the usecases like
> osidle, mp3, gallary are already very low power and that might be
> the reason I didn't notice major mA delta. Though the perf
> traces did show some filtering even at 25 % load. i tried upto
> 50 % threshold to see the effectiveness and there was more
> improvement and hence the suggestion about aggressiveness.
>
> May be you can try some of these use-cases on your setup instead of
> synthetic workload and see the results.
>
> Regards
> Santosh
>
>
>
>
> _______________________________________________
> linaro-dev mailing list
> [email protected]
> http://lists.linaro.org/mailman/listinfo/linaro-dev
>
Hi Vincent,
I have experienced suboptimal buddy selection on a dual cluster setup
(ARM TC2) if SD_SHARE_POWERLINE is enabled at MC level and disabled at
CPU level. This seems to be the correct flag settings for a system with
only cluster level power gating.
To me it looks like update_packing_domain() is not doing the right
thing. See inline comments below.
On Sun, Oct 07, 2012 at 08:43:55AM +0100, Vincent Guittot wrote:
> During sched_domain creation, we define a pack buddy CPU if available.
>
> On a system that share the powerline at all level, the buddy is set to -1
>
> On a dual clusters / dual cores system which can powergate each core and
> cluster independantly, the buddy configuration will be :
> | CPU0 | CPU1 | CPU2 | CPU3 |
> -----------------------------------
> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
>
> Small tasks tend to slip out of the periodic load balance.
> The best place to choose to migrate them is at their wake up.
>
> Signed-off-by: Vincent Guittot <[email protected]>
> ---
> kernel/sched/core.c | 1 +
> kernel/sched/fair.c | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++
> kernel/sched/sched.h | 1 +
> 3 files changed, 111 insertions(+)
>
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index dab7908..70cadbe 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
> rcu_assign_pointer(rq->sd, sd);
> destroy_sched_domains(tmp, cpu);
>
> + update_packing_domain(cpu);
> update_top_cache_domain(cpu);
> }
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 4f4a4f6..8c9d3ed 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
> update_sysctl();
> }
>
> +
> +/*
> + * Save the id of the optimal CPU that should be used to pack small tasks
> + * The value -1 is used when no buddy has been found
> + */
> +DEFINE_PER_CPU(int, sd_pack_buddy);
> +
> +/* Look for the best buddy CPU that can be used to pack small tasks
> + * We make the assumption that it doesn't wort to pack on CPU that share the
> + * same powerline. We looks for the 1st sched_domain without the
> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the lowest
> + * power per core based on the assumption that their power efficiency is
> + * better */
> +void update_packing_domain(int cpu)
> +{
> + struct sched_domain *sd;
> + int id = -1;
> +
> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
> + if (!sd)
> + sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
> + else
> + sd = sd->parent;
sd is the highest level where SD_SHARE_POWERLINE is enabled so the sched
groups of the parent level would represent the power domains. If get it
right, we want to pack inside the cluster first and only let first cpu
of the cluster do packing on another cluster. So all cpus - except the
first one - in the current sched domain should find its buddy within the
domain and only the first one should go to the parent sched domain to
find its buddy.
I propose the following fix:
- sd = sd->parent;
+ if (cpumask_first(sched_domain_span(sd)) == cpu
+ || !sd->parent)
+ sd = sd->parent;
> +
> + while (sd) {
> + struct sched_group *sg = sd->groups;
> + struct sched_group *pack = sg;
> + struct sched_group *tmp = sg->next;
> +
> + /* 1st CPU of the sched domain is a good candidate */
> + if (id == -1)
> + id = cpumask_first(sched_domain_span(sd));
There is no guarantee that id is in the sched group pointed to by
sd->groups, which is implicitly assumed later in the search loop. We
need to find the sched group that contains id and point sg to that
instead. I haven't found an elegant way to find that group, but the fix
below should at least give the right result.
+ /* Find sched group of candidate */
+ tmp = sd->groups;
+ do {
+ if (cpumask_test_cpu(id, sched_group_cpus(tmp)))
+ {
+ sg = tmp;
+ break;
+ }
+ } while (tmp = tmp->next, tmp != sd->groups);
+
+ pack = sg;
+ tmp = sg->next;
Regards,
Morten
> +
> + /* loop the sched groups to find the best one */
> + while (tmp != sg) {
> + if (tmp->sgp->power * sg->group_weight <
> + sg->sgp->power * tmp->group_weight)
> + pack = tmp;
> + tmp = tmp->next;
> + }
> +
> + /* we have found a better group */
> + if (pack != sg)
> + id = cpumask_first(sched_group_cpus(pack));
> +
> + /* Look for another CPU than itself */
> + if ((id != cpu)
> + || ((sd->parent) && !(sd->parent->flags && SD_LOAD_BALANCE)))
> + break;
> +
> + sd = sd->parent;
> + }
> +
> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
> + per_cpu(sd_pack_buddy, cpu) = id;
> +}
> +
> #if BITS_PER_LONG == 32
> # define WMULT_CONST (~0UL)
> #else
> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct *p, int target)
> return target;
> }
>
> +static inline bool is_buddy_busy(int cpu)
> +{
> + struct rq *rq = cpu_rq(cpu);
> +
> + /*
> + * A busy buddy is a CPU with a high load or a small load with a lot of
> + * running tasks.
> + */
> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
> + rq->avg.runnable_avg_period);
> +}
> +
> +static inline bool is_light_task(struct task_struct *p)
> +{
> + /* A light task runs less than 25% in average */
> + return ((p->se.avg.usage_avg_sum << 2) < p->se.avg.runnable_avg_period);
> +}
> +
> +static int check_pack_buddy(int cpu, struct task_struct *p)
> +{
> + int buddy = per_cpu(sd_pack_buddy, cpu);
> +
> + /* No pack buddy for this CPU */
> + if (buddy == -1)
> + return false;
> +
> + /*
> + * If a task is waiting for running on the CPU which is its own buddy,
> + * let the default behavior to look for a better CPU if available
> + * The threshold has been set to 37.5%
> + */
> + if ((buddy == cpu)
> + && ((p->se.avg.usage_avg_sum << 3) < (p->se.avg.runnable_avg_sum * 5)))
> + return false;
> +
> + /* buddy is not an allowed CPU */
> + if (!cpumask_test_cpu(buddy, tsk_cpus_allowed(p)))
> + return false;
> +
> + /*
> + * If the task is a small one and the buddy is not overloaded,
> + * we use buddy cpu
> + */
> + if (!is_light_task(p) || is_buddy_busy(buddy))
> + return false;
> +
> + return true;
> +}
> +
> /*
> * sched_balance_self: balance the current task (running on cpu) in domains
> * that have the 'flag' flag set. In practice, this is SD_BALANCE_FORK and
> @@ -3098,6 +3204,9 @@ select_task_rq_fair(struct task_struct *p, int sd_flag, int wake_flags)
> if (p->nr_cpus_allowed == 1)
> return prev_cpu;
>
> + if (check_pack_buddy(cpu, p))
> + return per_cpu(sd_pack_buddy, cpu);
> +
> if (sd_flag & SD_BALANCE_WAKE) {
> if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p)))
> want_affine = 1;
> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
> index a95d5c1..086d8bf 100644
> --- a/kernel/sched/sched.h
> +++ b/kernel/sched/sched.h
> @@ -875,6 +875,7 @@ static inline void idle_balance(int cpu, struct rq *rq)
>
> extern void sysrq_sched_debug_show(void);
> extern void sched_init_granularity(void);
> +extern void update_packing_domain(int cpu);
> extern void update_max_interval(void);
> extern void update_group_power(struct sched_domain *sd, int cpu);
> extern int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu);
> --
> 1.7.9.5
>
>
> _______________________________________________
> linaro-dev mailing list
> [email protected]
> http://lists.linaro.org/mailman/listinfo/linaro-dev
>
On 2 November 2012 12:00, Santosh Shilimkar <[email protected]> wrote:
> On Monday 29 October 2012 06:58 PM, Vincent Guittot wrote:
>>
>> On 24 October 2012 17:21, Santosh Shilimkar <[email protected]>
>> wrote:
>>>
>>> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>>>
>>>>
>>>> The ARM platforms take advantage of packing small tasks on few cores.
>>>> This is true even when the cores of a cluster can't be powergated
>>>> independently.
>>>>
>>>>
>>>> Signed-off-by: Vincent Guittot <[email protected]>
>>>> ---
>>>> arch/arm/kernel/topology.c | 5 +++++
>>>> 1 file changed, 5 insertions(+)
>>>>
>>>> diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c
>>>> index 26c12c6..00511d0 100644
>>>> --- a/arch/arm/kernel/topology.c
>>>> +++ b/arch/arm/kernel/topology.c
>>>> @@ -226,6 +226,11 @@ static inline void update_cpu_power(unsigned int
>>>> cpuid, unsigned int mpidr) {}
>>>> */
>>>> struct cputopo_arm cpu_topology[NR_CPUS];
>>>>
>>>> +int arch_sd_share_power_line(void)
>>>> +{
>>>> + return 0*SD_SHARE_POWERLINE;
>>>> +}
>>>
>>>
>>>
>>> Making this selection of policy based on sched domain will better. Just
>>> gives the flexibility to choose a separate scheme for big and little
>>> systems which will be very convenient.
>>
>>
>> I agree that it would be more flexible to be able to set it for each level
>>
> Will you be addressing that in next version then ?
Hi Santosh,
yes, I will try to address this point for the next version.
Vincent
>
> Regards
> santosh
>
On 2 November 2012 11:53, Santosh Shilimkar <[email protected]> wrote:
> On Monday 29 October 2012 06:42 PM, Vincent Guittot wrote:
>>
>> On 24 October 2012 17:20, Santosh Shilimkar <[email protected]>
>> wrote:
>>>
>>> Vincent,
>>>
>>> Few comments/questions.
>>>
>>>
>>> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>>>>
>>>>
>>>> During sched_domain creation, we define a pack buddy CPU if available.
>>>>
>>>> On a system that share the powerline at all level, the buddy is set to
>>>> -1
>>>>
>>>> On a dual clusters / dual cores system which can powergate each core and
>>>> cluster independantly, the buddy configuration will be :
>>>> | CPU0 | CPU1 | CPU2 | CPU3 |
>>>> -----------------------------------
>>>> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
>>>
>>>
>>> ^
>>> Is that a typo ? Should it be CPU2 instead of
>>> CPU0 ?
>>
>>
>> No it's not a typo.
>> The system packs at each scheduling level. It starts to pack in
>> cluster because each core can power gate independently so CPU1 tries
>> to pack its tasks in CPU0 and CPU3 in CPU2. Then, it packs at CPU
>> level so CPU2 tries to pack in the cluster of CPU0 and CPU0 packs in
>> itself
>>
> I get it. Though in above example a task may migrate from say
> CPU3->CPU2->CPU0 as part of packing. I was just thinking whether
> moving such task from say CPU3 to CPU0 might be best instead.
We pack in the cluster then at CPU level. Tasks could sometimes
migrate directly to CPU0 but we would miss the case where CPU0 is busy
but CPU2 is not
Vincent
>
>
>>>
>>>> Small tasks tend to slip out of the periodic load balance.
>>>> The best place to choose to migrate them is at their wake up.
>>>>
>>> I have tried this series since I was looking at some of these packing
>>> bits. On Mobile workloads like OSIdle with Screen ON, MP3, gallary,
>>> I did see some additional filtering of threads with this series
>>> but its not making much difference in power. More on this below.
>>
>>
>> Can I ask you which configuration you have used ? how many cores and
>> cluster ? Can they be power gated independently ?
>>
> I have been trying with couple of setups. Dual Core ARM machine and
> Quad core X86 box with single package thought most of the mobile
> workload analysis I was doing on ARM machine. On both setups
> CPUs can be gated independently.
>
>
>>>
>>>
>>>> Signed-off-by: Vincent Guittot <[email protected]>
>>>> ---
>>>> kernel/sched/core.c | 1 +
>>>> kernel/sched/fair.c | 109
>>>> ++++++++++++++++++++++++++++++++++++++++++++++++++
>>>> kernel/sched/sched.h | 1 +
>>>> 3 files changed, 111 insertions(+)
>>>>
>>>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>>>> index dab7908..70cadbe 100644
>>>> --- a/kernel/sched/core.c
>>>> +++ b/kernel/sched/core.c
>>>> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct
>>>> root_domain *rd, int cpu)
>>>> rcu_assign_pointer(rq->sd, sd);
>>>> destroy_sched_domains(tmp, cpu);
>>>>
>>>> + update_packing_domain(cpu);
>>>> update_top_cache_domain(cpu);
>>>> }
>>>>
>>>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>>>> index 4f4a4f6..8c9d3ed 100644
>>>> --- a/kernel/sched/fair.c
>>>> +++ b/kernel/sched/fair.c
>>>> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
>>>> update_sysctl();
>>>> }
>>>>
>>>> +
>>>> +/*
>>>> + * Save the id of the optimal CPU that should be used to pack small
>>>> tasks
>>>> + * The value -1 is used when no buddy has been found
>>>> + */
>>>> +DEFINE_PER_CPU(int, sd_pack_buddy);
>>>> +
>>>> +/* Look for the best buddy CPU that can be used to pack small tasks
>>>> + * We make the assumption that it doesn't wort to pack on CPU that
>>>> share
>>>> the
>>>
>>>
>>> s/wort/worth
>>
>>
>> yes
>>
>>>
>>>> + * same powerline. We looks for the 1st sched_domain without the
>>>> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the
>>>> lowest
>>>> + * power per core based on the assumption that their power efficiency
>>>> is
>>>> + * better */
>>>
>>>
>>> Commenting style..
>>> /*
>>> *
>>> */
>>>
>>
>> yes
>>
>>> Can you please expand the why the assumption is right ?
>>> "it doesn't wort to pack on CPU that share the same powerline"
>>
>>
>> By "share the same power-line", I mean that the CPUs can't power off
>> independently. So if some CPUs can't power off independently, it's
>> worth to try to use most of them to race to idle.
>>
> In that case I suggest we use different word here. Power line can be
> treated as voltage line, power domain.
> May be SD_SHARE_CPU_POWERDOMAIN ?
>
>
>>>
>>> Think about a scenario where you have quad core, ducal cluster system
>>>
>>> |Cluster1| |cluster 2|
>>> | CPU0 | CPU1 | CPU2 | CPU3 | | CPU0 | CPU1 | CPU2 | CPU3 |
>>>
>>>
>>> Both clusters run from same voltage rail and have same PLL
>>> clocking them. But the cluster have their own power domain
>>> and all CPU's can power gate them-self to low power states.
>>> Clusters also have their own level2 caches.
>>>
>>> In this case, you will still save power if you try to pack
>>> load on one cluster. No ?
>>
>>
>> yes, I need to update the description of SD_SHARE_POWERLINE because
>> I'm afraid I was not clear enough. SD_SHARE_POWERLINE includes the
>> power gating capacity of each core. For your example above, the
>> SD_SHARE_POWERLINE shoud be cleared at both MC and CPU level.
>>
> Thanks for clarification.
>
>
>>>
>>>
>>>> +void update_packing_domain(int cpu)
>>>> +{
>>>> + struct sched_domain *sd;
>>>> + int id = -1;
>>>> +
>>>> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
>>>> + if (!sd)
>>>> + sd =
>>>> rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
>>>> + else
>>>> + sd = sd->parent;
>>>> +
>>>> + while (sd) {
>>>> + struct sched_group *sg = sd->groups;
>>>> + struct sched_group *pack = sg;
>>>> + struct sched_group *tmp = sg->next;
>>>> +
>>>> + /* 1st CPU of the sched domain is a good candidate */
>>>> + if (id == -1)
>>>> + id = cpumask_first(sched_domain_span(sd));
>>>> +
>>>> + /* loop the sched groups to find the best one */
>>>> + while (tmp != sg) {
>>>> + if (tmp->sgp->power * sg->group_weight <
>>>> + sg->sgp->power *
>>>> tmp->group_weight)
>>>> + pack = tmp;
>>>> + tmp = tmp->next;
>>>> + }
>>>> +
>>>> + /* we have found a better group */
>>>> + if (pack != sg)
>>>> + id = cpumask_first(sched_group_cpus(pack));
>>>> +
>>>> + /* Look for another CPU than itself */
>>>> + if ((id != cpu)
>>>> + || ((sd->parent) && !(sd->parent->flags &&
>>>> SD_LOAD_BALANCE)))
>>>
>>>
>>> Is the condition "!(sd->parent->flags && SD_LOAD_BALANCE)" for
>>> big.LITTLE kind of system where SD_LOAD_BALANCE may not be used ?
>>
>>
>> No, packing small tasks is part of the load balance so if the
>> LOAD_BALANCE flag is cleared, we will not try to pack which is a kind
>> of load balance. There is no link with big.LITTLE
>>
> Now it make sense to me.
>
>
>>>
>>>
>>>> + break;
>>>> +
>>>> + sd = sd->parent;
>>>> + }
>>>> +
>>>> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
>>>> + per_cpu(sd_pack_buddy, cpu) = id;
>>>> +}
>>>> +
>>>> #if BITS_PER_LONG == 32
>>>> # define WMULT_CONST (~0UL)
>>>> #else
>>>> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct
>>>> *p, int target)
>>>> return target;
>>>> }
>>>>
>>>> +static inline bool is_buddy_busy(int cpu)
>>>> +{
>>>> + struct rq *rq = cpu_rq(cpu);
>>>> +
>>>> + /*
>>>> + * A busy buddy is a CPU with a high load or a small load with a
>>>> lot of
>>>> + * running tasks.
>>>> + */
>>>> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
>>>> + rq->avg.runnable_avg_period);
>>>
>>>
>>> I agree busy CPU is the one with high load, but many small threads may
>>> not make CPU fully busy, right ? Should we just stick to the load
>>> parameter alone here ?
>>
>>
>> IMO, the busy state of a CPU isn't only the load but also how many
>> threads are waiting for running on it. This formula tries to take into
>> account both inputs. If you have dozen of small tasks on a CPU, the
>> latency can be large even if the tasks are small.
>>
> Sure. Your point is to avoid throttling and probably use race for
> idle.
>
>
>>>
>>>
>>>> +}
>>>> +
>>>> +static inline bool is_light_task(struct task_struct *p)
>>>> +{
>>>> + /* A light task runs less than 25% in average */
>>>> + return ((p->se.avg.usage_avg_sum << 2) <
>>>> p->se.avg.runnable_avg_period);
>>>> +}
>>>
>>>
>>> Since the whole packing logic relies on the light threads only, the
>>> overall effectiveness is not significant. Infact with multiple tries on
>>> Dual core system, I didn't see any major improvement in power. I think
>>> we need to be more aggressive here. From the cover letter, I noticed
>>> that, you were concerned about any performance drop due to packing and
>>> may be that is the reason you chose the conservative threshold. But the
>>> fact is, if we want to save meaningful power, there will be slight
>>> performance drop which is expected.
>>
>>
>> I think that everybody agrees that packing small tasks will save power
>> whereas it seems to be not so obvious for heavy task. But I may have
>> set the threshold a bit too low
>>
> I agree on packing saves power part for sure.
>
>
>> Up to which load, you would like to pack on 1 core of your dual core
>> system ?
>> Can you provide more details of your load ? Have you got a trace that
>> you can share ?
>>
> More than how much load to pack, I was more looking from the power
> savings delta we can achieve by doing it. Some of the usecases like
> osidle, mp3, gallary are already very low power and that might be
> the reason I didn't notice major mA delta. Though the perf
> traces did show some filtering even at 25 % load. i tried upto
> 50 % threshold to see the effectiveness and there was more
> improvement and hence the suggestion about aggressiveness.
>
> May be you can try some of these use-cases on your setup instead of
> synthetic workload and see the results.
Yes, I will do that. The main advantage of synthetic workload is the
reproducibility and the interdependency against the framework that is
used. But I'm going to use more realistic use-cases.
Regards,
Vincent
>
> Regards
> Santosh
>
>
>
On 9 November 2012 17:46, Morten Rasmussen <[email protected]> wrote:
> On Fri, Nov 02, 2012 at 10:53:47AM +0000, Santosh Shilimkar wrote:
>> On Monday 29 October 2012 06:42 PM, Vincent Guittot wrote:
>> > On 24 October 2012 17:20, Santosh Shilimkar <[email protected]> wrote:
>> >> Vincent,
>> >>
>> >> Few comments/questions.
>> >>
>> >>
>> >> On Sunday 07 October 2012 01:13 PM, Vincent Guittot wrote:
>> >>>
>> >>> During sched_domain creation, we define a pack buddy CPU if available.
>> >>>
>> >>> On a system that share the powerline at all level, the buddy is set to -1
>> >>>
>> >>> On a dual clusters / dual cores system which can powergate each core and
>> >>> cluster independantly, the buddy configuration will be :
>> >>> | CPU0 | CPU1 | CPU2 | CPU3 |
>> >>> -----------------------------------
>> >>> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
>> >>
>> >> ^
>> >> Is that a typo ? Should it be CPU2 instead of
>> >> CPU0 ?
>> >
>> > No it's not a typo.
>> > The system packs at each scheduling level. It starts to pack in
>> > cluster because each core can power gate independently so CPU1 tries
>> > to pack its tasks in CPU0 and CPU3 in CPU2. Then, it packs at CPU
>> > level so CPU2 tries to pack in the cluster of CPU0 and CPU0 packs in
>> > itself
>> >
>> I get it. Though in above example a task may migrate from say
>> CPU3->CPU2->CPU0 as part of packing. I was just thinking whether
>> moving such task from say CPU3 to CPU0 might be best instead.
>
> To me it seems suboptimal to pack the task twice, but the alternative is
> not good either. If you try to move the task directly to CPU0 you may
> miss packing opportunities if CPU0 is already busy, while CPU2 might
> have enough capacity to take it. It would probably be better to check
> the business of CPU0 and then back off and try CPU2 if CP0 is busy. This
> would require a buddy list for each CPU rather just a single buddy and
> thus might become expensive.
>
>>
>> >>
>> >>> Small tasks tend to slip out of the periodic load balance.
>> >>> The best place to choose to migrate them is at their wake up.
>> >>>
>> >> I have tried this series since I was looking at some of these packing
>> >> bits. On Mobile workloads like OSIdle with Screen ON, MP3, gallary,
>> >> I did see some additional filtering of threads with this series
>> >> but its not making much difference in power. More on this below.
>> >
>> > Can I ask you which configuration you have used ? how many cores and
>> > cluster ? Can they be power gated independently ?
>> >
>> I have been trying with couple of setups. Dual Core ARM machine and
>> Quad core X86 box with single package thought most of the mobile
>> workload analysis I was doing on ARM machine. On both setups
>> CPUs can be gated independently.
>>
>> >>
>> >>
>> >>> Signed-off-by: Vincent Guittot <[email protected]>
>> >>> ---
>> >>> kernel/sched/core.c | 1 +
>> >>> kernel/sched/fair.c | 109
>> >>> ++++++++++++++++++++++++++++++++++++++++++++++++++
>> >>> kernel/sched/sched.h | 1 +
>> >>> 3 files changed, 111 insertions(+)
>> >>>
>> >>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> >>> index dab7908..70cadbe 100644
>> >>> --- a/kernel/sched/core.c
>> >>> +++ b/kernel/sched/core.c
>> >>> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct
>> >>> root_domain *rd, int cpu)
>> >>> rcu_assign_pointer(rq->sd, sd);
>> >>> destroy_sched_domains(tmp, cpu);
>> >>>
>> >>> + update_packing_domain(cpu);
>> >>> update_top_cache_domain(cpu);
>> >>> }
>> >>>
>> >>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> >>> index 4f4a4f6..8c9d3ed 100644
>> >>> --- a/kernel/sched/fair.c
>> >>> +++ b/kernel/sched/fair.c
>> >>> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
>> >>> update_sysctl();
>> >>> }
>> >>>
>> >>> +
>> >>> +/*
>> >>> + * Save the id of the optimal CPU that should be used to pack small tasks
>> >>> + * The value -1 is used when no buddy has been found
>> >>> + */
>> >>> +DEFINE_PER_CPU(int, sd_pack_buddy);
>> >>> +
>> >>> +/* Look for the best buddy CPU that can be used to pack small tasks
>> >>> + * We make the assumption that it doesn't wort to pack on CPU that share
>> >>> the
>> >>
>> >> s/wort/worth
>> >
>> > yes
>> >
>> >>
>> >>> + * same powerline. We looks for the 1st sched_domain without the
>> >>> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the
>> >>> lowest
>> >>> + * power per core based on the assumption that their power efficiency is
>> >>> + * better */
>> >>
>> >> Commenting style..
>> >> /*
>> >> *
>> >> */
>> >>
>> >
>> > yes
>> >
>> >> Can you please expand the why the assumption is right ?
>> >> "it doesn't wort to pack on CPU that share the same powerline"
>> >
>> > By "share the same power-line", I mean that the CPUs can't power off
>> > independently. So if some CPUs can't power off independently, it's
>> > worth to try to use most of them to race to idle.
>> >
>> In that case I suggest we use different word here. Power line can be
>> treated as voltage line, power domain.
>> May be SD_SHARE_CPU_POWERDOMAIN ?
>>
>
> How about just SD_SHARE_POWERDOMAIN ?
It looks better than SD_SHARE_POWERLINE. I will replace the name
>
>> >>
>> >> Think about a scenario where you have quad core, ducal cluster system
>> >>
>> >> |Cluster1| |cluster 2|
>> >> | CPU0 | CPU1 | CPU2 | CPU3 | | CPU0 | CPU1 | CPU2 | CPU3 |
>> >>
>> >>
>> >> Both clusters run from same voltage rail and have same PLL
>> >> clocking them. But the cluster have their own power domain
>> >> and all CPU's can power gate them-self to low power states.
>> >> Clusters also have their own level2 caches.
>> >>
>> >> In this case, you will still save power if you try to pack
>> >> load on one cluster. No ?
>> >
>> > yes, I need to update the description of SD_SHARE_POWERLINE because
>> > I'm afraid I was not clear enough. SD_SHARE_POWERLINE includes the
>> > power gating capacity of each core. For your example above, the
>> > SD_SHARE_POWERLINE shoud be cleared at both MC and CPU level.
>> >
>> Thanks for clarification.
>>
>> >>
>> >>
>> >>> +void update_packing_domain(int cpu)
>> >>> +{
>> >>> + struct sched_domain *sd;
>> >>> + int id = -1;
>> >>> +
>> >>> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
>> >>> + if (!sd)
>> >>> + sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
>> >>> + else
>> >>> + sd = sd->parent;
>> >>> +
>> >>> + while (sd) {
>> >>> + struct sched_group *sg = sd->groups;
>> >>> + struct sched_group *pack = sg;
>> >>> + struct sched_group *tmp = sg->next;
>> >>> +
>> >>> + /* 1st CPU of the sched domain is a good candidate */
>> >>> + if (id == -1)
>> >>> + id = cpumask_first(sched_domain_span(sd));
>> >>> +
>> >>> + /* loop the sched groups to find the best one */
>> >>> + while (tmp != sg) {
>> >>> + if (tmp->sgp->power * sg->group_weight <
>> >>> + sg->sgp->power *
>> >>> tmp->group_weight)
>> >>> + pack = tmp;
>> >>> + tmp = tmp->next;
>> >>> + }
>> >>> +
>> >>> + /* we have found a better group */
>> >>> + if (pack != sg)
>> >>> + id = cpumask_first(sched_group_cpus(pack));
>> >>> +
>> >>> + /* Look for another CPU than itself */
>> >>> + if ((id != cpu)
>> >>> + || ((sd->parent) && !(sd->parent->flags &&
>> >>> SD_LOAD_BALANCE)))
>> >>
>> >> Is the condition "!(sd->parent->flags && SD_LOAD_BALANCE)" for
>> >> big.LITTLE kind of system where SD_LOAD_BALANCE may not be used ?
>> >
>> > No, packing small tasks is part of the load balance so if the
>> > LOAD_BALANCE flag is cleared, we will not try to pack which is a kind
>> > of load balance. There is no link with big.LITTLE
>> >
>> Now it make sense to me.
>>
>> >>
>> >>
>> >>> + break;
>> >>> +
>> >>> + sd = sd->parent;
>> >>> + }
>> >>> +
>> >>> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
>> >>> + per_cpu(sd_pack_buddy, cpu) = id;
>> >>> +}
>> >>> +
>> >>> #if BITS_PER_LONG == 32
>> >>> # define WMULT_CONST (~0UL)
>> >>> #else
>> >>> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct
>> >>> *p, int target)
>> >>> return target;
>> >>> }
>> >>>
>> >>> +static inline bool is_buddy_busy(int cpu)
>> >>> +{
>> >>> + struct rq *rq = cpu_rq(cpu);
>> >>> +
>> >>> + /*
>> >>> + * A busy buddy is a CPU with a high load or a small load with a
>> >>> lot of
>> >>> + * running tasks.
>> >>> + */
>> >>> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
>> >>> + rq->avg.runnable_avg_period);
>> >>
>> >> I agree busy CPU is the one with high load, but many small threads may
>> >> not make CPU fully busy, right ? Should we just stick to the load
>> >> parameter alone here ?
>> >
>> > IMO, the busy state of a CPU isn't only the load but also how many
>> > threads are waiting for running on it. This formula tries to take into
>> > account both inputs. If you have dozen of small tasks on a CPU, the
>> > latency can be large even if the tasks are small.
>> >
>> Sure. Your point is to avoid throttling and probably use race for
>> idle.
>>
>> >>
>> >>
>> >>> +}
>> >>> +
>> >>> +static inline bool is_light_task(struct task_struct *p)
>> >>> +{
>> >>> + /* A light task runs less than 25% in average */
>> >>> + return ((p->se.avg.usage_avg_sum << 2) <
>> >>> p->se.avg.runnable_avg_period);
>> >>> +}
>> >>
>> >> Since the whole packing logic relies on the light threads only, the
>> >> overall effectiveness is not significant. Infact with multiple tries on
>> >> Dual core system, I didn't see any major improvement in power. I think
>> >> we need to be more aggressive here. From the cover letter, I noticed
>> >> that, you were concerned about any performance drop due to packing and
>> >> may be that is the reason you chose the conservative threshold. But the
>> >> fact is, if we want to save meaningful power, there will be slight
>> >> performance drop which is expected.
>> >
>> > I think that everybody agrees that packing small tasks will save power
>> > whereas it seems to be not so obvious for heavy task. But I may have
>> > set the threshold a bit too low
>> >
>> I agree on packing saves power part for sure.
>>
>
> I'm not fully convinced that packing always saves power. For systems
> with multiple cpu clusters where each cluster is a power domain and the
> cpus have no individual power saving states it would probably be more
> power efficient to spread the tasks and hope for more opportunities for
> hitting cluster shutdown. If all tasks are packed on one cpu it will
> keep the whole cluster up, while the remaining cpus are idling without
> possibility for entering efficient power states.
>
> Regards,
> Morten
>
>> > Up to which load, you would like to pack on 1 core of your dual core system ?
>> > Can you provide more details of your load ? Have you got a trace that
>> > you can share ?
>> >
>> More than how much load to pack, I was more looking from the power
>> savings delta we can achieve by doing it. Some of the usecases like
>> osidle, mp3, gallary are already very low power and that might be
>> the reason I didn't notice major mA delta. Though the perf
>> traces did show some filtering even at 25 % load. i tried upto
>> 50 % threshold to see the effectiveness and there was more
>> improvement and hence the suggestion about aggressiveness.
>>
>> May be you can try some of these use-cases on your setup instead of
>> synthetic workload and see the results.
>>
>> Regards
>> Santosh
>>
>>
>>
>>
>> _______________________________________________
>> linaro-dev mailing list
>> [email protected]
>> http://lists.linaro.org/mailman/listinfo/linaro-dev
>>
>
On 9 November 2012 18:13, Morten Rasmussen <[email protected]> wrote:
> Hi Vincent,
>
> I have experienced suboptimal buddy selection on a dual cluster setup
> (ARM TC2) if SD_SHARE_POWERLINE is enabled at MC level and disabled at
> CPU level. This seems to be the correct flag settings for a system with
> only cluster level power gating.
>
> To me it looks like update_packing_domain() is not doing the right
> thing. See inline comments below.
Hi Morten,
Thanks for testing the patches.
It seems that I have too optimized the loop and remove some use cases.
>
> On Sun, Oct 07, 2012 at 08:43:55AM +0100, Vincent Guittot wrote:
>> During sched_domain creation, we define a pack buddy CPU if available.
>>
>> On a system that share the powerline at all level, the buddy is set to -1
>>
>> On a dual clusters / dual cores system which can powergate each core and
>> cluster independantly, the buddy configuration will be :
>> | CPU0 | CPU1 | CPU2 | CPU3 |
>> -----------------------------------
>> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
>>
>> Small tasks tend to slip out of the periodic load balance.
>> The best place to choose to migrate them is at their wake up.
>>
>> Signed-off-by: Vincent Guittot <[email protected]>
>> ---
>> kernel/sched/core.c | 1 +
>> kernel/sched/fair.c | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++
>> kernel/sched/sched.h | 1 +
>> 3 files changed, 111 insertions(+)
>>
>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> index dab7908..70cadbe 100644
>> --- a/kernel/sched/core.c
>> +++ b/kernel/sched/core.c
>> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
>> rcu_assign_pointer(rq->sd, sd);
>> destroy_sched_domains(tmp, cpu);
>>
>> + update_packing_domain(cpu);
>> update_top_cache_domain(cpu);
>> }
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index 4f4a4f6..8c9d3ed 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
>> update_sysctl();
>> }
>>
>> +
>> +/*
>> + * Save the id of the optimal CPU that should be used to pack small tasks
>> + * The value -1 is used when no buddy has been found
>> + */
>> +DEFINE_PER_CPU(int, sd_pack_buddy);
>> +
>> +/* Look for the best buddy CPU that can be used to pack small tasks
>> + * We make the assumption that it doesn't wort to pack on CPU that share the
>> + * same powerline. We looks for the 1st sched_domain without the
>> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the lowest
>> + * power per core based on the assumption that their power efficiency is
>> + * better */
>> +void update_packing_domain(int cpu)
>> +{
>> + struct sched_domain *sd;
>> + int id = -1;
>> +
>> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
>> + if (!sd)
>> + sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
>> + else
>> + sd = sd->parent;
> sd is the highest level where SD_SHARE_POWERLINE is enabled so the sched
> groups of the parent level would represent the power domains. If get it
> right, we want to pack inside the cluster first and only let first cpu
You probably wanted to use sched_group instead of cluster because
cluster is only a special use case, didn't you ?
> of the cluster do packing on another cluster. So all cpus - except the
> first one - in the current sched domain should find its buddy within the
> domain and only the first one should go to the parent sched domain to
> find its buddy.
We don't want to pack in the current sched_domain because it shares
power domain. We want to pack at the parent level
>
> I propose the following fix:
>
> - sd = sd->parent;
> + if (cpumask_first(sched_domain_span(sd)) == cpu
> + || !sd->parent)
> + sd = sd->parent;
We always look for the buddy in the parent level whatever the cpu
position in the mask is.
>
>
>> +
>> + while (sd) {
>> + struct sched_group *sg = sd->groups;
>> + struct sched_group *pack = sg;
>> + struct sched_group *tmp = sg->next;
>> +
>> + /* 1st CPU of the sched domain is a good candidate */
>> + if (id == -1)
>> + id = cpumask_first(sched_domain_span(sd));
>
> There is no guarantee that id is in the sched group pointed to by
> sd->groups, which is implicitly assumed later in the search loop. We
> need to find the sched group that contains id and point sg to that
> instead. I haven't found an elegant way to find that group, but the fix
> below should at least give the right result.
>
> + /* Find sched group of candidate */
> + tmp = sd->groups;
> + do {
> + if (cpumask_test_cpu(id, sched_group_cpus(tmp)))
> + {
> + sg = tmp;
> + break;
> + }
> + } while (tmp = tmp->next, tmp != sd->groups);
> +
> + pack = sg;
> + tmp = sg->next;
I have a new loop which solves your issue and others. I will use it
for the next version
+ while (sd) {
+ struct sched_group *sg = sd->groups;
+ struct sched_group *pack = sg;
+ struct sched_group *tmp;
+
+ /* The 1st CPU of the local group is a good candidate */
+ id = cpumask_first(sched_group_cpus(pack));
+
+ /* loop the sched groups to find the best one */
+ for (tmp = sg->next; tmp != sg; tmp = tmp->next) {
+ if (tmp->sgp->power * pack->group_weight >
+ pack->sgp->power * tmp->group_weight)
+ continue;
+
+ if ((tmp->sgp->power * pack->group_weight ==
+ pack->sgp->power * tmp->group_weight)
+ && (cpumask_first(sched_group_cpus(tmp)) >= id))
+ continue;
+
+ /* we have found a better group */
+ pack = tmp;
+
+ /* Take the 1st CPU of the new group */
+ id = cpumask_first(sched_group_cpus(pack));
+ }
+
+ /* Look for another CPU than itself */
+ if ((id != cpu)
+ || ((sd->parent) && !(sd->parent->flags && SD_LOAD_BALANCE)))
+ break;
+
+ sd = sd->parent;
+ }
Regards,
Vincent
>
> Regards,
> Morten
>
>> +
>> + /* loop the sched groups to find the best one */
>> + while (tmp != sg) {
>> + if (tmp->sgp->power * sg->group_weight <
>> + sg->sgp->power * tmp->group_weight)
>> + pack = tmp;
>> + tmp = tmp->next;
>> + }
>> +
>> + /* we have found a better group */
>> + if (pack != sg)
>> + id = cpumask_first(sched_group_cpus(pack));
>> +
>> + /* Look for another CPU than itself */
>> + if ((id != cpu)
>> + || ((sd->parent) && !(sd->parent->flags && SD_LOAD_BALANCE)))
>> + break;
>> +
>> + sd = sd->parent;
>> + }
>> +
>> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
>> + per_cpu(sd_pack_buddy, cpu) = id;
>> +}
>> +
>> #if BITS_PER_LONG == 32
>> # define WMULT_CONST (~0UL)
>> #else
>> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct *p, int target)
>> return target;
>> }
>>
>> +static inline bool is_buddy_busy(int cpu)
>> +{
>> + struct rq *rq = cpu_rq(cpu);
>> +
>> + /*
>> + * A busy buddy is a CPU with a high load or a small load with a lot of
>> + * running tasks.
>> + */
>> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
>> + rq->avg.runnable_avg_period);
>> +}
>> +
>> +static inline bool is_light_task(struct task_struct *p)
>> +{
>> + /* A light task runs less than 25% in average */
>> + return ((p->se.avg.usage_avg_sum << 2) < p->se.avg.runnable_avg_period);
>> +}
>> +
>> +static int check_pack_buddy(int cpu, struct task_struct *p)
>> +{
>> + int buddy = per_cpu(sd_pack_buddy, cpu);
>> +
>> + /* No pack buddy for this CPU */
>> + if (buddy == -1)
>> + return false;
>> +
>> + /*
>> + * If a task is waiting for running on the CPU which is its own buddy,
>> + * let the default behavior to look for a better CPU if available
>> + * The threshold has been set to 37.5%
>> + */
>> + if ((buddy == cpu)
>> + && ((p->se.avg.usage_avg_sum << 3) < (p->se.avg.runnable_avg_sum * 5)))
>> + return false;
>> +
>> + /* buddy is not an allowed CPU */
>> + if (!cpumask_test_cpu(buddy, tsk_cpus_allowed(p)))
>> + return false;
>> +
>> + /*
>> + * If the task is a small one and the buddy is not overloaded,
>> + * we use buddy cpu
>> + */
>> + if (!is_light_task(p) || is_buddy_busy(buddy))
>> + return false;
>> +
>> + return true;
>> +}
>> +
>> /*
>> * sched_balance_self: balance the current task (running on cpu) in domains
>> * that have the 'flag' flag set. In practice, this is SD_BALANCE_FORK and
>> @@ -3098,6 +3204,9 @@ select_task_rq_fair(struct task_struct *p, int sd_flag, int wake_flags)
>> if (p->nr_cpus_allowed == 1)
>> return prev_cpu;
>>
>> + if (check_pack_buddy(cpu, p))
>> + return per_cpu(sd_pack_buddy, cpu);
>> +
>> if (sd_flag & SD_BALANCE_WAKE) {
>> if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p)))
>> want_affine = 1;
>> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
>> index a95d5c1..086d8bf 100644
>> --- a/kernel/sched/sched.h
>> +++ b/kernel/sched/sched.h
>> @@ -875,6 +875,7 @@ static inline void idle_balance(int cpu, struct rq *rq)
>>
>> extern void sysrq_sched_debug_show(void);
>> extern void sched_init_granularity(void);
>> +extern void update_packing_domain(int cpu);
>> extern void update_max_interval(void);
>> extern void update_group_power(struct sched_domain *sd, int cpu);
>> extern int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu);
>> --
>> 1.7.9.5
>>
>>
>> _______________________________________________
>> linaro-dev mailing list
>> [email protected]
>> http://lists.linaro.org/mailman/listinfo/linaro-dev
>>
>
Hi Vincent,
On Mon, Nov 12, 2012 at 01:51:00PM +0000, Vincent Guittot wrote:
> On 9 November 2012 18:13, Morten Rasmussen <[email protected]> wrote:
> > Hi Vincent,
> >
> > I have experienced suboptimal buddy selection on a dual cluster setup
> > (ARM TC2) if SD_SHARE_POWERLINE is enabled at MC level and disabled at
> > CPU level. This seems to be the correct flag settings for a system with
> > only cluster level power gating.
> >
> > To me it looks like update_packing_domain() is not doing the right
> > thing. See inline comments below.
>
> Hi Morten,
>
> Thanks for testing the patches.
>
> It seems that I have too optimized the loop and remove some use cases.
>
> >
> > On Sun, Oct 07, 2012 at 08:43:55AM +0100, Vincent Guittot wrote:
> >> During sched_domain creation, we define a pack buddy CPU if available.
> >>
> >> On a system that share the powerline at all level, the buddy is set to -1
> >>
> >> On a dual clusters / dual cores system which can powergate each core and
> >> cluster independantly, the buddy configuration will be :
> >> | CPU0 | CPU1 | CPU2 | CPU3 |
> >> -----------------------------------
> >> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
> >>
> >> Small tasks tend to slip out of the periodic load balance.
> >> The best place to choose to migrate them is at their wake up.
> >>
> >> Signed-off-by: Vincent Guittot <[email protected]>
> >> ---
> >> kernel/sched/core.c | 1 +
> >> kernel/sched/fair.c | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++
> >> kernel/sched/sched.h | 1 +
> >> 3 files changed, 111 insertions(+)
> >>
> >> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> >> index dab7908..70cadbe 100644
> >> --- a/kernel/sched/core.c
> >> +++ b/kernel/sched/core.c
> >> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
> >> rcu_assign_pointer(rq->sd, sd);
> >> destroy_sched_domains(tmp, cpu);
> >>
> >> + update_packing_domain(cpu);
> >> update_top_cache_domain(cpu);
> >> }
> >>
> >> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> >> index 4f4a4f6..8c9d3ed 100644
> >> --- a/kernel/sched/fair.c
> >> +++ b/kernel/sched/fair.c
> >> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
> >> update_sysctl();
> >> }
> >>
> >> +
> >> +/*
> >> + * Save the id of the optimal CPU that should be used to pack small tasks
> >> + * The value -1 is used when no buddy has been found
> >> + */
> >> +DEFINE_PER_CPU(int, sd_pack_buddy);
> >> +
> >> +/* Look for the best buddy CPU that can be used to pack small tasks
> >> + * We make the assumption that it doesn't wort to pack on CPU that share the
> >> + * same powerline. We looks for the 1st sched_domain without the
> >> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the lowest
> >> + * power per core based on the assumption that their power efficiency is
> >> + * better */
> >> +void update_packing_domain(int cpu)
> >> +{
> >> + struct sched_domain *sd;
> >> + int id = -1;
> >> +
> >> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
> >> + if (!sd)
> >> + sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
> >> + else
> >> + sd = sd->parent;
> > sd is the highest level where SD_SHARE_POWERLINE is enabled so the sched
> > groups of the parent level would represent the power domains. If get it
> > right, we want to pack inside the cluster first and only let first cpu
>
> You probably wanted to use sched_group instead of cluster because
> cluster is only a special use case, didn't you ?
>
> > of the cluster do packing on another cluster. So all cpus - except the
> > first one - in the current sched domain should find its buddy within the
> > domain and only the first one should go to the parent sched domain to
> > find its buddy.
>
> We don't want to pack in the current sched_domain because it shares
> power domain. We want to pack at the parent level
>
Yes. I think we mean the same thing. The packing takes place at the
parent sched_domain but the sched_group that we are looking at only
contains the cpus of the level below.
> >
> > I propose the following fix:
> >
> > - sd = sd->parent;
> > + if (cpumask_first(sched_domain_span(sd)) == cpu
> > + || !sd->parent)
> > + sd = sd->parent;
>
> We always look for the buddy in the parent level whatever the cpu
> position in the mask is.
>
> >
> >
> >> +
> >> + while (sd) {
> >> + struct sched_group *sg = sd->groups;
> >> + struct sched_group *pack = sg;
> >> + struct sched_group *tmp = sg->next;
> >> +
> >> + /* 1st CPU of the sched domain is a good candidate */
> >> + if (id == -1)
> >> + id = cpumask_first(sched_domain_span(sd));
> >
> > There is no guarantee that id is in the sched group pointed to by
> > sd->groups, which is implicitly assumed later in the search loop. We
> > need to find the sched group that contains id and point sg to that
> > instead. I haven't found an elegant way to find that group, but the fix
> > below should at least give the right result.
> >
> > + /* Find sched group of candidate */
> > + tmp = sd->groups;
> > + do {
> > + if (cpumask_test_cpu(id, sched_group_cpus(tmp)))
> > + {
> > + sg = tmp;
> > + break;
> > + }
> > + } while (tmp = tmp->next, tmp != sd->groups);
> > +
> > + pack = sg;
> > + tmp = sg->next;
>
>
> I have a new loop which solves your issue and others. I will use it
> for the next version
>
> + while (sd) {
> + struct sched_group *sg = sd->groups;
> + struct sched_group *pack = sg;
> + struct sched_group *tmp;
> +
> + /* The 1st CPU of the local group is a good candidate */
> + id = cpumask_first(sched_group_cpus(pack));
You make the assumption that the first sched_group in the list always contains
the current cpu. I think that is always the case, but I haven't verified
it. Maybe a comment about this would help people to understand the code
easier.
> +
> + /* loop the sched groups to find the best one */
> + for (tmp = sg->next; tmp != sg; tmp = tmp->next) {
> + if (tmp->sgp->power * pack->group_weight >
> + pack->sgp->power * tmp->group_weight)
> + continue;
> +
> + if ((tmp->sgp->power * pack->group_weight ==
> + pack->sgp->power * tmp->group_weight)
> + && (cpumask_first(sched_group_cpus(tmp)) >= id))
> + continue;
> +
> + /* we have found a better group */
> + pack = tmp;
> +
> + /* Take the 1st CPU of the new group */
> + id = cpumask_first(sched_group_cpus(pack));
> + }
> +
Works great on my setup.
> + /* Look for another CPU than itself */
> + if ((id != cpu)
> + || ((sd->parent) && !(sd->parent->flags && SD_LOAD_BALANCE)))
> + break;
If I understand correctly the last part of this check should avoid
selecting a buddy in a sched_group that is not load balanced with the
current one. In that case, I think that this check (or a similar check)
should be done before the loop as well. As it is, the first iteration of
the loop will always search all the groups of the first domain where
SD_SHARE_POWERLINE is disabled regardless of the state of
SD_LOAD_BALANCE flag. So if they are both disabled at the same level
packing will happen across groups that are not supposed to be
load-balanced.
Regards,
Morten
> +
> + sd = sd->parent;
> + }
>
> Regards,
> Vincent
>
> >
> > Regards,
> > Morten
> >
> >> +
> >> + /* loop the sched groups to find the best one */
> >> + while (tmp != sg) {
> >> + if (tmp->sgp->power * sg->group_weight <
> >> + sg->sgp->power * tmp->group_weight)
> >> + pack = tmp;
> >> + tmp = tmp->next;
> >> + }
> >> +
> >> + /* we have found a better group */
> >> + if (pack != sg)
> >> + id = cpumask_first(sched_group_cpus(pack));
> >> +
> >> + /* Look for another CPU than itself */
> >> + if ((id != cpu)
> >> + || ((sd->parent) && !(sd->parent->flags && SD_LOAD_BALANCE)))
> >> + break;
> >> +
> >> + sd = sd->parent;
> >> + }
> >> +
> >> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
> >> + per_cpu(sd_pack_buddy, cpu) = id;
> >> +}
> >> +
> >> #if BITS_PER_LONG == 32
> >> # define WMULT_CONST (~0UL)
> >> #else
> >> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct *p, int target)
> >> return target;
> >> }
> >>
> >> +static inline bool is_buddy_busy(int cpu)
> >> +{
> >> + struct rq *rq = cpu_rq(cpu);
> >> +
> >> + /*
> >> + * A busy buddy is a CPU with a high load or a small load with a lot of
> >> + * running tasks.
> >> + */
> >> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
> >> + rq->avg.runnable_avg_period);
> >> +}
> >> +
> >> +static inline bool is_light_task(struct task_struct *p)
> >> +{
> >> + /* A light task runs less than 25% in average */
> >> + return ((p->se.avg.usage_avg_sum << 2) < p->se.avg.runnable_avg_period);
> >> +}
> >> +
> >> +static int check_pack_buddy(int cpu, struct task_struct *p)
> >> +{
> >> + int buddy = per_cpu(sd_pack_buddy, cpu);
> >> +
> >> + /* No pack buddy for this CPU */
> >> + if (buddy == -1)
> >> + return false;
> >> +
> >> + /*
> >> + * If a task is waiting for running on the CPU which is its own buddy,
> >> + * let the default behavior to look for a better CPU if available
> >> + * The threshold has been set to 37.5%
> >> + */
> >> + if ((buddy == cpu)
> >> + && ((p->se.avg.usage_avg_sum << 3) < (p->se.avg.runnable_avg_sum * 5)))
> >> + return false;
> >> +
> >> + /* buddy is not an allowed CPU */
> >> + if (!cpumask_test_cpu(buddy, tsk_cpus_allowed(p)))
> >> + return false;
> >> +
> >> + /*
> >> + * If the task is a small one and the buddy is not overloaded,
> >> + * we use buddy cpu
> >> + */
> >> + if (!is_light_task(p) || is_buddy_busy(buddy))
> >> + return false;
> >> +
> >> + return true;
> >> +}
> >> +
> >> /*
> >> * sched_balance_self: balance the current task (running on cpu) in domains
> >> * that have the 'flag' flag set. In practice, this is SD_BALANCE_FORK and
> >> @@ -3098,6 +3204,9 @@ select_task_rq_fair(struct task_struct *p, int sd_flag, int wake_flags)
> >> if (p->nr_cpus_allowed == 1)
> >> return prev_cpu;
> >>
> >> + if (check_pack_buddy(cpu, p))
> >> + return per_cpu(sd_pack_buddy, cpu);
> >> +
> >> if (sd_flag & SD_BALANCE_WAKE) {
> >> if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p)))
> >> want_affine = 1;
> >> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
> >> index a95d5c1..086d8bf 100644
> >> --- a/kernel/sched/sched.h
> >> +++ b/kernel/sched/sched.h
> >> @@ -875,6 +875,7 @@ static inline void idle_balance(int cpu, struct rq *rq)
> >>
> >> extern void sysrq_sched_debug_show(void);
> >> extern void sched_init_granularity(void);
> >> +extern void update_packing_domain(int cpu);
> >> extern void update_max_interval(void);
> >> extern void update_group_power(struct sched_domain *sd, int cpu);
> >> extern int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu);
> >> --
> >> 1.7.9.5
> >>
> >>
> >> _______________________________________________
> >> linaro-dev mailing list
> >> [email protected]
> >> http://lists.linaro.org/mailman/listinfo/linaro-dev
> >>
> >
>
On 20 November 2012 15:28, Morten Rasmussen <[email protected]> wrote:
> Hi Vincent,
>
> On Mon, Nov 12, 2012 at 01:51:00PM +0000, Vincent Guittot wrote:
>> On 9 November 2012 18:13, Morten Rasmussen <[email protected]> wrote:
>> > Hi Vincent,
>> >
>> > I have experienced suboptimal buddy selection on a dual cluster setup
>> > (ARM TC2) if SD_SHARE_POWERLINE is enabled at MC level and disabled at
>> > CPU level. This seems to be the correct flag settings for a system with
>> > only cluster level power gating.
>> >
>> > To me it looks like update_packing_domain() is not doing the right
>> > thing. See inline comments below.
>>
>> Hi Morten,
>>
>> Thanks for testing the patches.
>>
>> It seems that I have too optimized the loop and remove some use cases.
>>
>> >
>> > On Sun, Oct 07, 2012 at 08:43:55AM +0100, Vincent Guittot wrote:
>> >> During sched_domain creation, we define a pack buddy CPU if available.
>> >>
>> >> On a system that share the powerline at all level, the buddy is set to -1
>> >>
>> >> On a dual clusters / dual cores system which can powergate each core and
>> >> cluster independantly, the buddy configuration will be :
>> >> | CPU0 | CPU1 | CPU2 | CPU3 |
>> >> -----------------------------------
>> >> buddy | CPU0 | CPU0 | CPU0 | CPU2 |
>> >>
>> >> Small tasks tend to slip out of the periodic load balance.
>> >> The best place to choose to migrate them is at their wake up.
>> >>
>> >> Signed-off-by: Vincent Guittot <[email protected]>
>> >> ---
>> >> kernel/sched/core.c | 1 +
>> >> kernel/sched/fair.c | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++
>> >> kernel/sched/sched.h | 1 +
>> >> 3 files changed, 111 insertions(+)
>> >>
>> >> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> >> index dab7908..70cadbe 100644
>> >> --- a/kernel/sched/core.c
>> >> +++ b/kernel/sched/core.c
>> >> @@ -6131,6 +6131,7 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
>> >> rcu_assign_pointer(rq->sd, sd);
>> >> destroy_sched_domains(tmp, cpu);
>> >>
>> >> + update_packing_domain(cpu);
>> >> update_top_cache_domain(cpu);
>> >> }
>> >>
>> >> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> >> index 4f4a4f6..8c9d3ed 100644
>> >> --- a/kernel/sched/fair.c
>> >> +++ b/kernel/sched/fair.c
>> >> @@ -157,6 +157,63 @@ void sched_init_granularity(void)
>> >> update_sysctl();
>> >> }
>> >>
>> >> +
>> >> +/*
>> >> + * Save the id of the optimal CPU that should be used to pack small tasks
>> >> + * The value -1 is used when no buddy has been found
>> >> + */
>> >> +DEFINE_PER_CPU(int, sd_pack_buddy);
>> >> +
>> >> +/* Look for the best buddy CPU that can be used to pack small tasks
>> >> + * We make the assumption that it doesn't wort to pack on CPU that share the
>> >> + * same powerline. We looks for the 1st sched_domain without the
>> >> + * SD_SHARE_POWERLINE flag. Then We look for the sched_group witht the lowest
>> >> + * power per core based on the assumption that their power efficiency is
>> >> + * better */
>> >> +void update_packing_domain(int cpu)
>> >> +{
>> >> + struct sched_domain *sd;
>> >> + int id = -1;
>> >> +
>> >> + sd = highest_flag_domain(cpu, SD_SHARE_POWERLINE);
>> >> + if (!sd)
>> >> + sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd);
>> >> + else
>> >> + sd = sd->parent;
>> > sd is the highest level where SD_SHARE_POWERLINE is enabled so the sched
>> > groups of the parent level would represent the power domains. If get it
>> > right, we want to pack inside the cluster first and only let first cpu
>>
>> You probably wanted to use sched_group instead of cluster because
>> cluster is only a special use case, didn't you ?
>>
>> > of the cluster do packing on another cluster. So all cpus - except the
>> > first one - in the current sched domain should find its buddy within the
>> > domain and only the first one should go to the parent sched domain to
>> > find its buddy.
>>
>> We don't want to pack in the current sched_domain because it shares
>> power domain. We want to pack at the parent level
>>
>
> Yes. I think we mean the same thing. The packing takes place at the
> parent sched_domain but the sched_group that we are looking at only
> contains the cpus of the level below.
>
>> >
>> > I propose the following fix:
>> >
>> > - sd = sd->parent;
>> > + if (cpumask_first(sched_domain_span(sd)) == cpu
>> > + || !sd->parent)
>> > + sd = sd->parent;
>>
>> We always look for the buddy in the parent level whatever the cpu
>> position in the mask is.
>>
>> >
>> >
>> >> +
>> >> + while (sd) {
>> >> + struct sched_group *sg = sd->groups;
>> >> + struct sched_group *pack = sg;
>> >> + struct sched_group *tmp = sg->next;
>> >> +
>> >> + /* 1st CPU of the sched domain is a good candidate */
>> >> + if (id == -1)
>> >> + id = cpumask_first(sched_domain_span(sd));
>> >
>> > There is no guarantee that id is in the sched group pointed to by
>> > sd->groups, which is implicitly assumed later in the search loop. We
>> > need to find the sched group that contains id and point sg to that
>> > instead. I haven't found an elegant way to find that group, but the fix
>> > below should at least give the right result.
>> >
>> > + /* Find sched group of candidate */
>> > + tmp = sd->groups;
>> > + do {
>> > + if (cpumask_test_cpu(id, sched_group_cpus(tmp)))
>> > + {
>> > + sg = tmp;
>> > + break;
>> > + }
>> > + } while (tmp = tmp->next, tmp != sd->groups);
>> > +
>> > + pack = sg;
>> > + tmp = sg->next;
>>
>>
>> I have a new loop which solves your issue and others. I will use it
>> for the next version
>>
>> + while (sd) {
>> + struct sched_group *sg = sd->groups;
>> + struct sched_group *pack = sg;
>> + struct sched_group *tmp;
>> +
>> + /* The 1st CPU of the local group is a good candidate */
>> + id = cpumask_first(sched_group_cpus(pack));
>
> You make the assumption that the first sched_group in the list always contains
> the current cpu. I think that is always the case, but I haven't verified
> it. Maybe a comment about this would help people to understand the code
> easier.
yes, the first sched_group contains the cpu. I will add a comment
>
>> +
>> + /* loop the sched groups to find the best one */
>> + for (tmp = sg->next; tmp != sg; tmp = tmp->next) {
>> + if (tmp->sgp->power * pack->group_weight >
>> + pack->sgp->power * tmp->group_weight)
>> + continue;
>> +
>> + if ((tmp->sgp->power * pack->group_weight ==
>> + pack->sgp->power * tmp->group_weight)
>> + && (cpumask_first(sched_group_cpus(tmp)) >= id))
>> + continue;
>> +
>> + /* we have found a better group */
>> + pack = tmp;
>> +
>> + /* Take the 1st CPU of the new group */
>> + id = cpumask_first(sched_group_cpus(pack));
>> + }
>> +
>
> Works great on my setup.
>
>> + /* Look for another CPU than itself */
>> + if ((id != cpu)
>> + || ((sd->parent) && !(sd->parent->flags && SD_LOAD_BALANCE)))
>> + break;
>
> If I understand correctly the last part of this check should avoid
> selecting a buddy in a sched_group that is not load balanced with the
> current one. In that case, I think that this check (or a similar check)
> should be done before the loop as well. As it is, the first iteration of
> the loop will always search all the groups of the first domain where
> SD_SHARE_POWERLINE is disabled regardless of the state of
> SD_LOAD_BALANCE flag. So if they are both disabled at the same level
> packing will happen across groups that are not supposed to be
> load-balanced.
you're right, i'm going to fix it
Thanks
>
> Regards,
> Morten
>
>> +
>> + sd = sd->parent;
>> + }
>>
>> Regards,
>> Vincent
>>
>> >
>> > Regards,
>> > Morten
>> >
>> >> +
>> >> + /* loop the sched groups to find the best one */
>> >> + while (tmp != sg) {
>> >> + if (tmp->sgp->power * sg->group_weight <
>> >> + sg->sgp->power * tmp->group_weight)
>> >> + pack = tmp;
>> >> + tmp = tmp->next;
>> >> + }
>> >> +
>> >> + /* we have found a better group */
>> >> + if (pack != sg)
>> >> + id = cpumask_first(sched_group_cpus(pack));
>> >> +
>> >> + /* Look for another CPU than itself */
>> >> + if ((id != cpu)
>> >> + || ((sd->parent) && !(sd->parent->flags && SD_LOAD_BALANCE)))
>> >> + break;
>> >> +
>> >> + sd = sd->parent;
>> >> + }
>> >> +
>> >> + pr_info(KERN_INFO "CPU%d packing on CPU%d\n", cpu, id);
>> >> + per_cpu(sd_pack_buddy, cpu) = id;
>> >> +}
>> >> +
>> >> #if BITS_PER_LONG == 32
>> >> # define WMULT_CONST (~0UL)
>> >> #else
>> >> @@ -3073,6 +3130,55 @@ static int select_idle_sibling(struct task_struct *p, int target)
>> >> return target;
>> >> }
>> >>
>> >> +static inline bool is_buddy_busy(int cpu)
>> >> +{
>> >> + struct rq *rq = cpu_rq(cpu);
>> >> +
>> >> + /*
>> >> + * A busy buddy is a CPU with a high load or a small load with a lot of
>> >> + * running tasks.
>> >> + */
>> >> + return ((rq->avg.usage_avg_sum << rq->nr_running) >
>> >> + rq->avg.runnable_avg_period);
>> >> +}
>> >> +
>> >> +static inline bool is_light_task(struct task_struct *p)
>> >> +{
>> >> + /* A light task runs less than 25% in average */
>> >> + return ((p->se.avg.usage_avg_sum << 2) < p->se.avg.runnable_avg_period);
>> >> +}
>> >> +
>> >> +static int check_pack_buddy(int cpu, struct task_struct *p)
>> >> +{
>> >> + int buddy = per_cpu(sd_pack_buddy, cpu);
>> >> +
>> >> + /* No pack buddy for this CPU */
>> >> + if (buddy == -1)
>> >> + return false;
>> >> +
>> >> + /*
>> >> + * If a task is waiting for running on the CPU which is its own buddy,
>> >> + * let the default behavior to look for a better CPU if available
>> >> + * The threshold has been set to 37.5%
>> >> + */
>> >> + if ((buddy == cpu)
>> >> + && ((p->se.avg.usage_avg_sum << 3) < (p->se.avg.runnable_avg_sum * 5)))
>> >> + return false;
>> >> +
>> >> + /* buddy is not an allowed CPU */
>> >> + if (!cpumask_test_cpu(buddy, tsk_cpus_allowed(p)))
>> >> + return false;
>> >> +
>> >> + /*
>> >> + * If the task is a small one and the buddy is not overloaded,
>> >> + * we use buddy cpu
>> >> + */
>> >> + if (!is_light_task(p) || is_buddy_busy(buddy))
>> >> + return false;
>> >> +
>> >> + return true;
>> >> +}
>> >> +
>> >> /*
>> >> * sched_balance_self: balance the current task (running on cpu) in domains
>> >> * that have the 'flag' flag set. In practice, this is SD_BALANCE_FORK and
>> >> @@ -3098,6 +3204,9 @@ select_task_rq_fair(struct task_struct *p, int sd_flag, int wake_flags)
>> >> if (p->nr_cpus_allowed == 1)
>> >> return prev_cpu;
>> >>
>> >> + if (check_pack_buddy(cpu, p))
>> >> + return per_cpu(sd_pack_buddy, cpu);
>> >> +
>> >> if (sd_flag & SD_BALANCE_WAKE) {
>> >> if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p)))
>> >> want_affine = 1;
>> >> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
>> >> index a95d5c1..086d8bf 100644
>> >> --- a/kernel/sched/sched.h
>> >> +++ b/kernel/sched/sched.h
>> >> @@ -875,6 +875,7 @@ static inline void idle_balance(int cpu, struct rq *rq)
>> >>
>> >> extern void sysrq_sched_debug_show(void);
>> >> extern void sched_init_granularity(void);
>> >> +extern void update_packing_domain(int cpu);
>> >> extern void update_max_interval(void);
>> >> extern void update_group_power(struct sched_domain *sd, int cpu);
>> >> extern int update_runtime(struct notifier_block *nfb, unsigned long action, void *hcpu);
>> >> --
>> >> 1.7.9.5
>> >>
>> >>
>> >> _______________________________________________
>> >> linaro-dev mailing list
>> >> [email protected]
>> >> http://lists.linaro.org/mailman/listinfo/linaro-dev
>> >>
>> >
>>
>