Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <YD9dUkGhlRT8vvcy@hirez.programming.kicks-ass.net> <20210303224653.2579656-1-joshdon@google.com>
In-Reply-To: <20210303224653.2579656-1-joshdon@google.com>
From:   Nick Desaulniers <ndesaulniers@google.com>
Date:   Thu, 4 Mar 2021 09:34:29 -0800
Message-ID: <CAKwvOdmijctJfM3gNfwEVjaQyp3LZkhnAwgsT7EBhsSBJyfLAA@mail.gmail.com>
Subject: Re: [PATCH v2] sched: Optimize __calc_delta.
To:     Josh Don <joshdon@google.com>
Cc:     Ingo Molnar <mingo@redhat.com>,
        Peter Zijlstra <peterz@infradead.org>,
        Juri Lelli <juri.lelli@redhat.com>,
        Vincent Guittot <vincent.guittot@linaro.org>,
        Dietmar Eggemann <dietmar.eggemann@arm.com>,
        Steven Rostedt <rostedt@goodmis.org>,
        Ben Segall <bsegall@google.com>, Mel Gorman <mgorman@suse.de>,
        Daniel Bristot de Oliveira <bristot@redhat.com>,
        Nathan Chancellor <nathan@kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        clang-built-linux <clang-built-linux@googlegroups.com>,
        Clement Courbet <courbet@google.com>,
        Oleg Rombakh <olegrom@google.com>,
        Bill Wendling <morbo@google.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Mar 3, 2021 at 2:48 PM Josh Don <joshdon@google.com> wrote:
>
> From: Clement Courbet <courbet@google.com>
>
> A significant portion of __calc_delta time is spent in the loop
> shifting a u64 by 32 bits. Use `fls` instead of iterating.
>
> This is ~7x faster on benchmarks.
>
> The generic `fls` implementation (`generic_fls`) is still ~4x faster
> than the loop.
> Architectures that have a better implementation will make use of it. For
> example, on X86 we get an additional factor 2 in speed without dedicated
> implementation.
>
> On gcc, the asm versions of `fls` are about the same speed as the
> builtin. On clang, the versions that use fls are more than twice as
> slow as the builtin. This is because the way the `fls` function is
> written, clang puts the value in memory:
> https://godbolt.org/z/EfMbYe. This bug is filed at
> https://bugs.llvm.org/show_bug.cgi?id=3D49406.

Hi Josh, Thanks for helping get this patch across the finish line.
Would you mind updating the commit message to point to
https://bugs.llvm.org/show_bug.cgi?id=3D20197?

>
> ```
> name                                   cpu/op
> BM_Calc<__calc_delta_loop>             9.57ms =C2=B112%
> BM_Calc<__calc_delta_generic_fls>      2.36ms =C2=B113%
> BM_Calc<__calc_delta_asm_fls>          2.45ms =C2=B113%
> BM_Calc<__calc_delta_asm_fls_nomem>    1.66ms =C2=B112%
> BM_Calc<__calc_delta_asm_fls64>        2.46ms =C2=B113%
> BM_Calc<__calc_delta_asm_fls64_nomem>  1.34ms =C2=B115%
> BM_Calc<__calc_delta_builtin>          1.32ms =C2=B111%
> ```
>
> Signed-off-by: Clement Courbet <courbet@google.com>
> Signed-off-by: Josh Don <joshdon@google.com>
> ---
>  kernel/sched/fair.c  | 19 +++++++++++--------
>  kernel/sched/sched.h |  1 +
>  2 files changed, 12 insertions(+), 8 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 8a8bd7b13634..a691371960ae 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -229,22 +229,25 @@ static void __update_inv_weight(struct load_weight =
*lw)
>  static u64 __calc_delta(u64 delta_exec, unsigned long weight, struct loa=
d_weight *lw)
>  {
>         u64 fact =3D scale_load_down(weight);
> +       u32 fact_hi =3D (u32)(fact >> 32);
>         int shift =3D WMULT_SHIFT;
> +       int fs;
>
>         __update_inv_weight(lw);
>
> -       if (unlikely(fact >> 32)) {
> -               while (fact >> 32) {
> -                       fact >>=3D 1;
> -                       shift--;
> -               }
> +       if (unlikely(fact_hi)) {
> +               fs =3D fls(fact_hi);
> +               shift -=3D fs;
> +               fact >>=3D fs;
>         }
>
>         fact =3D mul_u32_u32(fact, lw->inv_weight);
>
> -       while (fact >> 32) {
> -               fact >>=3D 1;
> -               shift--;
> +       fact_hi =3D (u32)(fact >> 32);
> +       if (fact_hi) {
> +               fs =3D fls(fact_hi);
> +               shift -=3D fs;
> +               fact >>=3D fs;
>         }
>
>         return mul_u64_u32_shr(delta_exec, fact, shift);
> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
> index 10a1522b1e30..714af71cf983 100644
> --- a/kernel/sched/sched.h
> +++ b/kernel/sched/sched.h
> @@ -36,6 +36,7 @@
>  #include <uapi/linux/sched/types.h>
>
>  #include <linux/binfmts.h>
> +#include <linux/bitops.h>

This hunk of the patch is curious.  I assume that bitops.h is needed
for fls(); if so, why not #include it in kernel/sched/fair.c?
Otherwise this potentially hurts compile time for all TUs that include
kernel/sched/sched.h.

>  #include <linux/blkdev.h>
>  #include <linux/compat.h>
>  #include <linux/context_tracking.h>
> --
> 2.30.1.766.gb4fecdf3b7-goog
>


--=20
Thanks,
~Nick Desaulniers