Hi,
Since I got my new machine I noticed it seemed to be running slower than
I expected for a duel core machine including a lot of stuttering. After
tweaking the BIOS settings from "Legacy" to "AHCI" I measured a doubling
of read performance with hdparm but heavy IO still makes the machine
sluggish, with top showing ~80% of the time in the wait state (and
loadavg shooting up). This seems like a DMA problem because I was under
the impression a task demanding IO should be able to sleep on a DMA
completion rather than blocking everything else.
Looking with hdparm I get the following info on my disk:
root@pitcairn:/home/ajb# hdparm -v -i /dev/sda
/dev/sda:
IO_support = 0 (default 16-bit)
readonly = 0 (off)
readahead = 256 (on)
geometry = 19457/255/63, sectors = 312581808, start = 0
Model=ST3160815AS , FwRev=3.AAD , SerialNo= 6RA38XMD
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs RotSpdTol>.5% }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=4
BuffType=unknown, BuffSize=8192kB, MaxMultSect=16, MultSect=?16?
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=268435455
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2 udma3 udma4 udma5
AdvancedPM=no WriteCache=enabled
Drive conforms to: Unspecified: ATA/ATAPI-1,2,3,4,5,6,7
* signifies the current active mode
Which seems to indicate DMA is supported but not on. Attempting to
switch it on with -d just fails:
root@pitcairn:/home/ajb# hdparm -v -d 1 /dev/sda
/dev/sda:
setting using_dma to 1 (on)
HDIO_SET_DMA failed: Inappropriate ioctl for device
IO_support = 0 (default 16-bit)
readonly = 0 (off)
readahead = 256 (on)
geometry = 19457/255/63, sectors = 312581808, start = 0
I'm currently running an Ubuntu (7.10) distro kernel although I can get
a test kernel running if it helps to debug:
root@pitcairn:/home/ajb# uname -a
Linux pitcairn 2.6.22-14-generic #1 SMP Tue Feb 12 02:46:46 UTC 2008
x86_64 GNU/Linux
The (hopefully) relevant chunks of lspci show this:
root@pitcairn:/home/ajb# lspci -v
00:1f.0 ISA bridge: Intel Corporation 82801IH (ICH9DH) LPC Interface Controller (rev 02)
Subsystem: Intel Corporation Unknown device 5044
Flags: bus master, medium devsel, latency 0
Capabilities: [e0] Vendor Specific Information
00:1f.2 SATA controller: Intel Corporation 82801IR/IO/IH (ICH9R/DO/DH) 6 port SATA AHCI Controller (rev 02) (prog-if 01 [AHCI 1.0])
Subsystem: Intel Corporation Unknown device 5044
Flags: bus master, 66MHz, medium devsel, latency 0, IRQ 21
I/O ports at 3428 [size=8]
I/O ports at 3434 [size=4]
I/O ports at 3420 [size=8]
I/O ports at 3430 [size=4]
I/O ports at 3020 [size=32]
Memory at 53225000 (32-bit, non-prefetchable) [size=2K]
Capabilities: [80] Message Signalled Interrupts: Mask- 64bit- Queue=0/4 Enable-
Capabilities: [70] Power Management version 3
Capabilities: [a8] #12 [0010]
Capabilities: [b0] Vendor Specific Information
00:1f.3 SMBus: Intel Corporation 82801I (ICH9 Family) SMBus Controller (rev 02)
Subsystem: Intel Corporation Unknown device 5044
Flags: medium devsel, IRQ 10
Memory at 53226000 (64-bit, non-prefetchable) [size=256]
I/O ports at 3000 [size=32]
03:00.0 IDE interface: Marvell Technology Group Ltd. 88SE6101
single-port PATA133 interface (rev b2) (prog-if 8f [Master SecP SecO
PriP PriO])
Subsystem: Marvell Technology Group Ltd. 88SE6101 single-port
PATA133 interface
Flags: bus master, fast devsel, latency 0, IRQ 17
I/O ports at 1018 [size=8]
I/O ports at 1024 [size=4]
I/O ports at 1010 [size=8]
I/O ports at 1020 [size=4]
I/O ports at 1000 [size=16]
Memory at 53100000 (32-bit, non-prefetchable) [size=512]
Capabilities: [48] Power Management version 2
Capabilities: [50] Message Signalled Interrupts: Mask- 64bit-
Queue=0/0 Enable-
Capabilities: [e0] Express Legacy Endpoint IRQ 0
And the (snipped) dmesg output:
[ 30.478794] ahci 0000:00:1f.2: AHCI 0001.0200 32 slots 6 ports 3 Gbps
0x3f impl SATA mode
[ 30.478797] ahci 0000:00:1f.2: flags: 64bit ncq led clo pmp pio slum
part
[ 30.478801] PCI: Setting latency timer of device 0000:00:1f.2 to 64
[ 30.479035] scsi0 : ahci
[ 30.479066] scsi1 : ahci
[ 30.479090] scsi2 : ahci
[ 30.479112] scsi3 : ahci
[ 30.479135] scsi4 : ahci
[ 30.479158] scsi5 : ahci
[ 30.479228] ata1: SATA max UDMA/133 cmd 0xffffc200005f8100 ctl
0x0000000000000000 bmdma 0x0000000000000000 irq 21
[ 30.479231] ata2: SATA max UDMA/133 cmd 0xffffc200005f8180 ctl
0x0000000000000000 bmdma 0x0000000000000000 irq 21
[ 30.479234] ata3: SATA max UDMA/133 cmd 0xffffc200005f8200 ctl
0x0000000000000000 bmdma 0x0000000000000000 irq 21
[ 30.479237] ata4: SATA max UDMA/133 cmd 0xffffc200005f8280 ctl
0x0000000000000000 bmdma 0x0000000000000000 irq 21
[ 30.479240] ata5: SATA max UDMA/133 cmd 0xffffc200005f8300 ctl
0x0000000000000000 bmdma 0x0000000000000000 irq 21
[ 30.479243] ata6: SATA max UDMA/133 cmd 0xffffc200005f8380 ctl
0x0000000000000000 bmdma 0x0000000000000000 irq 21
[ 30.670632] usb 7-1.1: new high speed USB device using ehci_hcd and
address 5
[ 30.746315] ieee1394: Host added: ID:BUS[0-00:1023]
GUID[0090270001eac2c3]
[ 30.764496] usb 7-1.1: configuration #1 chosen from 1 choice
[ 30.961793] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
[ 30.962030] usb 7-1.2: new high speed USB device using ehci_hcd and
address 6
[ 30.999213] ata1.00: ATA-7: ST3160815AS, 3.AAD, max UDMA/133
[ 30.999216] ata1.00: 312581808 sectors, multi 16: LBA48 NCQ (depth
31/32)
[ 31.054773] usb 7-1.2: configuration #1 chosen from 1 choice
[ 31.055515] usbcore: registered new interface driver libusual
[ 31.057399] ata1.00: configured for UDMA/133
[ 31.059152] Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
[ 31.059155] ide: Assuming 33MHz system bus speed for PIO modes;
override with idebus=xx
[ 31.060168] Initializing USB Mass Storage driver...
[ 31.293123] usb 2-1: new low speed USB device using uhci_hcd and
address 2
[ 31.368971] ata2: SATA link down (SStatus 0 SControl 300)
[ 31.475149] usb 2-1: configuration #1 chosen from 1 choice
[ 31.680343] ata3: SATA link down (SStatus 0 SControl 300)
[ 31.716270] usb 2-2: new low speed USB device using uhci_hcd and
address 3
[ 31.898263] usb 2-2: configuration #1 chosen from 1 choice
[ 31.901635] scsi6 : SCSI emulation for USB Mass Storage devices
[ 31.901668] usb-storage: device found at 5
[ 31.901670] usb-storage: waiting for device to settle before scanning
[ 31.901694] scsi7 : SCSI emulation for USB Mass Storage devices
[ 31.901718] usb-storage: device found at 6
[ 31.901719] usb-storage: waiting for device to settle before scanning
[ 31.901734] usbcore: registered new interface driver usb-storage
[ 31.901736] usbcore: registered new interface driver hiddev
[ 31.901738] USB Mass Storage support registered.
[ 31.916372] input: Microsoft Microsoft 5-Button Mouse with
IntelliEye(TM) as /class/input/input1
[ 31.916388] input: USB HID v1.10 Mouse [Microsoft Microsoft 5-Button
Mouse with IntelliEye(TM)] on usb-0000:00:1a.1-1
[ 31.930335] input: Microsoft Natural� Ergonomic Keyboard 4000
as /class/input/input2
[ 31.930341] input: USB HID v1.11 Keyboard [Microsoft Natural�
Ergonomic Keyboard 4000] on usb-0000:00:1a.1-2
[ 31.951194] input: Microsoft Natural� Ergonomic Keyboard 4000
as /class/input/input3
[ 31.951199] input: USB HID v1.11 Device [Microsoft Natural� Ergonomic
Keyboard 4000] on usb-0000:00:1a.1-2
[ 31.951207] usbcore: registered new interface driver usbhid
[
31.951209] /build/buildd/linux-source-2.6.22-2.6.22/drivers/hid/usbhid/hid-core.c: v2.6:USB HID core driver
[ 31.991722] ata4: SATA link down (SStatus 0 SControl 300)
[ 32.303093] ata5: SATA link down (SStatus 0 SControl 300)
[ 32.614466] ata6: SATA link down (SStatus 0 SControl 300)
[ 32.614550] scsi 0:0:0:0: Direct-Access ATA ST3160815AS
3.AA PQ: 0 ANSI: 5
[ 32.614825] ACPI: PCI Interrupt 0000:03:00.0[A] -> GSI 17 (level,
low) -> IRQ 17
[ 32.614847] PCI: Setting latency timer of device 0000:03:00.0 to 64
[ 32.614870] scsi8 : pata_marvell
[ 32.614998] scsi9 : pata_marvell
[ 32.615114] ata7: PATA max UDMA/100 cmd 0x0000000000011018 ctl
0x0000000000011026 bmdma 0x0000000000011000 irq 17
[ 32.615116] ata8: DUMMY
[ 32.615866] BAR5:00:00 01:7F 02:22 03:CA 04:00 05:00 06:00 07:00
08:00 09:00 0A:00 0B:00 0C:01 0D:00 0E:00 0F:00
[ 32.622523] sd 0:0:0:0: [sda] 312581808 512-byte hardware sectors
(160042 MB)
[ 32.622533] sd 0:0:0:0: [sda] Write Protect is off
[ 32.622535] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 32.622545] sd 0:0:0:0: [sda] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
[ 32.622580] sd 0:0:0:0: [sda] 312581808 512-byte hardware sectors
(160042 MB)
[ 32.622586] sd 0:0:0:0: [sda] Write Protect is off
[ 32.622588] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 32.622598] sd 0:0:0:0: [sda] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
[ 32.622601] sda: sda1 sda2 < sda5 >
[ 32.666726] sd 0:0:0:0: [sda] Attached SCSI disk
[ 32.668697] sd 0:0:0:0: Attached scsi generic sg0 type 0
[ 32.861928] Attempting manual resume
[ 32.861931] swsusp: Resume From Partition 8:5
[ 32.861932] PM: Checking swsusp image.
[ 32.862067] PM: Resume from disk failed.
[ 32.898435] kjournald starting. Commit interval 5 seconds
[ 32.898444] EXT3-fs: mounted filesystem with ordered data mode.
[ 32.934174] ata7.00: ATAPI: TSSTcorpDVD-ROM SH-D162D, SB00, max
UDMA/33
[ 33.105820] ata7.00: configured for UDMA/33
[ 33.106844] scsi 8:0:0:0: CD-ROM TSSTcorp DVD-ROM SH-D162D
SB00 PQ: 0 ANSI: 5
[ 33.106894] scsi 8:0:0:0: Attached scsi generic sg1 type 5
[ 36.890025] usb-storage: device scan complete
[ 36.890101] usb-storage: device scan complete
[ 36.890525] scsi 7:0:0:0: Direct-Access WD 1200BEVExternal
1.02 PQ: 0 ANSI: 0
[ 36.890668] scsi 6:0:0:0: Direct-Access Generic USB SD Reader
1.00 PQ: 0 ANSI: 0
[ 36.891139] scsi 6:0:0:1: Direct-Access Generic USB CF Reader
1.01 PQ: 0 ANSI: 0
[ 36.891506] sd 7:0:0:0: [sdb] 234441648 512-byte hardware sectors
(120034 MB)
[ 36.891636] scsi 6:0:0:2: Direct-Access Generic USB SM Reader
1.02 PQ: 0 ANSI: 0
[ 36.892138] scsi 6:0:0:3: Direct-Access Generic USB MS Reader
1.03 PQ: 0 ANSI: 0
[ 36.892253] sd 7:0:0:0: [sdb] Write Protect is off
[ 36.892255] sd 7:0:0:0: [sdb] Mode Sense: 00 00 00 00
[ 36.892257] sd 7:0:0:0: [sdb] Assuming drive cache: write through
[ 36.893125] sd 7:0:0:0: [sdb] 234441648 512-byte hardware sectors
(120034 MB)
[ 36.893746] sd 7:0:0:0: [sdb] Write Protect is off
[ 36.893748] sd 7:0:0:0: [sdb] Mode Sense: 00 00 00 00
[ 36.893749] sd 7:0:0:0: [sdb] Assuming drive cache: write through
[ 36.893753] sdb:<6>pci_hotplug: PCI Hot Plug PCI Core version: 0.5
[ 39.011940] shpchp: Standard Hot Plug PCI Controller Driver version:
0.4
[ 39.042055] input: PC Speaker as /class/input/input4
[ 39.146879] Intel(R) PRO/1000 Network Driver - version 7.6.5-NAPI
[ 39.146882] Copyright (c) 1999-2007 Intel Corporation.
[ 39.146920] ACPI: PCI Interrupt 0000:00:19.0[A] -> GSI 20 (level,
low) -> IRQ 20
[ 39.146930] PCI: Setting latency timer of device 0000:00:19.0 to 64
[ 39.193869] e1000: 0000:00:19.0: e1000_probe: (PCI
Express:2.5Gb/s:Width x1) 00:19:d1:7c:f2:69
[ 39.223347] heci: Intel(R) AMT Management Interface - version
3.1.0.31
[ 39.223350] heci: Copyright (c) 2003 - 2007 Intel Corporation.
[ 39.263051] e1000: eth0: e1000_probe: Intel(R) PRO/1000 Network
Connection
[ 39.263092] ACPI: PCI Interrupt 0000:00:03.0[A] -> GSI 16 (level,
low) -> IRQ 16
[ 39.263096] PCI: Setting latency timer of device 0000:00:03.0 to 64
[ 39.264034] heci: link layer has been established.
[ 39.285949] nvidia: module license 'NVIDIA' taints kernel.
[ 39.539111] heci: heci driver initialization successful.
[ 39.543742] ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level,
low) -> IRQ 16
[ 39.543750] PCI: Setting latency timer of device 0000:01:00.0 to 64
[ 39.543823] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 100.14.19
Wed Sep 12 14:08:38 PDT 2007
[ 39.554103] usbcore: registered new interface driver xpad
[
39.554106] /build/buildd/linux-source-2.6.22-2.6.22/drivers/input/joystick/xpad.c: driver for Xbox controllers v0.1.6
[ 39.745530] sr0: scsi3-mmc drive: 48x/48x cd/rw xa/form2 cdda tray
[ 39.745534] Uniform CD-ROM driver Revision: 3.20
[ 39.745699] sr 8:0:0:0: Attached scsi CD-ROM sr0
[ 39.959429] ACPI: PCI Interrupt 0000:00:1b.0[A] -> GSI 22 (level,
low) -> IRQ 22
[ 39.960191] PCI: Setting latency timer of device 0000:00:1b.0 to 64
[ 40.167253] hda_codec: Unknown model for ALC883, trying auto-probe
from BIOS...
[ 40.501814] sdb1
[ 40.501866] sd 7:0:0:0: [sdb] Attached SCSI disk
[ 40.501899] sd 7:0:0:0: Attached scsi generic sg2 type 0
[ 40.502634] sd 6:0:0:0: [sdc] Attached SCSI removable disk
[ 40.502657] sd 6:0:0:0: Attached scsi generic sg3 type 0
[ 40.503677] sd 6:0:0:1: [sdd] Attached SCSI removable disk
[ 40.503701] sd 6:0:0:1: Attached scsi generic sg4 type 0
[ 40.506604] sd 6:0:0:2: [sde] Attached SCSI removable disk
[ 40.506639] sd 6:0:0:2: Attached scsi generic sg5 type 0
[ 40.510621] sd 6:0:0:3: [sdf] Attached SCSI removable disk
[ 40.510649] sd 6:0:0:3: Attached scsi generic sg6 type 0
As you can see I'm running the NVida kernel module (needed for duel
screen), but I can test without it if needed. Finally the lsmod output:
root@pitcairn:/home/ajb# lsmod
Module Size Used by
rfcomm 47656 2
l2cap 28672 11 rfcomm
bluetooth 63876 4 rfcomm,l2cap
ppdev 11272 0
button 10400 0
sbs 21520 0
ac 7304 0
video 21140 0
container 6400 0
dock 12264 0
battery 12424 0
acpi_cpufreq 10632 0
cpufreq_stats 8160 0
cpufreq_userspace 6048 0
cpufreq_ondemand 10896 2
freq_table 6464 3 acpi_cpufreq,cpufreq_stats,cpufreq_ondemand
cpufreq_conservative 9608 0
cpufreq_powersave 3072 0
ipv6 317192 20
nfs 272984 2
lockd 76336 2 nfs
sunrpc 198536 3 nfs,lockd
parport_pc 41896 0
lp 15048 0
parport 44172 3 ppdev,parport_pc,lp
loop 21764 0
snd_hda_intel 337192 4
snd_pcm_oss 50048 0
snd_mixer_oss 20096 1 snd_pcm_oss
snd_pcm 94344 3 snd_hda_intel,snd_pcm_oss
snd_seq_dummy 5380 0
snd_seq_oss 36864 0
sr_mod 19876 0
cdrom 41768 1 sr_mod
snd_seq_midi 11008 0
snd_rawmidi 29824 1 snd_seq_midi
snd_seq_midi_event 9984 2 snd_seq_oss,snd_seq_midi
snd_seq 62496 6 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_seq_midi_event
snd_timer 27272 3 snd_pcm,snd_seq
xpad 11400 0
snd_seq_device 10260 5 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_rawmidi,snd_seq
nvidia 7013492 36
psmouse 45596 0
ata_generic 9988 0
i2c_core 30208 1 nvidia
heci 66064 0
serio_raw 9092 0
e1000_ich9 211008 0
snd 69288 15 snd_hda_intel,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_seq_oss,snd_rawmidi,snd_seq,snd_timer,snd_seq_device
soundcore 10272 1 snd
snd_page_alloc 12560 2 snd_hda_intel,snd_pcm
pcspkr 4608 0
shpchp 38300 0
pci_hotplug 36612 1 shpchp
intel_agp 30624 0
joydev 13440 0
evdev 13056 4
ext3 146576 2
jbd 69360 1 ext3
mbcache 11272 1 ext3
sg 41384 0
sd_mod 32512 5
usbhid 32576 0
hid 33408 1 usbhid
usb_storage 81728 1
ide_core 141200 1 usb_storage
libusual 22824 1 usb_storage
ahci 27012 2
pata_marvell 9472 0
ohci1394 38984 0
ieee1394 109528 1 ohci1394
libata 138928 3 ata_generic,ahci,pata_marvell
scsi_mod 172856 5 sr_mod,sg,sd_mod,usb_storage,libata
ehci_hcd 40076 0
uhci_hcd 29600 0
usbcore 161584 7 xpad,usbhid,usb_storage,libusual,ehci_hcd,uhci_hcd
thermal 16528 0
processor 36232 2 acpi_cpufreq,thermal
fan 6920 0
fuse 52528 1
apparmor 47008 0
commoncap 9472 1 apparmor
Any pointers?
--
Alex, homepage: http://www.bennee.com/~alex/
Other restrictions may apply.
> of read performance with hdparm but heavy IO still makes the machine
> sluggish, with top showing ~80% of the time in the wait state (and
> loadavg shooting up). This seems like a DMA problem because I was under
Load average includes wait states for disk.
> [ 30.999213] ata1.00: ATA-7: ST3160815AS, 3.AAD, max UDMA/133
> [ 30.999216] ata1.00: 312581808 sectors, multi 16: LBA48 NCQ (depth
> 31/32)
> [ 31.054773] usb 7-1.2: configuration #1 chosen from 1 choice
> [ 31.055515] usbcore: registered new interface driver libusual
> [ 31.057399] ata1.00: configured for UDMA/133
So it is running at UDMA133
> [ 32.934174] ata7.00: ATAPI: TSSTcorpDVD-ROM SH-D162D, SB00, max
> UDMA/33
> [ 33.105820] ata7.00: configured for UDMA/33
CD-ROM at UDMA 33
> [ 39.285949] nvidia: module license 'NVIDIA' taints kernel
And any problems beyond that point are ones you need to take up with
Nvidia as they have all the source code but we don't have theirs.
Alex Bennee wrote:
> Hi,
>
> Since I got my new machine I noticed it seemed to be running slower than
> I expected for a duel core machine including a lot of stuttering. After
> tweaking the BIOS settings from "Legacy" to "AHCI" I measured a doubling
> of read performance with hdparm but heavy IO still makes the machine
> sluggish, with top showing ~80% of the time in the wait state (and
> loadavg shooting up). This seems like a DMA problem because I was under
> the impression a task demanding IO should be able to sleep on a DMA
> completion rather than blocking everything else.
That's not what IOwait means. It basically means "nothing better to do
than wait for IO to complete". If you have only one running task which
is blocked waiting for IO you will always have high IOwait.
On Wed, 2008-03-26 at 23:20 -0600, Robert Hancock wrote:
> Alex Bennee wrote:
> > Hi,
> >
> > Since I got my new machine I noticed it seemed to be running slower than
> > I expected for a duel core machine including a lot of stuttering. After
> > tweaking the BIOS settings from "Legacy" to "AHCI" I measured a doubling
> > of read performance with hdparm but heavy IO still makes the machine
> > sluggish, with top showing ~80% of the time in the wait state (and
> > loadavg shooting up). This seems like a DMA problem because I was under
> > the impression a task demanding IO should be able to sleep on a DMA
> > completion rather than blocking everything else.
>
> That's not what IOwait means. It basically means "nothing better to do
> than wait for IO to complete". If you have only one running task which
> is blocked waiting for IO you will always have high IOwait.
So if my loadavg shoots up at the same time (indicating more than one
task wanting to run) does that infer that most of my tasks are IO
starved and waiting for the disk to catch up with them?
The main problem is I'm not sure if my disk subsystem is running as fast
as it should be. What sort of data rates should I be seeing from a
modern SATA type setup?
As Alan pointed out (and I missed) the dmesg shows DMA is enabled, it's
just hdparam doesn't seem to be able to infer the fact (new IOTCLS for
newer disk systems?).
I could be I've already peaked in my performance and I'm just making
unrealistic demands on memory usage (I have been running cvsps after
all ;-).
--
Alex, homepage: http://www.bennee.com/~alex/
Debian is the Jedi operating system: "Always two there are, a master and
an apprentice". -- Simon Richter on debian-devel
Alex Bennee wrote:
> On Wed, 2008-03-26 at 23:20 -0600, Robert Hancock wrote:
>> Alex Bennee wrote:
>>> Hi,
>>>
>>> Since I got my new machine I noticed it seemed to be running slower than
>>> I expected for a duel core machine including a lot of stuttering. After
>>> tweaking the BIOS settings from "Legacy" to "AHCI" I measured a doubling
>>> of read performance with hdparm but heavy IO still makes the machine
>>> sluggish, with top showing ~80% of the time in the wait state (and
>>> loadavg shooting up). This seems like a DMA problem because I was under
>>> the impression a task demanding IO should be able to sleep on a DMA
>>> completion rather than blocking everything else.
>> That's not what IOwait means. It basically means "nothing better to do
>> than wait for IO to complete". If you have only one running task which
>> is blocked waiting for IO you will always have high IOwait.
>
> So if my loadavg shoots up at the same time (indicating more than one
> task wanting to run) does that infer that most of my tasks are IO
> starved and waiting for the disk to catch up with them?
>
> The main problem is I'm not sure if my disk subsystem is running as fast
> as it should be. What sort of data rates should I be seeing from a
> modern SATA type setup?
50-70MB/second per disk, if you are doing single sequential reads/writes (alone,
not together-if you are doing both at the same time or more than one at the sime
time it will be worse), if you are doing a lot of seeks it can be pretty much
any number below that. The smaller the pieces of data that are being sent to
disk, the worse things will be. The speed is dependent mostly on the type of
disk, each disk has a underlying inside/outside platter bits per second, and
that is the limit that you won't ever exceed.
>
> As Alan pointed out (and I missed) the dmesg shows DMA is enabled, it's
> just hdparam doesn't seem to be able to infer the fact (new IOTCLS for
> newer disk systems?).
>
> I could be I've already peaked in my performance and I'm just making
> unrealistic demands on memory usage (I have been running cvsps after
> all ;-).
>