| View previous topic :: View next topic |
| Author |
Message |
kshade
n00b
Joined: 13 Jan 2004
Posts: 60
Location: Germany
|
 Posted: Mon Apr 23, 2007 10:09 am Post subject: SATA: Slow cached reads |
 |
|
Hi,
I just realized that most apps load faster inside the virtual machine I set up (running Archlinux) than directly on the host (Gentoo, of course). Firefox for example takes about 15 seconds on the host and only 8 inside the VM. While investigating that strange behavior I did some performance tests with hdparm, here's the (averaged) result:
| Code: |
root@gm ~ # hdparm -tT /dev/sda
/dev/sda:
Timing cached reads: 670 MB in 2.00 seconds = 334.43 MB/sec
Timing buffered disk reads: 180 MB in 3.01 seconds = 59.75 MB/sec |
The buffered reads are OK, but the cached reads are way too low.
Some informations about my system:
lspci
| Code: |
00:00.0 Host bridge: VIA Technologies, Inc. K8T800Pro Host Bridge
00:00.1 Host bridge: VIA Technologies, Inc. K8T800Pro Host Bridge
00:00.2 Host bridge: VIA Technologies, Inc. K8T800Pro Host Bridge
00:00.3 Host bridge: VIA Technologies, Inc. K8T800Pro Host Bridge
00:00.4 Host bridge: VIA Technologies, Inc. K8T800Pro Host Bridge
00:00.7 Host bridge: VIA Technologies, Inc. K8T800Pro Host Bridge
00:01.0 PCI bridge: VIA Technologies, Inc. VT8237 PCI bridge [K8T800/K8T890 South]
00:0c.0 Multimedia audio controller: Creative Labs SB Live! EMU10k1 (rev 0a)
00:0c.1 Input device controller: Creative Labs SB Live! Game Port (rev 0a)
00:0f.0 IDE interface: VIA Technologies, Inc. VIA VT6420 SATA RAID Controller (rev 80)
00:0f.1 IDE interface: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06)
00:10.0 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 81)
00:10.1 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 81)
00:10.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 81)
00:10.3 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 81)
00:10.4 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 86)
00:11.0 ISA bridge: VIA Technologies, Inc. VT8237 ISA bridge [KT600/K8T800/K8T890 South]
00:11.5 Multimedia audio controller: VIA Technologies, Inc. VT8233/A/8235/8237 AC97 Audio Controller (rev 60)
00:13.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139/8139C/8139C+ (rev 10)
00:18.0 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] HyperTransport Technology Configuration
00:18.1 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Address Map
00:18.2 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] DRAM Controller
00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron] Miscellaneous Control
01:00.0 VGA compatible controller: nVidia Corporation NV43 [GeForce 6600/GeForce 6600 GT] (rev a2)
|
hdparm -I /dev/sda
| Code: |
/dev/sda:
ATA device, with non-removable media
Model Number: SAMSUNG SP1614C
Serial Number: 0696J1FX201693
Firmware Revision: SW100-27
Standards:
Used: ATA/ATAPI-7 T13 1532D revision 0
Supported: 7 6 5 4
Configuration:
Logical max current
cylinders 16383 16383
heads 16 16
sectors/track 63 63
--
CHS current addressable sectors: 16514064
LBA user addressable sectors: 268435455
LBA48 user addressable sectors: 312579695
device size with M = 1024*1024: 152626 MBytes
device size with M = 1000*1000: 160040 MBytes (160 GB)
Capabilities:
LBA, IORDY(can be disabled)
Standby timer values: spec'd by Standard, no device specific minimum
R/W multiple sector transfer: Max = 16 Current = 16
Recommended acoustic management value: 254, current value: 254
DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6 udma7
Cycle time: min=120ns recommended=120ns
PIO: pio0 pio1 pio2 pio3 pio4
Cycle time: no flow control=120ns IORDY flow control=120ns
Commands/features:
Enabled Supported:
SMART feature set
Security Mode feature set
* Power Management feature set
* Write cache
* Look-ahead
* Host Protected Area feature set
* WRITE_BUFFER command
* READ_BUFFER command
* DOWNLOAD_MICROCODE
SET_MAX security extension
* Automatic Acoustic Management feature set
* 48-bit Address feature set
* Device Configuration Overlay feature set
* Mandatory FLUSH_CACHE
* FLUSH_CACHE_EXT
* SMART error logging
* SMART self-test
* SATA-I signaling speed (1.5Gb/s)
Security:
Master password revision code = 65534
supported
not enabled
not locked
not frozen
not expired: security count
supported: enhanced erase
56min for SECURITY ERASE UNIT. 56min for ENHANCED SECURITY ERASE UNIT.
Checksum: correct
|
Thanks in advance.
_________________
Get an operating system!
Last edited by kshade on Tue Apr 24, 2007 10:16 am; edited 2 times in total |
|
| Back to top |
|
 |
guero61
l33t
Joined: 14 Oct 2002
Posts: 811
Location: Behind you
|
| Posted: Mon Apr 23, 2007 6:30 pm Post subject: |
|
|
I assume you're asking for advice on speeding things up, but for the record your post simply reads as a statement, not as a request for help.
That said, on most vanilla Linux systems the "cached reads" section is actually a measure of your processor-memory I/O (man 8 hdparm) and not disk I/O or even interaction with the disk's cache. There isn't much tuning you can do to alter that short of overclocking. 334MB/s seems to translate roughly into the realm of PC-133 memory (1/2 bus bandwidth); is that about right?
Depending on how they're set up (loop device, raw partition, LVM, etc), Xen domU clients can benefit from the host's disk cache as well as their own, likely explaining your boost in speed within a VM. The only reliable way to test disk I/O within a VM or anything else is not by seat-of-the-pants measurement, but by ensuring your caches are flushed, using accurate load timing, and documenting your methodology so it is repeatable. |
|
| Back to top |
|
|
kshade
n00b
Joined: 13 Jan 2004
Posts: 60
Location: Germany
|
| Posted: Mon Apr 23, 2007 8:18 pm Post subject: |
|
|
| guero61 wrote: | | I assume you're asking for advice on speeding things up, but for the record your post simply reads as a statement, not as a request for help. |
You're right, sorry. Shouldn't have posted just after getting up.
| guero61 wrote: | | That said, on most vanilla Linux systems the "cached reads" section is actually a measure of your processor-memory I/O (man 8 hdparm) and not disk I/O or even interaction with the disk's cache. There isn't much tuning you can do to alter that short of overclocking. |
Thanks, didn't knew that.
| guero61 wrote: | | 334MB/s seems to translate roughly into the realm of PC-133 memory (1/2 bus bandwidth); is that about right? |
No, it's actually 3x 512MB DDR-400 RAM. My old PC (which has PC-133 RAM) performs about the same when it comes to cached reads but falls far behind when comparing buffered disk reads. I found various benchmark results on machines similar to mine and those indicated that I should get about 1000 MB/sec.
| guero61 wrote: | | Depending on how they're set up (loop device, raw partition, LVM, etc), Xen domU clients can benefit from the host's disk cache as well as their own, likely explaining your boost in speed within a VM. |
It's Virtualbox and the HDD's just a file. I don't care too much about the HDD performance in the VM really, it just got me investigating the low performance of the host.
Some memory benchmarks:
xfbsuite.pl -r
| Code: | RAM: 1518
*** MB/s ***
Function
Copy: 893
Scale: 894
Add: 1002
Triad: 998 |
bashmark --just-mem-test
| Code: | : T E S T : :S C O R E : : R A T I O:
:-----------------------------------------------------:
:Memory r/w (cached) : : 1936: : +61%:
:Memory de-/alloc : : 585: : -11%: |
_________________
Get an operating system! |
|
| Back to top |
|
|
kshade
n00b
Joined: 13 Jan 2004
Posts: 60
Location: Germany
|
| Posted: Mon Apr 23, 2007 9:19 pm Post subject: |
|
|
Dang, looks like I those 3 DIMMs don't get along well, memtest told me that they were running at PC100 speed. I removed one of the three and got those way better results:
| Code: | /dev/sda:
Timing cached reads: 1004 MB in 2.00 seconds = 501.73 MB/sec
Timing buffered disk reads: 182 MB in 3.02 seconds = 60.23 MB/sec |
| Code: | RAM: 1010
*** MB/s ***
Function
Copy: 1783
Scale: 1816
Add: 1937
Triad: 2062 |
| Code: | : T E S T : :S C O R E : : R A T I O:
:-----------------------------------------------------:
:Memory r/w (cached) : : 1973: : +64%:
:Memory de-/alloc : : 676: : +3%: |
_________________
Get an operating system! |
|
| Back to top |
|
|
drescherjm
Advocate
Joined: 05 Jun 2004
Posts: 2790
Location: Pittsburgh, PA, USA
|
| Posted: Mon Apr 23, 2007 9:41 pm Post subject: |
|
|
| Quote: | | Dang, looks like I those 3 DIMMs don't get along well, memtest told me that they were running at PC100 speed. I removed one of the three and got those way better results |
The slow down is normal with a lot of DDR systems. The motherboard bios will generally run the memory at slower bus speed when a channel has more than one double sided dimm installed to minimize stability issues caused by the increased noise on the bus.
_________________
John
My gentoo overlay
Instructons for overlay |
|
| Back to top |
|
|
kshade
n00b
Joined: 13 Jan 2004
Posts: 60
Location: Germany
|
| Posted: Tue Apr 24, 2007 10:16 am Post subject: |
|
|
| drescherjm wrote: | | Quote: | | Dang, looks like I those 3 DIMMs don't get along well, memtest told me that they were running at PC100 speed. I removed one of the three and got those way better results |
The slow down is normal with a lot of DDR systems. The motherboard bios will generally run the memory at slower bus speed when a channel has more than one double sided dimm installed to minimize stability issues caused by the increased noise on the bus. |
Hum, the mainboards manual states that it will run at full speed with all banks occupied.
I ran some test on my notebook (Pentium M 1.5 GHz, 512 MB RAM, Intel 82801 chipset, IDE drive) and got those results:
| Code: | Timing cached reads: 1264 MB in 2.00 seconds = 631.84 MB/sec
Timing buffered disk reads: 72 MB in 3.06 seconds = 23.56 MB/sec |
Faster than my desktop (Athlon 64 3200+, 1024 MB RAM, VIA K8T800+ chipset) for some reasons.
Bashmark on the other hand yields those numbers:
| Code: | : T E S T : :S C O R E : : R A T I O:
:-----------------------------------------------------:
:Memory r/w (cached) : : 690: : -43%:
:Memory de-/alloc : : 412: : -37%: |
OK for a notebook I think and clearly below the desktop machine.
Here's the NBs drive information: | Code: |
ATA device, with non-removable media
Model Number: FUJITSU MHT2060AT
Serial Number: NN78T4814BNM
Firmware Revision: 0022
Standards:
Used: ATA/ATAPI-6 T13 1410D revision 3a
Supported: 6 5 4
Configuration:
Logical max current
cylinders 16383 16383
heads 16 16
sectors/track 63 63
--
CHS current addressable sectors: 16514064
LBA user addressable sectors: 117210240
device size with M = 1024*1024: 57231 MBytes
device size with M = 1000*1000: 60011 MBytes (60 GB)
Capabilities:
LBA, IORDY(cannot be disabled)
Standby timer values: spec'd by Standard, no device specific minimum
R/W multiple sector transfer: Max = 16 Current = 16
Advanced power management level: 128 (0x80)
Recommended acoustic management value: 254, current value: 254
DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 *udma5
Cycle time: min=120ns recommended=120ns
PIO: pio0 pio1 pio2 pio3 pio4
Cycle time: no flow control=240ns IORDY flow control=120ns
Commands/features:
Enabled Supported:
* SMART feature set
Security Mode feature set
* Power Management feature set
* Write cache
* Look-ahead
* Host Protected Area feature set
* WRITE_BUFFER command
* READ_BUFFER command
* DOWNLOAD_MICROCODE
* Advanced Power Management feature set
Power-Up In Standby feature set
* SET_MAX security extension
* Automatic Acoustic Management feature set
* Device Configuration Overlay feature set
* Mandatory FLUSH_CACHE
* SMART error logging
* SMART self-test
Security:
Master password revision code = 65534
supported
not enabled
not locked
frozen
not expired: security count
not supported: enhanced erase
60min for SECURITY ERASE UNIT.
HW reset results:
CBLID- above Vih
Device num = 0 determined by the jumper
Checksum: correct
|
_________________
Get an operating system! |
|
| Back to top |
|
|
guero61
l33t
Joined: 14 Oct 2002
Posts: 811
Location: Behind you
|
| Posted: Tue Apr 24, 2007 2:49 pm Post subject: |
|
|
| kshade wrote: |
| guero61 wrote: | | That said, on most vanilla Linux systems the "cached reads" section is actually a measure of your processor-memory I/O (man 8 hdparm) and not disk I/O or even interaction with the disk's cache. There isn't much tuning you can do to alter that short of overclocking. |
Thanks, didn't knew that.
|
Neither did I, until recently. That's what so many of us get for blindly running commands someone else posts without reading the man page and actually understanding what's going on.
I'm sure the mainboard's manual stating that it will run at "full speed with all banks occupied" is probably one of those weasel statements - the kind with microscopic footnotes that say, "unless you hold your mouth wrong, or use more than one double-sided DIMM, or..."
It's unfortunate, but if it's really tied to 2 v. 3 DIMMs, you may just have to choose between being slower and having less RAM, unless you purchase a pristine set of matching single-sided DIMMS. |
|
| Back to top |
|
|
kshade
n00b
Joined: 13 Jan 2004
Posts: 60
Location: Germany
|
| Posted: Tue Apr 24, 2007 3:39 pm Post subject: |
|
|
| guero61 wrote: | I'm sure the mainboard's manual stating that it will run at "full speed with all banks occupied" is probably one of those weasel statements - the kind with microscopic footnotes that say, "unless you hold your mouth wrong, or use more than one double-sided DIMM, or..."
It's unfortunate, but if it's really tied to 2 v. 3 DIMMs, you may just have to choose between being slower and having less RAM, unless you purchase a pristine set of matching single-sided DIMMS. |
Seems so. Just looked that table up in the manual, it says that 3 double sided modules can run at DDR333 speed and 3 single sided modules at DDR400 speed.
It runs at DDR-400 speed now but the cached reads from HD are still way too slow. I'll try tinkering with some BIOS setting, but I doubt that will help. Any other suggestions?
_________________
Get an operating system! |
|
| Back to top |
|
|
luminoso
n00b
Joined: 13 Oct 2005
Posts: 60
Location: Portugal, Aveiro
|
| Posted: Mon May 07, 2007 11:32 am Post subject: |
|
|
did u found any solution?
_________________
luminoso @ irc.ptnet.org |
|
| Back to top |
|
|
kshade
n00b
Joined: 13 Jan 2004
Posts: 60
Location: Germany
|
| Posted: Mon May 07, 2007 11:38 am Post subject: |
|
|
Not really. Bought a new HDD (wanted some more space anyway) and reinstalled Gentoo, the system feels a lot more responsive now and actually is measurable quicker (but not that much).
_________________
Get an operating system! |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Kernel & Hardware
