View previous topic :: View next topic |
Author |
Message |
Pointer n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
Joined: 10 Sep 2003 Posts: 16
|
Posted: Fri Feb 20, 2004 10:43 pm Post subject: Performance Tips for Kernel 2.6 |
|
|
Purpose of this thread is to give performance related hints for
kernel configuration and parametrisation. Please share your tricks!
I like to select only what I need and compile all into the kernel.
S E L E C T / U N S E L E C T
S = select, U = unselect
Processor type and features:
S: HPET Timer Support , not supported by all hardware
S: ..Provide RTC interrupt (HPET_EMULATE_RTC)
S: MTRR
S: Local APIC support
S: ..IO-APIC support, (not supported by all hardware)
S: Enable kernel irq balancing (IRQBALANCE)
http://lwn.net/Articles/33344/ (scroll down)
S: Use register arguments (EXPERIMENTAL) (REGPARM)
https://forums.gentoo.org/viewtopic.php?t=183954&highlight=register+arguments
U: Generic x86 support
U: Preemptible Kernel (for server use)
Bus options:
S: Message Signaled Interrupts (MSI and MSI-X) (PCI_MSI)
Related forum post
Kernel hacking:
S: Use 4Kb for kernel stacks instead of 8Kb (4KSTACKS)
http://lwn.net/Articles/84583/
Device Drivers:
-->ATA/ATAPI/...
S: Include IDE/ATA-2 DISK support
S: Use multi-mode default
U: IDE Taskfile Access
U: IDE Taskfile IO
U: generic/default IDE chipset support
-->PCI IDE chipset support
U: Sharing PCI IDE interrupts support !!
S: Generic PCI bus-master DMA support (BLK_DEV_IDEDMA_PCI)
S: Use PCI DMA by default when available (IDEDMA_PCI_AUTO)
S: The chipset driver of your board
Networking support:
-->Packet socket
S: Packet socket:mmapped IO
Profiling support:
U: All
Gentoo needs:
http://www.gentoo.org/doc/en/handbook/handbook-x86.xml?part=1&chap=7
File systems --->
Pseudo Filesystems --->
[*] /proc file system support
[*] Virtual memory file system support (former shm fs)
(Select one or more of the following options as needed by your system)
<*> Reiserfs support
<*> Ext3 journalling file system support
<*> JFS filesystem support
<*> Second extended fs support
<*> XFS filesystem support
CFLAGS for kernel can be set to:
.../linux/Makefile
.../linux/arch/i386/Makefile
The options of make
make help - Lists all options of Makefile
For example these are nice!
make allnoconfig - New minimal config (a good starting point )
make defconfig - New config with default answer to all options
Another tips:
Native Posix Thread Library
http://linuxdevices.com/articles/AT6753699732.html
Related forum post
Select the scheduler that is best for your task, AS(default) is only one possibility.
https://forums.gentoo.org/viewtopic-t-462268-highlight-.html
HDPARM-options
http://gentoo-wiki.com/HOWTO_Use_hdparm_to_improve_IDE_device_performance
Swappiness, if you think that there is too much swapping going on try this.
http://kerneltrap.org/node/view/3000
Good reading for everyone:
Red Hat Linux Manuals
Red Hat Enterprise Linux Documentation
The Relation of kernel headers and kernel source (good to know)
http://uwsg.iu.edu/hypermail/linux/kernel/0007.3/0587.html
Related news group post
ps.
Gentoo should have a performance optimization and benchmarking HOWTO.
Last edited by Pointer on Thu Jun 01, 2006 2:39 pm; edited 15 times in total |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
ectospasm l33t
![l33t l33t](/images/ranks/rank_rect_4.gif)
![](images/avatars/178538108643c6afee61ed2.gif)
Joined: 19 Feb 2003 Posts: 711 Location: Mobile, AL, USA
|
Posted: Fri Feb 20, 2004 11:00 pm Post subject: Re: Performance Tips for Kernel 2.6 |
|
|
Pointer wrote: |
Unselect:
Preemptible Kernel, standard kernel is very responsive even without this, at least in my machine. |
Note that the preemptive kernel options will make your machine even more responsive. And some programs, notably JACK, require this. Unless you're running a lot of CPU intensive applications that use a lot of kernel space functions, you will generally want a preemptable kernel. _________________ Join the adopt an unanswered post initiative today
Join the EFF!
Join the Drug Policy Alliance! |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
zieloo Veteran
![Veteran Veteran](/images/ranks/rank_rect_5_vet.gif)
![](images/avatars/93921736942c6aa75a0a93.png)
Joined: 28 Mar 2004 Posts: 1337
|
Posted: Mon Mar 14, 2005 9:40 am Post subject: |
|
|
I have this option disabled as there were some problems with it and my -ck kernel. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
GaMMa l33t
![l33t l33t](/images/ranks/rank_rect_4.gif)
![](images/avatars/2567983364490c9ae52395.jpg)
Joined: 23 Aug 2002 Posts: 684 Location: USA
|
Posted: Mon Mar 14, 2005 6:29 pm Post subject: |
|
|
Nice guide, my system is a lot more responsive now. I always used the "Preemptible Kernel" and didn't know about some of the other things.. _________________ Ubuntu Linux Dapper Drake running Gnome-2.14.1
[Website | Screenshot | Portage Guide] |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
frilled Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/15648281494375ee83ca682.png)
Joined: 15 Mar 2004 Posts: 386 Location: Atlantis, inner city ring
|
Posted: Thu Oct 06, 2005 6:44 am Post subject: |
|
|
GaMMa wrote: | Nice guide, my system is a lot more responsive now. I always used the "Preemptible Kernel" and didn't know about some of the other things.. |
I have always been using preemption on workstations as soon as I could get a grip on it, but interestingly, on this hyperthreaded P4 from which I'm typing this I decided to turn it off after some months, since responsiveness is actually better without preemption on this particular machine. I don't really have an explanation for this, but keep in mind that you might want to try both options to see what fits you best. _________________ "Failure is not an option!"
"Sir, we are out of further options." |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
madmango Guru
![Guru Guru](/images/ranks/rank_rect_3.gif)
![](images/avatars/160262239540153adaebb0a.gif)
Joined: 15 Jul 2003 Posts: 507 Location: PA, USA
|
Posted: Thu Oct 06, 2005 5:33 pm Post subject: |
|
|
wgi wrote: | GaMMa wrote: | Nice guide, my system is a lot more responsive now. I always used the "Preemptible Kernel" and didn't know about some of the other things.. |
I have always been using preemption on workstations as soon as I could get a grip on it, but interestingly, on this hyperthreaded P4 from which I'm typing this I decided to turn it off after some months, since responsiveness is actually better without preemption on this particular machine. I don't really have an explanation for this, but keep in mind that you might want to try both options to see what fits you best. |
Placebo? Honestly, the docs say that preemption is for desktop use because it decreases latency, but it also decreases throughput, which is what you need for servers. _________________ word. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
frilled Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/15648281494375ee83ca682.png)
Joined: 15 Mar 2004 Posts: 386 Location: Atlantis, inner city ring
|
Posted: Thu Oct 06, 2005 8:26 pm Post subject: |
|
|
Quote: | Placebo? Honestly, the docs say that preemption is for desktop use because it decreases latency, but it also decreases throughput, which is what you need for servers. |
Well, I specifically stated "workstation". No problems with my production servers
I have reproduced the effect on another P4 workstation with HT, although less dramatically. It seems that the worst part on the aforementioned workstation seems to be the (ATA) disk on the ICH5 controller. Somehow NOT using preemption seems to make things better. It is less notable on the other workstation which also uses an ICH5, but with SATA disks.
And, as I said, this is just meant to make you try both settings, for it seems there's no absolute truth here. I have been using the preemption patches for a long time since it naturally seemd A Good Thing [TM] to do, but it isn't in all cases. That's my experience at least. _________________ "Failure is not an option!"
"Sir, we are out of further options." |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
helmholtz n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
![](images/avatars/54796032641f80bced9e57.jpg)
Joined: 20 Aug 2004 Posts: 6
|
Posted: Thu Oct 06, 2005 8:35 pm Post subject: |
|
|
Quote: | I have always been using preemption on workstations as soon as I could get a grip on it, but interestingly, on this hyperthreaded P4 from which I'm typing this I decided to turn it off after some months, since responsiveness is actually better without preemption on this particular machine. I don't really have an explanation for this, but keep in mind that you might want to try both options to see what fits you best. |
I wonder if could be cause hyperthreading is in itself a form of "hardware preemption". The second logical processor, somewhat acting like a real second processor, can "immediately" pick up a new request. Of course, since there really is only one processor, your throughput will suffer, just like with kernel preemption. So adding kernel preemption on top of the hardware preemption really doesn't help any, and in some cases can hurt.
I'm certainly no engineer, but if I was in front of a firing squad, that's the hypothesis I'd put forward. ![Wink ;)](images/smiles/icon_wink.gif) |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
frilled Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/15648281494375ee83ca682.png)
Joined: 15 Mar 2004 Posts: 386 Location: Atlantis, inner city ring
|
Posted: Thu Oct 06, 2005 8:49 pm Post subject: |
|
|
The benefits of HT are without doubt debatable; I still don't really "feel" any major difference. I won't retell the whole seremon, but in the end it all depends. Given the right environment there may be benefits, on the other hand there's the overhead and sharing of limited ressources that work against it. Nevertheless, preemption should be completely independend of that, because it works on a different layer. I did not really dive into this yet, but if it was related to HT/SMP I think it would be an unwanted side effect. I'd rather say it's got something to do with either I/O scheduling [I'm currently testing whether the different schedulers give me really different results] or maybe even something on a lower (=hardware) level with the chipset drivers. I have no clue yet, and I guess by the time I dug into the sources somebody else will already have an explanation for this ![Cool 8)](images/smiles/icon_cool.gif) _________________ "Failure is not an option!"
"Sir, we are out of further options." |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
TheThaWav n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
Joined: 04 Jun 2003 Posts: 6
|
Posted: Mon Oct 17, 2005 7:47 am Post subject: |
|
|
I'd suggest setting the option section Processor Type and Features/Timer Frequency to 1000hertz. that'll speed up the boot process. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
frilled Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/15648281494375ee83ca682.png)
Joined: 15 Mar 2004 Posts: 386 Location: Atlantis, inner city ring
|
Posted: Mon Oct 17, 2005 2:19 pm Post subject: |
|
|
TheThaWav wrote: | I'd suggest setting the option section Processor Type and Features/Timer Frequency to 1000hertz. that'll speed up the boot process. |
Not if you use SMP or HT. You'll get sick desktop effects; at least I do. _________________ "Failure is not an option!"
"Sir, we are out of further options." |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
Tazok Guru
![Guru Guru](/images/ranks/rank_rect_3.gif)
Joined: 25 Oct 2003 Posts: 310
|
Posted: Mon Oct 17, 2005 5:13 pm Post subject: |
|
|
wgi wrote: | TheThaWav wrote: | I'd suggest setting the option section Processor Type and Features/Timer Frequency to 1000hertz. that'll speed up the boot process. |
Not if you use SMP or HT. You'll get sick desktop effects; at least I do. |
For example? |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
frilled Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/15648281494375ee83ca682.png)
Joined: 15 Mar 2004 Posts: 386 Location: Atlantis, inner city ring
|
Posted: Mon Oct 17, 2005 5:20 pm Post subject: |
|
|
Tazok wrote: | wgi wrote: | TheThaWav wrote: | I'd suggest setting the option section Processor Type and Features/Timer Frequency to 1000hertz. that'll speed up the boot process. |
Not if you use SMP or HT. You'll get sick desktop effects; at least I do. |
For example? |
Recurring delays. Mouse chopping, applications grinding to a halt for a fraction of a second once every couple of seconds. Just try it, it's massively evil. It's easily reproducable on my boxes. _________________ "Failure is not an option!"
"Sir, we are out of further options." |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
Tazok Guru
![Guru Guru](/images/ranks/rank_rect_3.gif)
Joined: 25 Oct 2003 Posts: 310
|
Posted: Mon Oct 17, 2005 11:27 pm Post subject: |
|
|
So, which frequency do you suggest? |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
frilled Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/15648281494375ee83ca682.png)
Joined: 15 Mar 2004 Posts: 386 Location: Atlantis, inner city ring
|
Posted: Tue Oct 18, 2005 3:54 am Post subject: |
|
|
Tazok wrote: | So, which frequency do you suggest? |
100. IIRC it was mentioned somewhere, I guess LKML. But I'd be interested to hear if you get good results with something else! _________________ "Failure is not an option!"
"Sir, we are out of further options." |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
moosh Tux's lil' helper
![Tux's lil' helper Tux's lil' helper](/images/ranks/rank_rect_1.gif)
![](images/avatars/205533894848a1fefb54d14.jpg)
Joined: 17 Sep 2004 Posts: 148
|
Posted: Thu Oct 20, 2005 2:38 am Post subject: |
|
|
Hi wgi,
I moved from gentoo-sources-2.6.12-r10 to gentoo-sources-2.6.13-r3 I noticed a considerable slowdown. For example: booting time (until the KDE login screen) was 1 minute 25 seconds, and this becamse after the upgrade, 2 minutes and 10 seconds. The system was choppy and screen redrawing was slow.
After reading your post, I tried changing the frequency from the 1000Hz setting to the 100Hz setting. This however, caused my system to be even slower, WAAAAAAAAAAY slower. Bootup time increased to 8 minutes and 20 seconds!!!! and the screen redrawing was painfully slow. It's as if I need more than 1000Hz to get the performance I had in gentoo-sources-2.6.12-r10. (I have pentium4 with HT BTW) Did any of you get the same slowdown in 2.6.13? |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
raf Apprentice
![Apprentice Apprentice](/images/ranks/rank_rect_2.gif)
![](images/avatars/gallery/Star Wars/movie_star_wars_emperor.gif)
Joined: 16 Jan 2005 Posts: 158
|
Posted: Thu Oct 20, 2005 3:09 am Post subject: |
|
|
I have the exact same performance as I had before upgrading from 2.6.12.5-vanilla to 2.6.13-gentoo-r3. I tried both 100Hz and 1000Hz with no noticable performance hit/boost.
Did you use the EXACT same configuration as your previous kernel? ie make oldconfig and say no to all new options? _________________ -Raf |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
frilled Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/15648281494375ee83ca682.png)
Joined: 15 Mar 2004 Posts: 386 Location: Atlantis, inner city ring
|
Posted: Thu Oct 20, 2005 3:11 am Post subject: |
|
|
There's definitely something wrong there. I don't have any noticable difference between 2.6.12 and 2.6.13. In fact, even with the frequency setting set to 100 this is by far the smoothest box I have. The setting _shouldn't_ affect overall speed ... must be some evil side effect. Like something that really hangs and completely hogs its timeslice. 8 minutes is completely insane. I haven't timed it but I doubt my box needs more than a minute to boot to the desktop.
Here's what I have under Processor type and features for my P4-3GHz with HT:
Code: |
Subarchitecture Type (PC-compatible) --->
Processor family (Pentium-4/Celeron(P4-based)/Pentium-4 M/Xeon) --->
[ ] Generic x86 support
[*] HPET Timer Support
[*] Symmetric multi-processing support
(2) Maximum number of CPUs (2-255)
[*] SMT (Hyperthreading) scheduler support
Preemption Model (Voluntary Kernel Preemption (Desktop)) --->
[*] Preempt The Big Kernel Lock
[*] Machine Check Exception
<*> Check for non-fatal errors on AMD Athlon/Duron / Intel Pentium 4
[*] check for P4 thermal throttling interrupt.
< > Toshiba Laptop support
< > Dell laptop support
[ ] Enable X86 board specific fixups for reboot
< > /dev/cpu/microcode - Intel IA32 CPU microcode support
< > /dev/cpu/*/msr - Model-specific register support
< > /dev/cpu/*/cpuid - CPU information support
Firmware Drivers --->
High Memory Support (4GB) --->
Memory model (Flat Memory) --->
[ ] Allocate 3rd-level pagetables from highmem
[ ] Math emulation
[*] MTRR (Memory Type Range Register) support
[ ] Boot from EFI support (EXPERIMENTAL)
[*] Enable kernel irq balancing
[ ] Use register arguments (EXPERIMENTAL)
[*] Enable seccomp to safely compute untrusted bytecode
Timer frequency (100 HZ) --->
[ ] kexec system call (EXPERIMENTAL)
|
... and I use the cfq I/O scheduler.
Things you might want to consider are:
- not using preemption at all (I had ill effects with another P4 with Intel Chipset and parallel ATA harddrive [this one uses SATA])
- not using kernel IRQ balancing (also had strange effects with this on another machine)
- not using SMP/SMT at all. You may well set the timer to 250 or 1000 then (although I think 250 is really enough, but YMMV - as it might with all the other stuff)
Double check your dmesg and chipset support! _________________ "Failure is not an option!"
"Sir, we are out of further options." |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
moosh Tux's lil' helper
![Tux's lil' helper Tux's lil' helper](/images/ranks/rank_rect_1.gif)
![](images/avatars/205533894848a1fefb54d14.jpg)
Joined: 17 Sep 2004 Posts: 148
|
Posted: Thu Oct 20, 2005 4:01 am Post subject: |
|
|
Hi all,
Thanks for the really quick replies. Yes, I did use make oldconfig. Afterwards, seeing everything is so slow I started doing a methodical cleaning of everything which I did not need.
wgi, if I compare your configuration to mine I have but minor differences:
1) The preemption model is "preemtible kernel"
2) I don't have high memory support (since I don't have more than 1GB)
(there's some other stuff which is clearly unrelated, like Dell laptop support compiled as a module, and /dev/cpu/... stuff compiled in)
I'll try changing the preemption mode and see what happens. I already tried irq balancing on/off and nothing changed. dmesg shows nothing out of the ordinary (I'll post it in case nothing helps, maybe you can see something).
Thanks |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
moosh Tux's lil' helper
![Tux's lil' helper Tux's lil' helper](/images/ranks/rank_rect_1.gif)
![](images/avatars/205533894848a1fefb54d14.jpg)
Joined: 17 Sep 2004 Posts: 148
|
Posted: Thu Oct 20, 2005 4:49 am Post subject: |
|
|
Well, nothing really helped, so here is my dmesg:
Code: | Linux version 2.6.13-gentoo-r3 (root@rama) (gcc version 3.3.6 (Gentoo 3.3.6, ssp-3.3.6-1.0, pie-8.7.8)) #9 SMP Wed Oct 19 21:38:17 PDT 2005
BIOS-provided physical RAM map:
BIOS-e820: 0000000000000000 - 000000000009f000 (usable)
BIOS-e820: 000000000009f000 - 00000000000a0000 (reserved)
BIOS-e820: 0000000000100000 - 000000001ffaa800 (usable)
BIOS-e820: 000000001ffaa800 - 0000000020000000 (reserved)
BIOS-e820: 00000000fec00000 - 00000000fec10000 (reserved)
BIOS-e820: 00000000fecf0000 - 00000000fecf1000 (reserved)
BIOS-e820: 00000000fed20000 - 00000000fed90000 (reserved)
BIOS-e820: 00000000feda0000 - 00000000fee10000 (reserved)
BIOS-e820: 00000000ffb00000 - 0000000100000000 (reserved)
511MB LOWMEM available.
On node 0 totalpages: 130986
DMA zone: 4096 pages, LIFO batch:1
Normal zone: 126890 pages, LIFO batch:31
HighMem zone: 0 pages, LIFO batch:1
DMI 2.3 present.
ACPI: RSDP (v000 DELL ) @ 0x000fde90
ACPI: RSDT (v001 DELL CPi R 0x27d50308 ASL 0x00000061) @ 0x1ffefbcd
ACPI: FADT (v001 DELL CPi R 0x27d50308 ASL 0x00000061) @ 0x1fff0400
ACPI: MADT (v001 DELL CPi R 0x27d50308 ASL 0x00000047) @ 0x1fff0c00
ACPI: SSDT (v001 PmRef CpuPm 0x00003000 INTL 0x20030522) @ 0x1ffefbfd
ACPI: DSDT (v001 INT430 SYSFexxx 0x00001001 MSFT 0x0100000e) @ 0x00000000
ACPI: Local APIC address 0xfee00000
ACPI: LAPIC (acpi_id[0x00] lapic_id[0x00] enabled)
Processor #0 15:3 APIC version 20
ACPI: LAPIC (acpi_id[0x01] lapic_id[0x01] enabled)
Processor #1 15:3 APIC version 20
ACPI: LAPIC_NMI (acpi_id[0x00] high edge lint[0x1])
ACPI: LAPIC_NMI (acpi_id[0x01] high edge lint[0x1])
ACPI: IOAPIC (id[0x02] address[0xfec00000] gsi_base[0])
IOAPIC[0]: apic_id 2, version 32, address 0xfec00000, GSI 0-23
ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 high level)
ACPI: IRQ0 used by override.
ACPI: IRQ2 used by override.
ACPI: IRQ9 used by override.
Enabling APIC mode: Flat. Using 1 I/O APICs
Using ACPI (MADT) for SMP configuration information
Allocating PCI resources starting at 20000000 (gap: 20000000:dec00000)
Built 1 zonelists
Kernel command line: root=/dev/hda7 video=vesafb:1024x768-16@60,nomtrr splash=silent,theme:livecd-2005.1 quiet CONSOLE=/dev/tty1
mapped APIC to ffffd000 (fee00000)
mapped IOAPIC to ffffc000 (fec00000)
Initializing CPU#0
PID hash table entries: 2048 (order: 11, 32768 bytes)
Detected 2793.498 MHz processor.
Using tsc for high-res timesource
Console: colour VGA+ 80x25
Dentry cache hash table entries: 131072 (order: 7, 524288 bytes)
Inode-cache hash table entries: 65536 (order: 6, 262144 bytes)
Memory: 512032k/523944k available (3218k kernel code, 11404k reserved, 1229k data, 248k init, 0k highmem)
Checking if this processor honours the WP bit even in supervisor mode... Ok.
Calibrating delay using timer specific routine.. 5590.67 BogoMIPS (lpj=2795337)
Mount-cache hash table entries: 512
CPU: After generic identify, caps: bfebfbff 00000000 00000000 00000000 0000441d 00000000 00000000
CPU: After vendor identify, caps: bfebfbff 00000000 00000000 00000000 0000441d 00000000 00000000
monitor/mwait feature present.
using mwait in idle threads.
CPU: Trace cache: 12K uops, L1 D cache: 16K
CPU: L2 cache: 1024K
CPU: Physical Processor ID: 0
CPU: After all inits, caps: bfebfbff 00000000 00000000 00000080 0000441d 00000000 00000000
Intel machine check architecture supported.
Intel machine check reporting enabled on CPU#0.
CPU0: Intel P4/Xeon Extended MCE MSRs (12) available
CPU0: Thermal monitoring enabled
mtrr: v2.0 (20020519)
Enabling fast FPU save and restore... done.
Enabling unmasked SIMD FPU exception support... done.
Checking 'hlt' instruction... OK.
CPU0: Intel(R) Pentium(R) 4 CPU 2.80GHz stepping 04
Booting processor 1/1 eip 2000
Initializing CPU#1
Calibrating delay using timer specific routine.. 5585.31 BogoMIPS (lpj=2792657)
CPU: After generic identify, caps: bfebfbff 00000000 00000000 00000000 0000441d 00000000 00000000
CPU: After vendor identify, caps: bfebfbff 00000000 00000000 00000000 0000441d 00000000 00000000
monitor/mwait feature present.
CPU: Trace cache: 12K uops, L1 D cache: 16K
CPU: L2 cache: 1024K
CPU: Physical Processor ID: 0
CPU: After all inits, caps: bfebfbff 00000000 00000000 00000080 0000441d 00000000 00000000
Intel machine check architecture supported.
Intel machine check reporting enabled on CPU#1.
CPU1: Intel P4/Xeon Extended MCE MSRs (12) available
CPU1: Thermal monitoring enabled
CPU1: Intel(R) Pentium(R) 4 CPU 2.80GHz stepping 04
Total of 2 processors activated (11175.98 BogoMIPS).
ENABLING IO-APIC IRQs
..TIMER: vector=0x31 pin1=2 pin2=-1
checking TSC synchronization across 2 CPUs:
CPU#0 had 0 usecs TSC skew, fixed it up.
CPU#1 had 0 usecs TSC skew, fixed it up.
Brought up 2 CPUs
checking if image is initramfs... it is
Freeing initrd memory: 1235k freed
NET: Registered protocol family 16
ACPI: bus type pci registered
PCI: PCI BIOS revision 2.10 entry at 0xfcc7e, last bus=2
PCI: Using configuration type 1
ACPI: Subsystem revision 20050408
ACPI: Interpreter enabled
ACPI: Using IOAPIC for interrupt routing
ACPI: PCI Root Bridge [PCI0] (0000:00)
PCI: Probing PCI hardware (bus 00)
ACPI: Assume root bridge [\_SB_.PCI0] segment is 0
ACPI: Assume root bridge [\_SB_.PCI0] bus is 0
PCI: Ignoring BAR0-3 of IDE controller 0000:00:1f.1
Boot video device is 0000:01:00.0
PCI: Transparent bridge - 0000:00:1e.0
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
ACPI: PCI Interrupt Link [LNKA] (IRQs 9 10 *11)
ACPI: PCI Interrupt Link [LNKB] (IRQs 5 7) *11
ACPI: PCI Interrupt Link [LNKC] (IRQs 9 10 *11)
ACPI: PCI Interrupt Link [LNKD] (IRQs 5 7 9 10 *11)
ACPI: PCI Interrupt Link [LNKE] (IRQs 3 4 5 6 7 9 10 11 12 14 15) *0, disabled.
ACPI: PCI Interrupt Link [LNKH] (IRQs 3 4 5 6 7 9 10 *11 12 14 15)
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.AGP_._PRT]
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PCIE._PRT]
Linux Plug and Play Support v0.97 (c) Adam Belay
pnp: PnP ACPI init
pnp: PnP ACPI: found 10 devices
SCSI subsystem initialized
usbcore: registered new driver usbfs
usbcore: registered new driver hub
PCI: Using ACPI for IRQ routing
PCI: If a device doesn't work, try "pci=routeirq". If it helps, post a report
pnp: 00:01: ioport range 0x4d0-0x4d1 has been reserved
pnp: 00:01: ioport range 0x1000-0x1005 could not be reserved
pnp: 00:01: ioport range 0x1008-0x100f could not be reserved
pnp: 00:01: ioport range 0x800-0x80f has been reserved
pnp: 00:02: ioport range 0xf400-0xf4fe has been reserved
pnp: 00:02: ioport range 0x1006-0x1007 has been reserved
pnp: 00:02: ioport range 0x1010-0x105f could not be reserved
pnp: 00:02: ioport range 0x1060-0x107f has been reserved
pnp: 00:02: ioport range 0x1080-0x10bf has been reserved
pnp: 00:02: ioport range 0x10c0-0x10df has been reserved
pnp: 00:07: ioport range 0x900-0x97f has been reserved
PCI: Bridge: 0000:00:01.0
IO window: c000-cfff
MEM window: fc000000-fdffffff
PREFETCH window: f0000000-f7ffffff
PCI: Bus 3, cardbus bridge: 0000:02:01.0
IO window: 0000e000-0000efff
IO window: 00002000-00002fff
PREFETCH window: 20000000-21ffffff
MEM window: 24000000-25ffffff
PCI: Bridge: 0000:00:1e.0
IO window: e000-efff
MEM window: fa000000-fbffffff
PREFETCH window: 20000000-21ffffff
PCI: Setting latency timer of device 0000:00:1e.0 to 64
PCI: Enabling device 0000:02:01.0 (0000 -> 0003)
ACPI: PCI Interrupt 0000:02:01.0[A] -> GSI 19 (level, low) -> IRQ 169
Machine check exception polling timer started.
IA-32 Microcode Update Driver: v1.14 <tigran@veritas.com>
audit: initializing netlink socket (disabled)
audit(1129758039.963:1): initialized
Squashfs 2.2 (released 2005/07/03) (C) 2002-2005 Phillip Lougher
Installing knfsd (copyright (C) 1996 okir@monad.swb.de).
Initializing Cryptographic API
Real Time Clock Driver v1.12
Linux agpgart interface v0.101 (c) Dave Jones
agpgart: Detected an Intel 865 Chipset.
agpgart: AGP aperture is 128M @ 0xe8000000
vesafb: ATI Technologies Inc., P11 , 01.00 (OEM: ATI MOBILITY RADEON 9700 )
vesafb: VBE version: 2.0
vesafb: protected mode interface info at c000:5acf
vesafb: pmi: set display start = c00c5b63, set palette = c00c5baf
vesafb: pmi: ports = ec10 ec16 ec54 ec38 ec3c ec5c ec00 ec04 ecb0 ecb2 ecb4
vesafb: monitor limits: vf = 0 Hz, hf = 0 kHz, clk = 0 MHz
vesafb: scrolling: redraw
Console: switching to colour frame buffer device 128x48
fbsplash: console 0 using theme 'livecd-2005.1'
fbsplash: switched splash state to 'on' on console 0
vesafb: framebuffer at 0xf0000000, mapped to 0xe0880000, using 15000k, total 65472k
fb0: VESA VGA frame buffer device
ACPI: AC Adapter [AC] (on-line)
ACPI: Battery Slot [BAT0] (battery present)
ACPI: Lid Switch [LID]
ACPI: Power Button (CM) [PBTN]
ACPI: Sleep Button (CM) [SBTN]
ACPI: Video Device [VID] (multi-head: yes rom: no post: no)
Using specific hotkey driver
ACPI: CPU0 (power states: C1[C1] C2[C2] C3[C3])
ACPI: Processor [CPU0] (supports 8 throttling states)
ACPI: CPU1 (power states: C1[C1] C2[C2] C3[C3])
ACPI: Processor [CPU1] (supports 8 throttling states)
ACPI: Thermal Zone [THM] (64 C)
PNP: PS/2 Controller [PNP0303:KBC,PNP0f13:PS2M] at 0x60,0x64 irq 1,12
serio: i8042 AUX port at 0x60,0x64 irq 12
serio: i8042 KBD port at 0x60,0x64 irq 1
Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing disabled
ACPI: PCI Interrupt 0000:00:1f.6[B] -> GSI 17 (level, low) -> IRQ 177
ACPI: PCI interrupt for device 0000:00:1f.6 disabled
mice: PS/2 mouse device common for all mice
input: PC Speaker
io scheduler noop registered
io scheduler cfq registered
RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
loop: loaded (max 8 devices)
pktcdvd: v0.2.0a 2004-07-14 Jens Axboe (axboe@suse.de) and petero2@telia.com
b44.c:v0.95 (Aug 3, 2004)
ACPI: PCI Interrupt 0000:02:00.0[A] -> GSI 18 (level, low) -> IRQ 185
eth0: Broadcom 4400 10/100BaseT Ethernet 00:0f:1f:22:87:38
PPP generic driver version 2.4.2
PPP Deflate Compression module registered
Linux video capture interface: v1.00
Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
ICH5: IDE controller at PCI slot 0000:00:1f.1
PCI: Enabling device 0000:00:1f.1 (0005 -> 0007)
ACPI: PCI Interrupt 0000:00:1f.1[A] -> GSI 16 (level, low) -> IRQ 193
ICH5: chipset revision 2
ICH5: not 100% native mode: will probe irqs later
ide0: BM-DMA at 0xbfa0-0xbfa7, BIOS settings: hda:DMA, hdb:pio
ide1: BM-DMA at 0xbfa8-0xbfaf, BIOS settings: hdc:DMA, hdd:pio
Probing IDE interface ide0...
input: AT Translated Set 2 keyboard on isa0060/serio0
hda: FUJITSU MHT2040AH, ATA DISK drive
input: PS/2 Mouse on isa0060/serio1
input: AlpsPS/2 ALPS GlidePoint on isa0060/serio1
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Probing IDE interface ide1...
hdc: SAMSUNG CDRW/DVD SN-324S, ATAPI CD/DVD-ROM drive
ide1 at 0x170-0x177,0x376 on irq 15
hda: max request size: 128KiB
hda: 78140160 sectors (40007 MB) w/8192KiB Cache, CHS=65535/16/63, UDMA(100)
hda: cache flushes supported
hda: hda1 hda2 hda3 hda4 < hda5 hda6 hda7 >
hdc: ATAPI 24X DVD-ROM CD-R/RW drive, 2048kB Cache, UDMA(33)
Uniform CD-ROM driver Revision: 3.20
libata version 1.12 loaded.
ohci1394: $Rev: 1299 $ Ben Collins <bcollins@debian.org>
ACPI: PCI Interrupt 0000:02:01.1[A] -> GSI 19 (level, low) -> IRQ 169
ohci1394: fw-host0: OHCI-1394 1.1 (PCI): IRQ=[169] MMIO=[faffd800-faffdfff] Max Packet=[2048]
ieee1394: raw1394: /dev/raw1394 device initialized
NFTL driver: nftlcore.c $Revision: 1.97 $, nftlmount.c $Revision: 1.40 $
INFTL: inftlcore.c $Revision: 1.18 $, inftlmount.c $Revision: 1.16 $
ACPI: PCI Interrupt 0000:02:01.0[A] -> GSI 19 (level, low) -> IRQ 169
Yenta: CardBus bridge found at 0000:02:01.0 [1028:0159]
Yenta: Using CSCINT to route CSC interrupts to PCI
Yenta: Routing CardBus interrupts to PCI
Yenta TI: socket 0000:02:01.0, mfunc 0x012c1202, devctl 0x64
Yenta: ISA IRQ mask 0x0cf8, PCI irq 169
Socket status: 30000086
pcmcia: parent PCI bridge I/O window: 0xe000 - 0xefff
pcmcia: parent PCI bridge Memory window: 0xfa000000 - 0xfbffffff
pcmcia: parent PCI bridge Memory window: 0x20000000 - 0x21ffffff
ACPI: PCI Interrupt 0000:00:1d.7[D] -> GSI 23 (level, low) -> IRQ 201
PCI: Setting latency timer of device 0000:00:1d.7 to 64
ehci_hcd 0000:00:1d.7: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB2 EHCI Controller
ehci_hcd 0000:00:1d.7: debug port 1
ehci_hcd 0000:00:1d.7: new USB bus registered, assigned bus number 1
ehci_hcd 0000:00:1d.7: irq 201, io mem 0xf8fffc00
PCI: cache line size of 128 is not supported by device 0000:00:1d.7
ehci_hcd 0000:00:1d.7: USB 2.0 initialized, EHCI 1.00, driver 10 Dec 2004
hub 1-0:1.0: USB hub found
hub 1-0:1.0: 8 ports detected
ohci_hcd: 2005 April 22 USB 1.1 'Open' Host Controller (OHCI) Driver (PCI)
USB Universal Host Controller Interface driver v2.3
ACPI: PCI Interrupt 0000:00:1d.0[A] -> GSI 16 (level, low) -> IRQ 193
PCI: Setting latency timer of device 0000:00:1d.0 to 64
uhci_hcd 0000:00:1d.0: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #1
uhci_hcd 0000:00:1d.0: new USB bus registered, assigned bus number 2
uhci_hcd 0000:00:1d.0: irq 193, io base 0x0000bf80
hub 2-0:1.0: USB hub found
hub 2-0:1.0: 2 ports detected
ACPI: PCI Interrupt 0000:00:1d.1[B] -> GSI 19 (level, low) -> IRQ 169
PCI: Setting latency timer of device 0000:00:1d.1 to 64
uhci_hcd 0000:00:1d.1: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #2
uhci_hcd 0000:00:1d.1: new USB bus registered, assigned bus number 3
uhci_hcd 0000:00:1d.1: irq 169, io base 0x0000bf60
hub 3-0:1.0: USB hub found
hub 3-0:1.0: 2 ports detected
ACPI: PCI Interrupt 0000:00:1d.2[C] -> GSI 18 (level, low) -> IRQ 185
PCI: Setting latency timer of device 0000:00:1d.2 to 64
uhci_hcd 0000:00:1d.2: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI #3
uhci_hcd 0000:00:1d.2: new USB bus registered, assigned bus number 4
uhci_hcd 0000:00:1d.2: irq 185, io base 0x0000bf40
hub 4-0:1.0: USB hub found
hub 4-0:1.0: 2 ports detected
ACPI: PCI Interrupt 0000:00:1d.3[A] -> GSI 16 (level, low) -> IRQ 193
PCI: Setting latency timer of device 0000:00:1d.3 to 64
uhci_hcd 0000:00:1d.3: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #4
uhci_hcd 0000:00:1d.3: new USB bus registered, assigned bus number 5
uhci_hcd 0000:00:1d.3: irq 193, io base 0x0000bf20
hub 5-0:1.0: USB hub found
hub 5-0:1.0: 2 ports detected
Initializing USB Mass Storage driver...
usbcore: registered new driver usb-storage
USB Mass Storage support registered.
usb 3-1: new full speed USB device using uhci_hcd and address 2
usbcore: registered new driver hiddev
usbcore: registered new driver usbhid
drivers/usb/input/hid-core.c: v2.01:USB HID core driver
i2c /dev entries driver
NET: Registered protocol family 2
IP route cache hash table entries: 8192 (order: 3, 32768 bytes)
TCP established hash table entries: 32768 (order: 7, 524288 bytes)
TCP bind hash table entries: 32768 (order: 6, 393216 bytes)
TCP: Hash tables configured (established 32768 bind 32768)
TCP reno registered
ip_conntrack version 2.1 (4093 buckets, 32744 max) - 216 bytes per conntrack
ip_tables: (C) 2000-2002 Netfilter core team
ipt_recent v0.3.1: Stephen Frost <sfrost@snowman.net>. http://snowman.net/projects/ipt_recent/
ClusterIP Version 0.7 loaded successfully
arp_tables: (C) 2002 David S. Miller
TCP bic registered
Initializing IPsec netlink socket
NET: Registered protocol family 1
NET: Registered protocol family 17
NET: Registered protocol family 15
Starting balanced_irq
Using IPI Shortcut mode
BIOS EDD facility v0.16 2004-Jun-25, 1 devices found
kjournald starting. Commit interval 5 seconds
EXT3-fs: mounted filesystem with ordered data mode.
VFS: Mounted root (ext3 filesystem) readonly.
Freeing unused kernel memory: 248k freed
usb 5-1: new low speed USB device using uhci_hcd and address 2
ieee1394: Host added: ID:BUS[0-00:1023] GUID[374fc00031d948a1]
input: USB HID v1.10 Mouse [Logitech Optical USB Mouse] on usb-0000:00:1d.3-1
Adding 1004020k swap on /dev/hda6. Priority:-1 extents:1
EXT3 FS on hda7, internal journal
fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.
[fglrx] Maximum main memory to use for locked dma buffers: 429 MBytes.
ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 193
[fglrx] module loaded - fglrx 8.14.13 [Jun 8 2005] on minor 0
ACPI: PCI Interrupt 0000:00:1f.5[B] -> GSI 17 (level, low) -> IRQ 177
PCI: Setting latency timer of device 0000:00:1f.5 to 64
intel8x0_measure_ac97_clock: measured 49485 usecs
intel8x0: clocking to 48000
ACPI: PCI Interrupt 0000:00:1f.6[B] -> GSI 17 (level, low) -> IRQ 177
PCI: Setting latency timer of device 0000:00:1f.6 to 64
b44: eth0: Link is up at 100 Mbps, full duplex.
b44: eth0: Flow control is off for TX and off for RX.
NETDEV WATCHDOG: eth0: transmit timed out
b44: eth0: transmit timed out, resetting
b44: eth0: Link is down.
b44: eth0: Link is up at 100 Mbps, full duplex.
b44: eth0: Flow control is off for TX and off for RX.
kjournald starting. Commit interval 5 seconds
EXT3 FS on hda3, internal journal
EXT3-fs: mounted filesystem with ordered data mode.
[fglrx] Kernel AGP support doesn't provide agplock functionality.
[fglrx] AGP detected, AgpState = 0x1f004a1b (hardware caps of chipset)
agpgart: Found an AGP 3.0 compliant device at 0000:00:00.0.
agpgart: Putting AGP V3 device at 0000:00:00.0 into 8x mode
agpgart: Putting AGP V3 device at 0000:01:00.0 into 8x mode
[fglrx] AGP enabled, AgpCommand = 0x1f004312 (selected caps)
[fglrx] free AGP = 121909248
[fglrx] max AGP = 121909248
[fglrx] free LFB = 43110400
[fglrx] max LFB = 43110400
[fglrx] free Inv = 0
[fglrx] max Inv = 0
[fglrx] total Inv = 0
[fglrx] total TIM = 0
[fglrx] total FB = 0
[fglrx] total AGP = 32768
|
I guess it's still a mystery why 2.6.13-r3 works so slowly for me. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
moosh Tux's lil' helper
![Tux's lil' helper Tux's lil' helper](/images/ranks/rank_rect_1.gif)
![](images/avatars/205533894848a1fefb54d14.jpg)
Joined: 17 Sep 2004 Posts: 148
|
Posted: Thu Oct 20, 2005 5:35 am Post subject: |
|
|
Hi everyone,
wgi, I've read your posting again and noticed the HT removal suggestion. I tried it and it worked! Once I turn SMP and HT in the 2.6.13-r3 kernel everything works just fine (only without HT of course). I guess something is broken in this kernel.
Thanks |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
frilled Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/15648281494375ee83ca682.png)
Joined: 15 Mar 2004 Posts: 386 Location: Atlantis, inner city ring
|
Posted: Thu Oct 20, 2005 10:58 am Post subject: |
|
|
Hm, well, it's not completely broken, at least. I'm running like ten boxes with HT. Maybe something to do with your specific board or APIC. I'll have a look at your dmesg, but I doubt I'll find something useful
You don't really suffer a lot without HT support, IMHO. I never bothered to benchmark it, but I guess it is heavily situational, anyway.
(Edit:)
What strikes me is that quirky Code: | checking TSC synchronization across 2 CPUs:
CPU#0 had 0 usecs TSC skew, fixed it up.
CPU#1 had 0 usecs TSC skew, fixed it up. |
Mine says Code: | checking TSC synchronization across 2 CPUs: passed.
Brought up 2 CPUs |
Now, 0µs is not that much ![Very Happy :D](images/smiles/icon_biggrin.gif) _________________ "Failure is not an option!"
"Sir, we are out of further options." |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
moosh Tux's lil' helper
![Tux's lil' helper Tux's lil' helper](/images/ranks/rank_rect_1.gif)
![](images/avatars/205533894848a1fefb54d14.jpg)
Joined: 17 Sep 2004 Posts: 148
|
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
das_leid n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
![](images/avatars/15689159240e09b9a7bdc6.jpg)
Joined: 11 Apr 2004 Posts: 63 Location: Bochum / Germany
|
Posted: Sun Oct 30, 2005 12:26 pm Post subject: Re: Performance Tips for Kernel 2.6 |
|
|
Pointer wrote: | Purpose of this thread is to give performance related hints for
kernel configuration and parametrisation. Please share your tricks!
|
Well, I'd like to say thanks. Using your guide my overall network performance increased from 2.9MB/s to 6.6MB/s.
I wonder if this is the practical maximum of a SCP connection in a 100Mbit network?!
Currently I use Kernel 2.6.13-r5 which seems to be stable. At least for me. However I will try plain 2.6.14 soon. Seems to have some nice features.
Cheers
Das_Leid
PS: NEVER EVER I am going to use a 3com 3c905B-TX in my network. They play ping-pong with your interrupts and effectively increase your RX error rate. _________________ emerge vim
OH EVIL ! |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
frilled Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/15648281494375ee83ca682.png)
Joined: 15 Mar 2004 Posts: 386 Location: Atlantis, inner city ring
|
Posted: Mon Oct 31, 2005 6:01 am Post subject: Re: Performance Tips for Kernel 2.6 |
|
|
Quote: | Well, I'd like to say thanks. Using your guide my overall network performance increased from 2.9MB/s to 6.6MB/s.
I wonder if this is the practical maximum of a SCP connection in a 100Mbit network?!
|
No. I get 10.5 - 11 MB/s between our Servers/Workstations (Full Duplex). But you need some horsepower for ssh. Since there's virtually nothing slower than 2.4 GHz x86 (at least of those who speak ssh I almost always get full speed. But with a slow machine the protocol and encryption becomes the bottleneck. _________________ "Failure is not an option!"
"Sir, we are out of further options." |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
|
|
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
|
|