Re: HTT on current

2003-08-28 Thread Jens Rehsack
Bruce Evans wrote: On Tue, 26 Aug 2003, John Baldwin wrote: On 26-Aug-2003 Yamada Ken Takeshi wrote: [...] One test is not sufficient. -current is also not the best place to test. :) When I first implemented HTT in -current The above times seem slow enough to be partly the result of debugging

Re: HTT on current

2003-08-27 Thread Bruce Evans
On Tue, 26 Aug 2003, John Baldwin wrote: > On 26-Aug-2003 Yamada Ken Takeshi wrote: > > JYI, > > I tested machdep.hlt_logical_cpus=0/1, buildworld, > > and found no particular reason to disable HTT > > as below with Zeon 2.8Ghz x 2, 1GBmem, Slow IDE HDD. > > > > It may not be the common case,

Re: HTT on current

2003-08-26 Thread John Baldwin
On 26-Aug-2003 Yamada Ken Takeshi wrote: > JYI, > I tested machdep.hlt_logical_cpus=0/1, buildworld, > and found no particular reason to disable HTT > as below with Zeon 2.8Ghz x 2, 1GBmem, Slow IDE HDD. > > It may not be the common case, so just for your info. > ># sysctl machdep.hlt_log

Re: HTT on current

2003-08-26 Thread Yamada Ken Takeshi
JYI, I tested machdep.hlt_logical_cpus=0/1, buildworld, and found no particular reason to disable HTT as below with Zeon 2.8Ghz x 2, 1GBmem, Slow IDE HDD. It may not be the common case, so just for your info. # sysctl machdep.hlt_logical_cpus=0 # /usr/bin/time make -j32 buildworld 191

Re: HTT on current

2003-08-25 Thread John Baldwin
On 23-Aug-2003 [EMAIL PROTECTED] wrote: >> On Fri, Aug 22, 2003 at 07:03:03PM -0400, [EMAIL PROTECTED] wrote: >>> > That is not an SMP kernel. An SMP kernel (with APIC_IO) would not >>> print >>> > out >>> > the pcib0 interrupt routing messages. >>> > >>> > -- >>> >>> So whats the problem here? >

Re: HTT on current

2003-08-25 Thread David O'Brien
On Mon, Aug 25, 2003 at 02:14:18PM -0300, Daniel C. Sobral wrote: > >I've heard this several times and don't doubt it, but it would be nice to > >know more about the issue. What type of cases? What benchmarks have > >been run showing this? > > Well, I haven't actually seen any case where there w

Re: HTT on current

2003-08-25 Thread Tim Kientzle
Garrett Wollman wrote: The problem is that the P4 is not very wide to begin with, and it's very hard to optimize well for that 23-stage pipeline. I'll say. I spent months tuning some assembly code for P3 and P4 and was quite disappointed that the P4 consistently required more CPU cycles for the sa

Re: HTT on current

2003-08-25 Thread Garrett Wollman
< said: > There are two problems with HTT. First, L1/L2 cache issues. Second, the > virtual CPUs are not independent, and there are many cases where > instructions in one virtual CPU stall the other. So take, for example, > the case of a userland application on CPU0 stalling the kernel on CPU1.

Re: HTT on current

2003-08-25 Thread Daniel C. Sobral
David O'Brien wrote: On Mon, Aug 25, 2003 at 08:27:46AM -0600, Scott Long wrote: Since HTT can lead to performance degradation in some (many?) cases, the second logical CPU's are halted by default. They are enabled, however, in order for interrupt routing to work right. Work is ongoing to make a

Re: HTT on current

2003-08-25 Thread David O'Brien
On Mon, Aug 25, 2003 at 08:27:46AM -0600, Scott Long wrote: > Since HTT can lead to performance degradation in some (many?) cases, > the second logical CPU's are halted by default. They are enabled, > however, in order for interrupt routing to work right. Work is ongoing > to make an HTT-aware sc

Re: HTT on current

2003-08-25 Thread Daniel C. Sobral
, August 24, 2003 12:49 PM To: Kris Kennaway Cc: Yamada Ken Takeshi; [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: Re: HTT on current yes: ganymede# grep SMP /usr/src/sys/i386/conf/kernel options SMP # Symmetric MultiProcessor Kernel and: CPU: Intel(R) Pentium(

Re: HTT on current

2003-08-25 Thread Scott Long
ROTECTED]; [EMAIL PROTECTED] Subject: Re: HTT on current yes: ganymede# grep SMP /usr/src/sys/i386/conf/kernel options SMP # Symmetric MultiProcessor Kernel and: CPU: Intel(R) Pentium(R) 4 CPU 2.40GHz (2393.19-MHz 686-class CPU) Origin = "GenuineIntel" Id = 0

RE: HTT on current

2003-08-25 Thread Mike Jakubik
PROTECTED]; [EMAIL PROTECTED] > Subject: Re: HTT on current > > > > yes: > > ganymede# grep SMP /usr/src/sys/i386/conf/kernel > options SMP # Symmetric MultiProcessor Kernel > > and: > > CPU: Intel(R) Pentium(R) 4 CPU 2.40GHz (2393.19-MHz

RE: HTT on current

2003-08-25 Thread Mike Jakubik
> -Original Message- > From: Sang Woo Shim [mailto:[EMAIL PROTECTED] > Sent: Saturday, August 23, 2003 1:58 AM > To: [EMAIL PROTECTED] > Cc: [EMAIL PROTECTED] > Subject: Re: HTT on current > > > On Fri, Aug 22, 2003 at 10:23:52PM -0400, [EMAIL PROTECTED] wrote

Re: HTT on current

2003-08-25 Thread Yamada Ken Takeshi
Strange Mine (Supermicro X5DAL-G) goes like ; tyd3# dmesg Copyright (c) 1992-2003 The FreeBSD Project. Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved. FreeBSD 5.1-CURRENT #30: Sun Aug 24 20:25:

Re: HTT on current

2003-08-24 Thread Kris Kennaway
On Sun, Aug 24, 2003 at 01:06:28PM -0300, Marc G. Fournier wrote: > > > ganymede# sysctl machdep.hlt_logical_cpus > sysctl: unknown oid 'machdep.hlt_logical_cpus' > ganymede# uname -a > FreeBSD ganymede.hub.org 5.1-CURRENT FreeBSD 5.1-CURRENT #1: Sat Aug 23 00:08:54 ADT > 2003 [EMAIL PROTECT

Re: HTT on current

2003-08-24 Thread Michael W. Oliver
+--- On Saturday, August 23, 2003 21:18, | [EMAIL PROTECTED] proclaimed: | | > On Fri, Aug 22, 2003 at 10:23:52PM -0400, [EMAIL PROTECTED] wrote: | >> Well i've enabled all SMP options, recompiled and rebooted. The CPUs | >> now properly show up in dmesg and i can see the C header in top. | >> Howe

Re: HTT on current

2003-08-24 Thread Yamada Ken Takeshi
Did you do "sysctl machdep.hlt_logical_cpus=0"? Default setting halt logical CPUs, and won't start HTT. This is a kind of FAQ, and you will see a lot in the past mailing-list. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinf

Re: HTT on current

2003-08-24 Thread mikej
> On Fri, Aug 22, 2003 at 10:23:52PM -0400, [EMAIL PROTECTED] wrote: > >> Well i've enabled all SMP options, recompiled and rebooted. The CPUs now >> properly show up in dmesg and i can see the C header in top. However no >> processes seem to be being assigned to cpu 1. Why is the schedueler only >

Re: HTT on current

2003-08-22 Thread Kris Kennaway
On Fri, Aug 22, 2003 at 10:23:52PM -0400, [EMAIL PROTECTED] wrote: > Well i've enabled all SMP options, recompiled and rebooted. The CPUs now > properly show up in dmesg and i can see the C header in top. However no > processes seem to be being assigned to cpu 1. Why is the schedueler only > using

Re: HTT on current

2003-08-22 Thread mikej
> On Fri, Aug 22, 2003 at 07:03:03PM -0400, [EMAIL PROTECTED] wrote: >> > That is not an SMP kernel. An SMP kernel (with APIC_IO) would not >> print >> > out >> > the pcib0 interrupt routing messages. >> > >> > -- >> >> So whats the problem here? > > See above. >> How come the CPUs dont show up in

Re: HTT on current

2003-08-22 Thread Kris Kennaway
On Fri, Aug 22, 2003 at 07:03:03PM -0400, [EMAIL PROTECTED] wrote: > > That is not an SMP kernel. An SMP kernel (with APIC_IO) would not print > > out > > the pcib0 interrupt routing messages. > > > > -- > > So whats the problem here? See above. > How come the CPUs dont show up in top. See abov

RE: HTT on current

2003-08-22 Thread mikej
>> That is not an SMP kernel. An SMP kernel (with APIC_IO) would not print >> out >> the pcib0 interrupt routing messages. >> >> -- > > So whats the problem here? How come the CPUs dont show up in top. How do i > realy know the system sees 2, and actually utilizes them? Is there > anything else i

RE: HTT on current

2003-08-22 Thread mikej
> That is not an SMP kernel. An SMP kernel (with APIC_IO) would not print > out > the pcib0 interrupt routing messages. > > -- So whats the problem here? How come the CPUs dont show up in top. How do i realy know the system sees 2, and actually utilizes them? Is there anything else i have to do o

RE: HTT on current

2003-08-22 Thread John Baldwin
That is not an SMP kernel. An SMP kernel (with APIC_IO) would not print out the pcib0 interrupt routing messages. -- John Baldwin <[EMAIL PROTECTED]> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/

RE: HTT on current

2003-08-22 Thread Mike Jakubik
-Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] Behalf Of Kenneth Culver > Sent: Friday, August 22, 2003 4:33 PM > To: Mike Jakubik > Cc: [EMAIL PROTECTED] > Subject: Re: HTT on current > > > > Hi, > > > > I have a HTT capa

Re: HTT on current

2003-08-22 Thread Kenneth Culver
> Hi, > > I have a HTT capabale PCU on an Intel MB with the 875P chipset. I have > enabled HTT in the BIOS and compiled my kernel with the required SMP > options, however i dont think the system is really running in SMP mode. Top > does not display CPU numbers. Here is my dmesg: > > --- > > F

HTT on current

2003-08-22 Thread Mike Jakubik
Hi, I have a HTT capabale PCU on an Intel MB with the 875P chipset. I have enabled HTT in the BIOS and compiled my kernel with the required SMP options, however i dont think the system is really running in SMP mode. Top does not display CPU numbers. Here is my dmesg: --- FreeBSD 5.1-CURR