[9fans] semaphore testing help

[9fans] semaphore testing help

[erik quanstrom]

unfortunately, i only have a pretty narrow selection of
rather new intel xeons and and rather old intel atoms.
if anyone with an x86-based machine would mind sending
me the output of:

	9fs sources
	/n/sources/contrib/quanstro/locktest/doalltests

that would be a great help.  output from a pentium iii would
be very helpful.

- erik

Re: [9fans] semaphore testing help

[Nick Owens]

[-- Attachment #1: Type: text/plain, Size: 51 bytes --]

is there source code available for these tests?


[-- Attachment #2: Type: application/pgp-signature, Size: 851 bytes --]

Re: [9fans] semaphore testing help

[erik quanstrom]

On Tue Sep 24 12:24:30 EDT 2013, mischief@9.offblast.org wrote:

> is there source code available for these tests?

yes, but i realized while i was making the request that they
will compile incorrectly on machines that have already upgraded
the system libraries because semaphores will be used where locks
are wanted.

i'll fix this as soon as possible.

- erik

Re: [9fans] semaphore testing help

[sl]

This is from a Thinkpad T43p:

term% sysinfo | grep cpu0
cpu0: 1995MHz GenuineIntel P6 (cpuid: AX 0x06D8 CX 0x0180 DX 0xAFE9FBFF)
term% 9fs sources
post...
term% /n/sources/contrib/quanstro/locktest/doalltests
P6 1996
1
0.25u 0.82s 3.04r 	 8.lpingpong
0.22u 2.37s 5.54r 	 8.sempingpong

0.00u 0.00s 7.89r 	 8.chantestsem -O0 -R1 -T1
0.00u 0.00s 17.28r 	 8.chantestsem -O0 -R1 -T2
0.00u 0.00s 34.79r 	 8.chantestsem -O0 -R1 -T4
0.00u 0.00s 69.91r 	 8.chantestsem -O0 -R1 -T8
0.00u 0.00s 155.70r 	 8.chantestsem -O0 -R1 -T16
0.00u 0.00s 7.28r 	 8.chantest -O0 -R1 -T1
0.00u 0.00s 14.87r 	 8.chantest -O0 -R1 -T2
0.00u 0.00s 29.70r 	 8.chantest -O0 -R1 -T4
0.00u 0.00s 61.62r 	 8.chantest -O0 -R1 -T8
0.00u 0.00s 130.01r 	 8.chantest -O0 -R1 -T16
term%

Re: [9fans] semaphore testing help

[erik quanstrom]

On Tue Sep 24 12:24:30 EDT 2013, mischief@9.offblast.org wrote:

> is there source code available for these tests?

/n/sources/contrib/quanstro/locktest.

i've discovered that machines with 24 or more cores are
particularly interesting.

- erik

Re: [9fans] semaphore testing help

[erik quanstrom]

On Tue Sep 24 22:54:57 EDT 2013, sl@9front.org wrote:
> This is from a Thinkpad T43p:
>

thanks!  interesting processor.  iirc that was the first one from
the group that got us off the netburst arch.

- erik

Re: [9fans] semaphore testing help

[BurnZeZ]

[-- Attachment #1: Type: text/plain, Size: 43 bytes --]

I don't have a p3, but I hope this helps.

[-- Attachment #2: Type: text/plain, Size: 1149 bytes --]

term% aux/cpuid
vendor AuthenticAMD
procmodel 00200f31 / 01020800
features fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat
features pse36 clflush mmx fxsr sse sse2 ht
features pni cx16
extmodel 00200f31 / 20001400
extfeatures fpu vme de pse tsc msr pae mce cx8 apic syscall mtrr pge mca cmov pat
extfeatures pse36 nx mmx+ mmx fxsr ffxsr tscp lm 3dnow!+ 3dnow!
extfeatures ahf64 cmp svm eas cr8d 3dnow!p osvw skinit
procname AMD Athlon(tm) X2 Dual-Core QL-64
physbits 40
virtbits 48

term% doalltests
fcp: threadsem is a directory
fcp: threadtas is a directory
unknown 2100
2
0.30u 0.20s 0.50r 	 8.lpingpong
0.44u 3.68s 6.61r 	 8.sempingpong

0.00u 0.00s 10.85r 	 8.chantestsem -O0 -R1 -T1
0.00u 0.00s 24.62r 	 8.chantestsem -O0 -R1 -T2
0.00u 0.00s 64.22r 	 8.chantestsem -O0 -R1 -T4
0.01u 0.00s 146.50r 	 8.chantestsem -O0 -R1 -T8
0.00u 0.00s 324.50r 	 8.chantestsem -O0 -R1 -T16
0.00u 0.00s 9.80r 	 8.chantest -O0 -R1 -T1
0.00u 0.00s 19.33r 	 8.chantest -O0 -R1 -T2
0.00u 0.01s 39.26r 	 8.chantest -O0 -R1 -T4
0.00u 0.00s 82.72r 	 8.chantest -O0 -R1 -T8
0.01u 0.00s 194.29r 	 8.chantest -O0 -R1 -T16

[-- Attachment #3: Type: text/plain, Size: 1150 bytes --]

neinchan% aux/cpuid
vendor GenuineIntel
procmodel 000206a7 / 00100800
features fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat
features pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe
features pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 popcnt
features tscdeadline aes xsave avx
features xsaveopt
extmodel 00000000 / 00000000
extfeatures tscp lm
extfeatures ahf64
procname Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz
physbits 36
virtbits 48

neinchan% doalltests
fcp: threadsem is a directory
fcp: threadtas is a directory
P6 3397
4
0.21u 0.08s 0.26r 	 8.lpingpong
0.31u 1.68s 3.72r 	 8.sempingpong

0.00u 0.00s 4.97r 	 8.chantestsem -O0 -R1 -T1
0.00u 0.00s 7.06r 	 8.chantestsem -O0 -R1 -T2
0.00u 0.00s 28.17r 	 8.chantestsem -O0 -R1 -T4
0.00u 0.00s 65.24r 	 8.chantestsem -O0 -R1 -T8
0.00u 0.00s 144.70r 	 8.chantestsem -O0 -R1 -T16
0.00u 0.00s 4.82r 	 8.chantest -O0 -R1 -T1
0.00u 0.00s 6.60r 	 8.chantest -O0 -R1 -T2
0.00u 0.00s 12.63r 	 8.chantest -O0 -R1 -T4
0.00u 0.00s 27.02r 	 8.chantest -O0 -R1 -T8
0.00u 0.00s 60.55r 	 8.chantest -O0 -R1 -T16

Re: [9fans] semaphore testing help

[erik quanstrom]

On Tue Sep 24 23:19:26 EDT 2013, brz-systemd-dev@intma.in wrote:

> I don't have a p3, but I hope this helps.

very useful.

- erik