New linpack from intel

A

azrael.arach

Limp Gawd
Joined
Feb 17, 2011
Messages
152
Intel released newest linpack. Holy crap huge gain in Gflops but wow the temps get high.

http://registrationcenter.intel.com/irc_nas/3234/w_lpk_p_11.0.5.009.zip

To run this under the dist folder there is a libiomp5md.dll that you need to put in the same direcetory as linpack under benchmarks.

Here is my results. I5 4670K at 4.4ghz
Size LDA Align. Time(s) GFlops Residual Residual(norm) Check
10000 10000 4 3.840 173.6707 9.436774e-011 3.327502e-002 pass
10000 10000 4 3.809 175.0745 9.436774e-011 3.327502e-002 pass
15000 15000 4 12.193 184.5728 2.169435e-010 3.416896e-002 pass
15000 15000 4 12.160 185.0664 2.440928e-010 3.844503e-002 pass
20000 20000 4 29.110 183.2389 3.504283e-010 3.102058e-002 pass
20000 20000 4 27.449 194.3278 3.504283e-010 3.102058e-002 pass
25000 25000 4 53.033 196.4415 5.569678e-010 3.167276e-002 pass
 
Good, finally I can go back to using/trusting a decent program to do burn tests, and not Prime95. Thanks.
 
azrael.arach said:
No other stress test got me over 90c, this new linpack I hit 99c quick.
Click to expand...

Yeah, well when everyone's recommending a stress test utility made in 1989 (slight exaggeration) of course it's not going to stress the chip enough compared to a custom modification of software (that's better to begin with) by Intel to specifically utilize as much of their chips as possible. Glad to hear this one works well for you also... I'll give it a shot on my home machine tonight as I've been having some stability issues that have been difficult to pin down.
 
Last edited:
Thanks for sharing. :)


Is this for any particular OS?
 
Last edited:
SonDa5 said:
The LinX link that I posted is really easy to use. Did you try it?
Click to expand...

Yup! crashed after 30 seconds at 4.5 @ 1.325v. Back to 4.4 @ 1.28v. 85/91/89/85 swiftech h20 delid. 1.28 crashed after 1 run. No BSOD, Upping the Volts.

After a few more runs I'm not even stable at stock :(

Everything default in Bios including RAM to 1600mhz. Stops in between #1 and #2 around 3 minutes

I need to now know why my system isn't stable: CPU/Mobo/Ram?

Intel(R) LINPACK 64-bit data - LinX 0.6.4

Current date/time: Tue Jul 09 21:36:30 2013

CPU frequency: 3.898 GHz
Number of CPUs: 1
Number of cores: 4
Number of threads: 8

Parameters are set to:

Number of tests: 1
Number of equations to solve (problem size) : 30587
Leading dimension of array : 30600
Number of trials to run : 20
Data alignment value (in Kbytes) : 4

Maximum memory requested that can be used=3193346400, at the size=30587

=================== Timing linear equation system solver ===================

Size LDA Align. Time(s) GFlops Residual Residual(norm) Check
30587 30600 4 136.645 139.6264 7.788427e-010 2.962937e-002 pass
 
Last edited:
rfisher983 said:
Yup! crashed after 30 seconds at 4.5 @ 1.325v. Back to 4.4 @ 1.28v. 85/91/89/85 swiftech h20 delid. 1.28 crashed after 1 run. No BSOD, Upping the Volts.

After a few more runs I'm not even stable at stock :(

Everything default in Bios including RAM to 1600mhz. Stops in between #1 and #2 around 3 minutes

I need to now know why my system isn't stable: CPU/Mobo/Ram?

Intel(R) LINPACK 64-bit data - LinX 0.6.4

Current date/time: Tue Jul 09 21:36:30 2013

CPU frequency: 3.898 GHz
Number of CPUs: 1
Number of cores: 4
Number of threads: 8

Parameters are set to:

Number of tests: 1
Number of equations to solve (problem size) : 30587
Leading dimension of array : 30600
Number of trials to run : 20
Data alignment value (in Kbytes) : 4

Maximum memory requested that can be used=3193346400, at the size=30587

=================== Timing linear equation system solver ===================

Size LDA Align. Time(s) GFlops Residual Residual(norm) Check
30587 30600 4 136.645 139.6264 7.788427e-010 2.962937e-002 pass
Click to expand...

yeah for me the same problem.. but it stop due to error between minutes 10 - 10:30 at stock configuration. with the extra 4 bins oc'd to turbo [email protected] it stop the run between minutes 9:30 - 9:50 due to error..
 
Araxie said:
yeah for me the same problem.. but it stop due to error between minutes 10 - 10:30 at stock configuration. with the extra 4 bins oc'd to turbo [email protected] it stop the run between minutes 9:30 - 9:50 due to error..
Click to expand...

Better than me! I cant last over 3 minutes. It's not the ram, swapped another pair. I'm guessing CPU. Did you delid?
 
For those using a 4770K and the Linpack program, it won't stress your CPU correctly. The first 4 threads are being stressed and the other 4 are left at idle during the various modes.
 
SolidBladez said:
For those using a 4770K and the Linpack program, it won't stress your CPU correctly. The first 4 threads are being stressed and the other 4 are left at idle during the various modes.
Click to expand...



The Gflops and heat on the 4 cores is still greater than any other benchmark I have seen. Loooks like a work in progress. Impressive to see such high Gflop though. Mine went over 200.
 
SonDa5 said:
The Gflops and heat on the 4 cores is still greater than any other benchmark I have seen. Loooks like a work in progress. Impressive to see such high Gflop though. Mine went over 200.
Click to expand...

True. I did notice a 10°C increase over Prime95 which only utilizes AVX. I just wish it would stress all the cores/threads. Guess we have to wait for another update.
 
SolidBladez said:
For those using a 4770K and the Linpack program, it won't stress your CPU correctly. The first 4 threads are being stressed and the other 4 are left at idle during the various modes.
Click to expand...

My little CPU monitoring app shows all 8 threads active, as does Windows Task Manager. Did you check the settings for the number of threads to use?

linpack.jpg
 
Forceman said:
My little CPU monitoring app shows all 8 threads active, as does Windows Task Manager. Did you check the settings for the number of threads to use?
Click to expand...


Are you able to get it work with LinX? I couldn't get it to work. It would have an error after 1 test.


If you got it to work please share what you did to get it to work. :)
 
SonDa5 said:
Are you able to get it work with LinX? I couldn't get it to work. It would have an error after 1 test.


If you got it to work please share what you did to get it to work. :)
Click to expand...

LOL,why were you posting a link to the file in the other thread proclaiming its dominance as the haswell testing tool if you hadnt run it?!
 
Hamateer said:
LOL,why were you posting a link to the file in the other thread proclaiming its dominance as the haswell testing tool if you hadnt run it?!
Click to expand...




I ran it and I got an error right away and I thought at first my over clock was unstable but I was blown away by the 200+ GFLOP output from the first test and I got excited and wanted to share. :D
I edited my posts. :D



Anyhow... It's still the best stress test that know of to date and it was designed by Intel and uses the AVX2.

I still don't know how to get it to work with LinX or Intel Burn Test but I have figured out how to get it to work from the download on the first post.

This benchmark is the only true standard to measure stability and performance for Haswell CPUs.
 
Last edited:
You have to add libiomp5md.dll file from the compiler folder to the linpack benchmark folder.

You go into the linpack benchmark folder and click runme_xeon64 (for 64 bit OS) and the benchmark will start. Open up a temp monitor app before hand. I use real Temp. It will appear nothing is happening but you can see the load jump all over the place and temps shoot all over the place. After about 10 minutes or so you will see the temp monitor idle with no load and the benchmark is over.

In the benchmark linpack folder you can open a text file titled win_xeon64 to see the results.

There is a file titled xhelp in the benchmark linpack folder and you can open it with MS notepad. Lots of information in there about configuring it. I'm still learning.
 
Last edited:
SonDa5 said:
You have to add libiomp5md.dll file from the compiler folder to the linpack benchmark folder.

You go into the linpack benchmark folder and click runme_xeon64 (for 64 bit OS) and the benchmark will start. Open up a temp monitor app before hand. I use real Temp. It will appear nothing is happening but you can see the load jump all over the place and temps shoot all over the place. After about 10 minutes or so you will see the temp monito idle with no load and the benchmark is over.

In the benchmark linpack folder you can open a text file titled win_xeon64 to see the results.

There is a file titled xhelp in the benchmark linpack folder and you can open it with MS notepad. Lots of information in there about configuring it. I'm still learning.
Click to expand...

thx i was able to run too.. i made for fun with my humble i7 2600 with HT OFF.. you double my flops XD my max 110GFLOPS and your max 226FLOPS huge diference..

hnw8.png




http://valid.canardpc.com/2859332
 
From Intel's xhelp file....


LINPACK makes use of SMP parallelism and by default generates as many
threads of execution as the number of the processors of the SMP system. One can
override the default number of threads by setting the environment variable
OMP_NUM_THREADS to any positive number. Note that performance is usually
maximized when the number of the threads is equal to the number of cores (so,
for a two socket dual-core system, 4 would be ideal.)
Click to expand...


So to get the maximum performance testing of this benchmark from a 4770k it is imperative to turn off Hyper Threading. If you don't turn off HT the benchmark will not stress the cpu on all 4 cores.
 
Araxie said:
thx i was able to run too.. i made for fun with my humble i7 2600 with HT OFF.. you double my flops XD my max 110GFLOPS and your max 226FLOPS huge diference..
Click to expand...

Thanks for the info. I see you have HT disabled. Cherry on top would be to see how your core temps compare. :)
 
SonDa5 said:
Thanks for the info. I see you have HT disabled. Cherry on top would be to see how your core temps compare. :)
Click to expand...

the temps diference are huge too i got: via core temp RC5: 56/61/60/57, room temp 27C... tomorrow I remake the test including core temp in the screen now im bit tired here 3:15am..
 
SonDa5 said:
From Intel's xhelp file....





So to get the maximum performance testing of this benchmark from a 4770k it is imperative to turn off Hyper Threading. If you don't turn off HT the benchmark will not stress the cpu on all 4 cores.
Click to expand...

You DO NOT need to turn off HT man..we went through this when I posted my IBT numbers, and my 3770K's Gflops were lower then your 3570Ks..I talked to IdontCare @ Anandtech, and he said you have to manually set the core affinity before you start the test to Cores, 0, 2, 4, 6,..that will lock the tests to the physical cores and will stress the cpu to its maximum..Try it yourself and see..
 
ccityinstaller said:
You DO NOT need to turn off HT man..we went through this when I posted my IBT numbers, and my 3770K's Gflops were lower then your 3570Ks..I talked to IdontCare @ Anandtech, and he said you have to manually set the core affinity before you start the test to Cores, 0, 2, 4, 6,..that will lock the tests to the physical cores and will stress the cpu to its maximum..Try it yourself and see..
Click to expand...


When you do that you are manually turning off HT for that instance of the application running.

Intel stated it clear in the xhelp file. Linpack runs best with 1 thread per core. 4770k has only 4 cores.

I run this from directory restore boot in windows 7 to limit other applications running. Ideally you want the most processing power directed to this benchmark to get the highest Gflop performance and heat.
 
Lulz at the guys who can't even run this thing stable at stock. Knowing how bad my sample is, I'm guessing I won't be able to either. I guess I should try that as well before I keep progressively lowering my overclock until I eventually am back to stock and realize I wasn't even stable in the first place! :rolleyes:
 
If you want to run it properly on a 4 core cpu with HT enabled then change the contents of "runme_xeon64.bat" to the following:
Code: 
@echo off
SETLOCAL

rem Setting up affinity for better threading performance
set KMP_AFFINITY=nowarnings,compact,granularity=fine
rem Setting path to OpenMP library
set PATH=..\..\..\redist\intel64\compiler;%PATH%

echo Running linpack_xeon64.exe. Output can be found in win_xeon64.txt.
start /b /affinity 55 linpack_xeon64.exe lininput_xeon64 > win_xeon64.txt

echo.
echo When this window closes the calculation is done.

ENDLOCAL

The important piece was adding "start /b /affinity 55", the rest is mostly cosmetical.

Edit: I might aswell paste the numbers i got running it, i7-4770K (HT on) @ 4200MHz, ram at 1600MHz, only touched the multiplier on the CPU. Running fixed Vcore at 1.215V, Noctua DH14, max temp: 93 degrees Celsius.
Code: 
Intel(R) Optimized LINPACK Benchmark data

Current date/time: Wed Jul 10 16:04:37 2013

CPU frequency:    4.194 GHz
Number of CPUs: -1
Number of cores: 1
Number of threads: 4

Parameters are set to:

Number of tests: 12
Number of equations to solve (problem size) : 1000  2000  3000  4000  5000  10000 15000 20000 25000 30000 35000 40000
Leading dimension of array                  : 1000  2000  3000  4000  5000  10000 15000 20000 25000 30000 35000 40000
Number of trials to run                     : 4     4     4     4     4     2     2     2     2     1     1     1    
Data alignment value (in Kbytes)            : 4     4     4     4     4     4     4     4     4     4     4     4    

Maximum memory requested that can be used=4210869504, at the size=40000

=================== Timing linear equation system solver ===================

Size   LDA    Align. Time(s)    GFlops   Residual     Residual(norm) Check
1000   1000   4      0.006      113.2085 1.194739e-012 4.074366e-002   pass
1000   1000   4      0.006      112.8564 1.194739e-012 4.074366e-002   pass
1000   1000   4      0.006      117.0292 1.194739e-012 4.074366e-002   pass
1000   1000   4      0.005      129.1411 1.194739e-012 4.074366e-002   pass
2000   2000   4      0.044      121.9598 4.536926e-012 3.946570e-002   pass
2000   2000   4      0.042      128.0727 4.536926e-012 3.946570e-002   pass
2000   2000   4      0.042      128.5855 4.536926e-012 3.946570e-002   pass
2000   2000   4      0.042      128.6219 4.536926e-012 3.946570e-002   pass
3000   3000   4      0.129      139.9257 8.334888e-012 3.209566e-002   pass
3000   3000   4      0.126      143.0048 8.334888e-012 3.209566e-002   pass
3000   3000   4      0.131      137.8053 8.334888e-012 3.209566e-002   pass
3000   3000   4      0.132      136.6785 8.334888e-012 3.209566e-002   pass
4000   4000   4      0.279      152.9169 1.519912e-011 3.312792e-002   pass
4000   4000   4      0.275      155.5023 1.519912e-011 3.312792e-002   pass
4000   4000   4      0.275      155.4596 1.519912e-011 3.312792e-002   pass
4000   4000   4      0.274      155.7175 1.519912e-011 3.312792e-002   pass
5000   5000   4      0.531      156.9110 2.471656e-011 3.446525e-002   pass
5000   5000   4      0.512      162.8201 2.471656e-011 3.446525e-002   pass
5000   5000   4      0.509      163.9433 2.471656e-011 3.446525e-002   pass
5000   5000   4      0.514      162.2297 2.471656e-011 3.446525e-002   pass
10000  10000  4      3.541      188.3203 9.436774e-011 3.327502e-002   pass
10000  10000  4      3.553      187.6777 9.436774e-011 3.327502e-002   pass
15000  15000  4      11.374     197.8607 2.169435e-010 3.416896e-002   pass
15000  15000  4      11.378     197.7890 2.169435e-010 3.416896e-002   pass
20000  20000  4      26.093     204.4265 3.504283e-010 3.102058e-002   pass
20000  20000  4      26.077     204.5550 3.504283e-010 3.102058e-002   pass
25000  25000  4      51.053     204.0617 5.194889e-010 2.954147e-002   pass
25000  25000  4      50.811     205.0333 5.194889e-010 2.954147e-002   pass
30000  30000  4      86.056     209.1864 7.133177e-010 2.811906e-002   pass
35000  35000  4      135.915    210.3207 1.085449e-009 3.150893e-002   pass
40000  40000  4      202.414    210.8051 1.338275e-009 2.976370e-002   pass

Performance Summary (GFlops)

Size   LDA    Align.  Average  Maximal
1000   1000   4       118.0588 129.1411
2000   2000   4       126.8100 128.6219
3000   3000   4       139.3536 143.0048
4000   4000   4       154.8991 155.7175
5000   5000   4       161.4760 163.9433
10000  10000  4       187.9990 188.3203
15000  15000  4       197.8248 197.8607
20000  20000  4       204.4907 204.5550
25000  25000  4       204.5475 205.0333
30000  30000  4       209.1864 209.1864
35000  35000  4       210.3207 210.3207
40000  40000  4       210.8051 210.8051

Residual checks PASSED

End of tests
 
Last edited:
SonDa5 said:
Thanks for the info. I see you have HT disabled. Cherry on top would be to see how your core temps compare. :)
Click to expand...

here are a new test, this time my room temp was 25C results.. almost the same, i never saw another thing that push my temps to that numbers while HT OFF, thats like 5-7C over P95.. few days ago i made a p95 27.7 Run same clock with HT on for 30min in a 22C ambient room and my hottest core was 50C. O.O.

2k5.png
 
SonDa5 said:
Are you able to get it work with LinX? I couldn't get it to work. It would have an error after 1 test.
Click to expand...

same here, always an error after every test (if you uncheck 'stop on error'). tested on multiple systems, some that aren't even overclocked.

probably a bug between linx and the latest version of linpack.
 
my little i5-2500K @4.3Ghz chugging thru a 20x run (temps pretty good considering ambient temps 25C ) cooler hyper 212+ w/ Push /Pull Gelid Silent 12 PWM fans

 
HMB said:
If you want to run it properly on a 4 core cpu with HT enabled then change the contents of "runme_xeon64.bat" to the following:

*snip*
Click to expand...

Thanks! Here's my results:

i1wjh4.jpg


Note this is with my 4770K under custom water cooling with a quad rad. :eek:
 
SolidBladez said:
Thanks! Here's my results:

Note this is with my 4770K under custom water cooling with a quad rad. :eek:
Click to expand...

are you using custom problem size?.. check your SIZE and LDA are much lower than rest of test posted here O_O..
 
You must log in or register to reply here.
Back
Top