Lets put your CPU into REAL test! Intel Linpack 64bit

[Cronos]

Seeing how the importance of proper stability testing is slowly getting more evident among XS members , i want to invite everyone to put their Intel CPUs, primarily new Core2, into really stressing stability test. Introducing the Intel Linpack benchmark, the most stressing and demanding benchmark/stability test ever!

The primary goal of this benchmark is to measure Gigaflops rating of (Intel) CPU(s). Cluster Linpack version can be used to measure performance of computational clusters, and even huge Super-Computers.
The software solves mathematical problem - arbitrary large set of linear equations, and displays the rate of solving, measured in Gigaflops.

The software can be downloaded from Intel:
http://www.intel.com/cd/software/pro...eng/363184.htm

Here is the instruction on how to properly use Linpack for stability testing:
Open the file lininput_xeon(32,64) pertaining to you OS arch., 32 bit or 64 bit. Determine the maximum problem size your memory can hold. The rule is simple : take the whole amount of memory you see in your OS, subtract 500MB for the OS needs (may need more for memory-hungry Vista), the rest can be assigned to Linpack job. In any case, do not let the test run from disk swap! The amount of memory Linpack uses is equal 8x (Problem size)^2. So, for the problem size 10000 the amount of memory would be 800MB. For 4GB memory (needs 64bit OS to see), 3.5GB can be used for Linpack test, this amounts to sqrt(3 500 000 000/8) ~ 21000 problem size.

Set 1 for #number of tests field
Set only the maximum problem size in the # problem sizes field
Set the same number in the next field #leading dimensions
Set 100-200 for # times to run a test
(It may take several hours for the benchmark to complete. Set less for quick test, set 1 for only one pass)
Leave the last #alignment value as is.

The example file for 8GB RAM under Win X64 (Linux x64 actually, but the file is the same) will look like this:

Sample Intel(R) LINPACK data file (lininput_xeon64)
Intel(R) LINPACK data
1 # number of tests
30000 # problem sizes
30000 # leading dimensions
200 #times to run a test
4 # alignment values (in KBytes)

For 4GB, substitute 30000 with 21000.
For 2GB, *--------------------------* 13700.
For 1GB, *--------------------------* 7000.

The problems size for the test may be set arbitrary small, however bigger problem size will produce higher benchmark result. Small problem size can be used to test CPU, or large problem size can be used to test both CPU ( more extensively) and all available memory.

To run the test, run runme_xeon(32,64).bat. The best way is to run it from Far manager (google Far manager and download it, its free) or similar program.
The results will be in file with the name lin_xeon(32,64).txt. The progress can be conveniently monitored in another Far instance with F3 viewer.

Here is the example part of the output file for problem size 31500:

CPU frequency: 3.744 GHz
Number of CPUs: 4
Number of threads: 4
Parameters are set to:

Number of tests : 1
Number of equations to solve (problem size) : 31500
Leading dimension of array : 31500
Number of trials to run : 100
Data alignment value (in Kbytes) : 4

Maximum memory requested that can be used = 7938634096, at the size = 31500
============= Timing linear equation system solver =================

Size LDA Align. Time(s) GFlops Residual Residual(norm)
Error: info returned = 1
31500 31500 4 446.411 46.6817 1.098303e-09 3.939015e-02
Error: info returned = 1
31500 31500 4 445.847 46.7407 1.098303e-09 3.939015e-02
Error: info returned = 1
***************************************
31500 31500 4 445.789 46.7468 1.098303e-09 3.939015e-02
Error: info returned = 1
31500 31500 4 445.792 46.7466 1.098303e-09 3.939015e-02
Error: info returned = 1
31500 31500 4 445.908 46.7343 1.098303e-09 3.939015e-02

Performance Summary (GFlops)

Size LDA Align. Average Maximal
31500 31500 4 46.7362 46.7495

End of tests
Wed Jan 30 10:39:30 NOVT 2008

The program incorrectly determine my Q6600's speed, which is 3.328GHz (8x416). This does not affect the results in any way.

For each tests ran, there will be a string with 7 fields, like this
31500 31500 4 445.908 46.7343 1.098303e-09 3.939015e-02
Never mind the Error: info returned = 1 string, this is not actually an error and harmless. The returned value may differ from 1.

The most interesting fields for us are: the fifth value, which is Gigaflops rating obtained in this run, and the seventh, which is Residual (norm) value. The last is most important part. It should be small, on the order of 10-02 (0.01),
and should not change from run to run with the same problems size. Any abnormal value here indicates
Linpack fails the test, though it will continue to work. This is very sensitive test, actually.

The 32bit version produces good load and is also good for error catching, but to really stress Core2 CPU, 64bit version should be used. It has 40-50% higher rate of solving (gives higher gigaflops rating), produces much much higher temps (considerably higher than TAT, and much higher than Prime) and is exceptionally good for error catching, both for CPU and memory errors.

[dogsx2]

It saddens me to see how XS is becoming XT(XtremeTesting).

All you guys that think a benching rig should be primed or whatever for hours on end can rip me all you want. I just hate to see the forum go this direction.

Maybe there should be a section at XS for all those who what to show how many hours of prime or whatever there rig can do. Instead of spi or 3D benching... it could be whoever can prime or do Linpack the longest has the most stable computer.

[dinos22]

i only need 1 hour prime and to use my machine for whatever i normally use it for and if it doesn't do anything stupid i'm happy with it

i also don't feel the need to really push my 24/7 hardware or have something superfast

heck i run an E6300 @ 3GHz with 8800GT
i'm happy with that

[Cronos]

Some of us do something really useful with our computers, not just benching.
I refuse to admit i am the only one here

[dinos22]

yeah 24/7 machines are not benching machines and we all have them and use them for "something else"

[Cronos]

As the most extreme CPU and memory benchmark/test available, Linpack perfectly fits into XS philosophy .

[JAG87]

this thing made my cpu 2c hotter than 8K FFT prime. not bad at all

[Cronos]

this thing made my cpu 2c hotter than 8K FFT prime. not bad at all

What "thing", exactly? 32bit or 64bit version? These are VERY different in load, benchmark results and temps.

[kiwi]

Nothing wrong with stability test but please note that most applications don't stress and heat cpu so much. So it might be an overkill

[KTE]

Don't really bother with it unless I'm running crucial equipment and testing them at work but I wish I could run it for my own testing purposes on an AMD rig.

[Cronos]

It may be useful to explain a bit more how Linpack works and is optimized for given platform. The benchmark is distributed as source code, and can be compiled and ran on any platform, including non -x86, like IBM Power G5, Cell, Intel IA64 (Itanium) so on. The code uses calls to specific mathematical library of functions, called BLAS (Basic Linear Algebra Subprograms), which contains most frequently used operations with matrices and vectors. This BLAS library is heavily platform -optimized and distributed by each CPU vendor.

For example, Intel has MKL for their x86 and Itanium CPUs,
AMD has ACML. There are also non-vendor versions which
have optimizations for different CPUs (Goto BLAS, Atlas).

Then Linpack exe file is compiled from source, it is linked with some specific version of BLAS, thus achieving high optimization level for this specific CPU.

The above Intel Linpack is compiled and linked with some BLAS version. The BLAS used to link with 32bit version is very general and should run even on older Intel CPUs, 64bit version is linked with latest version of MKL, which is very heavily optimized for latest Intel CPUs. Hence the large difference both in performance and load on Core2.

The BLAS libraries are frequently linked with application which require high performance, for example mathematical packets like Mathematica, Maple, Mathcad , some encoders/decoders, distributed projects like Folding&Home and so on.

If Linpack fails, all these applications may also fail, though it may take (significantly) more time, and may go largely unnoticed, as most CPU errors do not lead to any abnormal termination.

[sofarfrome]

i only need 1 hour prime and to use my machine for whatever i normally use it for and if it doesn't do anything stupid i'm happy with it

i also don't feel the need to really push my 24/7 hardware or have something superfast

heck i run an E6300 @ 3GHz with 8800GT
i'm happy with that

I'm with you dude...if I can fold or do the WCG and play UT2004 and UT3 without crashing then I am quite happy. I also use my rig for DL'ing and editing music/vid files and to date I have had ZERO issues with any of the tasks it is intended for. All of this Prime95 for 24hours with small ffts is crazy. The stress on the board is at least as bad as cranking the volts and doing 10min 32M Pi passes and 10s 1M Pi passes for several minutes, if not worse.

[arfett]

Every computer I've ever built I have overclocked. Every single computer would blue screen me randomly or when running certain programs if I was not at least 8 hours prime stable. It might be weekly, bi-weekly, monthly, but it would happen. I always do a minimum of 8 hours prime 95 on small fft and blend and nowdays I go for 12 hours before I call it quits or else I know at some point I'll get a blue screen or random program crash.

The stress on the board is at least as bad as cranking the volts and doing 10min 32M Pi passes and 10s 1M Pi passes for several minutes, if not worse.

Can't tell you how many times a 5 hour prime stable computer has crashed on me while playing UT2004 that was 32M Pi "stable". Even for general use you must prime or do some sort of stress testing if you want to be assured no errors/crashing. If it works for you then that's cool, but if you're overclocking and you didn't stress it then you're not stable in my opinion and at some point I'd imagine it will get you.

[dogsx2]

Every computer I've ever built I have overclocked. Every single computer would blue screen me randomly or when running certain programs if I was not at least 8 hours prime stable. It might be weekly, bi-weekly, monthly, but it would happen. I always do a minimum of 8 hours prime 95 on small fft and blend and nowdays I go for 12 hours before I call it quits or else I know at some point I'll get a blue screen or random program crash.


Can't tell you how many times a 5 hour prime stable computer has crashed on me while playing UT2004 that was 32M Pi "stable". Even for general use you must prime or do some sort of stress testing if you want to be assured no errors/crashing. If it works for you then that's cool, but if you're overclocking and you didn't stress it then you're not stable in my opinion and at some point I'd imagine it will get you.

Sorry to hear your bad luck with building computers.

[arfett]

Sorry to hear your bad luck with building computers.

Not bad luck, just overclocking too much. Drop the mhz down a bit and get a 8-12 hour run of prime and stable with no blue screens for a year. I've primed this q6600 with small fft and blend for 12 hours (in addition to some Pi 32M runs) and I can basically guarantee you I will never have a blue screen unless it's some crappy vista driver problem.

[Cronos]

Linpack64 vs. 4x TAT on Q6600@3600 1.4V, watercooled:

4x TAT:


Linpack64:


4x Linpack 64 bit - 72C
4x TAT ~Linpack 32bit - 66C
4x Prime Small FFTs - 59C
OCCT cpu test - 55C

[-n7-]

Thanx for explaining how to configure this properly.

I'd already grabbed it thanx to reading your posts about this from before, but didn't know how to set its parameters.

[aoch88]

I think Orthos does a good job ensuring stability. Well, everyone has a different intepretion of "overclocking". Some are happy when the thing can do sky high Mhz and complete Super PI but look again, it's not stable. If it's a screenshot and numbers is what you're looking at - this is only what you need to do.

To me, stability is important because I wouldn't want those pretty 5Ghz screenshot. It serves no purpose for me as I can't enjoy the speed. Hence, stability test is important. At the same time, I don't understand the point of tellign the whole world "I got my E8400 @ 6Ghz" but when it can't even run Windows Solitaire successfully

[dogsx2]

I think Orthos does a good job ensuring stability. Well, everyone has a different intepretion of "overclocking". Some are happy when the thing can do sky high Mhz and complete Super PI but look again, it's not stable. If it's a screenshot and numbers is what you're looking at - this is only what you need to do.

To me, stability is important because I wouldn't want those pretty 5Ghz screenshot. It serves no purpose for me as I can't enjoy the speed. Hence, stability test is important. At the same time, I don't understand the point of tellign the whole world "I got my E8400 @ 6Ghz" but when it can't even run Windows Solitaire successfully

Then you do not understand a big part of XS.

[aoch88]

Then you do not understand a big part of XS.

I do understand some people would spend $1,000+ for a QX chip and capture screenshot of lets say, 6Ghz?

Things eventually changes so XS today might have to cater for stable overclocks too.

[Cronos]

I guess very few can understand BOTH parts of XS at the same time

[chew*]

actually i run mine a tad slower than his and its a core 2 quad and its at stock volts and yes I would use it to solve a cancer problem on one of my own family members lives.......I however have other means of testing stability besides prime or linpax..........Having used an athlon 64 with a finicky damn mem controller and a manta board with over 200+ mem settings........intel is a piece of cake.........

specs asus maxximus formula Stock volts across the board ( overvolts ram to 2.0 )
q6600 @ 8x400 1.225v (stock vid)
corsair ram ddr2 800 5-5-5-18 (i run 4-4-4-9)

FYI the best way to test stability is a buttload of benches then games then 2 fresh installs of windows.....1 stock 1 oced..........I have a program i use to compare files/folders.......if the OS is the slightest bit different your not stable.......Installing windows OC'ed is the best way to find corruption.......

[STEvil]

Looks like this thread could use a cleaning

Guys, this is really simple to explain.

Some benchmarks require a short time period to run.
Some cars go down a 1/4 mile track in under 4 seconds.

Some benchmarks require long periods of time to run.
Some cars race for 24 hours or more at a time.

Build what you want for the desired outcome. Dont rain on everyone's parade because you dont want to build for their competing requirements.

[KTE]

Well put STEvil.

I'll try it on my C2Q in a few hours to see how much load it does put and if it's actually effective in finding instabilities. Always nice to know a setting can pass ex.high server loads on the desktop, afterall, they are testing in similar ways to CPUs are after the fabs. But it's nothing new in my testing either.

[bachus_anonym]

For now, it's just a 10 mins test run to get a feel of the benchmark... Good idea for a challenge, and I'm sure it will fit nicely into XS spirit Thanks Cronos I will add a longer run later...

A note regarding the screenshot - this is my 24/7 configuration which is normally running WCG on all 4 cylinders, used for gaming and the whole nine yards... It's cooled with Tuniq Tower 120 on low fan speed. However, Linpack bump LOAD temperatures another 10C over what WCG, and 5C over ORTHOS. I'm on Windows XP Professional 32-bit, with 4GB (3.25GB usable), therefore only 1.8GB can be utilized by this benchmark...




A side note... Please do not turn this thread into yet another "what one calls stable" debate - we've had a fair share of those and it never ends... Run the test, post a screenshot, otherwise don't derail the thread - plain and simple. Thanks for understanding guys