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

Quote:

---

Originally Posted by 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.


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.

---

I had a rig (an E6600 at 3.71GHz) that passed prime95 for 8+ hours while looping 3DMark03 in the back ground that would crash after 10 minutes of UT2004. Even though I passed all acceptable stability tests I still needed a little more vcore to keep it stable when running UT2004. I am done killing my boards. 1 hour OCCT or P95 while looping a 3DMark program and it is off to some folding or gaming for the real test. Anyone that runs 24hours+ of P95 or Orthos to gain confidence in the OC is shortening the life of the board and other components and nothing more. What task does a PC perform that utilizes 100% CPU resources?

Quote:

---

Originally Posted by 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 :up: 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...

http://www.clockmehigh.com/cooler/sc...inpack_8x5.png


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 :yepp:

---

mint... you play PES, I believe everything you say.


anyway, this proggy seems to produce a heck of a lot of heat, but I don't know how well it can find small errors like prime does. one thing is for sure, if you are not stable, running linpack is sure to reboot your computer. but if you are very close to being stable, prime would give you a rounding errors, I dont know about linpack...

Im just happy if mine turns on....

Lol

Nah, I find if I can get through 1 batch of large FFT's and one of small FFT's then it will never ever crash no many how many things I have running in the background. I'm not going to bother with this thing.

Quote:

---

Originally Posted by 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 :D with 8800GT
i'm happy with that :)

---

Hell, I don't even do that. If it Small FFT's for 5 minutes I'm set. I iron out the issues as they come along, if any come along. I'm usually good @ getting it stable first try.

Quote:

---

Originally Posted by Cronos

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.

---

Just put your statements here to the test and I was hoping you weren't going to sway them in any way to favor Linpack. The following part isn't true at all:

"...produces much much higher temps (considerably higher than TAT, and much higher than Prime)"

http://img230.imageshack.us/img230/7...litynf1.th.png http://img221.imageshack.us/img221/2...ytatfp5.th.png

I tried this (10minutes each/10 trials) on my stock 1.2V Q6600 2.4G. Problem with Linpack is, the highest temps were there for a slight millisecond only, most of the temps/load was fluctuating and low compared to the rest of the apps producing consistent static loads. 32bit, but that's what I use day to day and that's all that counts to me.

Testing load (maximum seen):

EV FPU Julia: 165W AC (consistent)
P95 Small FFT: 165W AC (consistent)
P95 Large FFT: 177W AC (consistent)
Intel TAT: 179W AC (consistent)
Linpack: 162-176W AC (millisecond fluctuations)

Testing load temps (maximum seen in °C):

EV FPU Julia: 39 37 37 39 (consistent)
P95 Small FFT: 46 43 43 46 (consistent)
P95 Large FFT: 45 42 42 45 (consistent)
Intel TAT: 50 46 46 50 (consistent)
Linpack: 46 45 45 46 (millisecond fluctuations)

So Prime/TAT are extremely close if not better stability testers for C2Q, higher load and TAT specifically is higher load and produces higher temps.

I just finished "100x" run, which took little over 2 hours to complete. Still, only WinXP Pro 32-bit and 4GB RAM...

http://www.clockmehigh.com/cooler/sc...0x_resized.png

From what I noticed, my system running Linpack pulls 480W from DC outlet, which is:

- 12% more power than Prime95 LargeFFT and WCG (a distributed computing project I run 24/7 on this system),
- 8% more power than Prime95 SmallFFT.

Quote:

---

Originally Posted by JAG87

mint... you play PES, I believe everything you say

---

Hehe, PES 2008 rocks :up:

pretty toasty temps there Michal

how long has that chip been running at full blast EEK

Quote:

---

Originally Posted by dinos22

pretty toasty temps there Michal

how long has that chip been running at full blast EEK

---

Hey dino! This system has been running WCG 100%/24/7 @ 3.6GHz for over 3 months straight now. It's also my gaming and everyday PC. I might have rebooted it maybe 3-5 times during that time, for usual maintenance (new drivers, updates, etc). Those temps are not "normal". As I mentioned above, Linpack heats that quad pretty good, 10-12C over WCG running full blast or gaming.

oh i see what you mean

damn this software really does stess it hard

WEEEEEEEEEEEL to the OP and others
it seems half an hour of this stuff and you're good to go hahahah

i also do some mandatory superpi of course :D

Just ran the 64-bit version for 30mins, problem sizes 12,000....

Temps: 47 45 45 47
Power: 178-190W AC

:confused:

Seems to me much better at inducing ex.high core loads although the temps are no different here.

Quote:

---

Originally Posted by NickS

Hell, I don't even do that. If it Small FFT's for 5 minutes I'm set. I iron out the issues as they come along, if any come along. I'm usually good @ getting it stable first try.

---

Here here; I think you can get the overclock 95% stable in 5 - 15 minutes. Then, address any problems that come up later.

Now for the "extreme" overclocking, that last 5%, is there really any true stability achieved?

In other words, if you think it is 100% stable, then is it really "extreme" overclocking? .....Of coarse not.

Can a liquid nitrogen system be considered stable?....

Quote:

---

Originally Posted by KTE

Just ran the 64-bit version for 30mins, problem sizes 12,000....

Temps: 47 45 45 47
Power: 178-190W AC

:confused:

Seems to me much better at inducing ex.high core loads although the temps are no different here.

---

Your results are not consistent with my own and others i have seen.
May i suggest you increase the problem size and try to run it on overclocked CPU?
Your CPU power at this speed/voltage combination is ~ 2x times less than mine at 3.6GHz/1.4V,
this may explain low temps. and low temps. difference. Besides, to get more even load, you have to use bigger problem size.
I also hope there was no swaping activity during the test

Quote:

---

Originally Posted by bachus_anonym

I just finished "100x" run, which took little over 2 hours to complete. Still, only WinXP Pro 32-bit and 4GB RAM...

---

Sorry to disappoint you, but looking at those results, Linpack fails. You have at least one abnormal value for Residual(norm).

Quote:

---

Originally Posted by Cronos

Your results are not consistent with my own and others i have seen.
May i suggest you increase the problem size and try to run it on overclocked CPU?
Your CPU power at this speed/voltage combination is ~ 2x times less than mine at 3.6GHz/1.4V,
this may explain low temps. and low temps. difference. Besides, to get more even load, you have to use bigger problem size.
I also hope there was no swaping activity during the test

---

Yeah your results are not matching my Q6600 results. I would've tried higher MHz mainly but I switched the PC on after a week and it failed to oc even 1MHz. :(
Coldboot issues, very oddly. Needed to bench 2hrs before I could get back and set lower volts for 3.6G again and it now needs much higher volts for stability than 2 weeks back. Volts the cooling I have spare at 30C ambient can't handle. :confused:

I gave it a quick run anyway.. it does detect instability pretty good though, so far on par with P95. :)

RAM size, 1.48GB taken up is enough of 2GB frankly, I had applications open I needed which are memory hogs (even Firefox).

TAT looked like this running:

http://img146.imageshack.us/img146/355/tatwq2.th.png

Linpack running and output file was like this:

http://img103.imageshack.us/img103/4226/linrx3.th.png

http://img352.imageshack.us/img352/5853/linpww2.png

I stopped it at that, gave it the same time as others. Will run it for longer later when system is free.

Idle (AC / °C):
126W - 35 30 30 35

Load (AC / °C):
259W - 67 62 62 67 - P95 small FFT
263W - 68 62 62 68 - P95 large FFT
258-278W 73 69 69 73 - Linpack
293W - 75 71 71 75 - Intel TAT

This time the differences were larger between them, especially Large FFT vs the rest and TAT was considerable higher load and temps.

KTE
You said you tested Linpack64, i.e. you have Windows 64bit. Is there any reason you again testing under Windows 32bit ?

Set problem size 13700 and run Linpack64 for several hours.

Just noticed, your results are very weird anyway. All Residual(norm) values are large ( indicates all tests failed),
but the most strange thing is they all the same! I can't explain this...

Quote:

---

Originally Posted by sofarfrome

I had a rig (an E6600 at 3.71GHz) that passed prime95 for 8+ hours while looping 3DMark03 in the back ground that would crash after 10 minutes of UT2004. Even though I passed all acceptable stability tests I still needed a little more vcore to keep it stable when running UT2004. I am done killing my boards. 1 hour OCCT or P95 while looping a 3DMark program and it is off to some folding or gaming for the real test. Anyone that runs 24hours+ of P95 or Orthos to gain confidence in the OC is shortening the life of the board and other components and nothing more. What task does a PC perform that utilizes 100% CPU resources?

---

Gulp. Is this 100% accurate? I understand while stress testing you are obviously putting the system under stress (duh), but are motherboards/ram/cpu that fragile that I should be worried about them giving up the ghost from putting them under extreme pressure? I am worried now as I have HOT ambient temps of 30c, and have ran my system for weeks at 35+h straight runs under prime95/OCCT etc. Temps were quite high across the board for those long hours. Though I have not noticed any issues from using my system, I am worried I may have done some damage after reading the above statement. :confused:

Quote:

---

Originally Posted by cantankerous

Gulp. Is this 100% accurate? I understand while stress testing you are obviously putting the system under stress (duh), but are motherboards/ram/cpu that fragile that I should be worried about them giving up the ghost from putting them under extreme pressure? I am worried now as I have HOT ambient temps of 30c, and have ran my system for weeks at 35+h straight runs under prime95/OCCT etc. Temps were quite high across the board for those long hours. Though I have not noticed any issues from using my system, I am worried I may have done some damage after reading the above statement. :confused:

---

No, he's just been burned by buying crap components and expecting too much from them.

The reason I was irritated after hearing it stated that Linpack 64 was "very sensitive" at detecting memory and cpu instability, is I did not find that to be true.

problem: BSOD's and reboots after 1-3 hours of orthos large FFT's and seemingly random stops, reboots in small ffts.

Linpack 64, runs of 50 to 100 trials of problem size 10000, which takes 1.3 to 3 hours, never an error of any kind, all residual norms same.

Memtest tells me the problem in 2 minutes. RAM is bad, and unstable at any settings.

I dont think I would call Linpack a sensitive test for memory, when 1.3hrs plus shows no error, and memtest tells me in 2 minutes.

To me a real machine overclocked is not some benchy numbers ..

It's a machine that can run 24/7 as well as it's stock speed with 4 GB/ 2 GB minimum ram liquid or air cooled and running Simulation like FSX or others games or heavy calculation softwares with out crashing..

The rest is just pictures in the wall or tests to maximum possible strengh harware to know teh limit and finally make a machine stable not 10 seconds ..

I still say windows itself is the best test......BSOD and corruption if not stable and this is amplified moreso in vista....Also its a good idea to find your maxx "stable" then downclock....since this is my only rig atm and i don't have a benching rig persay.....it did 3350 @ stock voltage which is the max i'm willing to go due to heat and i need this pc for the long haul.....clocked it down to 3.2 because I like round #rs and it gives me a cushion in the stability range.....fwiw I just bought crysis last night...I have a q6600, 2 gigs ram, x1900xt flashed xtx and vista premium so basically its handcapped for the game.....I played it from 9pm till 4am then i crashed due to exhaustion......woke up pc was still running crysis.......

Give me a little while Cronos. I'll retest but this chip is bad oc for air. Temps are going to sky rocket with any tool. :(

Quote:

---

Originally Posted by rge

The reason I was irritated after hearing it stated that Linpack 64 was "very sensitive" at detecting memory and cpu instability, is I did not find that to be true.

problem: BSOD's and reboots after 1-3 hours of orthos large FFT's and seemingly random stops, reboots in small ffts.

Linpack 64, runs of 50 to 100 trials of problem size 10000, which takes 1.3 to 3 hours, never an error of any kind, all residual norms same.

Memtest tells me the problem in 2 minutes. RAM is bad, and unstable at any settings.

I dont think I would call Linpack a sensitive test for memory, when 1.3hrs plus shows no error, and memtest tells me in 2 minutes.

---

i always run superpi as well in windows

Spi is not fullproof but you can't trust any of these programs as per your example either

best stability test really is some simple orthos/spi stuff and then user trial.....if you use it for a month and no issues with whatever you use it for that's a stable machine right there heh

Quote:

---

Originally Posted by rge

I dont think I would call Linpack a sensitive test for memory, when 1.3hrs plus shows no error, and memtest tells me in 2 minutes.

---

You tested half of the memory with Linpack 32bit.
Try bigger problem size and run Linpack 64bit.

Linpack is not a substitute for memtest86+ anyway. You always should test your memory with memtest for at least 10+ complete passes.

There is something not right in your way of thinking....

Quote:

---

Originally Posted by STEvil

No, he's just been burned by buying crap components and expecting too much from them.

---

I sure hope this is the case. I would like to think the parts in my sig are of good quality and that they could in fact hold up to the many long weeks of extreme temps I was submitting them too. I hate dead parts, not a fun time to deal with them.

I take a moderate stand between those who say you need 12+ hours prime, to those who prime for 5 minutes or post meaningless ;) superpi shots that they'd never in their life run 24/7.. usually if it can prime 3-5 hours I consider it stable... if I have problems down the road, I'll prime it longer or reprime it.. usually doesn't happen.