DivX 5.1.1 as the best Stability Test?!?

cbuchach

Golden Member
Nov 5, 2000
1,164
1
81
So I recently have upgraded to using DivX 5.1.1 in my encoding of my DVD's. I have a P4 2.8 running at 3.25 at 1.6V on air cooling which I consider a decent but by no means great overclock. My system has up until this point been absolutely stable, able to run hours and hours of dual Prime95 tests, many video encodes with DivX 5.05 all with no problems. I thought I had a good stable overclock.

So I two days ago loaded up DivX 5.1.1 and started encoding in Slowest mode. Well I came back to my machine after a few hours and it was completely locked up. I did the same thing and the machine again locked after about and hour of encoding. I tried one more time and and came back to the machine with a Virtual Dub error. Hmm, I thought that this was strange as my machine NEVER locks up and I had never had problems like this before. So last night I clocked the processor at 3006 and woke up with a perfectly completed DivX encode. SO I am now again testing to see where my machine is stable using DivX and what voltage to use. I am currently running at 3206 and all is good so far.

I guess I am glad that I found out that machine was not 100% stable, but I am a bit pissed that I might have to back down on my already modest overclock. And I have tried the Standard mode in DivX 5.1.1 and it seems to run fine, it is only when I have run the codec in Slowest mode where I have noticed the problems.

I should add that the system is running on an Abit IC7-G with a 5:4 FSB:memory ratio. The memory has been tested extensively with Memtest and has never had any errors. My temps are around 40-45C at idle and get up to around 59-60 when encoding in Slowest mode with Seti running in the background. My case temps have always been what I consider abnormally high around 30-36 in the winter and 35-42 in the summer. I have two 120 fans and a 92mm side fan blowing directly on the CPU. I think the problem is that I have my hard drive rack sitting directly in front of my front 120mm intake and this must impact airflow though the case
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,920
32,169
146
We discussed this at length a good 6-8months ago and some of us concluded the same thing; That there are better stability tests than P95TT and memtest86. I've had overclocks pass both easily only to have a FoldingatHome Gromacs work unit crash out in under an hour. Duvie and Thugsrook have had DIVX lay the same smackdown on a P95/memtest stable system. The contemporary advice is to include P95&memtest in your stability suite but also use other methods too. DIVX and Gromacs WU's are great ones, POVRAY and PiFast are no slouches either.
 

cbuchach

Golden Member
Nov 5, 2000
1,164
1
81
Actually i remember reading those threads and took their advice in initially testing my system with DivX 5.05. I thought everything was OK. It seems at least in my case that DivX 5.1.1 Slowest mode is somehow more strenous and more sensitive to overclocking for whatever reason. I knew that DivX was a much better test than P95. P95 is so useless on my system that I can clock it to 3400 and have P95 work flawlessly until the system locks up. Also I recenly was running P95 in the background to generate more heat when testing with DivX 5.11 and again it never gave me an error.

Oh, and do you have a link to Gromac's folding at home WU?
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,920
32,169
146
You'll need to join the folding team and run F@H, apply the proper switches and wait for it to give you a Gromacs, sorry but Gromacs pick you, you don't pick them :p
 

slag

Lifer
Dec 14, 2000
10,473
81
101
I use tmpgenc for stability testing. I can run my chip at 252 fsb all day long with prime 95 and a loop of 3dmark2k1 going without any problems, but I can run tmpgenc for 30 mins and lock up my system hard.

After dropping back down to 250 mhz fsb, its fine.
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,920
32,169
146
Originally posted by: slag
I use tmpgenc for stability testing. I can run my chip at 252 fsb all day long with prime 95 and a loop of 3dmark2k1 going without any problems, but I can run tmpgenc for 30 mins and lock up my system hard.

After dropping back down to 250 mhz fsb, its fine.
I used to use that with DVD2SVCD before I switched to CCE and it's definitely a tough one too :beer:
 

Duvie

Elite Member
Feb 5, 2001
16,215
0
71
Originally posted by: cbuchach
Actually i remember reading those threads and took their advice in initially testing my system with DivX 5.05. I thought everything was OK. It seems at least in my case that DivX 5.1.1 Slowest mode is somehow more strenous and more sensitive to overclocking for whatever reason. I knew that DivX was a much better test than P95. P95 is so useless on my system that I can clock it to 3400 and have P95 work flawlessly until the system locks up. Also I recenly was running P95 in the background to generate more heat when testing with DivX 5.11 and again it never gave me an error.

Oh, and do you have a link to Gromac's folding at home WU?



Actually one thing that needs to be taken into account is the fact that divx and other freeware apps often used to run these things are buggy..only go to the foruns and notice the constant revisions....Try the 5.1.0 version....I undersatnd the move away from the 5.05 codec as it is piss slow....

One thing that also needs to be mentioned is when I had my prime stable 2.4b@3.24ghz and yet it failed in Divx 5.03pro Gknot I didn't take its word only. I tried other apps like besweet audio encoding and had similar results...i had errors in complexed autocadd renderings....So I basically conrfirmed it with multiple apps...


IN this case I would try a program like POV-Ray, and even another encoder like TMPGenc 2.52 free version...If oyu have issues this cpu intensive apps will as well....

Last but not least prime95 I still don't consider a great memory tester adn thus maybe some of the issues may lie in the memory subsystem with timing, speeds, heat, stress, etc....

You may play with the ram and even boost the vdimm and vcore up a notch and if you gain stability you know the errors are likely related to your overclocking and you have an unstable system...


I trust no ONE program....I test many, and as Thugs suggest nothing like running a good gaming app cause it stresses the system probably the most complete since it has video, sound, with heavy cpu and memory usage...