The Prime 95 fallacy

[lookin4dlz]

I've been running my computer 24x7 for weeks. Daily I run a very intensive 8MB Excel spreadsheet that crunches statistics on its content for a continuous 1.5 hours - even multi-tasking during this with games, etc. has no effect on stability.

So, what gives? Are certain chips just less stable than others so that when using them P95 actually does predict stability?

 

[big4x4]

I have never used prime.....I just loop tests 1 and 3 on 3dmark 01. If that does not crash or freeze, its stable! I have never had a problem with this method!

 

[fishmonger12]

try clocking back your hardware to stock speeds then running prime 95.

 

[cbehnken]

If it's not stable running Prime95, it's not stable. The only reason Prime95 fails is because your computer cannot calculate numbers as accurately as they should be.

 

[xbassman]

Originally posted by: cbehnken
If it's not stable running Prime95, it's not stable. The only reason Prime95 fails is because your computer cannot calculate numbers as accurately as they should be.

agreed...
I believe the only reason your computer should ever crash is due to bad programming!

 

[rogue1979]

If you can't run Prime95, something in your computer is not stable, most likely the cpu or memory.

It may be stable enough for all other applications including gaming, but it still isn't right.

It will come back to nip you in the butt at a later time if you neglect this......

 

[big4x4]

Originally posted by: rogue1979
If you can't run Prime95, something in your computer is not stable, most likely the cpu or memory.

It may be stable enough for all other applications including gaming, but it still isn't right.

It will come back to nip you in the butt at a later time if you neglect this......

I don't buy into that prime stuff. I used it once, and never used it again. I would just loop 3dmark01 tests 1 and 3. Then, just do your everyday stuff and if it doesn't crash / freeze you are golden! Prime might be useful to some, but I know many people who use it and their computer fails it eventhough it is stable in everything else they do!

 

[gururu]

I stopped using prime95 a long time ago. games and movies have always let me know when something was amiss. fine tuning an overclock usually takes me a good couple of weeks as I believe it should.
IMO, the reality is that prime95 only tells you that your system is stable enough for prime95 but doesn't guarantee that your system is absolutely stable. as for me, my system only needs to be stable enough for the apps I use.

 

[Somniferum]

Originally posted by: lookin4dlz
I've been running my computer 24x7 for weeks. Daily I run a very intensive 8MB Excel spreadsheet that crunches statistics on its content for a continuous 1.5 hours - even multi-tasking during this with games, etc. has no effect on stability.

So, what gives? Are certain chips just less stable than others so that when using them P95 actually does predict stability?

If Prime95 is crashing you can be sure that rounding errors are occuring with your CPU. When you crunch those stats with Excel, depending on the margin of error you may or may not see a difference in the results. I would be wary of trusting the numbers you get out of Excel under those conditions if I were you.

 

[TheStu]

Originally posted by: gururu
I stopped using prime95 a long time ago. games and movies have always let me know when something was amiss. fine tuning an overclock usually takes me a good couple of weeks as I believe it should.
IMO, the reality is that prime95 only tells you that your system is stable enough for prime95 but doesn't guarantee that your system is absolutely stable. as for me, my system only needs to be stable enough for the apps I use.

If you think that then you should run Prime95.... almost a benchmark test, it will also stop if the values are wrong.... my roommate is neither Prime stable nor can he successfully run SuperPi... oh yeah his games crash quite often....

 

[imported_NoGodForMe]

Great thread. Similar situation here.
My memory is also failing Prime95 at 2225 timings with a rounding error. Runs fine at 2336.
This is an ongoing problem with the new Corsair XMS PC3200XLPro memory that is touted with those 2225 timings. Many using an Asus A8V are reporting problems. I tested the sticks one at a time, they pass Memtest86+, but one fails Prime95 after 14 minutes with the rounding error.
So I set the timings to 2336 and everything runs fine.

I've posted a message in the CorsairMicro forums, and others are reporting the same problems as me. Some lucky people can get those 2225 timings, but most can't, so we're thinking it's the RAM.

Oh, and get a load of this story. I'm been running my Abit machine for 6 months on PC3200LLPro at 2326. Machine runs great, play D3, runs all benchmarks. Just last week I ran Memtest on it for the first time, and it failed. That memory is being replaced now. So there ya go, memory can fail tests, yet the machine can run fine.

 

[slag]

Originally posted by: TheStu

Originally posted by: gururu
I stopped using prime95 a long time ago. games and movies have always let me know when something was amiss. fine tuning an overclock usually takes me a good couple of weeks as I believe it should.
IMO, the reality is that prime95 only tells you that your system is stable enough for prime95 but doesn't guarantee that your system is absolutely stable. as for me, my system only needs to be stable enough for the apps I use.

If you think that then you should run Prime95.... almost a benchmark test, it will also stop if the values are wrong.... my roommate is neither Prime stable nor can he successfully run SuperPi... oh yeah his games crash quite often....

You overclock your roomate?

 

[Lithan]

If you can't pass prime it's not because somethings wrong with prime it's because your computer ISN'T STABLE. It doesn't matter if you can run 3dmark, if you can't pass prime (or anything else that's properly coded), then your machine ISN'T STABLE. You may be happy with a machine that ISN'T STABLE. But you can't call it stable because it ISN'T STABLE. You might as well say that a table missing a leg is perfectly stable because you never bump it or put anything on it, so it doesn't tip over. Always funny seeing people sell cpu's as stable @ XXXXmhz, then someone buys the cpu and it can't even get close to those speeds. The sellers excuse? Well it ran so and so a program just fine. You can't call something stable unless it runs EVERYTHING. And so far Prime is the best program for testing cpu stability. Get the picture?

 

[TheStu]

Originally posted by: slag

Originally posted by: TheStu

Originally posted by: gururu
I stopped using prime95 a long time ago. games and movies have always let me know when something was amiss. fine tuning an overclock usually takes me a good couple of weeks as I believe it should.
IMO, the reality is that prime95 only tells you that your system is stable enough for prime95 but doesn't guarantee that your system is absolutely stable. as for me, my system only needs to be stable enough for the apps I use.

If you think that then you should run Prime95.... almost a benchmark test, it will also stop if the values are wrong.... my roommate is neither Prime stable nor can he successfully run SuperPi... oh yeah his games crash quite often....

You overclock your roomate?

No but I really want to.... unfortunately his ram won't handle it, he has a gig of PC2700, my friend and I went into his BIOS and adjusted his voltages a little while ago and it helped alot, he used to have to reboot windows 2-3 times a day and now he almost never has to.

See the reason I want to OC his proc is that he has an 0330 Barton 2500 so.... unlocked multi.... yum!

 

[cbehnken]

The difference between Prime95 passing and not (assuming everything else runs stable) is usually only around 20-40Mhz anyway. Why not just clock down till it's Prime95 stable and rest more easily?

I guess I'm more a stickler for not loosing important work than increasing my computer's speed an imperceptable amount.(Hence the Raid 1)

 

[implicit1]

This is the way I look at it...

A CPU is used for calculating, that's why you bought it. If Prime95 (or any properly written program for that matter) is failing, then the CPU isn't properly calculating. If it isn't properly calculating then what use is it? Running 8MB Spreadsheets on a CPU that can't properly calculate prime numbers is risky to say the least.

It also makes people think they can get more mhz out of X chip than they really can. If someone is touting an "almost stable" overclock in their sig it will make others think that it is rock solid, and they'll think that their chip should do the same (or close to it).

 

[gururu]

I think if you are running operations such as missile defense, even prime95 isn't going to give me comfort. Same goes if I were a pilot. I wouldn't want some engineer telling me he overclocked my computer system and that it was stable because prime95 could run on it.

 

[Zebo]

I don't buy into that prime stuff.

You all are only cheating yourself. Excel can make a round errors. Games and apps will be more prone to lock up when makeing math errors. etc. I personally go for about 6 hours on a game machine but office machine should be at least 24 hour prime stable. Mission critial systems should be forever prime stable which is why they use ECC/REG mem.

 

[Concillian]

Exactly. My machines aren't considered stable unless they prime for 24+ hours. My fileserver gets prime 24+ hours at 0.05 less volts than what I plan on running it 24x7.

I find Battlefield 1942 and BFV like to crash to desktop even with Prime95 is stable. So I generally drop a few MHz on top of being Prime stable.

Stability > ultimate speed

 

[Lithan]

Prime95 is really only good for stressing cpu. If you crash in games with prime stability it's usually a memory/memory controller issue. I've yet to find anything that actually tests the CPU better than prime.

 

[belazeru]

has any one had there system crash on P95 while running everthing stock?

 

[onemorekornkid]

Personally, I feel that my computer is stable after the first test (the 8 kb one or whatever). I've had my system like that forever and it has only once come to bite me in the ass (cod crashes after like an hour).


I would like to have 100% stability but if it means the difference between 2.3 ghrz and 2.4 ghrz than I will choose speed over stability.

I know I sound like an ass but that's just how I think of it and I am a huge fps whore.

 

[gururu]

Originally posted by: onemorekornkid
I am a huge fps whore.

you are not the only one my friend. every couple months I try for higher, knowing that it is no longer possible.

 

[Richardito]

I swer by Prime95, but be aware that they've had some buggy versions and that there are some incompatiblities in some systems. DO not ask mne why, but I've seen systems that can never pass P95, even when not overclocked. P95 usually gives errors if your memory is getting corrupted and the second cause is an unstable CPU.

 

[classy]

Originally posted by: big4x4
I have never used prime.....I just loop tests 1 and 3 on 3dmark 01. If that does not crash or freeze, its stable! I have never had a problem with this method!

Thats the ticket. 3dmark 01 is the best for stability. I never liked Prime 95. For me the outside of 3dmark 01, the 2nd best was looping original Unreal flyby. They put strees on the entire system, memory, cpu, and graphics card. Prime just seems to test cpu. Just my opinion.