Printable View
i missed this post yesterday....Quote:
Originally Posted by blazin-asian
i don't think its better, i've learned to tweak this mobo a lot in the last few days, and i'm starting to get some more performance out of it. i could have gone just as far with the 0610.
Finally someone gets it right. Pay attention to Itznfb, he does know what he's talking about. A lot of people don't seem to understand that the stability is mostly temperature bound on these coldbugged Opties @ high clocks.Quote:
Originally Posted by itznfb
Prime erroring doesn't necessarily make the CPU unstable. Think about it this way, your CPU idles at 35c and loads at 50c. During dual prime95 load at 50c, after 25 minutes you get an error. You take off prime 95 and you're back to 35c, 15 degrees cooler. Idling at 35c you are perfectly stable. Loading internet, MS Office, maybe encode something and put a workload on 1 CPU and you'll be around 42c. Still stable. Run a game, it loads 1 core to the max, say you get to 45c. Still stable.
The only reason prime95 fails is because it gets hotter than everything else. That is why i've run my computer 24/7 for DAYS without a single problem at a speed that has failed Prime95 in less than 10 minutes.
IMO as a general thing about 30 minutes of prime is good. Past 30 minutes of prime you are stable enough so that you can chalk up the remainder to temperature. Maybe it fails at 35 minutes, well say you take the side off your case, drop CPU load 2.5c, and it primes for 3+ hours and doesn't stop...
basically after a certain point it's all temperature bound and it's :banana::banana::banana::banana:ing pointless to prime all day and night. It's a waste of time and energy, i sure as hell didn't build a 2K PC to run prime and i'm sure most others didn't either. So people asking for 8+ whatever hours prime, realize that those are YOUR standards of stability and don't necessarily mean anything to Xtremesystems on a whole, and we aren't required to run Prime for any set amount of time. Any half decent OCer has run their CPU at enough speed/voltage settings to understand exactly what it's capable of and where it's stable.
Next thing you know we'll have to have an input for ambient temp, not quite fair if one person has 15c and another has 25c. And then IHS makes a 10c difference :rolleyes: where does it end?
thank you Absolute_0. as this was basically the point i was trying to make. we are here to max these chips out, not see how long we can run a program that stresses the cpu beyond any other app we are going to use (besides folding). if its stable for normal everyday use, then its stable.Quote:
Originally Posted by Absolute_0
WELL PUT, I SECOND THAT , ABSOLUTE O.
I ve tried a resembling discussion on the DFI forum and was cussed by some
prime fundamentalists :)
I am not that experienced in overclocking, but the above stated implies to me as well.
I had 3GHZ at 1.41V C, failed in Oblivion, bumped it a notch and wham -OCCT SPI Prime 1 hour ok.
At 1.536 Prime large FFTS goes for 3 hours or so -but temps despite the WC are around 50C :(
As said before in the thread (I think it's growing so much as the previous one for newcomers to read it throughfully :D) agree 100% with Absolute_0 and itznfb. Prime failing is due to temps. You will never, ever, get your chip as hot with any other app (or apps) as priming.
If not, think about it this way: What's the better cooling for? Not only for even higher OCs, but also gives stability at higher volt/mhz where air craps priming.
Why does a chip crap at 1.525, 334x9 with a TT BT and the same chip with the same config runs OK with a good WC kit? (assuming it's not coldbugged). Because the chips runs fine with that config while it doesn't get higher than X degrees. Something that only happens when priming.
Someone could argue with this, OK, but then what are you going to do? run prime 24/7 for one week to be sure of what? yeah it's 1week prime stable, but and if it fails at 1week and 1 minute? :D:D:D
I hope this post helps the discussion to calm down and not keep the primewars flooding the thread. Would be sad to see this closed too :(
Alright, everyone shutup about priming, kiss and makeup ASAP
:) ;)
Sure, now the thread is crowded, I wil re-re-re-ask something I'm yet to get answered: (sorry if with the XS downtimes I've lost some answer :p:)
Does anyone know if lighter liquid (from zippo for example) can help my died 0610? I read someone with top-notch cooling did a bath to all his chips. I think I killed her when removing the IHS. No, I didn't cut any SMD, it was prolly some AS5, but I doubt it, I don't really know what happened. The chip booted fine 2 or 3 times, but had higher temps and worse OC, so I remounted twice reapplying AS5. 3rd time I plugged it none of the 4 DFI SLI-DR lights turned off -> dead cpu. I cleaned it with 100% pure acetone and with AS5 2xcleaning liquids (remover and purifier) but didn't help.
Any idea? Big Thanks
does the motherboard still work?Quote:
Originally Posted by mursaat
Yes I plugged my 0551 and everything's OK. Everything but the speed :D
haha. and the chip isn't burned anywhere?Quote:
Originally Posted by mursaat
i wonder if you put too much pressure on the core :stick:
itz whats ur batch on CCB1E 0609 FPMW
I just got one of those from monarch and my batch is 0148 (last 4 digits)
Also, im eagerly awaiting ur OC'ing results, as I dont have a mobo to test this right now.
i should have some results tonight. i don't know what the batch is, i'll check when i get home. i got the chip yesterday i haven't even looked at it yet. i was too busy trying to find the max on my 0615 before i switched over.
Don't think so, can I harm the core without leaving any visible trace over the core? I put a little more pressure than previous times, only to try. But I've read a lot of times TTBT+noIHS use the pressure you can with your fingers. Always geting the IHS base balanced. I glued with epoxy the H metal plate to the IHS so it's easier.Quote:
Originally Posted by itznfb
i'm not 100% sure, but i think you can damage the core without any visible damage. more than likely you would see where you chipped or cracked it, but i'm sure its possible to not do any visible damage.
God, all you people talk about stability and what it is, but the thing is, you can push all you want with 30 min stability or 24 hour stability.
So today you tell us you're 30 min prime stable, and all you care about is being able to game and it won't crash, but is this a setting you would use 24/7? You can push 3.2 GHz stable for 30 min, but are you going to use this setting when you have to get a video project done or something important? Are you going to risk everything you do on this setting? More often than not, you're going to be one of those people with corrupted data, who needs to reformat, and buy a new CPU, etc etc. Don't tell me you can get away by backing up and everything. This is like saying you abuse your car, but you can afford a new car every month. Who cares? Point is that your "overclocking" really isn't much of an achievement.
Then there are those of you who argue 24 hr stability doesn't matter because 1/10th of a degree of fluctuation and change in weather screws everything up.
This is precisely why I bench now, and I don't leave my windows wide open for air to flow in. I bench with my case closed because this is what I know I'll be runnign like 24/7. I also unplug my side 120mm fans so I'm down to 2x120mm fans (front + back). I also run my fan controller at what I run at 24/7 (not max) so I know I can turn the fans up if necessary. My side fan drops temps a good 2 - 3 deg on the CPU side and definitely more for the PWMIC and 7800GT, but I leave it off. Why? Because a summer heat wave can screw me over one day. Plus, this gives me more leeway.
And why do I do this? I know I can't hit "24 hr stability" like some people do. My goal isn't to have a 3.2 GHz chip that I can show off for one day that will crash the next because I rigged a test setup that can get me this one time certificate that I won't be able to realistically achieve everyday because I don't want some loud@$$ system that sounds like a jet engine that doesnt let me sleep.
Sure prime erroring may let you get away with working in Windows. You might be able to get away with a lot of things.
But would you make your system a web server and advertise 99.9% uptime the next day? I think not. What's the point of having something that's not 100% stable in every way you know it? It's just like strapping a bomb to your chest. Armed or not, I'm sure it's not a very comfortable feeling.
Maybe my goal is different from everyone. I want an OCed system that I'm comfortable using 24/7 and that is JUST AS STABLE if not MORE STABLE than my stock settings.
Dude, we were done discussing it. Please stop...seriously.
Yeah really :lol: it was dead 3 pages ago :lol:
I dont know where you guys are coming up with this information from but its not based on any scientific or engineering standpoint of a processor whatsoever. Temperature is just one factor in overclocking. Just one. Voltage is another factor. The fact that a processor has its limits as to how high it will even overclock is another factor and is the most common reason for Prime95 failing. Even at 2.8GHz these Opteron 165 processors are pushed 1GHz past their initial clock speed. Each processor has a limit as to how far its possible to push its internal clock before it reaches a point to where it can no longer carry out normal operations. Low temps and increased voltage are means by which to increase this potential clock but nevertheless there is a limit.Quote:
Originally Posted by Absolute_0
Prime erroring doesn't necessarily make the CPU unstable? If it wasn't unstable it WOULD NOT cause an error. Just because you only load one core and the max temp is only 45C does NOT mean that one core is going to be stable. I can run ONE instance of Prime95 and if that core is not stable it will fail all by itself. You dont have to run dual Prime to get the temps way up there to fail. The core may not be getting enough voltage or it may just not be capable of running those speeds. The purpose in running dual Prime is so you test both cores working together at max load. Temps are just an added factor in the mix.Quote:
Originally Posted by Absolute_0
We are also moving forward into an age of SMP aware applications and games. Most encodeing software such as DVD Shrink is multithreaded. Alot of productivity software like 3ds max is multithreaded. This means its going to push your CPU to 100% load just like Prime95. And even when your only running single threaded applications you are still running 100% on that single core. Core 0 is normally the core that fails first with these Opterons, at least it has for all the Opterons i've had to this point. Thats the primary core you will be running all your applications off of. And if you guys think Prime95 pushes your system further then anything else will, try running S&M. Your system will most likely lockup because temps get higher with it then any other application i have seen.
http://www.xtremesystems.org/forums/...d.php?t=100302
Nope, thats not true at all. Temperatures are only ONE reason Prime95 will fail.Quote:
Originally Posted by Absolute_0
You sound exactly like one of the scenarios presented in the Prime95 stress FAQ lol.
I doubt very many of you even understand how Prime95 works. Prime95 doesn't constantly run the exact same set of numbers so that you have to push it for 9 hours under 100% load before it breaks. It takes hours to completely run through all the different combinations that could possibly result in an error.Quote:
Q) A forum member said "Don't bother with prime95, it always pukes on me,
and my system is stable!. What do you make of that?"
or
"We had a server at work that ran for 2 MONTHS straight, without a reboot
I installed Prime95 on it and ran it - a couple minutes later I get an error.
You are going to tell me that the server wasn't stable?"
A) These users obviously do not subscribe to the 100% rock solid
school of thought. THEIR MACHINES DO HAVE HARDWARE PROBLEMS.
But since they are not presently running any programs that reveal
the hardware problem, the machines are quite stable. As long as
these machines never run a program that uncovers the hardware problem,
then the machines will continue to be stable.
http://www.mersenneforum.org/showthread.php?t=5676Quote:
A torture test runs a certain number of L-L iterations on each of many different exponents (actually, Mersenne numbers with different exponents, of course). The table of test exponents (and the corresponding correct answer for each one) is fairly long - many dozens of different exponents, covering a wide range. I think it may take more than 24 hours to cycle through the complete set of test exponents on at least some systems, so letting the torture test run for long periods is the best way to push as many different bit patterns through as possible, so as to maximize the probability of hitting upon a pattern that is defectively processed.
And once again, just because your system doesn't CRASH or LOCKUP does NOT mean your system is stable. Errors can be happening in the background without you even knowing it. What happens when your memory is causing errors? Do you normally know it from just running Windows and playing games? It depends on how bad it is. You may never notice it until you run Memtest86 or Windows Memtest or Prime95 Blend that your RAM even has a problem. Thats the whole point behind these stability testing programs. If your system was stable you could pass all of them without a problem.
You only have to run Prime while your testing your overclock. Once you establish your max overclock your done. Its not something you sit around and run every day just for the fun of it lol. Some of you are just INCREDIBLY lazy. I run my own business, i'm in college, i have a girlfriend, i play basketball a lot so i spend time weight training and practicing, and i have many other things going on in my life and yet i SOMEHOW manage to find time to click Start on a program and walk away lol.Quote:
Originally Posted by Absolute_0
And asking for 8+ hours isn't OUR standards. Those are the standards set forth by the creator of the program and by anyone who actually understand how it works and how long it takes to find stability problems. Its not a 10 minute job.
These 30 minute and 2 hour standards are YOUR standards that you're coming up with based on your own opinion of what stability is. Its not based on any actual facts or methods that can be proven to be true. Its not based on anything that the creator of the program has said himself. What i'm saying is based off of facts presented. Yours is based on theories. Theories are worthless until facts are introduced to support them.Quote:
Originally Posted by pulled from the Prime95 README again
I wouldn't even bother replying in this thread about this subject if it wasn't for all of this FUD being spread around about Prime95 and what constitutes stability. Some of the things said just completely astonish me.
And what has really got me going on this subject here lately is the morons on eBay selling processors advertised at being Prime stable at 3-3.1GHz by the dozens. I've been selling a few of my chips on there and i post honest results that have actually been thoroughly tested. I dont post some suicide screen at 3.1GHz just to lure in the noobs dumb enough to bid higher thinking they're getting some incredible overclocking processor.
Like i told someone earlier, XtremeSystems is suppost to present the pinnacle of overclocking knowledge and expertise. Any beginner with extremely limited computer knowledge can install a processor, hit a button and enter his BIOS, and change a handful of settings that duplicates the settings someone else has used. It takes a veteren overclocker to be able to do that same process of pushing the system to the limits all while taking the steps neccessary (extreme cooling, extreme voltage) to produce a rock solid machine.
We should start another thread on this topic if anyone cares to discuss it further. I would really love to see some actual tests and facts presented on the subject outside of just personal opinions of what is thought to be stable.
can someone please ban him
Why? He does make alot of good points.Quote:
Originally Posted by itznfb
Theories are worthless until facts are introduced to support them.
That being one of them.
Do you normally know it from just running Windows and playing games? It depends on how bad it is. You may never notice it until you run Memtest86 or Windows Memtest or Prime95 Blend that your RAM even has a problem. Thats the whole point behind these stability testing programs. If your system was stable you could pass all of them without a problem.
This is true aswell, sometimes you bluescreen, sometimes your pc locks up, sometimes your app errors out. 1 hour of prime is nothing, since the test does do alot of tests of diffrent combos, you can error out at any time. That's why sometimes you error out 11 hours later, not because of heat.
1. no he really doesn't.
2. because he refuses to stop the continuous BS and he is going to get the thread closed.
As soon as you care to stop replying about the topic in this thread i will be more then happy to procede with the continuous FACTS against your continuous BS in a more appropriate thread.Quote:
Originally Posted by itznfb
You guys miss my previous post?
Quote:
Alright, everyone shutup about priming, kiss and makeup ASAP
BurningGrave i addressed your concerns in dark red
I dont know where you guys are coming up with this information from but its not based on any scientific or engineering standpoint of a processor whatsoever. Temperature is just one factor in overclocking. Just one. Voltage is another factor. The fact that a processor has its limits as to how high it will even overclock is another factor and is the most common reason for Prime95 failing. Even at 2.8GHz these Opteron 165 processors are pushed 1GHz past their initial clock speed. Each processor has a limit as to how far its possible to push its internal clock before it reaches a point to where it can no longer carry out normal operations. Low temps and increased voltage are means by which to increase this potential clock but nevertheless there is a limit.
Buddy i'm talking about prime for 30 minutes versus prime for hours where heat IS the factor. if it fails in a second obviously you need more volts.
Prime erroring doesn't necessarily make the CPU unstable? If it wasn't unstable it WOULD NOT cause an error. Just because you only load one core and the max temp is only 45C does NOT mean that one core is going to be stable. I can run ONE instance of Prime95 and if that core is not stable it will fail all by itself. You dont have to run dual Prime to get the temps way up there to fail. The core may not be getting enough voltage or it may just not be capable of running those speeds. The purpose in running dual Prime is so you test both cores working together at max load. Temps are just an added factor in the mix.
I'm saying the instability stems from heat more so than the clock speed.
Obviously if you only run 1 instance of Prime and it fails, you are way unstable. I'm talking about a stable system that i only run prime for 30 minutes on... you're talkng about a completely unstable system.
We are also moving forward into an age of SMP aware applications and games. Most encodeing software such as DVD Shrink is multithreaded. Alot of productivity software like 3ds max is multithreaded. This means its going to push your CPU to 100% load just like Prime95. And even when your only running single threaded applications you are still running 100% on that single core. Core 0 is normally the core that fails first with these Opterons, at least it has for all the Opterons i've had to this point. Thats the primary core you will be running all your applications off of. And if you guys think Prime95 pushes your system further then anything else will, try running S&M. Your system will most likely lockup because temps get higher with it then any other application i have seen.
Why would i run S&M? I built my computer for Folding and gaming, not for software that applies an artificial load :rolleyes:
I'm guessing most other people are with me.
http://www.xtremesystems.org/forums/...d.php?t=100302
Nope, thats not true at all. Temperatures are only ONE reason Prime95 will fail.
You sound exactly like one of the scenarios presented in the Prime95 stress FAQ lol.
Again you assume a perfectly unstable situation. I never said prime was useless. My point is that running it for hours and hours is useless. Read my post.
I doubt very many of you even understand how Prime95 works. Prime95 doesn't constantly run the exact same set of numbers so that you have to push it for 9 hours under 100% load before it breaks. It takes hours to completely run through all the different combinations that could possibly result in an error.
http://www.mersenneforum.org/showthread.php?t=5676
And once again, just because your system doesn't CRASH or LOCKUP does NOT mean your system is stable. Errors can be happening in the background without you even knowing it. What happens when your memory is causing errors? Do you normally know it from just running Windows and playing games? It depends on how bad it is. You may never notice it until you run Memtest86 or Windows Memtest or Prime95 Blend that your RAM even has a problem. Thats the whole point behind these stability testing programs. If your system was stable you could pass all of them without a problem.
If i don't notice the errors, why should i care?
You only have to run Prime while your testing your overclock. Once you establish your max overclock your done. Its not something you sit around and run every day just for the fun of it lol. Some of you are just INCREDIBLY lazy. I run my own business, i'm in college, i have a girlfriend, i play basketball a lot so i spend time weight training and practicing, and i have many other things going on in my life and yet i SOMEHOW manage to find time to click Start on a program and walk away lol.
And asking for 8+ hours isn't OUR standards. Those are the standards set forth by the creator of the program and by anyone who actually understand how it works and how long it takes to find stability problems. Its not a 10 minute job.
You run 8 hours all you want, at Xtremesystems i think asking for 10 hours of prime is rubbish. I'm not giving up Rosetta cycles to convince other people that my overclock is "stable" when i'm perfectly content with it.
You have to accept that people have different standards of stability, and not hold others to your standards.
These 30 minute and 2 hour standards are YOUR standards that you're coming up with based on your own opinion of what stability is. Its not based on any actual facts or methods that can be proven to be true. Its not based on anything that the creator of the program has said himself. What i'm saying is based off of facts presented. Yours is based on theories. Theories are worthless until facts are introduced to support them.
There's no facts about prime stability. You could be 24 hour prime stable and fail something else. It happens all the time, Prime isn't the holy friggin grail.
I wouldn't even bother replying in this thread about this subject if it wasn't for all of this FUD being spread around about Prime95 and what constitutes stability. Some of the things said just completely astonish me.
What astonishes me is that after 3 people agree with me and 2 people say to leave the subject alone, you go and make this post. This thread is so much closer to being locked now : /
http://carcino.gen.nz/images/image.p...?cb=1115204527Quote:
Originally Posted by burningrave101