ORTHOS
- Thread starter Gwyn
- Start date
when you think you have a stable overclock, run it for 9 -12 hours, on priorty 7-8, i think 9 or 10 is the highest, on both cores.
Permabanned
- Joined
- 3 Oct 2007
- Posts
- 142
I usualy only run mine for about 4 hours but we are all different.
Soldato
- Joined
- 2 Dec 2006
- Posts
- 8,204
To be honest i still stand by the fact tah stability should be judged by your needs.
Your cpu will never be taxed out as much as orthos and there is allthese people saying dont get quad ull onyl use 2 cores. Then its definitly never going to be taxed that much or get taht hot.
Judge it on your needs. Play your games, if they crash ok its an inconveniance so you make sure you save regularly while testing. if it does crash take the overclock down a bit.
4 hours orthos stable is the longest id run it then do the rest with testing on games and my own needs.
If you need to build a bridge for just you to cross something. Why make it so it can be used by a lorry when it only needs to be stable for your body weight ?
Your cpu will never be taxed out as much as orthos and there is allthese people saying dont get quad ull onyl use 2 cores. Then its definitly never going to be taxed that much or get taht hot.
Judge it on your needs. Play your games, if they crash ok its an inconveniance so you make sure you save regularly while testing. if it does crash take the overclock down a bit.
4 hours orthos stable is the longest id run it then do the rest with testing on games and my own needs.
If you need to build a bridge for just you to cross something. Why make it so it can be used by a lorry when it only needs to be stable for your body weight ?
10 or 20 mins between increments should be fine, while you're searching for a target. when you hit that, you need to run a longer session to test for stability
goodness knows this is a contentious topic on the o/c forums and 8igdave makes some very good points - common sense in my opinion.
I tend to favour 10-12 hours small-ffts, and then same again on blend, but I need this for the kind of work I do on it.
If I was browsing/gaming I probably wouldn't bother. Then again, I leave the stress tests running overnight, so force of habit would probably apply if I were to build a browsing/gaming machine.
ya pays yer money, and so on...
goodness knows this is a contentious topic on the o/c forums and 8igdave makes some very good points - common sense in my opinion.
I tend to favour 10-12 hours small-ffts, and then same again on blend, but I need this for the kind of work I do on it.
If I was browsing/gaming I probably wouldn't bother. Then again, I leave the stress tests running overnight, so force of habit would probably apply if I were to build a browsing/gaming machine.
ya pays yer money, and so on...
Associate
- Joined
- 18 Jul 2006
- Posts
- 609
Totally agree mate!
The only way I test to see if my overclock is stable is by just doing all the things my pc will be doing. If I can game for hours and do everything I need to then thats good enough for me.
My last setup, in which I had a Pentium D950 at 4.25GHz, couldn't run Orthos for more than 30 mins without an error. But I never had a single problem with it in the 13-14 months I used it.
But if it makes you feel better about your overclock then theirs no harm. I just think it's excessive to stress your processors at 100% for hours and hours when in reality your pc will never have to do that anyway.
I wonder if any of you can help?
My system is very stable at 330/9 with full load temps of 50c
At 340/9 I get nothing on screen and so I have to clear the cmos.
This board ups the vcore automatically so what would be the next step?
Cheers
My system is very stable at 330/9 with full load temps of 50c
At 340/9 I get nothing on screen and so I have to clear the cmos.
This board ups the vcore automatically so what would be the next step?
Cheers
Soldato
- Joined
- 24 Jul 2006
- Posts
- 8,876
- Location
- Hoddesdon, London, UK
I don't usually say this as i hate confronatation, but these two posts are absolutely rubbish advice.
Running prime etc... DOES NOT only involve stressing components 100%, it also tests how accurate the hardware is at its speed. A voltage starved cpu or chipset might not overheat, but the lack of volts make everything it does inaccurate. The stability of the system in general is tested, just not under heat. The RAM is tested as well to an extent and undervolted or too tightly overclocked/timed RAM will produce errors. Some say 'well it folds for hours', well it might be folding but its producing garbage, it 'games' for hours.. well one day it might well lockup or cause some save game corruption cause a few bytes got lost in the instability blackhole. Your OS might be fine for weeks and one day cause your RAM or cpu was unstable it totally blows its partition table. As for testing two cores on a quad only... seriously.. like i said, its not only about load and heat. However even that plays a big part, you test two cores and one day you use a nice encoder that cranks everything 100% and things get toasty.. opps why did the rig crash??
I've been overclocking longer than most, not showing off but i've done everything under the sun, water, phase, Peltiers.. only thing i never really done much was LN, you name it and can tell ya, a system might be stable for weeks and suddenly cave in. Testing gives a good scope on how stable you are. Simply put, if your PC can't pass a test at 4Ghz but can do it for hours at 3.6Ghz then it ain't stable. In most reasonable peoples minds, a stable overclock is a PC, that while overclocked, can do everything it can at stock speeds.
Last edited:
Associate
- Joined
- 18 Jul 2006
- Posts
- 609
It wasn't advice, just my opinion on my experiences. In my eyes, if my overclock is stable for over a year and allows me to do everything I want, and gives me no problems, then it's stable.
But I'm not telling anyone not to do it.
Soldato
- Joined
- 2 Dec 2006
- Posts
- 8,204
You twisted or ignored a lot there. I never said test it on 2 cores. I said a lot of people say dont get q6600 because it will only use 2 cores as an example as to why it will never ever reach the stress levels of when u put it under prime or orthos. I didn't say test it on 2 cores.
Secondly i said id test it for a maxium of 4 hours. But i didn't say i wouldn't at all. If there are voltage starvations its going to show way before 4 hours.
Soldato
- Joined
- 24 Jul 2006
- Posts
- 8,876
- Location
- Hoddesdon, London, UK
Apologies, i thought you were the same guy that said he'd only test 2 cores on a quad. Vcore starvation might show earlier or later. I go for a full 24 hr cycle as i've seen it fail as much as 17+ hours into a test. However, that late in the game for a fail usually means its just about on the verge of the right vcore and one notch up usually sorts it.
Soldato
- Joined
- 2 Dec 2006
- Posts
- 8,204
Im the guy who said to all the people asying nothing used quad core:
"well if it only uses 2 cores and your saying ooo but i can ovreclock my duel core higher because of it. Then just overclock it for 2 cores as it wont be using the other two then."
They all complained abut stability which is contradicting their own argument of it not useing the other 2 cores. lol.
Can you dissable cores? if thats the case then id of got a q6600 and dissabled 2 and overclocked the othres like a dual.
Associate
- Joined
- 18 Oct 2002
- Posts
- 715
- Location
- Wales, UK
Its peoples choice, there is no hard and fast rule on this. Ya pays yer money, ya takes yer chances, as they say.
For me, I always stress my cpu in tests for minimum 9 hrs (usually 12 ). But I know for a fact that if stable in orthos for just 30mins can be usually good enough for low end workloads (browsing, some games, etc ). But because I use this PC for a lot of 3d rendering (up to 6 hr renders with both cores at 100% a lot of the time) I need the peace of mind that my cpu can handle it. But even with this safety margin I've given myself, there is still risk, albiet much less than if not. Moral is -Everything in life is a risk
Because I have water cooling, temps are never a real problem with me even for decent overclocks, but I'm clever enough not to push stabilty too far. Just the other day I had a clock that bombed out after 7 hrs stressing. I decided it wasn't good enough so turned the clock down/upped the volts a bit..
For me, I always stress my cpu in tests for minimum 9 hrs (usually 12 ). But I know for a fact that if stable in orthos for just 30mins can be usually good enough for low end workloads (browsing, some games, etc ). But because I use this PC for a lot of 3d rendering (up to 6 hr renders with both cores at 100% a lot of the time) I need the peace of mind that my cpu can handle it. But even with this safety margin I've given myself, there is still risk, albiet much less than if not. Moral is -Everything in life is a risk
Because I have water cooling, temps are never a real problem with me even for decent overclocks, but I'm clever enough not to push stabilty too far. Just the other day I had a clock that bombed out after 7 hrs stressing. I decided it wasn't good enough so turned the clock down/upped the volts a bit..
Associate
- Joined
- 18 Jul 2006
- Posts
- 609
Good points!