Bought an OCUK '4.8Ghz' chip, can barely get to 4.6Ghz.
- Thread starter joxang
- Start date
Associate
- Joined
- 11 Apr 2011
- Posts
- 710
- Location
- Somerset
How long did you test it for? Either it wasn't long or you have a spectacularly good chip.
If your talking about ChrisGTL you can see the test started at 19:57 and screenshot was at 20:43 so it was 46 mins
Associate
- Joined
- 11 Apr 2011
- Posts
- 710
- Location
- Somerset
If your talking about ChrisGTL you can see the test started at 19:57 and screenshot was at 20:43 so it was 46 mins
Yes but it wasn't stopped at the point of the screenshot so may have gone on longer.
Yes but it wasn't stopped at the point of the screenshot so may have gone on longer.
True but surely you'd take the screenshot at the end of the test otherwise you wouldn't know it would be stable....
Anywho, we'll have to wait to see what he says
Soldato
- Joined
- 22 Mar 2009
- Posts
- 5,769
I had now changed from Silver Arrow to Noctua D14, much even better!
When I take out silver arrow, I was horrific and shocked to see the cpu chip was almost spread of my IC Diamond 24 from center to the right side, nothing on the left side, very poor paste. Anyway I had decided to put Chill Factor III paste on the center of the cpu chip and put the Noctua D14 heatsink on it and twist it round before tight the screw.
4.60GHz before on silver arrow: temp 59, 62, 69, 63 @ 1.38vcore (bios) 1.35vcore (windows) with 2 x 140mm (thermalright fans)
4.60GHz now on Noctua D14: temp 58, 57, 59, 56 @ 1.38vcore (bios) 1.35 (windows) with 2 x 140mm (thermalright fans)
Much even better it look as Noctua D14 is very flat surface than silver arrow.
When I take out silver arrow, I was horrific and shocked to see the cpu chip was almost spread of my IC Diamond 24 from center to the right side, nothing on the left side, very poor paste. Anyway I had decided to put Chill Factor III paste on the center of the cpu chip and put the Noctua D14 heatsink on it and twist it round before tight the screw.
4.60GHz before on silver arrow: temp 59, 62, 69, 63 @ 1.38vcore (bios) 1.35vcore (windows) with 2 x 140mm (thermalright fans)
4.60GHz now on Noctua D14: temp 58, 57, 59, 56 @ 1.38vcore (bios) 1.35 (windows) with 2 x 140mm (thermalright fans)
Much even better
it look as Noctua D14 is very flat surface than silver arrow.True but surely you'd take the screenshot at the end of the test otherwise you wouldn't know it would be stable....
Anywho, we'll have to wait to see what he says
I was too impatient to wait 8 hours, do I need to leave it 8 hours for it to be classed as 100% stable?
Edit** My 2500k batch is L104A368
Last edited:
Associate
- Joined
- 31 Jan 2011
- Posts
- 2,313
- Location
- Hampshire
I have Never left mine going for 8 hours the max i have ever done is 1 hour, after one hour just go use the PC and see how it goes, I have always done this with every one of my builds and never had a unstable machine 
but personal preference i guess
but personal preference i guessI have Never left mine going for 8 hours the max i have ever done is 1 hour, after one hour just go use the PC and see how it goes, I have always done this with every one of my builds and never had a unstable machine
You can't know if had an unstable machine if you haven't stress tested it for that long?
Soldato
- Joined
- 26 Apr 2004
- Posts
- 9,801
- Location
- Milton Keynes
This is why I tend to use Linpack64 these days. Gets hotter, and tends to flake much quicker than Prime; if I can pass an hour of max stress Linpack; I'm almost certainly going to pass a good day or so of Prime.
Conversely I've had overclocks do several hours of Prime, but fail/bluescreen in about 5-10 mins of Linpack.
Conversely I've had overclocks do several hours of Prime, but fail/bluescreen in about 5-10 mins of Linpack.
Last edited:
Soldato
- Joined
- 22 Mar 2009
- Posts
- 7,754
- Location
- Cornwall
just adding to this, i have had 10-16hours of prime blend passing without error, then 5minutes into a game BSOD. i think the only way to be 100% sure is some stress testing mixed with some real world use.
Most definitely, I'm stable in prim95 but flash videos crash the system.
Soldato
- Joined
- 26 Dec 2009
- Posts
- 9,707
- Location
- North
Well OCUK very kindly gave me an RMA so I'm going to return this one and get a retail chip instead.
I think my components are all high quality (see OP) - otherwise, a definition of 'high quality' might help inform a buying decision. Or recommended boards, and ones to avoid too.
That was you selling that rig?!!! If you'd told me you could get 4.8 with 1.325 I'd have been on it like a dog in heat
Yea i got very lucky with that rigs components. My GTX 570 would also do 985MHz on core, is still the fastest GTX 570 on OCUK heaven tests beating a few 580's. Shame i had to sell rig.
Anyhow new rig build starting V soon, Lian Li X2000-B and a Quad rad!!
Last edited:
Soldato
- Joined
- 22 Mar 2009
- Posts
- 5,769
I think doing prime 95 blend test is a waste of time really. I do remember I usually passed 25 hours of blend test, passed 50x linpack, passed 1 hour OCCT and passed 25x IBT then passed memtest. Next few days later got BSD !
It really pointless and every stressed tests doesn't mean it will be guarantee 100% stable but mostly 0.01% will be crashed in BSD or freeze. It more likely hardware issues somewhere! It the same things if you purchase high end pc from elsewhere or high street shop, it will end up BSD anyway.
It really pointless and every stressed tests doesn't mean it will be guarantee 100% stable but mostly 0.01% will be crashed in BSD or freeze. It more likely hardware issues somewhere! It the same things if you purchase high end pc from elsewhere or high street shop, it will end up BSD anyway.
Soldato
- Joined
- 24 Jul 2006
- Posts
- 8,876
- Location
- Hoddesdon, London, UK
Highly disagree. Prime and IBT etc.. give you a clear indication of how stable or unstable you are. I've used prime95 torture test since god knows when, i think around 1998 i discovered it and its always resulted in a rock solid rig after tweaking to get a 24hr pass.
If you fail p95 etc.. within minutes or reboot etc.. it cleary indicates high instability.. whereas a rounding error after a coupla hours might mean you're close but no cigar.
I've had literally no bluescreens that were overclocking related and that weren't replicatable and all eventually lead back to bad hardware or buggy software/drivers as it did the same at stock. I use prime95, and since IBT came about i've used that + 3d benchmarking tools and build about 100 overclocked rigs a year with no issues related to overclocking.
To be fair my foot, if you sell an overclocked product you don't 'guess' its stable you bloody well do it right or don't sell it. Guessing can lead to downtime and unneeded annoyance for the customer, and they bought the darn thing because likely they didn't want to be hassled or didn't know much about overclocking.
well i had to put it a less harsh way. I was in the same boat as you when i got my bundle the oc failed and yet I had paid extra for the oc. took me all weekend to get it right which is what i paid oc for
The voltage you require will be dependant on the motherboard. From an OCUK point however you chip can run at 4.8 but just needs more volts.
Furthermore:
I know the P8P67 Pro's use less voltage than the Gigabytes (at least in my experience and I have owned both the P8P67 Pro, UD4 and UD7).
Last edited:
Especially with Sandy Prime Blend tends to reveal instability quite quickly. If you were infact 8 hours Prime Blend stable and then you received a BSOD then perhaps degradation was to blame.