Looking at my vantage scores over time my cpu score has steadily decreased by about 500 points over about 6 months. Is that normal considering the clocks are the same? Every few months I've had to raise voltages when I find that my previously stable overclock has become unstable (presumably due to raise in ambient temps). Would this reduce performance?
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
CPU performance decrease?
- Thread starter Devrij
- Start date
yeah it's 24/7, but wolfdales shouldn't degrade at 1.36V (which I started it off at, now at 1.38), as far as I was aware it was only likely above 1.4V. It had occurred to me though. Temps peak at 71C in Linx, but never anywhere near that in regular use.
Screw it, I think I'll drop it down to 4.2GHz to get the vcore down and hope I haven't got a degraded chip. I don't mind a bit of wear and tear over time as a result of OCing a chip, that's part of the risk, but after 6 months, I think I'll try to get some more longevity out of it!
I disabled it, assuming it would mean more vcore for a stable clock at those speeds, and I was trying to keep it as low as possible. Ironically, this may have caused more damage than simply enabling speedstep and bumping the max vcore up a notch.
Well, I can't speak for the i7 crowd, but disabling speedstep CAN help you get a more stable overclock, but leaving it on doesn't mean you won't be able to. Especially when running such high clocks and voltages, speedstep makes a lot of sense.
Hey Devrij, I'm not sure what's going on with your 4.5GHz clock but lowering the E8400's multi from stock [x9.0] to [8.5x] would overclock the NBCC from 500MHz to 562.5MHz which is an extra 12.5% that may account for the performance increase @ 4.25GHz vs 4.5GHz *if* the benchmark app was particularly sensitive to system bandwidth!
Have you tried using Fritz to determine the speed difference between 4250 & 4500?
Ah, pretty significant drop when comparing my old fritz benchies with this.
4.5
4.25
Superpi was 11.173 btw at this clock
Can you explain that bit about the NBCC? Always looking to expand my knowledge
Last edited:
Well, it's a natural process for any chip, but overclocking accelerates it. It was a problem particularly associated with wolfdale chips, but I was always under the assumption it was only a risk at higher voltages. I haven't seen any threads regarding i7 degradation so I would presume it's not such an issue with those chips.
That's excellent, I always wondered how fsb speeds affected performance, and I certainly learnt something new there: exactly the reason I frequent this forum!
So, assuming my board can handle it without frying my NB, I would want a higher fsb (to compensate for lower cpu multi), lower cpu multi, and dram divider that wasn't 1:1. Or conversely if I felt my NB was a cause of instability, I would want to raise the cpu multi, lower the fsb setting, and aim for a 1:1 ram divider to relieve stress on the NB?
As for the drop in the vantage cpu score: at the same clock (4.5GHz), I experienced approximately a 7% drop over the course of about 6 months. I'd attribute that to a poorly made synthetic benchmark and varying external conditions, but the drop seemed so consistent, a hundred or two points (from 8800 total) every month or two, never any spikes. That was what confused me. I can only guess it had to do with system instability caused by degredation, but as you can tell, I only know the basics of "how" but not "why".
So, assuming my board can handle it without frying my NB, I would want a higher fsb (to compensate for lower cpu multi), lower cpu multi, and dram divider that wasn't 1:1. Or conversely if I felt my NB was a cause of instability, I would want to raise the cpu multi, lower the fsb setting, and aim for a 1:1 ram divider to relieve stress on the NB?
As for the drop in the vantage cpu score: at the same clock (4.5GHz), I experienced approximately a 7% drop over the course of about 6 months. I'd attribute that to a poorly made synthetic benchmark and varying external conditions, but the drop seemed so consistent, a hundred or two points (from 8800 total) every month or two, never any spikes. That was what confused me. I can only guess it had to do with system instability caused by degredation, but as you can tell, I only know the basics of "how" but not "why".
Well, it WAS stable. Initially I made sure it was rock solid, 5 hrs orthos stable and 20 runs of LinX (temps got pretty toasty on that, but never over 81C), all without error and with pretty high ambient temps so I hoped it'd survive summer. I did find a couple months ago that it had somehow lost stability so had to up vcore a notch to make it LinX stable, and this time I wasn't able to get it stable at 4.5GHz without going beyong my comfort zone so dropped down to 4.25GHz, which is stable on both LinX and orthos (only ran for 2 hrs before getting bored though). I may do a backup of my os and see how that goes, sounds like a likely cause for the benchmark drops.
From what I understand from Wayne's post, yes. Of course, this means your NB will run hotter, but should provide more bandwidth.