[Hyoketsu]

Hello there. I have a few peculiarities floating around my system that I can't figure out, so I guess it's time to shoot some questions.
What made me notice the weird stuff is the PerformanceTest 7.0 benchmarking program. Basically, I get the feeling that some of my PC's components, namely the hard drive, can't keep up with the CPU/RAM overclock even though I considered them to be unrelated or at least have directly proportional performance boosts.
Now, let me explain. As I was overclocking the CPU and RAM (my RAM speed is tied to the CPU

), I saw the performance of the HDD rise a bit in accordance with the CPU boost. Then I stopped doing HDD benchmarks - thought I'd surprise myself later. However, once I found my system's limit on stock voltage (can't change it

) and ran a benchmark, I noticed that my HDD performance dropped even below the pre-overclock level. By performance I mean the sequential read/sequential write/random seek&RW speeds. The drop wasn't much, only around 1%, but considering it was going up earlier... Well, anyway. After that, I ran benchmarks every few days and compared the scores to the initial post-OC bencmmark. What I saw was a slight drop in CPU and memory performance. The physics area in the CPU was the only one that dropped and when it came to memory, well, all the scores went down (allocate small block, read cached, read uncached, write, large RAM). As to reassure my suspicions, slowly but surely, the HDD performance reached the pre-OC level. Now, out of the blue, the CPU/RAM scores are the same as they were after I had just OCed and the HDD score dropped once again.
So, my questions are:

• How is the performance of other components (namely the HDD) tied to the CPU/RAM performance?
• How come the HDD performance dropped, even though the CPU/RAM performance rose?(remember, it was rising at first)
• What could be the cause of the random performance power shift?

It looks like some sort of a performance bottleneck, but I don't get the reason.
Just taking a wild guess here: could it by any chance be the fault of insufficient voltage?

Thanks in advance for coping with my TL;DR-ness. I've provided most of my PC specs in the system-signature-thingy below the post. If you need anything else, just ask.

 

[kyle2194]

How big of a variation were these? Could they just be attributed to error?

 

[Hyoketsu]

No, those couldn't have been errors. I ran the tests several times and those weren't just fluctuations. As I said earlier, the RAM scores went down ~2% percent in a few days, whereas the margin of error would be ~0,1% at most.
And after the initial post-OC benchmark the HDD score went down even more than that. Where the pre-OC read/wride speed was ~45MB/s, it dropped to ~32MB/s after OCing. Now it's ~40.
Ah, right, after I OCed, the HDD also started to produce a silent high-pitched squeal. This occurs only when under full load during benchmarks, though.

 

[pursuinginsanity]

Quote:

Originally Posted by Hyoketsu No, those couldn't have been errors. I ran the tests several times and those weren't just fluctuations. As I said earlier, the RAM scores went down ~2% percent in a few days, whereas the margin of error would be ~0,1% at most. And after the initial post-OC benchmark the HDD score went down even more than that. Where the pre-OC read/wride speed was ~45MB/s, it dropped to ~32MB/s after OCing. Now it's ~40. Ah, right, after I OCed, the HDD also started to produce a silent high-pitched squeal. This occurs only when under full load during benchmarks, though.

Are you sure you're not hearing the motherboard's capacitors squealing? This is quite common with overclocks.

As for the performance loss, I say margin of error, or something else is happening. Unrelated to clock speeds IMO.

 

[Hyoketsu]

Mobo capacitors, huh? Never thought of that possibility, since the only time it happens is when the hdd is under load. If that really is the mobo... Is it dangerous? Should I lower the clock? The system is stable (no errors on both memtest86 & prime95, no game crashes), but if the OC puts so much stress on the mobo I guess I should give some ground, huh?
I'll stubbornly keep insisting that it wasn't due to the margin of error. The hdd performance dropped tremendously and went back up as other stuff's performance went down. After a few days the other stuff is back up and the hdd is suffering once again. Too much of a coincidence, I'd say. Thanks for the opinion, though.