Sign-in / Sign-up
Your question

A bunch o' questions for ya

Tags:
  • CPUs
  • RAM
  • Overclocking
Last response: in Overclocking
July 1, 2008 8:36:05 PM

CPU: Q6600@3.2ghz (355x9) (G0)
Mobo: Asus P5B
RAM: 4x1gb 800Mhz 4-4-4-12 (2 sets of Corsair XMS2)

I'm wondering if it'd be better to use 400x8 instead of 355x9 because that'd let me use my ram at 800mhz (does ram frequency even matter as long as it's 1:1?). However, that'd make the fsb 1600 IIRC, which leads me to the next question.

If the mobo only supports 1066mhz fsb (not sure about the p5b, anyone know?), would anything higher than 2.4ghz make any difference since the mobos fsb doesn't go that high? (266x9=2.4ghz, 266 with 4 cores is 1066 although I could be completely wrong with the fsb thing)

About my Vcore, I can't be bothered to check in the bios right now but the Vcore is over 1.4, probably 1.45 or so, anything less and it's not stable at all. Prime95 has been running for almost an hour and a half but with the high Vcore. Why do I need such a high Vcore compared to everyone else? Is it because I have 4x1gb instead of 2x2gb ram? I only have one type of Vcore to change in my P5B bios so I should probably mess with some other ones, but I can't see any others. What do I need to do? I'm pretty sure that I have the latest bios version.

One last question, ever since I updated my bios from the driver disc version it loads my CPU fan to 100% and I need to slow it down in SpeedFan. I didn't have to do this before, can I change it in the bios? (some special hidden menu or something)

Thanks in advance :) 

More about : bunch questions

July 1, 2008 8:40:02 PM

Right now the core temperatures will go as high as 69C (running Prime95) but "CPU" (SpeedFan) will only go to 52C. Which of these does the 71C Intel reccomendation apply to?
July 1, 2008 9:14:20 PM

You're completely wrong with the FSB thing...

1066 is the quad pumped value, because the FSB transmits data 4 times per cycle, so a 266 MHz bus , transmitting data 4 times per cycle is like a 1066 MHz bus transmitting data one time per cycle, as you can see, is the same thing, but marketing says 1066 to appeal the consumer.

Second, the fact that your Mobo FSB supports 1066 MHz means it's guaranteed to support that speed, but the moment you OCed it the FSB stopped being 1066 (266 Mhz real) and became 1420 (355 Mhz real).

Runing at 400*8 and the ram at 800 MHz (400 MHz real as DDR2 ram transmits data twice per cycle, hence the name dual data rate) will give you a performance boost over runing it at 710 Mhz (355 real) but it won't be (repeat: WON'T BE) noticeable except in synthetic memory benchmarks.

About the fan, maybe fan control got itself deactivated in BIOS, forcing the fan to run at 100% constantly, check the thermal control options for fan settings, the vCore is certainly high, but you need to watch the vDrop, load CPU-Z and note the vCore, then start Orthos and see how much the vCore drops, that could give you some ideas as to why the instability if you lower the voltage.

Also, with tha voltage I recommend you keep an eye in your temps, that CPU may be overheating if you're using the stock cooler.

All best.

Edit: Didn't see the second post, that temp is quite high, but still tolerable, it wouldn't hurt to get some affordable aftermarket cooler IMO, not something to fancy but a good 30~35 US$ one, as far as the 71C recommendation, I'd say it's the individual cores (what realtemp or Coretemp reports).
Related resources
July 1, 2008 9:22:57 PM

Thanks, that cleared things up a lot. I'll try running at 400x8 and play around with the Vcore, as long as it's not worse than 355x9 that should be a pretty good alternative. What should I base the 1.5V limit on? The bios setting? Vdrop/Vdroop?
July 1, 2008 9:24:35 PM

Hmm, why am I not allowed to edit my messages? I'll have a look at aftermarket coolers after I've tried 400x8. :) 
July 1, 2008 9:53:49 PM

I couldn't boot into windows until I upped to 1.5V, which is really too much for me. And still it was only stable for about 2 minutes so I set it back to 355x9 and 1.45 Vcore so I could post this message. Isn't it pretty weird that I have to go past 1.5V Vcore for 3.2ghz? If I use a 9x multiplier instead of 8x I'd still have to go past 1.5V for 3.4ghz I think.

Does anyone know why this is happening? If it's only happening to me then I've probably done some stupid mistake but what could that be?
July 2, 2008 1:20:09 AM

I've tried 3.4ghz, 3.33ghz and 3.3ghz with 1.5 Vcore with no success, I'm at 356x9 at 1.45V stable now. I can probably get a lower Vcore than that but I want to reach 3.4ghz now. :p  (I'm only using a 9x multiplier for the time being since a higher fsb requires a higher Vcore, if my observations are correct)

Something that I've noticed is that the Vcore goes down A LOT compared to the bios setting, can setting a very high Vcore in the bios to compensate for this ruin the cpu? As long as the Vcore in CPU-Z is fine (less than 1.5V) then the cpu should be alright, right? My 1.45V Vcore is only 1.312V when running Prime95, 1.416 at idle. I think the difference is bigger at 1.5V.

One more thing, my VID is 1.30V in case that's of any importance.
July 2, 2008 2:15:18 AM

I came across this thread, is it relevant to my interests? Should I try with the fsb at 401mhz? http://www.thetechrepository.com/showthread.php?t=30

Quote:
To finish...if you are clocking 1:1 I seriously suggest you skip 360 to 400fsb and push up from 401, the errors you see are NB related and not the memory in most cases.

July 2, 2008 2:35:01 AM

Nope, didn't work. I tried:

425x8 1.5V Vcore (3.4ghz)
401x8 1.5V Vcore (3.2ghz)
401x8 1.525V Vcore (the increase seemed to help a bit but I still got a BSOD and I can't afford a new Q6600 so I don't want to push it too hard)
July 2, 2008 2:54:45 AM

About the 3.2ghz stability, I can't go lower than 1.475V Vcore without core #2 crapping out in Prime95. Why is it always the #2? Why do I need such a high Vcore?! :( 
July 2, 2008 3:00:50 AM

Do you have any vdrop issues on the mobo? What is your VID?
July 2, 2008 3:09:58 AM

My VID is 1.30V. I'm running 1.475V right now, at idle-ish (just Firefox) and the CPU-Z Vcore is alternating between 1.392V and 1.40V.

It seems that I skipped 1.45V this time when testing stability, I must be getting tired... Just for the sake of correctness (yes, that's a real word apparently) I can probably run 3.2ghz 356x9 at 1.45V Vcore.
July 2, 2008 4:46:07 AM

Minor update, I'm still running at 3.2ghz 1.475V Vcore, 356x9. Core temps are 61/62, 61/62, 58 and 59/60 under Prime95 load. (according to SpeedFan)
July 2, 2008 4:51:46 AM

Core #3 broke this time, after 7 minutes. At least something different than the #2 crap. I'll see if I can get some sleep, it's 6:50 in the morning lol, but I'm still annoyed by the high Vcore, I'll need even more than 1.475 Vcore for just 3.2ghz... Any help would be very much appreciated.