Does increased heat necessitate raising your CPU voltage?

whisper

whisper

Limp Gawd
Joined
Mar 16, 2003
Messages
466
I've heard many times that bumping up your CPU voltage can result in a hotter chip. But if your chip is running hotter just due to the summertime, and is having some stability problems, is it likely that bumping up the voltage will stabilze things? For over a year, I was running my Athlon XP 2100+ at 192x11.5 with 1.65v on the CPU, and it was working great. Lately, i've been getting sudden blue screens followed by a restart, usually when doing something at least a little bit demanding. So far, that has included ripping songs from CD into AAC using iTunes, once during a Windows update, a couple tiems while playing Morrowind and Medal of Honor, and several times running 3DMark03. Windows error reporting says it was a serious error caused by the CPU. Although I never had to do it before, I decided to increase Vcore to 1.675v and now it seems to be doing better. It seems counterintuitive that rasing Vcore, which in theory should increase temp, is a remedy for a heat problem but what do I know? That's why I'm asking! :p
 
What are you saying here?

You were running an unstable computer, you raised the cpu core voltage, and now it is stable?


Great:)



whisper said:
That's why I'm asking! :p
Click to expand...

Asking what? I would like to try to help...later:)

Edit: Sorry, I forgot the question in the title... shame on me :p


If your CPU was running hotter than usual, it might get a little unstable. Did you prime it? Sounds like it might not have been totally stable from the start. But increasing the voltage should help, but also raises temps even more, all though, I am more farmiliar in the 1.8-1.9V range...later... uhhh, since when has there been a limit on how many smilies we can use?...later : )
 
raising the cpu voltage is usually only done to try and stabilize an overclock, but in your case, it just sounds like your stuff may be dying and pouring more juice into it is kinda like turning on the life support.

and increasing voltage is NOT a remedy for heat problems. it causes them.
 
ehZn said:
raising the cpu voltage is usually only done to try and stabilize an overclock, but in your case, it just sounds like your stuff may be dying and pouring more juice into it is kinda like turning on the life support.

and increasing voltage is NOT a remedy for heat problems. it causes them.
Click to expand...


That's what I was thinking as well. From everything I've heard, you can jack up the voltage if you're hitting a MHz limit and need a little more stability. But when stuff goes unstable due to heat it seemed weird to try to stabilize by increasing voltage, which could potentially lead to yet more heat. But nevertheless, it seemed to work for me.

I'm stupid for not mentioning this, but the other thing is that I recently added a Seagate SATA drive (now my system and boot drive), and I updated the bios on my mobo so that I could get the latest SiL SATA bios. For over a year, I was using the 1002uber bios and now I'm running the 1004uber with most recent SATA bios. Maybe there's something a little different about this bios that was destabilizing things slightly, and increasing Vcore to 1.675 could have helped.

Finally, even if my CPU is slowly burning out or something like that, 1.675v doesn't seem that high. People routinely run the 2100+ at 1.7, 1.8, or even 1.9v. I've always been totally stable at 1.65v on my old bios, but if I have to run at 1.675 now it's not the end of the world. Also, I put an A/C in my office where the computer is, so maybe that'll help to :)
 
what kind of temps are you getting? id like to think that a cpu would last longer than 1 year overclocked.
 
heres what i think

it was the BIOS update. ive noticed that the voltage put out by the board is ALWAYS 0.025 higher than what I set it to. when you put the new BIOS on, it started REALLY giving it 1.65 instead of 1.675, which it WAS doing before. so you had to set it to actually give 1.675, to match what it was getting before.
 
Yeah, that's the board. A7N8X deluxe revision 1.4. I prefer the uber bios versions though, as opposed to the ones offered from asus. That being said, I had initially tried using the 1007uber and I was getting the stability problem, so I switched to 1004uber because I heard a lot of people were happy with that version. It didn't help my problem at all, but it didn't really get worse either so I just left it with the 1004. It just so happens that all of this coincided with a dramatic increase in the ambient temp of my house (from an avg of around 68F to 75-80), with corresponding increase in CPU temps from around 35 idle/42 load to around 42 idle/48 load. That's measured using the asus probe that runs in windows. So it was difficult to separate the effects of increased temp with those arising from flashing the new bios.

kronchev: I like your explanation. When you say that the voltage is always 0.025 higher than what you set, which bios are you using? An older one I assume? So maybe the new bios versions are more precise?
 
Asus boards have a tendancy to overvolt the CPU. I run mine at 1.85 and every monitor I've used shows around 1.875 for the core voltage. This has happened on my A7V133, A7V266-e and A7N8x-Deluxe rev 1.0x.

I myself don't use the uber BIOS's. I used to use them but ended up going to a newer Asus BIOS and stuck with it. After running it for a while I found I had a higher stable overclock with the ASUS BIOS and everything ran fine. I also never got any type of advantage out of the extra settings the uber BIOS had. I don't remember what BIOS I am using at the moment though. I'd check but I'm at work.
 
i have the earlier rev asus board as well. i'm running the 1008 uber bios (found a link somewhere, the uber bios page doesn't have it for the non-E board) i have vcore set to 1.775 (max allowable on my mobile 2400 - doesn't matter, because my board won't let me do any better than 2500mhz, dammit) anyway it is reported by MBM5 as being 1.82. so yeah, the asus board overvolts.
 
You must log in or register to reply here.
Back
Top