Intel vs. AMD

Whose got the weiners and marshmallows? LOL

 

I still love my Gigabyte GA 965P-S3. It's very stable and nice to use. I'm definitely ready when I decide to make the jump to Core 2 Duo. I still use my Sempron as my LAN rig and I'll soon be using my P4 if all goes according to plan

 

Excuse the double post but I did some OC'ing tonight and I had my dual core to 2.8GHz stable. But when I tried pushing higher I got the BSOD. I need to adjust my bus speeds and lower my multi.

...Thought you guys might like the update.

 

0202. The thing is, with most boards you get crashes or BSODs when you push too far. This board just refuses to POST at all, and sometimes you can boot at a higher OC but not a lower one (e.g. 1300, but not 1220!)

 

@sam

how high are your temps with the freezer pro at your current oc?

 

At high speed about 35 idle, 42 or so load, at low fan speed about 40-42 idle and 47-48 load.

 

is that the cpu or each core?

im just wondering as i have now tonight just finished putting mine together but as yet its not had chance to idle so i cant give you mine as yet but i will let it calm down over night then see how it is tomorrow.

p5n e sli
e4300
2gb ddr800 cas4 dominator
freezer 7 pro

i know i said i was waiting till the end of the month but scan had all the bits in today when i looked on their webby so i went up and bought them early.

edit
also how did you change the speed?

 

Lol exactly the same build except my RAM is plain XMS2, not Dominator, but still 6400C4. I changed the speed via FSB early, nothing else. Nothing helps! The system only POSTs at a select few FSBs (whether or not they would be stable it seems), and they're so far apart that I don't think the gains achieved by increased voltage and latencies will help! For example the PC will fail to POST at numerous values in the 10xx FSB range, but will boot fine at 1220. I'm thinking BIOS flash at this stage, but why should that be necessary?

 

marsey99
If it's taking your system overnight to cool down to normal low idle temps, you've got a problem. Once the load is off, my CPU temps plummet back to normal lows. About the only thing you'll find shutting down long term is the start up temp of your system and that depends a lot on your room temps.

 

pacman i think it would of cooled down quickly i just wanted to go to bed as we are 5 hours ahead of est in the uk. at the time of posting i was running the xp update so although not a high load it was still constantly working. temps were only 2 or 3c above what they are now.

right now, its only been on 10 min, i hve a cpu temp of 25c and core 1 is 33c and core 2 is 40c. the temp in my case is 24c.the 7 pro i spinning at 2576rpm and i cant hear it over my 7600 fan at 3267 (still the loudest thing in my case).

is it normal to have such a difference in temps between cores?

 

You need to go to 0401 BIOS.
You're still using the release version which has been well documented to have issues.
Looks like you're suffering from the black hole effect on the FSB.
As for why you'd need to change the BIOS, its very simple, they didn’t get it right first time around, they never do.
If you’ve only got as far as 1220, then you’re still 113 below what the board can do in the book, let alone what it’s capable of with the right tweaks.
BTW if you simply rely on FSB and voltage tweaks, you’re never going to get anywhere with those boards, you need to adjust a whole host of features to unlock the full potential of the board with over-clocking, that’s where most people fall down, because they fail to understand the interaction of one feature on another, and it does go beyond memory as well.
If it was as simple as upping the FSB and applying voltage, everyone would be flying along with max’d out PC’s, it’s a lot harder than it looks to get it cranked right up and stable.
I only ever get a difference of 1`c on my two cores, on both the E6600 and the E4300.
Even If I run something like super-pi and manually set the affinity to one core only, I still only get a difference of 3`c between the loaded core and the idle one, maybe try Intel Thermal Analysis tool to check the readings.
Again, if you’re still running the release BIOS 0202 for some unknown reason, then upgrade it to 0401, I would avoid the version in between 0307 as it caused a whole load of issues as well as addressing some it was supposed to fix.

 

mines the 0307, so far its all been fine but i aint started to play with it yet. i wanted to get to know it at stock first.

out the box it does superpi 1m in 30.532s.
and after being on for a bit now doing nothing too heavy (just ran pi but you know) everest is reporting temps of,
CPU 25 °C (77 °F)
CPU #1 / Core #1 36 °C (97 °F)
CPU #1 / Core #2 41 °C (106 °F)

i will look for that itel temp thing and the new bios now.

edit

the tat util doesnt seem to like my setup and says it cant enumerate on demand clock modulation support?

edit 2

bios now flashed

 

Sorry had a brain lapse.
You're using the Nforce chipset with the CPU.
I forgot it was only made for use with Intel Chipsets and CPU in combo Sorry!!!

 

marsey99
In your post did you mean CPU 0 and CPU 1? Which brings up the question, If 0 and 1 are both above 36C, how do you get an overall CPU temp of 25C? Something doesn't look right.

Before BigDK went over things, you posted:
i hve a cpu temp of 25c and core 1 is 33c and core 2 is 40c
There again you had an overall temp lower than the 2 separate core temps.

Is the software stuck on 25C? I just noticed it had the same temp in both your posts with different core temps.

 

The main CPU temp is always lower than the core temps inside the CPU.

The lower temp will be from the reading obtained from the motherboard CPU sensor.

The other temps are those that are read fom within the CPU itself on the cores.

 

That must vary by system because I'm sitting here looking at a main CPU temp of 33C with core 1 at 27C and core 2 at 33C. I'm using the Everest Ultimate Edition. PC Probe II confirms the main temp.

 

Strange how the internal core temp is no higher than the outer sensor temp.
Heat on the cpu is generated by the cores, this then radiates out via conduction.

Looks like you have a problem with your sensor readings on the board.
Both PC Probe and Everest are good but are not perfect, and they will use the same sensors on the board.
If something is incorrect, it will be the same on both applications.

Simple physics dictates, that the source of heat will be hotter than the material around it unless ou have a 100% perfect conductor and 100% perfect insulation from heat loss [which you don't]. Therefore the radiating heat from the source will get lower as it spreads out.

 

BigDK
The main CPU temp is always lower than the core temps inside the CPU.
Guess I'll have to rush out a request to Asus to get this thing repaired. Here I was all these months thinking I had a good board. I was putting my trust in those sensors. The readings go up and down. The internal core readings will sometimes be higher or lower depending on whether the temps are rising or falling. I've done some decent OCs with the power turned up. If the CPU monitor wasn't working properly, I thought I would have fried the CPU by now.

Out of curiosity, where did you get your information on the sequence in which the sensors read and the temperature variance according to location and sensor position? I'd like to do a bit of light reading before I go after the boys at Asus.

BTW, I've had enough physics and practical experience to understand conductors, thermal compounds, and heat transfer. I've seen similar readings on boards besides this one, with an internal core reading lower than the other. I'd really not paid attention to it being a problem and the main temps never have been a problem. One thing I know is if a board has a main sensor 11C below the internal it isn't transfering the heat very well. If a decent conductor, the temp will rise at a slower rate, but should get close to the core temp. Internal temps should also be either higher or lower than the conductor depending on whether the temps are rising or lowering, i.e. starting Orthos and shutting it down.

 

On my way to work, but will dig out some stuff to link to when I get a chance.
Chances are the BIOS is just giving the problem.
There's quite a few ducuments showing that the sensors on boards cn be as much as +/- 15`c out from the actual temperature.
What chip are you using and what board, that will help me to get specific info on the issue?

 

This one is the Asus A8N-SLI Premium with the nF4 chipset. The BIOS is 1009 with the updated chipset drivers. I've had no problems with the BIOS (unless it's been reading incorrectly). The CPU gets cooled by a Zalman 9500, using Arctic Silver thermal compound.