VCore voltage with A8N32-SLi & Opteron 170

[Covered_in_bees]

This is my 1st attempt at overclocking, and I have managed to safely get to 2.6GHz on air.

However, the VCore voltage is confusing me a bit.

I read the AMD specs for this cpu here, and it's 1.3-1.35V. Not sure how correct this is, though, as it's also rated at 110W and 80A, which should make the voltage required 1.375 (W = V * A).

So, on my system without overclock, and everything set to auto, my VCore is reading 1.392V according to BIOS, CPU-Z and Speedfan (same sensor, so no surprise all are reading the same really )

Anyway, I did some tests and managed to get up to 2.643GHz 'stable' (20min test with Prime95 on both cores) using ClockGen in windows (anything beyond that failed with voltage at auto/standard). I decided to drop it back to 2.6GHz, and plan to do a 24Hr test at this speed.

However, in order to fix these settings, I have to manually set the VCore voltage due to the way the options are available in bios. Basically, this mobo has an option FID/HID change. This controls the CPU multiplier and VCore voltage. If left on auto, and I put the HTT up to 260, then the multiplier automatically drops to 8x.

So, knowing that I have to manually fix the multiplier at 10x, I need also to set the VCore.

This is where I'm a little unsure as to what to set for the VCore. I decided to set it to 1.35 (in an attempt to stay within specification for the cpu). However, the VCore still reports 1.392V in bios.

Is this just a perculiarity of the motherboard ? Or maybe the mobo is recognising that the CPU requires a minimum of 1.39V, and therefore supplies that ?

If anyone has had similar problems and would care to share their experiences, it would be appreciated.

Thx in advance

CPU-Z results
PSU is Tagan 580W 'Easycon' (TG580-U15)

 

[Quetz]

For some reason this mobo seems to overvolt. I would set your vcore to 1.2875 and see how you do and then go up from there if you need to.
Good luck with it it a great board once you get all the kinks worked out with it.

 

[fteoOpty64]

Some mobo Vcore settings are more flexible than others. My MSI Neo2 Plat can be set for 1.285v for the 3200+ and it runs nicely. The Foxxcomm 6150 K8MA would not go below 1.344v for my Opty165. It will happily do 1.55v with some rise in temp.

I have not swapped cpus on my boards as its too much work. Mine you the Opty is SOI technology so very clockable, i believe. Its on my Antec Aria SFF case so i am not letting CPU load temps over 46Deg C.

 

[EllisD420]

I have the A8R32-MVP Deluxe which is getting RMA'd due to really poor CPU performance, but messing around in the bios, i have found that it does overvolt a bit from what you set in the BIOS. When i had the voltage set to 1.35v (with the CPU overvoltage set to disabled) it would read 1.39-1.42v in windows as well as in the BIOS hardware monitor. In order to get it to 1.35v actual, i needed to set it to 1.30v, which got it to 1.34-1.36v. Keep this in mind while setting your voltage.

 

[fteoOpty64]

Hi Ellis,

"When i had the voltage set to 1.35v (with the CPU overvoltage set to disabled) it would read 1.39-1.42v in windows as well as in the BIOS". I would suspect either your board or your PSU causing such a large swing in the voltage. Not good. I guess you RMAed it for this reason ?.

My CoolMaster Real Power 450ACLY PSU is solid and does a swing of 1.344 to 1.348 (that 0.002 volts or or 0/1% variance!). I have ran my Opty at Vcore og 1.440 volts (again 0.1% swing only) for 3 days doing SANDRa95 bURN-IN , with Folding on, Itunes blasting with visualization fullon. In effect both cores maxing out 100%. No real problems there, Running 223Mhz FSB. Just mildly clocked. Now I put the Vcore back to 1.344v and it runs fine 24 X 7 everyday. My PSU gives me the confidence in voltage stability. I had the Antec TruPower 380S. But its 12V rails pulling 11.78v under load was a concern to me, although everything was functionaing very well.

With CM Real 450 under full load, my voltages are near perfect at : 3.30, 5.00 and 12.02volts. Both SpeedFan 3.28 and MBM5 concurs and so does Sandra 95. But Sandra gave erroneous minus voltages for some reason.

I am happy, there is no need for me to clock it further unless I load up 4GB Ram and do VMware with 3 guest Oses.....

 

[EllisD420]

Those were the figures that i remembered at the time. I dont know if they are actually fluctuating that much. But i when i had my ASRock 939dual, all the readings were bang on using MBM so i dont think its the PSU. The reason that im RMA'ing the board is because the move from my budget ASRock 939Dual to the "Top of the Line" A8R32-MVP actually decreased performance. My CPU scores in 3dmark06 were in the mid to high 1800's constantly, now, with this supposedely stellar board, they dropped a whopping 500pts and now the CPU score hovers in the low to mid 1300's. Also, on 4 seperate occasions it emitted a horrible high pitch noise that was only resolved by powering down the PC. I now have it back in the box waiting to have time to go out of my way to the retailer. Bah, enough about that.