A little undervolting adventure

[scaryjim]

6A each on those, so it shouldn't be. I can't find anything on google about the minimum current for each line on that motherboard though. Also I find it suspicious that it's only when I plug the 12V ATX in that I get the problem, so 12V line seems most likely: although as I say there *should* be enough current on it. it's *very* frustrating...

EDIT: thought I'd figured it out a few minutes ago: it occured to me that since I was so close to the load limit fo the PSU that it might be causing vdroop and with the undervolting in place that could be enough to cause the CPU to drop out. So I reset the CPU voltage to stock. Sadly, no difference

[claylomax]

Hello again,

thanks to some luck, I've had the chance to get a Phenom II X4 965 BE for an upgrade from previous X3 720 BE (though fully unlocked, but nevertheless).

That said, it's a 125W unit as opposed to my 95W 720, so wanted to undervolt that as well.

As a reference, I've registered that my unlocked X3 720 would happily run four cores at 2.8GHz with 1.275V (down 0.05V from stock voltage of 1.325V) with a temperature of 46C with the fan running at 1600RPM.


Sadly, my stock data of the new CPU is unreliable, since it was gathered with EasyEnergySaver enabled, which resulted in less voltage and as such lower temps and fan speed RPM, but I'll include it nonetheless.

So, here we go:

Stock voltage, 1.35V:
Average temperature: 51.7C
Average RPM: 1974

-0.025V, 1.325V:
Average temperature: 50.6C
Average RPM: 2287

-0.05V, 1.3V:
Average temperature: 50.3C
Average RPM: 2306

-0.075V, 1.275V:
Average temperature: 49.5C
Average RPM: 2277

-0.1V, 1.25V:
Average temperature: 48.5C
Average RPM: 2243

-0.125V, 1.225V:
Crash

So here we go, again the CPU runs along smoothly (at least for now, it's been on load for several hours, so probably will be OK) with 0.1V less than intended, which lowers the temperature by around 4 degrees. Currently I've re-enabled EES and am currently running the fan at 1900RPM (completely reasonable noise) with the temperature hovering around 48-49C.

Interesting thing about the difference a stepping/revision makes: at the moment I'm running the CPU at 600MHz more than the previous one with 0.025V less running through it. Win-win

So much for now.

Cheers,
Queelis


EDIT: nvm, glitch.

It's funny, 49c is the load temp of my Phenom II x4 965 ... with a voltage of 1.55 in the bios.

[scaryjim]

Using a stock cooler? With the same fan profile? Same ambient temperature? Keeping a Phenom II at 49deg on air cooling isn't exactly hard, it's all about the details!

Anyway, got my PC up and running with the FSP 1U PSU today: idle draw @ 1.325V: 44W!! Idle at 1.05V: 43W!! I also tested it out with a moderate CPU and GPU loading (playing a fairly busy section of the Neverwinter Nights OC) - 55W!! Don't know what that picopsu was complaining about, but as long as linitx don't try to mess me about with the DSR return I think I've ended up with a better - and cheaper - option anyway.

Given that the old PSU gave me 64W idle, I think it's fair to say it was well on its way out even a year ago, poor thing. I'll do a full set of new figures once I've modded the old PSU chassis into a fan mount + flexATX PSU mount.

[Queelis]

It's funny, 49c is the load temp of my Phenom II x4 965 ... with a voltage of 1.55 in the bios.

Exactly as scaryjim pointed out, it doesn't work from user to user, since you might have a case with 11 fans spinning at 2000RPM and a Scythe Orochi or some other hulk of a heatsink.

[Sputnik]

This is a very interesting topic. I have a 720BE which I run at 3.0 GHz @ stock volts. But, when I put it together originally my motherboard was running it at 1.425volts.

I've not bothered undervolting it as I have a huge cooler on it and it only run at 27*c @idle and gets to about 38*c on full load.

Although it is food for thought.

[Queelis]

This is a very interesting topic. I have a 720BE which I run at 3.0 GHz @ stock volts. But, when I put it together originally my motherboard was running it at 1.425volts.

I've not bothered undervolting it as I have a huge cooler on it and it only run at 27*c @idle and gets to about 38*c on full load.

Although it is food for thought.

Well, you can always do it for the sake of saving electricity - why waste it when you can save it (i.e. pay less). That's one view of eco-friendliness I can accept

[kalniel]

Meh, no need - we need new examples!

I encourage everyone to undervolt and tell us (or me, but I see I'm not the only one interested in that, thankfully), especially the folding folk who don't overclock - pay less for electricity/save the planet/whatever's your cup of tea

Well.. my Gigabyte mobo comes with something they call Dynamic Energy Saver 2, which undervolts and switches power phases down etc. to save power. I've been using it on the mildest settings and it takes my CPU voltage (i7 950) down from 1.248 to 1.216 fully loaded, and 0.896 instantaneous idle (before various sleep type stages kick in). Not a huge undervolt, granted!

Can't tell if it's had much effect on temps, and not sure whether to push the clock higher for this voltage (seeing as it cuts back so much on idle anyway) or whether to apply more aggressive DES settings or a -ve offset in BIOS.

[Queelis]

Well.. my Gigabyte mobo comes with something they call Dynamic Energy Saver 2, which undervolts and switches power phases down etc. to save power. I've been using it on the mildest settings and it takes my CPU voltage (i7 950) down from 1.248 to 1.216 fully loaded, and 0.896 instantaneous idle (before various sleep type stages kick in). Not a huge undervolt, granted!

Can't tell if it's had much effect on temps, and not sure whether to push the clock higher for this voltage (seeing as it cuts back so much on idle anyway) or whether to apply more aggressive DES settings or a -ve offset in BIOS.

Yup, mine has something similar as well. I've found that even using the most aggressive setting it undervolts the CPU to 1,275V loaded, so I'm doing a bit better anyway.

Granted, my PC did reboot all of a sudden today, but don't know whether it's a voltage thing, since I had no errors on my folding@home work unit (usually when the PC is not stable the core throws an error and a work unit is lost).

Should you decide you don't need more performance, you might think about undervolting

[claylomax]

Exactly as scaryjim pointed out, it doesn't work from user to user, since you might have a case with 11 fans spinning at 2000RPM and a Scythe Orochi or some other hulk of a heatsink.

Sorry to everybody; I thought I had listed my Thermolab Baram in my system specs. Voltage is 1.55 in the BIOS, 1.52 idle on Windows desktop and 1.47 load (OCCT, prime95); by the way last week I changed my case from an Antec Three Hundred to a Silverstone FT01B-W and temps are a bit higher, I also went from a push-pull configuration to a single blue led fan; however I still think that the problem might be that I didn't apply the TIM properly. Edit: I quoted the wrong post.

[Queelis]

How come your idle voltage is higher than load?

[kalniel]

Should you decide you don't need more performance, you might think about undervolting

Going both ways at the moment. Firstly I found prime 95 small FFTs did the best job of heating things up, getting me up to 78C on my hottest core (I have one hot one, 5C or so above the rest ) with my previous settings (1.216 with DES2.. only I found it was bumping to 1.232 on occasion during the prime test). Now using the higher levels of DES I'm down to 1.184v peak, and temps have dropped to 71C on the hottest core (CPU fan = slow/1600rpm). Might be able to do more in software still, using the custom sliders, but I should probably go to BIOS for more undervolting really.. just feel that using an inbuilt application is likely to be kinder on the warranty

[Queelis]

Going both ways at the moment. Firstly I found prime 95 small FFTs did the best job of heating things up, getting me up to 78C on my hottest core (I have one hot one, 5C or so above the rest ) with my previous settings (1.216 with DES2.. only I found it was bumping to 1.232 on occasion during the prime test). Now using the higher levels of DES I'm down to 1.184v peak, and temps have dropped to 71C on the hottest core (CPU fan = slow/1600rpm). Might be able to do more in software still, using the custom sliders, but I should probably go to BIOS for more undervolting really.. just feel that using an inbuilt application is likely to be kinder on the warranty

Well, undervolting shouldn't harm the CPU, since it's not too little current that does the damage. Besides, no one's going to know that you've been undervolting, I think.

Let us know how things go in the BIOS

[claylomax]

How come your idle voltage is higher than load?

It's called Vdroop and on my motherboard it's massive; that's why I need to set it so high in the BIOS otherwise it would BSOD

[kalniel]

Changed things around - dropped the multi by one (disabled turbo, basically), running RAM at lower multiplier to ease uncore frequencies, and bump bus speed. Now stable at 1.12V (and possibly lower, still just DES2) . Should be able to lower DRAM voltage at this speed too. 0.2 of a second slower 1M prime time though

[Attila the Bun]

I thought I would give undervolting ago this weekend as I'm going to have a girlfriend and children free saturday morning and I have a 965BE to play with as well.

Can I ask what software you used to monitor the various readings (any links would be appreciated) and where you look to find the stock voltage.

I thought I would do this before overclocking because I get to play with the system settings but I'm less likely to bugger up the CPU.

[kalniel]

CPUz

http://www.cpuid.com/softwares/cpu-z.html

Handles the vCore which is the main target. Stock voltage varies for each and every processor, so it'll just be the default before you change anything. You probably want to load up your computer with something like prime95 (http://www.mersenne.org/freesoft/) to get the 100% load voltage rather than idle.