Crazy BIOS Settings

[VirusType2]

Glad to hear all the good news, might drop MSI a line about the SATA port - it may be an on-going issue they plan to address in a future BIOS update...or even something they don't know about....can't hurt. Anyway, it's good all is good...Enjoy!

Yeah, I created a thread at MSI forums on May 1st, but apparently didn't give them enough information. I have been too frustrated to go back, but I will try to follow up on it soon. And thanks again for the advice.

[Tradesman]

No problem.....how's the weather out there, getting humid yet? Spent four years at Langley AFB way back when.

[VirusType2]

No problem.....how's the weather out there, getting humid yet? Spent four years at Langley AFB way back when.

It's actually pretty cool this morning. But you know, it gets nasty here in the summer. Miserable.

Hey, I'm glad I still have you here. The thing is, I seem to have lost some performance in one of my games overnight and was pulling out my hair messing with things trying to figure out what changed, or if it was just my imagination or something. Then it dawned on me, I lowered the voltage on the RAM since losing the performance, and I think that's the only thing that changed. Can this possibly change the performance?

I was under the impression that voltage does nothing for performance, but only allows setting higher clock speeds - making it more stable in most cases - at the cost of increased power consumption, higher temperatures, and decreased life expectancy. Therefore, I had lowered the RAM voltage from 1.55 to 1.50, since it is stable at that voltage. EDIT: and let me clarify what I mean by stable is that there were no noticeable problems/errors or shut downs/crashes. Everything seems fine. I did not actually test it with some memory test program.

TL;DR - is it possible that lowering the RAM voltage lowered my framerate in games?

Thanks

EDIT: also, bro - didn't you say that these MSI boards are notorious for giving the RAM/components lower voltage that specified? Do you have any more information on that?

My very high end PSU voltages are +/-.3 within exact spec, and are just over. (12.3; 5.03; 3.33), if that means anything.

[Tradesman]

Combination of things in play that can affect things - the BIOS may not be written/set as exact as it can be (happens w/ many mobos) so playing w/ voltage can affect performance i.e. 1.5 set on different mobos could vary from say 1.48 - 1.53 as an example. (Can check it to find exact voltage if you have a mind to or know an electronics wiz). DRAM is speced at a specific voltage, again say 1.5, but for performance, just like a mobo using FSB, the sticks may have a sweetspot, so with a mobo you might run a 400 FSB and be fine, it might actually bench better at say 398-399 if that's a sweet spot - same with DRAM, might perform better at 1.49 or 1.51 than it does at 1.5...or....on a 7-7-7-21 set you may get better performance with say 7-8-7-21 (another reason people who buy two sets of DRAM, same model/etc may have two sets that don't play well together)

When you take into consideration the various components that can be adjusted, you may get an idea of how some OCers end up spending hours, days, weeks fine tuning a box to pull the last NTH of performance out of them - or why there are actually teams of people that work together trying to set records with 3DMark benchmark testing (been there done that)

[VirusType2]

Good information, thanks a lot.

I spent a couple of hours researching my question. I knew it would be hard to find because I hadn't already come across it in the past. But I finally got a hit. It seems this guy knows what he's talking about. He's trying to cram as much information in the least amount of words. So he doesn't go into detail, but I think my question has been answered:

Power

Chips at higher speeds may need more power. Raising the vcore voltage on a CPU might enable it to go at slightly faster speeds but by doing so you add a lot more heat output from the CPU. The vcore of a processor is the voltage at which a chip is set to run at with the stock speed. This voltage may need to be changed when the multiplier is raised because otherwise the transistors in the chip won't switch fast enough - transistors switch faster the higher the supply voltage. If there is not enough voltage then the chip will begin to make mistakes and give bad data results. Good cooling is needed to keep the system stable at higher speeds. Raising the vcore too much may harm or shorten the life of your system. Raising the vcore can also greatly affect the stability of the system. This is where a high quality PSU will come into play. While many cheap, no-name brand PSU's will crash and die with more vcore, a good quality one will live to serve you for a long time.

http://en.wikipedia.org/wiki/Overclocking

So I think with the XMP overclock, increasing the voltage may have increased the performance of my computer under load even though it was stable at the nominal voltage. "It allows the transistors to switch faster." Well, that's what the guy says, and I don't have any reason to disbelieve it. I know I just could have experimented to find out if I notice any difference, but I just like to know if I'm not just imagining it or if it's possible. And it seems it is.

I will have to test it again. I'll let you all know how it goes, just FYI.

EDIT: Maybe I'm misunderstanding what he's saying. Hmmm. We'll see. Let me know if you have any corrections or clarifications.

Thanks again

[Tradesman]

What he says is accurate and keep in mind while you can raise the MC voltage (the MC is in the CPU) that doesn't raise the actual vCore, so at times you can get a problem if the MC voltage is good but the vCore is a bit low for the frequency you are running at

[VirusType2]

What he says is accurate and keep in mind while you can raise the MC voltage (the MC is in the CPU) that doesn't raise the actual vCore, so at times you can get a problem if the MC voltage is good but the vCore is a bit low for the frequency you are running at

But raising the voltage allows the transistors to switch faster in RAM as well? That's what I'm trying to figure out. I mean, it makes sense to me that it would apply to both.

This voltage may need to be changed when the multiplier is raised because otherwise the transistors in the chip won't switch fast enough - transistors switch faster the higher the supply voltage.

I put the DRAM voltage up a bit, how I had it before. I can't tell right away if the performance was better. I certainly didn't seem worse. I will have to test it some more.

Temperatures from my computer seem a few degrees C higher though. I mean, I am taking measurements but, it's kind of hard to tell outside of a controlled environment. Also because I did make a change in my cooling.

I increased the speed of the top fan, and sealed some holes in my case where I don't want hot or unfiltered air coming in. Anyway, this seemed to drastically lower the rear exhaust fan performance - whether it was previously pulling in just expelled (hot) air or if the top fan is just taking away air pressure from the adjacent rear fan, I haven't a clue.

But I will probably turn the top fan back down to medium. If not only because it's a little too loud for my taste, but also because it didn't seem to help - in fact, like I said, it seems to have made things worse.

The other thing is - with the settings changed as above - I did get a BSOD this morning with a _possible_ RAM related cause but I didn't have my system event settings right, so I couldn't write down the culprit fast enough before the computer reset. I changed the system event settings.

May have to bump the RAM down (or up??) again.

:P

[Tradesman]

Keep us informed