MSI 870A-G54

[mgcul8r]

This is the place i've had the most luck with, so this is where i figured i should post my problem. I just bought a MSI 870A-G54 and a x4 955 BE. I installed the mobo/processor and flashed bios to latest MSI release. problem is that I CANNOT vhange the vcore at all, i mean i can physically change it in the bios, but the change is only in the bios, regardless of what i set it to, it stays 1.4v ALWAYS! i cannot even attempt to use by B.E. processor because any raise of the multiplier causes blue screens or failure of ANY stress test or benchmark. I really hope someone has an idea, cuz i'd really hate to r.m.a. this board and pay the shipping, and receive another board with the same problem. seeing how this is my first MSI board, maybe i'm missing something really obvious, but there is no hint in the manual, and MSI's site is a joke, every link to report a claim, or get any help, is an error 404 file not found. I'm losing any faith i ever had with AMD, as all the rigs i've built- INTELS were easy/fast and AMD's were slow/problematic. PLEASE someone, help me with this board, and maybe help renew some faith in AMD. thanks in advance. i want to see if one of AMD'S best processors is worth the B.E. hype

 

[mgcul8r]

hmm... well i reinstalled windows and updated all the drivers, and reflashed to the newest BIOS with afudos and still a no go, still can't control vcore- seems to be locked at 1.4v and i can't think of what to do. Does ANYONE have any ideas?

 

[xoke]

hmm... well i reinstalled windows and updated all the drivers, and reflashed to the newest BIOS with afudos and still a no go, still can't control vcore- seems to be locked at 1.4v and i can't think of what to do. Does ANYONE have any ideas?

Is there an option in the BIOS to unlock overclocking....or set things to manaul?

Maybe take a picture of the BIOS and someone will be familiar with it.

 

[mgcul8r]

hmmm...

i think it actually might be a bug with voltage monitoring software, or maybe i'm just dumb. i think the monitoring software was reporting cpu vid, not vcore. and in the process of testing, and not seeing a change, then returning the voltage back to "stock", i probably raised the multiplier AFTER returning vcore to auto, and that was probably why i was having problem. i downloaded cpuid hw monitor, and it is now showing the voltage i set in bios, but cpu-z and AOD and Everest are all incorrectly reporting vid instead of vcore. chalk it up to me being a novice, and also me being scared to just set x amount of vcore and raising the multiplier without being to see how much voltage the processor is being fed, you know? anyway i think i have it ok for now, so thanks for your help, i'm currently running 206x18 3709 MHz, and its running fine. thanks!

 

[xoke]

i think it actually might be a bug with voltage monitoring software, or maybe i'm just dumb. i think the monitoring software was reporting cpu vid, not vcore. and in the process of testing, and not seeing a change, then returning the voltage back to "stock", i probably raised the multiplier AFTER returning vcore to auto, and that was probably why i was having problem. i downloaded cpuid hw monitor, and it is now showing the voltage i set in bios, but cpu-z and AOD and Everest are all incorrectly reporting vid instead of vcore. chalk it up to me being a novice, and also me being scared to just set x amount of vcore and raising the multiplier without being to see how much voltage the processor is being fed, you know? anyway i think i have it ok for now, so thanks for your help, i'm currently running 206x18 3709 MHz, and its running fine. thanks!

It happens, lol, no worries! Also, somewhere in the BIOS is a hardware monitor screen...you can set a voltage, reboot and it should be reflected in the BIOS hardware monitor. Might be called system health, but it's in there somewhere.

 

[coonmanx]

So you are setting it on the Cell Menu page where it says CPU voltage, correct? I don't see anything else that could affect voltage.

You may have a bad board if you can't even raise the speed of your BE CPU. That in itself should raise some flags.

 

[QuietIce]

This is the place i've had the most luck with, so this is where i figured i should post my problem. I just bought a MSI 870A-G54 and a x4 955 BE. I installed the mobo/processor and flashed bios to latest MSI release. problem is that I CANNOT vhange the vcore at all, i mean i can physically change it in the bios, but the change is only in the bios, regardless of what i set it to, it stays 1.4v ALWAYS! i cannot even attempt to use by B.E. processor because any raise of the multiplier causes blue screens or failure of ANY stress test or benchmark. I really hope someone has an idea, cuz i'd really hate to r.m.a. this board and pay the shipping, and receive another board with the same problem. seeing how this is my first MSI board, maybe i'm missing something really obvious, but there is no hint in the manual, and MSI's site is a joke, every link to report a claim, or get any help, is an error 404 file not found. I'm losing any faith i ever had with AMD, as all the rigs i've built- INTELS were easy/fast and AMD's were slow/problematic. PLEASE someone, help me with this board, and maybe help renew some faith in AMD. thanks in advance. i want to see if one of AMD'S best processors is worth the B.E. hype

I felt the same way about the Q6600 G0 I bought. Everybody raving about how "easy" it was to OC to 3.6 GHz and all that. Yea, sure - never did get it above 3.4 GHz on water. Not long after that AMD came out with their quads so for me it just wasn't worth the time/effort to bother with it anymore. Currently running stock speeds because the Q6600 just won't stay OC stable on air, even at 2.8 GHz ...

 

[mgcul8r]

thanks guys

hey, got the voltage prob figured out, now on to overclocking

 

[mgcul8r]

any idea of safe 24/7 cpu/nb voltage? this platform is REALLY slow on the ram side, my ripjaws 1333 cl7 were pushing 18000 19000 20000 r-w-c on my i7 860, but the same exact set of ram is doing 9454 8403 8724 r-w-c on this amd 955be. edit those were my 3700 mhz r-w-c results, they're actually 9686-8623-8821-46.3

 

[mgcul8r]

i'm at 3800 1.43v vcore, 1800 ht on auto voltage, and 2600 nb on 1.24v

thinking since each time i bumped up cpu/nb speeds resulted in increased ram mb/s, that maybe i could try 2800, so what sounds safe?

 

[mgcul8r]

btw... no stability problems so far at current settings, ran prime95 blend, small, and large fft's each for 3 hours. cores temps never exceeded 53c, cpu 49c, mb 35c

 

[QuietIce]

any idea of safe 24/7 cpu/nb voltage? this platform is REALLY slow on the ram side, my ripjaws 1333 cl7 were pushing 18000 19000 20000 r-w-c on my i7 860, but the same exact set of ram is doing 9454 8403 8724 r-w-c on this amd 955be. edit those were my 3700 mhz r-w-c results, they're actually 9686-8623-8821-46.3

cpuNB voltage range is 1.175-1.250v and increasing it can increase CPU temps, as you might expect.


What benchmark/version?

 

[xoke]

I felt the same way about the Q6600 G0 I bought. Everybody raving about how "easy" it was to OC to 3.6 GHz and all that. Yea, sure - never did get it above 3.4 GHz on water. Not long after that AMD came out with their quads so for me it just wasn't worth the time/effort to bother with it anymore. Currently running stock speeds because the Q6600 just won't stay OC stable on air, even at 2.8 GHz ...

Thats odd, the q6600 was almost guaranteed 3.6 on anything. I had 2 of them, 3.6ghz on air with everyone...I never ran one at stock speeds.

 

[mgcul8r]

not me

my q6600 wasn't stable after 3.2 , i had it at 3.3 but it would fail prime95 overnight. that was with a tp45hp which aren't too good with quads either

 

[mgcul8r]

cpuNB voltage range is 1.175-1.250v and increasing it can increase CPU temps, as you might expect.

hey Quietice, why do people say up to 1.4 - 1.5 is safe? not that i want to run it that high, i just want to know why the varying specs from different people on forums? maybe they have high end cooling. btw, its a x4 955 C3 stepping with a xiggy 1283V i took off my i7, mobo died. would cooling effect how high a "safe voltage" can be?

and you asked which benchmark i used to compare, it is everest ultimate cpu and cache benchmark

 

[coonmanx]

OP what was the issue with the voltage? When you figure it out it's a good idea to post what the solution was so we can know as well.

Edit: Sorry. I see that you said it was the monitoring software.

 

[QuietIce]

my q6600 wasn't stable after 3.2 , i had it at 3.3 but it would fail prime95 overnight. that was with a tp45hp which aren't too good with quads either

Mine was IP35 I think - it's been awhile since I messed with it. It was running fine at 3.3 GHz under water but everything went south when I switched to air.

hey Quietice, why do people say up to 1.4 - 1.5 is safe? not that i want to run it that high, i just want to know why the varying specs from different people on forums? maybe they have high end cooling. btw, its a x4 955 C3 stepping with a xiggy 1283V i took off my i7, mobo died. would cooling effect how high a "safe voltage" can be?

and you asked which benchmark i used to compare, it is everest ultimate cpu and cache benchmark

Don't run Ultimate - just thought I could compare with my 940BE.

With AMD clocking it's mostly load core temp that determines how high you can run the vCore (and cpuNB) - but individual preference plays a role as well. AMD says the 955BE C3 is good for 0.825-1.425v. After spec I go by what the load core temp is, though I'd stop at 1.50v regardless of temp unless it's water-cooled, then I'd go to 1.55 vCore. The board also makes a big difference. Quality power delivery is a must for high vCore and it's more of what people are reporting for results than any information the manufacturers decide to release. One thing for sure, though, if it doesn't have a MOSFET sink I wouldn't exceed 1.45v (because of board limitations).

Note when I say "water cooled" I don't mean an H50 or some lame kit but a full-blown loop with at least a 120x2 (i.e., "240") rad and a Swiftech (Laing) pump ...

 

[mgcul8r]

quietice said :" One thing for sure, though, if it doesn't have a MOSFET sink I wouldn't exceed 1.45v (because of board limitations)."

what do you mean by mofset heatsink? here's a pic of my mobo, is it the jeatsink below the cpu socket, but above the pci-e slots? (the one that says MSI)

http://www.newegg.com/Product/Image...870A-G54 AM3 AMD 870 SATA 6Gb/s USB 3.0 ATX A

If so, is that heatsink good enough?

 

[mgcul8r]

hey QuietIce, seems like every time i have a problem, you're the one who helps me, just wanted to thank you

 

[QuietIce]

The MOSFET are that bank of chips between the CPU and the backplate-I/O ports. Your board doesn't seem to have one. I'd be very, very careful using too much vCore ...