• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

***official amd 83x0 overclocking thread***

I tried increasing my cpu/nb to 2600 and got an ntoskrnl problem on boot up. My gigabyte 990FXA-UD5 is being awkward with my 8320. Turning off core performance boost stops the voltage and frequency jittering about but introduces a bug where it shuts off at boot-up and comes back on of it's own accord a few seconds later. (whether it's a restart or a cold boot)

So far I've managed 4.2GHz (200x21 @+ 0.025V LLC @ high) and temperatures via AVX linpack peak around 61. (but i'm using a noctua NH-L12 due to space constraints) Inital attempts at increasing bclk and working from there resulted in failed post.
 
I've heard good and bad about Gigabyte boards tbh. Mind you even though it's a sick clocker I'm having issues with my CHVFZ. Last night I pushed too hard and it refused to post. Getting an overclock back on was a bloody nightmare. No idea where I was going wrong tbh but it would have been one of the ten million bloody settings the board has :S

It also won't boot straight into Windows using my OCZ Revodrive. I have to enter bios and force boot from the SIL raid controller.

My old 8320 wouldn't go over 4.2ghz. Not only that but I had to do it all via the FSB too. Touch the multi and it would just bug out.

I blamed the board at the time (4+1 VRM) but others didn't have that issue on that board.
 
Kei said:
I tried increasing my cpu/nb to 2600 and got an ntoskrnl problem on boot up. My gigabyte 990FXA-UD5 is being awkward with my 8320. Turning off core performance boost stops the voltage and frequency jittering about but introduces a bug where it shuts off at boot-up and comes back on of it's own accord a few seconds later. (whether it's a restart or a cold boot)

So far I've managed 4.2GHz (200x21 @+ 0.025V LLC @ high) and temperatures via AVX linpack peak around 61. (but i'm using a noctua NH-L12 due to space constraints) Inital attempts at increasing bclk and working from there resulted in failed post.
Click to expand...

first off evening all :D

did you increase cpu/nb voltage and ht link while putting them at 2600?

cause i had blue screens for this increasing voltage cured this for me :)
 
Anyone one know what causes lock ups as opposed to BSOD? Had a few lock up recently. Or could it be one of a million things? (i suspect this is the case)
 
farmerboy-86 said:
first off evening all :D

did you increase cpu/nb voltage and ht link while putting them at 2600?

cause i had blue screens for this increasing voltage cured this for me :)
Click to expand...
Yes tried increasing the voltage, though not beyond the +0.100V offset as temperatures shoot up from there due to the cooler. Presently its running at 2400 on the cpu/nb and 2600 ht link at stock voltages.

I've dropped the clock speed back to 4.1 and set turbo to 4.3 which keeps the board happy to boot properly without the bug and seems to stop the severe fluctuations in clock speed & voltage i was getting before.
 
Cool Breeze said:
CPU NB/ and memory timings being out are the only 2 culprits I can think of
Click to expand...

Hmm, overheating definitely isn't the issue, core has been well below 50° each time.

CPU/NB clocks running at less than 2400, perhaps a voltage bump might be in order, although I aready have it raised it a bit.

My memory timings are currently 9-10-9-27-2t I think, could be wrong on that though, running below spec as well at about 1400Mhz.
 
Another interesting point, who's using the offset voltage feature and who's using manual? I've always used manual because I don't how the offset works and assumed manual would be better!
 
BSODs are faults that are detected by the OS. The processor can raise these if something goes wrong. If the CPU or the system gets a fault that cannot be picked up by the OS, you're going to get a lock up or reset. There's really not much between them. BSODs often indicate not enough voltage and so forth, where the CPU is mostly working ok.

Remember that with these processors, board permitting, you can often keep raising the voltage to overclock them until they get too hot. But the higher they go the more strain they put on the board. If it has a 4+1 phase design you'll be very lucky to get over 4.5GHz stable. You'll also get VRM thermal problems at high voltages and clocks, even if the CPU itself is quite cool - this is why the better boards have far improved heatsinks and heatpipes.
 
Evening all. I wonder if I could pick the brains of people with Gigabyte 990FXA UD3 boards.

I noticed when running intel burn test (I know you shouldn't for long but I run it around 5-10 cycles to quickly test stability) that after around the 7th run, the CPU started throttling and the Gflop scores reduced. I stuck some temp sensors on the MOS and NB heatsinks and they went to 40C + when loading all the cores. I am sure it was much more than this as to touch it was nearly burning. The heatsinks themselves on the board are very flimsy and easily movable.

Did some research and found that this is not uncommon with revision 3.0 UD3 990FXA boards (my board). I then stuck 2 mini fans from a Gskill RAM cooler directly on each of the NB and MOS heatsinks and the temps dropped by around 5 C at load. The throttling then disappeared but I only did 10 runs of IBT.

I have to say I am a bit peeved as this is clearly a design fault with this motherboard. I have somewhat lost respect for Gigabyte. I am actually tempted to return this now as it is less than 14 days since purchase and get another board. Any recommendations of a solid board which doesn't have any throttling issues with the 8-core Piledriver chips?
 
ALXAndy said:
4+1 does 4.2ghz absolute max. I made the mistake of ordering the Extreme 3 without doing the right research (I'd been off AMD for years).
Click to expand...

I'd be surprised at much higher than 4.2GHz, but a really good chip that clocks high with low voltage might manage a bit more. Still, the Asus midrange 6+2 design does better than MSI, ASrock and Gigabyte's midrange 8+2 boards.
 
Your VRMS will throttle for safety dude. Same thing happened on my Asrock. 4+1 would throttle the CPU (no way to stop it) when the VRMs got to about 42c.

That's why when my rig went bang due to a bad cable I went for the Formula Z because it's the only board with 8+2+2. Even the Saber has 8+2 only.

The Formula Z is slowly dropping in price. I paid £160 for mine.
 
alex_123_fra said:
Evening all. I wonder if I could pick the brains of people with Gigabyte 990FXA UD3 boards.

I noticed when running intel burn test (I know you shouldn't for long but I run it around 5-10 cycles to quickly test stability) that after around the 7th run, the CPU started throttling and the Gflop scores reduced. I stuck some temp sensors on the MOS and NB heatsinks and they went to 40C + when loading all the cores. I am sure it was much more than this as to touch it was nearly burning. The heatsinks themselves on the board are very flimsy and easily movable.

Did some research and found that this is not uncommon with revision 3.0 UD3 990FXA boards (my board). I then stuck 2 mini fans from a Gskill RAM cooler directly on each of the NB and MOS heatsinks and the temps dropped by around 5 C at load. The throttling then disappeared but I only did 10 runs of IBT.

I have to say I am a bit peeved as this is clearly a design fault with this motherboard. I have somewhat lost respect for Gigabyte. I am actually tempted to return this now as it is less than 14 days since purchase and get another board. Any recommendations of a solid board which doesn't have any throttling issues with the 8-core Piledriver chips?
Click to expand...

Yes, I've posted about this issue every time I've seen someone suggesting the UD3 board. It's very often reported on other forums, and you can resolve it usually by sticking a fan on them as you did. The Rev 4 board replaces the thermal design with the same as the UD5 - the UD3 R3 is not even good enough to certify the 9370 (which is just 4.4GHz base).

For the same money, the Asus 990FX Pro R2.0 is the best board.
 
Thanks Andy, Teppic and MxOrange. The Gigabyte site states that the UD3 Rev 3.0 has an 8+2 phase power design. Pretty poor I think for it to start throttling literally after around 2-3 minutes of heavy load. I am only OCing to 4.5GHz and the chip is not thermally limited and only takes around 1.38V to get there. If I want to watercool to get to 5GHz in future, it doesn't look like this board will cut it.
 
alex_123_fra said:
Any recommendations of a solid board which doesn't have any throttling issues with the 8-core Piledriver chips?
Click to expand...

Although I am now using a crosshair V formula, I started with an M5A99x evo rev. 1 for my FX8350. This was good for overclocks at least up to 4.8Ghz and I was clocking 5GHz for many benches.

I was able to do 10 consecutive runs of 100Gflops+ at 302Mhz bus x 16 = 4832GHz with 2400MHz HTand NB and 2012Mhz memory on the M5A99x.

I would recommend at least a sabretooth 990fx or the M5A99x for good overclocking.
 
You must log in or register to reply here.
Back
Top Bottom