** Official ** Gigabyte UD3 P45 Series -- EP45-UD3 / EP45-UD3R / EP45-UD3P

[nullface]

Hep hey, just got a secondhand P45-UD3P 1.0 and I'm on the hunt for 4.5GHz on a E8400.

But there are a load of settings to fiddle with compared to my old Commando so a few tips are more than welcome.

I can boot into windows no problem at 4.5 and run a few minutes of Orthos before it gives me a "Rounding was 0.5, expected less than 0.4" error.

Settings:

500x9
LLC enabled
CPU VCore: 1.375
CPU Termi: 1.240
CPU PLL: 1570
CPU Ref: Normal

MCH Core: 1.320
MCH Ref: 0.7650
MCH/DRAM Ref: Normal
ICH I/O: Normal
ICH Core:Normal

DRAM Voltage: 2.1
DRAM Termi: Auto

2.00D, X.M.P Disabled, Performance Enhance std. and SPD Normal.

Any suggestions where to start (what to touch and what to leave be) to find some stability?

[Falkentyne]

I was going to try and buy that board and slap some cpu's in it to gun it high, but after people were getting desperate on ebay and bidding over $200 for a new one, and then later, I saw a bid of over $200 for a ***USED*** one, and even saw one sell that had a couple of bent pins (I emailed the seller) for over $140, I was like....forget it.... can only trust XS sellers or new items if elsewhere...

[In-Fluence]

Makes me wonder why my USB3P would have gone for less than £50 about 6 months ago if I hadn't reserved it. $200 is a lot for an old gen board and I'm glad I kept hold of mine.

Anyone got any settings or hints for an e8x00 above 600fsb? I can do 1M PI runs but not 32M at 600x6 and 605fsb max. I've tried all combinations of cpu/nb skews which don't seem to make a huge difference. Refs have helped me get 1M stable even at low volts (1.34vtt and 1.38mch) but it always hangs or gives rounding errors around 10 secs of 32M even with v.high volts with various refs... Currently on F2d bios (not that there's a lot to choose from on the USB3P)...help!

[Falkentyne]

Getting above 600 FSB is more luck than anything else. The highest fsb I ever saw mentioned was like 613 FSB and that was on a ep45 ud3p. And board revisions mattered, too.

I guess you may have to subzero the NB and throw a lot of volts at it...

[nullface]

I was going to try and buy that board and slap some cpu's in it to gun it high, but after people were getting desperate on ebay and bidding over $200 for a new one, and then later, I saw a bid of over $200 for a ***USED*** one, and even saw one sell that had a couple of bent pins (I emailed the seller) for over $140, I was like....forget it.... can only trust XS sellers or new items if elsewhere...

Wow! I just got a almost new UD3P 1.0 for $72 with all the accessory's and the box.

[Don_Dan]

@nullface: Play around with voltages a bit, you should be able to get it stable if it's already running Orthos for a few minutes. Just make sure your NB isn't overheating. And maybe raise MCH ref a bit, to 0.776V.

@In-Fluence: Did you try higher Vmch? Make sure your NB is cooled properly, 32M stresses it a lot more than 1M.
Which memory kit and divider are you using?

[In-Fluence]

Hi Don_Dan. Currently on GSkill's eco 1600 rated 7-8-7-24 1t but can do plenty more on this board. 2.00D has worked best for me but all dividers work with varying success. I've tried plenty of MCH (and have before whilst running q9650 benches) and have plenty of big fans blowing away at it. I can try and re-mount the mch sink again which may offer better results, but the brick wall at around 600fsb has me a little stumped... tried PCIE frequency, lowered tread to 15, tried low & high vtts with varying ref values. Seems like it could be end of the line but will give it another go in a day or so. Thanks for the hints.

[Don_Dan]

Hi Don_Dan. Currently on GSkill's eco 1600 rated 7-8-7-24 1t but can do plenty more on this board. 2.00D has worked best for me but all dividers work with varying success. I've tried plenty of MCH (and have before whilst running q9650 benches) and have plenty of big fans blowing away at it. I can try and re-mount the mch sink again which may offer better results, but the brick wall at around 600fsb has me a little stumped... tried PCIE frequency, lowered tread to 15, tried low & high vtts with varying ref values. Seems like it could be end of the line but will give it another go in a day or so. Thanks for the hints.

To me it seems like you tried almost everything I can think of...
Maybe it's just the limit of your chip, my E8400 doesn't like 32M above 605MHz FSB on air as well, not even on the REX.

[Falkentyne]

Wow! I just got a almost new UD3P 1.0 for $72 with all the accessory's and the box.

How'd you manage that?
If it was on ebay, can you link me to the item number?
I hardly trust people there; you never know when you are going to get a dead board, and then get hassled with "no refunds." But I guess things are better than they were eight years ago. The last mainboard I bought on ebay was advertised as a "working pull", but was COMPLETELY dead, not only were some of the caps bad, the chipset itself was dead. Luckily Homie over at the defunct motherboardrepair.com fixed everything...

If something like this had happened now, I would have filed a chargeback for fraud....