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

[rockfella]

I am itching for any of the P or R. Any one noticed any difference OCing between the P & R? I know i will never croossfire even if 4850 gets half the price. I wonder how good is DOLBY sound compared to the normal on-board sound of the "R"?
TIA.

[Jor3lBR]

I am itching for any of the P or R. Any one noticed any difference OCing between the P & R? I know i will never croossfire even if 4850 gets half the price. I wonder how good is DOLBY sound compared to the normal on-board sound of the "R"?
TIA.

Get the P couple of bucks now will save you from headache latter...

[njl]

I'm currently running a Q9650@4.25.

Running OCCT as we speak and it *seems* stable, but if I download something I'm getting corrupt files, so somethings wrong.

Any suggestions to what I can do to fix this?

Edit: OCCT just crashed - guess that's why :-/

[fatguy1992]

I just got my Q6600 to 3.2GHz on my new UD3P. Heres my settings could someone please tell me if there is something wrong with my settings.



Thanks.

[technut36]

I am itching for any of the P or R. Any one noticed any difference OCing between the P & R? I know i will never croossfire even if 4850 gets half the price. I wonder how good is DOLBY sound compared to the normal on-board sound of the "R"?
TIA.

i doubt anyone here will suggest the r over the p for a few bucks so just buy the p

[technut36]

it will not allow at 4-4-4-12 it's to tight to at that oc on the cpu loos'en it up to 5-5-5-15 as 1066 memory and.I bet you get 469 and above..I managed to
with some Corsair XMS2 2GB (2 x 1GB) 240-Pin ddr2 sdramddr2 800
http://www.newegg.com/Product/Produc...82E16820145034

sorry about the shot Networking between Machines to take this shot here it is
from my antec 300 case with the ga-p35-ds3r machine with my E6750 alittle
oc'd had to 3.6ghz stable b4 turned it down sense was at 3.8ghz stable with
some Corsair Dominator 2GB (2 x 1GB) 240-Pin ddr2 SDRAM ddrs 1066 but took them out for another board...
best a luck my friend and hope it works out..I plan to get some of that memory also later on there's only a alittle differents between the p35 and p45
bio's setup..i got p45 gigabyte board under my bed ds3r..

Thanks for you comment bro.

So I played around last night with the pi blacks, did some research, tried cas 4 and cas 5.

I was able to do

CPU: 465x8.5 = 3.95ghz, 1.24v
RAM: 465x2= 930mhz @ 4-4-4-12, 2.1v
I couldn't push the ram anymore then 930 @ cas 4.


CPU: 500x8.5 = 4.25ghz, 1.24v
RAM: 500x2= 1000mhz @ 5-5-5-15, 2.1v
I couldn't push the ram anymore then 1000 @ cas 5, I have a feeling it will do more with higher voltages and also if I lower the cpu multiplier


So out of the above two settings, according to my research the cas 4 setting will be faster.

Give me some input guys let me know what you think. I am gonna post my bios settings in a few hours.

[fatguy1992]

I recently upgraded to a UD3P. When ever any BIOS settings is changed from default, it is fine when I apply the settings, it loads windows, runs perfect. But if I turn the computer off and then back on (no matter how long i leave it) it starts but when it should post it restarts and keeps going in that cycle. To fix it I have to switch it off at the PSU, and then switch it back on and it will work but the BIOS said that overclocking failed and its un-overclocked itself. First time I thought it must have been un-stable. But this time it was able to run wPrime 1024m, I know its not the most intense program, but if it could pass it, there is no way it could be so un-stable not to post. BTW if I restart using Windows it is fine, but it does the same thing if I use the button on my case.

So , is there some settings I need to change?

Thanks for any help.

[technut36]

I recently upgraded to a UD3P. When ever any BIOS settings is changed from default, it is fine when I apply the settings, it loads windows, runs perfect. But if I turn the computer off and then back on (no matter how long i leave it) it starts but when it should post it restarts and keeps going in that cycle. To fix it I have to switch it off at the PSU, and then switch it back on and it will work but the BIOS said that overclocking failed and its un-overclocked itself. First time I thought it must have been un-stable. But this time it was able to run wPrime 1024m, I know its not the most intense program, but if it could pass it, there is no way it could be so un-stable not to post. BTW if I restart using Windows it is fine, but it does the same thing if I use the button on my case.

So , is there some settings I need to change?

Thanks for any help.

what happens when the computer resets back to the optimized stock settings and then you turn it on from the button. Does it continue to go back to the cycle? or does it turn on fine from the button at the factory settings?

[fatguy1992]

It runs fine at 100% stock. But its way over volted and I like using a 3.2GHz Q6600 .

[technut36]

so that means it is definitely not the switch or your front panel plugs.

this is what i would do, start lowering you FSB and voltages in small increments, find out at what point does it boot fine.

maybe you should start at stock and work your way up. Also try resetting or switching ram or psu if you got spares.

hopefully someone else will come up with a faster solution but mean while get started.

[fatguy1992]

I'v used this CPU and RAM before on my P5E so I know what its capible of. But anyway atm its at 2.4GHz with 1.056v with all the power saving thingys off. It was always 100% stable at these settings on my P5E.

[technut36]

knowing your cpu and ram is one thing but how it perform can vary from mobo to mobo. anyhow good luck buddy hope this problem is solved soon for you.

[fatguy1992]

True, the only thing that confuses me is like I said, its wPrime 1024m stable. But as soon as I restart it won't post. WTF. If it wasn't stable enough to post, then why did it post the first time and not the second.

[pangingIII]

fatguy1992,

what you describe (it worked the first time but once you try to shutdown and restart the computer the bios won't post) happens a lot to Gigabyte boards, many times i find that the board didn't actually like all voltage settings i give it in BIOS, and when this happens i would reset cmos information and go back to all of the previous settings and retry by adding (sometime reducing) volts to cpu, mch, vtt, ram, sometime try using different cpu/mch/dram's GTLref value. i can't tell you exactly where to look in your settings.

i've not read every thing that you post, but for me, with an e8400 installed, if i use mch strap other than 333 straps (memory multi 2.0B, 2.40B) i normally get a double bios post (but this is normal the board is just resetting and to use other mch strap other than default), however, sometime i find that i need to use 333 strap to get a real stable system (and also get a more lower voltage use in the system). if i want to overclock ram and sometimes i have to use a different straps , then that's just for my ram overclocking fun and it's just another story.
i can't be much help, but hope to see you find and get a stable setting soon.

[fatguy1992]

Thanks, I just did another test and I got it to work Finally. Its at 3.0GHz. I'm going to reduce the volts later.

[flyfisher]

Here are my observations on this board. You really need to adjust CPU and MCH reference voltages (GTL's) on this board. Most seem to be using auto which is fine for stock. CPU termination and MCH voltages default values are 1.2 and 1.1 volts, respectively. At these voltages the reference voltages default to 0.760 volts. This corresponds to a CPU GTL reference of 63% (0.760/1.2 = 0.633) and an MCH GTL reference of 69% (0.760/1.1 = 0.691). These values may not be suitable for the overclock you are trying to obtain. Manually set these voltages then start raising the CPU termination voltage. You will see the CPU and MCH reference voltages rise also. Note that the ratio of CPU ref./CPU term. is always around 63%. This is what the auto setting does, changes the CPU reference as the CPU termination rises to keep the 63% GTL. While this is the generally accepted value for 45nm CPUs, a ratio is 67% is preferred for 65nm, especially quads. For example, if your CPU termination is 1.3 volts, choose the CPU reference of 0.863 volts (0.863/1.3 = 0.664.

Another problem arises with the MCH reference. If you leave it on auto the BIOS will set it to the same value as the CPU reference. So if you set the CPU termination to 1.3, the CPU and MCH reference will be 0.823 volts on the auto setting. But if you have set the MCH to 1.4 volts, your MCH GTL is 59% (0.823/1.4 = 0.587) which is likely too low. I found that while Linpack, Prime small FFT and Blend would all pass, large FFT would eventually cause a lock-up if the GTLs weren't set correctly. For my Q9550 on the UD3R I found that using 1.3 volt CPU termination and MCH with both references set to 0.863, giving a 67% GTL for each, gives me total stabilty at 475 x 8.5. Of course each board/CPU will be different so you just have to test.

[fatguy1992]

Thanks I think I understand all that

Also would you be able to tell me what CPU Clock Drive is?

As well as CPU and MCH Clock Skews?

Thanks

[fatguy1992]

I just overclocked to 3.2GHz (400x8). Could u take a look at my GTL settings.

CPU

0.863/1.3 = 0.66384 and so on

MCH

08.03/1.2 = 0.66916 and so on

So is that right?

Also what should I do with my DRAM Channel A and b references?

[technut36]

I just overclocked to 3.2GHz (400x8). Could u take a look at my GTL settings.

CPU

0.863/1.3 = 0.66384 and so on

MCH

08.03/1.2 = 0.66916 and so on

So is that right?

Also what should I do with my DRAM Channel A and b references?

hey fatty post your entire bios settings for more help

[fatguy1992]

K, whats in the brackects in what auto set it too.

Vcore: 1.275v
CPU Termination: 1.3v
CPU PLL: 1.5v
CPU Refences: 0.863

MCH Core: 1.2v
MCH Reference: 0.803
MCH/DRAM Reference: Auto (0.9)
ICH I/O: Auto (1.5)
ICH Core: 1.1v

DRAM Voltage: 2.0v
DRAM Termination: Auto (0.9)
Channel A Reference: Auto (0.9)
Channel B Reference: Auto (0.9)

Therefore

CPU

0.863/1.3 = 0.66384 and so on

MCH

08.03/1.2 = 0.66916 and so on

[technut36]

compare your settings to other q6600 on ud3p in this thread.

[flyfisher]

K, whats in the brackects in what auto set it too.

Vcore: 1.275v
CPU Termination: 1.3v
CPU PLL: 1.5v
CPU Refences: 0.863

MCH Core: 1.2v
MCH Reference: 0.803
MCH/DRAM Reference: Auto (0.9)
ICH I/O: Auto (1.5)
ICH Core: 1.1v

DRAM Voltage: 2.0v
DRAM Termination: Auto (0.9)
Channel A Reference: Auto (0.9)
Channel B Reference: Auto (0.9)

Therefore

CPU

0.863/1.3 = 0.66384 and so on

MCH

08.03/1.2 = 0.66916 and so on

Fatguy1992 it looks like you got the hang of my post. Other than Vdimm, I left the rest of the voltage settings on auto. I also left clock drive, skew, and channel A and B references on auto. At 400 fsb with a Q6600 you might need 1.3 MCH or more especially if you are running anything other than 1:1 memory. My old Q6600 did at least. What is your MCH frequency latch and memory multiplier? Also don't use the "extreme" memory performance enhance setting, I could never get anything stable with it. Use turbo or standard.

[Xvys]

I will try that ratio flyfisher. I just got a new UD3-R board, my first Gigabyte. I also have the endless reboot issue that FatGuy 1992 mentioned, having to turn off the power to get into bios. Frankly, I don't find this board as user-friendly as my Asus P5K-Deluxe which I got rid of to buy this board. I've had little luck bettering the stable 4.23 Ghz @ 1.34v (E3110 C0) of the old board. I still haven't matched it's memory latencies, yet, which has been a little disappointing.

[fatguy1992]

Its stable at only 1.2v for the MCH. I think my MCH frequ. latch is 200MHz (but i might be wrong) and i'm using the standard memory performance enhance setting.

Thanks for all ur help .

[flyfisher]

The endless reboot issue has me stumped. I know when I tried a setting that didn't work it would reboot three times then reset itself to stock and boot. If I had a setting really out of whack it would try to boot but never get to the BIOS; a CMOS clear would fix that. But never an endless loop of reboots. I hope you guys can come up with an answer.