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

[the_real_7]

ok im back after doing a nice case cleanup looks really hot and a nice cooling system , advanced my my wc and made in silent and 8 c cooler at 48 c load stable, you guys and Hoss331 helped me lock my system at 4050 at 4 4 4 10 and its been great
so i wanna tweak a lil bit should i update my bios from f8 and try that power saving stuff or should I find some new ram that can do 9 x 500 and hit 4500
or does neither matter and just run system another year without a upgrade to bios or ram you guys got me jealous with all this tweaking , and with a lil more oc I could justify not upgrading to i9

[godsa]

One thing I've noticed about overclockers especially old school. We look at everything a little deeply. Because with UD39 its so easy to overclock a ...ok I won't say it.


Simply, think basic when you touch this virgin. Keep it simple

[In-Fluence]

Uh, uh, I hope your CPU is fine! If it isn't, mine could have degraded too. I used my Q9650 at 1,6VTT and 1,55Vcore to bench at 9x515 at the MIIF As you already know, it needs insane volts to be able to bench at those clocks. Let me know how it goes! Hope you can get it figured out.

Well, after much messing around, I think it's down to mosfets overheating. Even though my house fan is pointing straight at the board, leaving it off overnight and running a test at 500x9 small ffts runs for about 10 minutes before the 4th core fails, but doesnt last much longer, no matter what the vcore is (with reasonable, and above previously stable volts for that speed). Then I tried it again immediately after, and it fails within seconds.
500x8.5 works fine at the same settings, yet the CPU temps remain about the same as 500x9, so I'm hoping that this is a sign that the CPU hasn't degraded too badly, but then again, not being able to reach the same speeds at the same settings is still a growing concern.
The mosfet sinks are too hot to touch during testing (the ones nearest the i/o backplate in particular), so I've gotta rip out the board and re-seat the sinks, or try to figure out a better solution. I'm really looking out for some mosfet waterblocks now, but I've only seen one rather unsatisfactory solution, but with a 12" house fan on it you'd think it would be enough.
I've still got some family errands to run over the next few days, but rebiulding and testing will resume and I'll hopefully resolve the issue. If not, I'll have to settle for some lower speeds.

[Hoss331]

Are you sure its not just cpu temps causing the errors, mid 70s is pretty warm to be priming at 4.5.

[Gameman]

I just recently installed a UD3R (WAITING ON A QUAD CHIP) with my old E8600 (500x7=3.5 ghz) and noticed the Northbridge "Gigabyte" Heatsink is hot to the touch, even though there appears to be no way to monitor this like on the X48...

do you think that's normal when you got the fsb at 500 and the MCH at 1.36 in order to do so, or should the MCH be much lower????

Cause all i know is 500 took at least 1.36 to 1.4 on the T2RsB plus...

I put my spare Dominator fan blowing on it (5000rpm)..... will this be sufficient??? or do Gigabyte Northbridges normally run hot cause of the high fsb attainable on them....
that's all thanks

edit: ps i don't remember the X48 T2RSB Plus Northbridge heatsink getting this hot , just warm

Merry Xmas and New Year

[andressergio]

Are you sure its not just cpu temps causing the errors, mid 70s is pretty warm to be priming at 4.5.

+1 Hoss

From my experience on the UD3P and so others IF CPU temp heats 75 constant (cores are around 70+) for a long time will lead to errors...if you can lower temps it'll pass

Cheers
Sergio

[truehighroller]

Mine was unstable until I stuck la little fan on the Mosfets right above the CPU. I had mentioned this a couple of times but I think no one believed me :P.

[godsa]

Mine was unstable until I stuck la little fan on the Mosfets right above the CPU. I had mentioned this a couple of times but I think no one believed me :P.

could always do a little reseating

btw sig shows what this did out of the box with only a few tweaks.

I'm pretty sure I can get to 4800 on water ..the problem is won't be as easy as this lol..

So I doubt ill go for it..not worth the headach3

[truehighroller]

I have mine at 4.2 now. It's fine once you stick a little fan on the ones right above the CPU. If you don't have one on yours and you do benching I'm willing to bet yours get hot as well or if not to bad right now, then when you start reaching 4.8GHz I bet they do. I posted my settings a page back. Here are my settings again since this is a new page and probably every one will miss them.

******************************
MB Intelligent Tweaker(M.I.T.)
******************************
Robust Graphics Booster ..............: [Turbo]
CPU Clock Ratio ......................: [8X]
x Fine CPU Clock Ratio................: +0.5

>>>>> Standard Clock Control
CPU Host Clock Control................: [Enabled]
x CPU Host Frequency(Mhz).............: [495]
PCI Express Frequency(Mhz)............: [100]
C.I.A.2...............................: [Disabled]

>>>>> Advanced Clock Control..........:
CPU Clock Drive.......................: [1000mV]
PCI Express Clock Drive...............: [800mV]
CPU Clock Skew........................: [ 0ps]
MCH Clock Skew........................: [ 50ps]

******** DRAM Performance Control ********

Performance Enhance...................: [Turbo]
MCH Frequency................: 333
System Memory Multiplier (SPD)........: [2.40B]
Memory Frequency(Mhz).................: 1188
DRAM Timing Selectable (SPD)..........: [Manual]

>>>>> Standard Timing Control
x CAS Latency Time................5...: [ 5]
x tRCD ...........................5...: [ 5]
x tRP.............................5...: [ 5]
x tRAS...........................18...: [18]

>>>>> Advanced Timing Control.........: [Press Enter]
(All Auto here)

>>>>> Channel A Timing Settings.......: [Press Enter]
x Static tRead Value............. (8)

(All Auto in Between)

x DDR Write Training..................: Disabled

>>>>> Channel A Driving Settings......: [Press Enter]
x Driving Strength Profile............: [1200MHz]

(All Auto in Between)

>>>>> Channel B Timing Settings.......: [Press Enter]
x Static tRead Value.............    (8)

(All Auto in Between)

x DDR Write Training..................: Disabled

>>>>> Channel B Driving Settings......: [Press Enter]
x Driving Strength Profile............: [1200MHz]

(All Auto in Between)

****** Motherboard Voltage Control ******


-----------------------------------------------
>>> CPU
Load-Line Calibration.................: [Enabled]
CPU Vcore.....................: [1.37500V] Actual Idle [1.344] Actual Load [1.328]
CPU Termination........1.200V.........: [1.420V]
CPU PLL................1.500V.........: [1.570V]
CPU Reference..........0.760V.........: [.939]
>>> MCH/ICH
MCH Core...............1.100V.........: [1.52V]
MCH Reference..........0.760V.........: [.899]
MCH/DRAM Reference.....0.900V.........: [.970]
ICH I/O................1.500V.........: [1.500V]
ICH Core...............1.100V.........: [1.200V]
>>> DRAM
DRAM Voltage...........1.800V.........: [2.00V] 
DRAM Termination.......0.900V.........: [.850]
Channel A Reference....0.900V.........: [Auto]
Channel B Reference....0.900V.........: [Auto]

(All CPU Features Turned Off)

[In-Fluence]

Are you sure its not just cpu temps causing the errors, mid 70s is pretty warm to be priming at 4.5.

It passed fine on the 8.5 multi, and the temps are pretty much the same (in fact they went higher than 9x test because it lasted longer). All settings were identical on both tests. I did notice that one of my mosfet sinks was a little wobbly, so I'm hoping another stab at remounting them shortly might help, but from what I can gather it's one of two things - mosfets overheating, or CPU degradation. One thing I do know is that the mosfets do get too hot for comfort during prime.

[maxwedge]

could always do a little reseating

btw sig shows what this did out of the box with only a few tweaks.

I'm pretty sure I can get to 4800 on water ..the problem is won't be as easy as this lol..

So I doubt ill go for it..not worth the headach3

Your sig result is with an air hs?

[In-Fluence]

I should probably correct my last statement for the record... 8.5 multi did run at lower temps but for some reason it ran hotter on one particular run

I'm gonna get the board out and go over it with a fine-toothed comb. I'm trying to think of some better solution to mosfet cooling, but without a decent number of fan headers (and no room for a drive-bay fan controller), I could get stuck unless I go for some thermalright sinks, or if I go for some waterblocks I just found. Bit risky if the board is flaky though...

[truehighroller]

I should probably correct my last statement for the record... 8.5 multi did run at lower temps but for some reason it ran hotter on one particular run

I'm gonna get the board out and go over it with a fine-toothed comb. I'm trying to think of some better solution to mosfet cooling, but without a decent number of fan headers (and no room for a drive-bay fan controller), I could get stuck unless I go for some thermalright sinks, or if I go for some waterblocks I just found. Bit risky if the board is flaky though...

Probably just because your reference settings are off a little CPU or MCH. If your settings are off some times it will run hotter then it really should and it is kind of a rare thing to witness, for me it has been far and between.. Usually if my CPU Ref is off by being one step to high is when it happens actually I think.

The board isn't flaky I think the Mosfets that they used were not really the best choice chip wise but your board is probably fine.

[Hoss331]

It passed fine on the 8.5 multi, and the temps are pretty much the same (in fact they went higher than 9x test because it lasted longer). All settings were identical on both tests. I did notice that one of my mosfet sinks was a little wobbly, so I'm hoping another stab at remounting them shortly might help, but from what I can gather it's one of two things - mosfets overheating, or CPU degradation. One thing I do know is that the mosfets do get too hot for comfort during prime.

The multi is not the issue, its the cpu speed, the higher you go the more sensitive it becomes to core temps. The 8.5 only gives you 4.25ghz so its obviously going to be far easier to pass than 4.5 and be less sensitive to the heat.

[maxwedge]

I just noticed today when using a cpu fan on the fan header that there is a delay before the cpu fan starts. Its like a second before the fan starts then it runs slowly and when the board beeps the fan comes on full speed.
I have cpu fan function disabled.

Is this normal?? seems odd

[dustin_ds3000]

I just noticed today when using a cpu fan on the fan header that there is a delay before the cpu fan starts. Its like a second before the fan starts then it runs slowly and when the board beeps the fan comes on full speed.
I have cpu fan function disabled.

Is this normal?? seems odd

my ep45-ud3r does the same thing

[Don_Dan]

I just noticed today when using a cpu fan on the fan header that there is a delay before the cpu fan starts. Its like a second before the fan starts then it runs slowly and when the board beeps the fan comes on full speed.
I have cpu fan function disabled.

Is this normal?? seems odd

Yeah, my UD3P does the same.

[winston856]

I have mine at 4.2 now. It's fine once you stick a little fan on the ones right above the CPU. If you don't have one on yours and you do benching I'm willing to bet yours get hot as well or if not to bad right now, then when you start reaching 4.8GHz I bet they do. I posted my settings a page back. Here are my settings again since this is a new page and probably every one will miss them.

******************************
MB Intelligent Tweaker(M.I.T.)
******************************
Robust Graphics Booster ..............: [Turbo]
CPU Clock Ratio ......................: [8X]
x Fine CPU Clock Ratio................: +0.5

>>>>> Standard Clock Control
CPU Host Clock Control................: [Enabled]
x CPU Host Frequency(Mhz).............: [495]
PCI Express Frequency(Mhz)............: [100]
C.I.A.2...............................: [Disabled]

>>>>> Advanced Clock Control..........:
CPU Clock Drive.......................: [1000mV]
PCI Express Clock Drive...............: [800mV]
CPU Clock Skew........................: [ 0ps]
MCH Clock Skew........................: [ 50ps]

******** DRAM Performance Control ********

Performance Enhance...................: [Turbo]
MCH Frequency................: 333
System Memory Multiplier (SPD)........: [2.40B]
Memory Frequency(Mhz).................: 1188
DRAM Timing Selectable (SPD)..........: [Manual]

>>>>> Standard Timing Control
x CAS Latency Time................5...: [ 5]
x tRCD ...........................5...: [ 5]
x tRP.............................5...: [ 5]
x tRAS...........................18...: [18]

>>>>> Advanced Timing Control.........: [Press Enter]
(All Auto here)

>>>>> Channel A Timing Settings.......: [Press Enter]
x Static tRead Value............. (8)

(All Auto in Between)

x DDR Write Training..................: Disabled

>>>>> Channel A Driving Settings......: [Press Enter]
x Driving Strength Profile............: [1200MHz]

(All Auto in Between)

>>>>> Channel B Timing Settings.......: [Press Enter]
x Static tRead Value.............    (8)

(All Auto in Between)

x DDR Write Training..................: Disabled

>>>>> Channel B Driving Settings......: [Press Enter]
x Driving Strength Profile............: [1200MHz]

(All Auto in Between)

****** Motherboard Voltage Control ******


-----------------------------------------------
>>> CPU
Load-Line Calibration.................: [Enabled]
CPU Vcore.....................: [1.37500V] Actual Idle [1.344] Actual Load [1.328]
CPU Termination........1.200V.........: [1.420V]
CPU PLL................1.500V.........: [1.570V]
CPU Reference..........0.760V.........: [.939]
>>> MCH/ICH
MCH Core...............1.100V.........: [1.52V]
MCH Reference..........0.760V.........: [.899]
MCH/DRAM Reference.....0.900V.........: [.970]
ICH I/O................1.500V.........: [1.500V]
ICH Core...............1.100V.........: [1.200V]
>>> DRAM
DRAM Voltage...........1.800V.........: [2.00V] 
DRAM Termination.......0.900V.........: [.850]
Channel A Reference....0.900V.........: [Auto]
Channel B Reference....0.900V.........: [Auto]

(All CPU Features Turned Off)

Did you test he DRAM Term. and the MCH/DRAM Term. up and down like you did the CPU/MCH Ref's?

[truehighroller]

Did you test he DRAM Term. and the MCH/DRAM Term. up and down like you did the CPU/MCH Ref's?

Yes I did The memory will start throwing errors when it is off so it is a little more time consuming getting those settings just right. More like find the out side points where it starts erroring out and then hit the middle point and see how it acts.

[maxwedge]

I just noticed today when using a cpu fan on the fan header that there is a delay before the cpu fan starts. Its like a second before the fan starts then it runs slowly and when the board beeps the fan comes on full speed.
I have cpu fan function disabled.

Is this normal?? seems odd

my ep45-ud3r does the same thing

Yeah, my UD3P does the same.

Really appreciate that guys, I was having thoughts of swapping out the pws to see if it would go away!
I only have a little Corsair VX450 in this on for now.

[truehighroller]

Really appreciate that guys, I was having thoughts of swapping out the pws to see if it would go away!
I only have a little Corsair VX450 in this on for now.

It's just the order the motherboard test everything it really isn't that big of a deal I think as far as the fan thing goes.

[maxwedge]

It's just the order the motherboard test everything it really isn't that big of a deal I think as far as the fan thing goes.

Thanks, I was thinking it could be normal but I hadn't noticed something like this with the 10-15 boards I've used in the past. Thanks

[RAMMAN]

my board doesnt have the same dividers as the ud3p on various straps.

266 strap = 4:5, 2:3, 1:2
200 strap = 3:4, 3:5, 1:2
333 strap = 1:1, 5:6, 5:8, 1:2
400 strap = 1:1, 3:4, 3:5

which is the best for high everest bandwidth numbers and is 1333 memory 500 fsb on 3:4 divider 400 strap with a quad a pipe dream without active cooling on the nb?

my best everest bandwidth results so far right here:

http://www.xtremesystems.org/forums/...=239757&page=3

edit: im also considering trying 266 strap 2:3 divider, is that divider+strap likely to be bugged or ok with my board?

[Gameman]

just got a cheap quad today for a cool 200.00 a lot cheaper than it used to be

I gotta hand it to you guys, taking these quad chips to high fsb's a breeze on these ud3 boards..

I wanna thank you all for continuing to recomend this board, it is as good as you say it is.....



just wondering now whether to exchange this q9400 for a q9550 and pay the difference or keep it... but either way 4.0ghz=4 cores should be a piece of cake

update - this was done with ease
i love this board...sure beats the t2rsb plus for fsb




HAPPY NEW YEAR !!!!

[truehighroller]

HAPPY NEW YEAR !!!!

Happy New Year!