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

[KRich]

Hey Krich, I'll see if I can give you a hand. First, I don't have a Quad so your ahead of me there. I can give you a couple suggestions you can see if they help or not. Until you get some stability at the higher speed set the ram timing to 5/5/5/15. Change it back later. The vdimm seems a bit high as well. I know that is good ram but I don't know if it needs that much esp. at this speed. For 1.4vtt, your mch ref seems high also. By a fair amount. Have you tried increasing the PLL to 1.57v when you try to boot at 4.3 or above? All these things on mine need to be "together". It takes some time but when you get things closer maybe you can get a bit lower vcore. Maybe someone with a Quad can help with your mch core or something relating to what I mentioned.

KRich,
My xeon has the same 1.225 vid so give these settings a try. I ran a quick prime to check these settings and had no errors. Also what kind of water cooling do you have, those temps seems pretty high for water with that voltage. Mine with 1.4625 bios only comes up to 60c under small fft with 25c ambient.

******Motherboard Intelligent Tweaker (M.I.T.)******
 
Robust Graphics Booster.........................[ Auto     ]
CPU Clock Ratio.................................[ 9        ]
Fine CPU Clock Ratio............................[ + 0.0    ]
CPU Frequency 4.365GHz..........................[ 485 x 9.0]
 
******Clock Chip Control******
 
CPU Host Clock Control..........................[ Enabled  ]
CPU Host Frequency (Mhz)........................[ 485      ]
PCI Express Frequency (Mhz).....................[ 101      ]
C.I.A 2.........................................[ Disabled ]
 
******Advanced Clock Control******
 
CPU Clock Drive.................................[ 800mv   ]
PCI Express Clock Drive.........................[ 900mv   ]
CPU Clock Skew..................................[ 0ps     ]
MCH Clock Skew..................................[ 50ps    ]
 
******DRAM Performance Control******
 
Performance Enhance.............................[ Standard ]
(G) MCH Frequency Latch.........................[ 266      ]
System Memory Multipler.........................[ 2.50A    ]
Memory Frequency 1066...........................[ 1212     ]
DRAM Timing Selectable..........................[ Manual   ]
 
CAS Latency Time................................[ 5       ]
tRCD............................................[ 5       ]
tRP.............................................[ 4       ]
tRAS............................................[ 10      ]
 
******Advanced Timing Control******
 
tRRD[3].........................................[ auto    ]
tWTR[3].........................................[ auto    ]
tWR[6]..........................................[ auto    ]
tRFC[54]........................................[ 55      ]
tRTP[3].........................................[ auto    ]
Command Rate (cmd)[0]...........................[ auto    ]
 
******Channel A Timing Settings******     
 
Static tRead Value        [ 8 ] ................[ 8       ]
tRD Phase0 Adjustment     [ 0 ] ................[ Auto    ]
tRD Phase1 Adjustment     [ 0 ] ................[ Auto    ]
tRD Phase2 Adjustment     [ 0 ] ................[ Auto    ]
tRD Phase3 Adjustment     [ 0 ] ................[ Auto    ]
 
tRD 2rd (Differnt Rank)   [ 6 ] ................[ Auto    ]
tWR 2wr (Differnt Rank)   [ 6 ] ................[ Auto    ]
tWR 2rd (Differnt Rank)   [ 5 ] ................[ Auto    ]
tRD 2wr (Same/Difft Rank) [ 8 ] ................[ Auto    ]
 
Dimm 1 Clock Skew Control.......................[ Auto    ]
Dimm 2 Clock Skew Control.......................[ Auto    ]
DDR Write Training..............................[ Auto    ]
 
******Channel A Driving Settings****** 
 
Drive Strength Profile..........................[ 1200    ]
Data Driving Pull Up Level......................[ Auto    ]
cmd Driving Pull Up Level.......................[ Auto    ]
ctrl Driving Pull Up Level......................[ Auto    ]
clk Driving Pull Up Level.......................[ Auto    ]
 
Data Driving Pull Down Level....................[ Auto    ]
cmd Driving Pull Down Level.....................[ Auto    ]
ctrl Driving Pull Down Level....................[ Auto    ]
clk Driving Pull Down Level.....................[ Auto    ]
 
******Channel B Timing Settings******     
 
Static tRead Value        [ 8 ] ................[ 8       ]
tRD Phase0 Adjustment     [ 0 ] ................[ Auto    ]
tRD Phase1 Adjustment     [ 0 ] ................[ Auto    ]
tRD Phase2 Adjustment     [ 0 ] ................[ Auto    ]
tRD Phase3 Adjustment     [ 0 ] ................[ Auto    ]
 
tRD 2rd (Differnt Rank)   [ 6 ] ................[ Auto    ]
tWR 2wr (Differnt Rank)   [ 6 ] ................[ Auto    ]
tWR 2rd (Differnt Rank)   [ 5 ] ................[ Auto    ]
tRD 2wr (Same/Difft Rank) [ 8 ] ................[ Auto    ]
 
Dimm 1 Clock Skew Control.......................[ Auto    ]
Dimm 2 Clock Skew Control.......................[ Auto    ]
DDR Write Training..............................[ Auto    ]
 
******Channel B Driving Settings****** 
 
Drive Strength Profile..........................[ 1200    ]
Data Driving Pull Up Level......................[ Auto    ]
cmd Driving Pull Up Level.......................[ Auto    ]
ctrl Driving Pull Up Level......................[ Auto    ]
clk Driving Pull Up Level.......................[ Auto    ]
 
Data Driving Pull Down Level....................[ Auto    ]
cmd Driving Pull Down Level.....................[ Auto    ]
ctrl Driving Pull Down Level....................[ Auto    ]
clk Driving Pull Down Level.....................[ Auto    ]
 
******Motherboard Voltage Control******
 
Load-Line Calibration...........................[ Enabled ]
CPU Vcore         [ 1.15    ]...................[ 1.400   ]
CPU Termination   [ 1.20    ]...................[ 1.300   ]
CPU PLL           [ 1.50    ]...................[ 1.500   ]
CPU Reference     [ 0.760   ]...................[ 0.848   ]
 
MCH Core          [ 1.100   ]...................[ 1.380   ]
MCH Reference     [ 0.760   ]...................[ 0.863   ]
MCH/DRAM Refernce [ 0.900   ]...................[ 1.250   ]
ICH I/O           [ 1.500   ]...................[ 1.500   ]
ICH Core          [ 1.100   ]...................[ 1.100   ]
 
DRAM Voltage      [ 1.800   ]...................[ 2.500   ]
DRAM Termination  [ 0.900   ]...................[ 1.250   ]
Channel A Ref     [ 0.900   ]...................[ 1.250   ]
Channel B Ref     [ 0.900   ]...................[ 1.250   ]

******Advanced BIOS Features******
 
Limit CPUID Max to 3............................[ Disabled ]
No-Execute memory Protect.......................[ Disabled ]
CPU Enhance Halt (CIE)..........................[ Disabled ]
C2/C2E State Support............................[ Disabled ]
C4/C4E State Support............................[ Disabled ]
CPU Thermal Monitor.............................[ Disabled ]
CPU EIST Function...............................[ Disabled ]
Virtualization Technology.......................[ Disabled ]

Thanks guys, I appreciate the input and advice. Thanks so much for your time

I tried the template you provided Hoss331 and firstly, my RAM doesn't seem to like the 2.50A multiplier at all as the system BSOD's during Windows load up and also HCI Memtest in DOS hangs. I've noticed this on all my testing over the past week and it is irrelevant to settings.

I therefore used the 2.40B multiplier and also tried the 2.66D and used all the settings as you suggested.

What I found is that Core 4 on Small FFT's fails immediately or within 1 to 2 minutes. It is either with a Rounding Error or a Fatal Error about resulting sums being different.

I brought into play with coop also kindly suggested and utilised this into the template and my own template too, but unfortunately, I didn't gain any stability.

I checked the +12v line with a DMM as Everest indicates that there is over 200W being drawn by the CPU and at maximum load, the DMM reads 12.17 which means the PSU is probably not a factor. It is an OCZ GamerXtreme 700W.

The water cooling is an Apogee GTX and an MCW55 with a MCW-655 pump and MCRES-Micro. It isn't done the nicest as it was my first proper water cooling project but a day ago, I did clean the rad out and get rid of all the dust, so temps did go down to what I quoted earlier.

Anything else you can offer by the way of settings or advice would be welcomed and valued

Cheers everyone

[andressergio]

Thanks guys, I appreciate the input and advice. Thanks so much for your time

I tried the template you provided Hoss331 and firstly, my RAM doesn't seem to like the 2.50A multiplier at all as the system BSOD's during Windows load up and also HCI Memtest in DOS hangs. I've noticed this on all my testing over the past week and it is irrelevant to settings.

I therefore used the 2.40B multiplier and also tried the 2.66D and used all the settings as you suggested.

What I found is that Core 4 on Small FFT's fails immediately or within 1 to 2 minutes. It is either with a Rounding Error or a Fatal Error about resulting sums being different.

I brought into play with coop also kindly suggested and utilised this into the template and my own template too, but unfortunately, I didn't gain any stability.

I checked the +12v line with a DMM as Everest indicates that there is over 200W being drawn by the CPU and at maximum load, the DMM reads 12.17 which means the PSU is probably not a factor. It is an OCZ GamerXtreme 700W.

The water cooling is an Apogee GTX and an MCW55 with a MCW-655 pump and MCRES-Micro. It isn't done the nicest as it was my first proper water cooling project but a day ago, I did clean the rad out and get rid of all the dust, so temps did go down to what I quoted earlier.

Anything else you can offer by the way of settings or advice would be welcomed and valued

Cheers everyone

Hey Krich

2.05A its a sucker for me, my gskills PI 1200 love 2.04B, but blades go fine on the other multis...i also tried other sets and seems the same...inmo try to stick to 2.04B and tune it.

other thing...try to make first your CPU stable as its failing on SmallFFT wich test mainly CPU, then dig on the mems

Cheers and best of the lucks
Sergio

[KRich]

Hey Krich

2.05A its a sucker for me, my gskills PI 1200 love 2.04B, but blades go fine on the other multis...i also tried other sets and seems the same...inmo try to stick to 2.04B and tune it.

other thing...try to make first your CPU stable as its failing on SmallFFT wich test mainly CPU, then dig on the mems

Cheers and best of the lucks
Sergio

Cheers Sergio

You're quite right. I am only trying with SmallFFTs now because I have found that if the SmallFFTs are not stable, it is failing on Blend right away too.

I had it running 4.4GHz with 1.47 actual and it failed after 2hrs 30mins on Small FFTs. I can't work out why this CPU wants so much more VCore than it should require.

[andressergio]

Cheers Sergio

You're quite right. I am only trying with SmallFFTs now because I have found that if the SmallFFTs are not stable, it is failing on Blend right away too.

I had it running 4.4GHz with 1.47 actual and it failed after 2hrs 30mins on Small FFTs. I can't work out why this CPU wants so much more VCore than it should require.

well yes that happens not all CPU are the same but if you pump vcore and its cold...i have a Q9650 low vid 1,1875 and im in water and still limits me for temps so there's some point you have to stop.

what temps are reporting at that vcore on LOAD SmallFFT ?

Cheers mate
Sergio

[KRich]

well yes that happens not all CPU are the same but if you pump vcore and its cold...i have a Q9650 low vid 1,1875 and im in water and still limits me for temps so there's some point you have to stop.

what temps are reporting at that vcore on LOAD SmallFFT ?

Cheers mate
Sergio

Cheers buddy

You're right indeed. I am getting low to mid 70s with running SmallFFT at that VCore so perhaps that is the problem.

The thing I find odd is that I know I was running 485 x 9 on the MFII and SmallFFT was 12+ hours stable at 1.40V set in the BIOS and on that board GTL 1 and 3 set at -45mv.

I don't know why I seem to require 0.05V more for the same speed on the UD3P. I know it isn't the board because I can run pretty much good at 515FSB with the 8.5 multiplier and though it isn't 12+ hours stable yet, it is getting there with tweaking. That is something the MFII couldn't do.

I guess I was hoping that the CPU would also fly, not just the FSB

[jas420221]

New BIOS:

GA-EP45T-UD3R - F7a
GA-EP45-UD3P - F10b
GA-EP45-UD3R - F10c

No info.
Seems to have new AHCI ROM.....

Where can these be found? Not on giga site??

[Hoss331]

Where can these be found? Not on giga site??

http://forums.tweaktown.com/f69/giga...st-bios-28441/

[Hoss331]

KRich,
I think you need to work on your cooling first. With the temps getting that high its going to be harder to get a good clock. What rad are you using?

[KRich]

KRich,
I think you need to work on your cooling first. With the temps getting that high its going to be harder to get a good clock. What rad are you using?

Cheers buddy

I am using a Swiftech MCR220.

You're right about the temps as when I up the VCore one notch from the 2hr run, it fails immediately in Prime so it has to be the temps.

Even at 1.37 actual I am pushing around 70 degrees and 1.34 actual is around 64 degrees.

[Hoss331]

Oh yea, you need a bigger rad. That 220 is hard pressed to cool that cpu and graphics card effectively.

[magnifico]

Hi , second you these voltages for the day use, are good or you can cause damages?
Sorry for my english...

[Hoss331]

You can probably drop PLL down to 1.50, but other than that they look fine.

[KRich]

Oh yea, you need a bigger rad. That 220 is hard pressed to cool that cpu and graphics card effectively.

Cheers

Any idea whether one of those stackable 220s would do the same as there is no room for a 3 x 120 at all as I don't want the rad outside the system and inside, even in my CM Stacker, there isn't any space to have a 3 x 120 internally.

I guess I could add another 2 x 120 rad so I could have two in one loop - How would that be for temps?

Cheers

[magnifico]

You can probably drop PLL down to 1.50, but other than that they look fine.

If I come down to 1,5 they give me error prime 95 after an hour on Core..

[coop]

When I first put my board together my temps were about 70 even at default. The problem was the H2O block was on the edge of one of the caps keeping it raised slightly. My block is from an Opteron type socket so it could be larger and it may be obvious this could not be an issue for your setup. seating good?

[KRich]

When I first put my board together my temps were about 70 even at default. The problem was the H2O block was on the edge of one of the caps keeping it raised slightly. My block is from an Opteron type socket so it could be larger and it may be obvious this could not be an issue for your setup. seating good?

Cheers buddy.

Truth be told, this WC setup has always been underperforming - Even my E8600 used to get up past 75 degrees celsius.

The seating is good. In fact, I got new screws from Swiftech because the old ones were worn down and I thought that was causing poor contact. However, the new screws are in and the mounting is good as I can tell when the block is removed.

The rad is getting warm to hot and there are good fans on it so I think it looks like the rad can't perhaps cope

[Hoss331]

Cheers

Any idea whether one of those stackable 220s would do the same as there is no room for a 3 x 120 at all as I don't want the rad outside the system and inside, even in my CM Stacker, there isn't any space to have a 3 x 120 internally.

I guess I could add another 2 x 120 rad so I could have two in one loop - How would that be for temps?

Cheers

The stacker add-on will help but not as good as two separate rads or one large rad. If you have room for 2 separate rads id go that route, also might want to into getting some better fans.

[KRich]

The stacker add-on will help but not as good as two separate rads or one large rad. If you have room for 2 separate rads id go that route, also might want to into getting some better fans.

I think I will add another 120.2 to the loop as it will be easier to do in my CM Stacker STC-T01 than a single 120.3. I've got the current 120.2 on the floor of the case which is getting fresh air from the through the mesh at the bottom and then I can add another 120.2 at the front which can get fresh air through the mesh there.

I think the 1 x MCW-655 will be good enough for both rads as I don't think I have room for two separate loops.

Cheers buddy

[LiquidReactor]

Trying for 488x9 stable, I keep getting this "a clock interrupt was not received on a secondary processor within the allocated time interval" BSOD, but strangely IBT can pass 20 times and no errors in Prime 95 result log before BSOD.

Settings
*************************************************

Robust Graphics Booster: Auto
CPU Clock Ratio: 9x
Fine CPU Clock Ratio: 0.0x

**********Clock Chip Control*************************

CPU Host Clock Control: Enabled
CPU Host Frequency (Mhz): 488
PCI Express Frequency (Mhz): 100
C.I.A 2: Disabled

*********Advanced Clock Control**********************

CPU Clock Drive:1000ps
PCI Express Clock Drive: 1000ps
CPU Clock Skew: 0ps
MCH Clock Skew: 50ps

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

Performance Enhance: Turbo
(G) MCH Frequency Latch 333MHz
System Memory Multipler 2.40B
Memory Frequency: 1171MHz
DRAM Timing Selectable: Manual

CAS Latency Time: 5
tRCD: 6
tRP: 5
tRAS: 15

*******Advanced Timing Control***********

tRRD: Auto
tWTR: Auto
tWR: Auto
tRFC: 54
tRTP: Auto
Command Rate (cmd): Auto

DRAM drive strength on both channels is OC-1200 and PL is set as 10.

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

Load-Line Calibration: Enabled
CPU Vcore: 1.425V
CPU Termination: 1.3400V
CPU PLL: 1.5000V
CPU Reference: 0.848V

MCH Core: 1.400V
MCH Reference: 0.888
MCH/DRAM Reference: 1.14000V
ICH I/O: 1.57000V
ICH Core: 1.2000V

DRAM Voltage: 2.2800V
DRAM Termination: 1.14000V
Channel A Ref: 1.14000V
Channel B Ref: 1.14000V

Is this because vdiim is too low?

[xguntherc]

hey Rich, ya it sounds like that water loop is for sure under performing. as I have my Q9650 at 4.0 Ghz with 1.304v and it idles under 30c and loads around 38c.. and when I've moved up to like 4.25 Ghz and stuff of that sort, and used even 1.38v or sometimes more, It's still never went over 35 or so c idle and I've never seen over low 51 or 52c yet..

not only is a better rad going to help, but what type of fans are on the case, are they the stock fans, you might need some better airflow on the case, and I'd actually say loose the stackable idea, and just get a really good 120.3 as it would perform better than stackable 120.2's... I'd throw a Feser TFC 360 outside the top of the case, and I bet your temps will be much better, or a XSPC RX360, but if you have to use a 120.2. I'd say either get another and stack, or try a top of the line 120.2 like the Feser.. they are great rads.. and use a good fan, one of the best RAD fans are the S-Flex fans. also try a reseat, or have you done that plenty of times.

[jmase]

hey, ive recently got hold of a UD3R and an E8400, seems pretty stable at the moment at 3.6, but whenever i try to push it further i get bsods quite frequently. Im not sure which settings i have to change to get it stabled at 4ghz - would be ideal for me, so im looking for a template for 4ghz if anyone has one.

Here are my current settings, (im using a AC 7 pro, 2 x 2gb OCZ 1066 Platinum) @ 3.6: 400 x 9, CPU Voltage 1.3V memory multiplier = 2.00D, DRAM voltage 2.1, all the stuff like speedstep, E1st disabled etc. Everything else at auto.

Now, when i try and go for 4ghz i use 450 x 9, i up cpu voltage to 1.4V, and leave everything else at auto seeing as im not really sure what i have to change to get it stable. Anyways, with these settings i can get into windows fine, but when i load a game / program i get random BSODS

Any help or templates much appreciated, thanks!

[kagula323]

@ radaja ... what vid is your cpu? I think I have the oportunity to buy a good E8400 Q841A171 ... I found here on xs that kemo has one Q841A200 that is vid 1.2125 (so pretty close to the one I can buy as fpo ) and kemo's one does 4200 @ 1.22 . Waiting for the seller to confirm the vid of the cpu :P. Hope it is 1.2125 and that is as good or even better than yours ... I want part of the fun u are having :P

and anyone who has E8400 FPO/BATCH Q841A171 or fpo close to this please share ur experiences

[KRich]

hey Rich, ya it sounds like that water loop is for sure under performing. as I have my Q9650 at 4.0 Ghz with 1.304v and it idles under 30c and loads around 38c.. and when I've moved up to like 4.25 Ghz and stuff of that sort, and used even 1.38v or sometimes more, It's still never went over 35 or so c idle and I've never seen over low 51 or 52c yet..

not only is a better rad going to help, but what type of fans are on the case, are they the stock fans, you might need some better airflow on the case, and I'd actually say loose the stackable idea, and just get a really good 120.3 as it would perform better than stackable 120.2's... I'd throw a Feser TFC 360 outside the top of the case, and I bet your temps will be much better, or a XSPC RX360, but if you have to use a 120.2. I'd say either get another and stack, or try a top of the line 120.2 like the Feser.. they are great rads.. and use a good fan, one of the best RAD fans are the S-Flex fans. also try a reseat, or have you done that plenty of times.

Hey buddy

I've done a reseat many a time across all the boards and CPUs I have had and really, the temps are usually like this.

I had my E8600 up to the high 80s and that was stable at 4.5GHz though but I have seen a lot of mention that the Q9650s aren't so forgiving and that they need to be kept under 65 degrees celsius.

I believe that the MCR-220 is good to dissipate upto 250 to 300W and if Everest is correct, the CPU is throwing out over 220W and when you add the HD4870 which apparently throws out over 100W at stock, you can see that the problem was obvious but I didn't see it until Hoss331 pointed it out.

I was actually going to add NB/SB and perhaps also Mosfet blocks to the same loop until I read the replies I got here

The fans in use are the stock ones which came with the MCR-220 as I got it as part of a Swiftech WC kit nearly 3 years back so each fan is capable of 75CFM each.

Thanks for all the tips and input which is very useful and will help alot. I am planning an overhaul as we speak and once temps are under control, I'll be sure to get the Q9650 higher

[Injuis]

Hi all. I've quietly lurked this thread for a long time and decided to register to ask for some help. I've got a q9300 M1 stepping and have tried to break the 3600 mark with little success. I've sort of given up on overclocking for the time being to try and figure out my temperature problems. I'm running an XSPC 360 radiator in an ATCS 840 case (2x 120mm pushing and 2x 230mm pulling). Is there any reason to believe that my idle temps of 50,45,50,50 are reasonable in this setup @ stock speeds? I hit about 62-65 C at load, which I find astonishing for my WC setup. Things get only slightly worse as I overclock - hitting about 67 C at load max. From what I've seen in this thread, nobody even comes close to these temps for a quad on water. What do you guys think? Faulty CPU or mobo?

-- EDIT --
Well I was able to put in my stock air cooler and my temps are as follows

57/52/58/58 Idle
80/72/80/81 Load

At least i know my water cooling was working. SHEESH! haha, burning up here!
I'm pretty sure the sensors on this CPU are faulty. The waterblock and this stock heatsink are barely warm to the touch.

[KRich]

Hey buddy

I've done a reseat many a time across all the boards and CPUs I have had and really, the temps are usually like this.

I had my E8600 up to the high 80s and that was stable at 4.5GHz though but I have seen a lot of mention that the Q9650s aren't so forgiving and that they need to be kept under 65 degrees celsius.

I believe that the MCR-220 is good to dissipate upto 250 to 300W and if Everest is correct, the CPU is throwing out over 220W and when you add the HD4870 which apparently throws out over 100W at stock, you can see that the problem was obvious but I didn't see it until Hoss331 pointed it out.

I was actually going to add NB/SB and perhaps also Mosfet blocks to the same loop until I read the replies I got here

The fans in use are the stock ones which came with the MCR-220 as I got it as part of a Swiftech WC kit nearly 3 years back so each fan is capable of 75CFM each.

Thanks for all the tips and input which is very useful and will help alot. I am planning an overhaul as we speak and once temps are under control, I'll be sure to get the Q9650 higher

I've switched out the 2 x 75CFM fans for a 1 x 120CFM and 1 x 100CFM (didn't have matching fans ) just to see what difference it made, and it took SmallFFT load temps down about 4 to 5 degrees celsius so now, after 10 mins of running, Core 1 & 2 have hit a maximum of 69 degrees and Core 3 & 4 have hit a maximum of 72 degrees celsius which though a lot better than the over 80 - 85 degrees celsius I was getting a day or two back, isn't yet good enough.

I read that my VGA can be pulling 200W at idle in Vista with Aero enabled so if the CPU is pulling over 200W during Small FFT's, that is over 400W that this MCR-220 is trying to dissipate so clearly that is causing the high temps.

Thanks for all the input and advice everyone