DFI LP UT X58-T3EH8 - First bench, bios screens etc

[faster3200]

I could reallyuse some help here...

I can't install quad SLi on this motherboard. I've installed OS with one GFX card and then added the other, but the nvidia driver didn't see it. I saw it caused a problem on one of the PCI-E ports, not enough ressources (kode 12)

Then I C-mos and tryed to install OS with quad installed, but then OS could'nt find my raid 0 - untill i removed one of the GFX.

Someone have an ideer on what to do?

Is there a PCI-E bios setting that controls the ports, maybe..?

Thanks!

My guess would be this:

Memory LowGap: is for mapping local memory (add-on-cards) to system memory address size. It most like AGP port Aperture size or PCIE1gen 250M fixed mapping size. Extend it can solve when if "add-on cards" request a bigger size as like we would like to enable Qual SLI on NV new 295, it must be extend over 2048M. Can confirm that with 2x HD4870x2 in Crossfire I had set my LowGap to around 2176 at least. With Highpoint 3520LF PCI-E x8 256MB cache controller added into HD4870x2 Crossfire, I needed at least LowGap of 2560 otherwise windows boots to a blank window and doesn't go any further. As to how this effects windows available memory, with single HD4870x2 installed and 3x1GB memory, if I set LowGap to 2560, WinXP Pro SP3 32bit sees 1.5GB available memory. If I set LowGap to 1024, WinXP Pro SP3 32bit sees 3.0GB available memory IIRC.

[PRJ]

My guess would be this:

Sounds interresting, I'll try that whn i return from work, is this a bios setting..?

Thanks!

did you try without the sound card?

Yes I did try without soundcard, did'nt help.

[faster3200]

Sounds interresting, I'll try that whn i return from work, is this a bios setting..?

I have a T3EH6 on an older bios but it is an option for me so I assume you have it as well unless they got rid of that option. IIRC it is under mem settings.

[PRJ]

I have a T3EH6 on an older bios but it is an option for me so I assume you have it as well unless they got rid of that option. IIRC it is under mem settings.

Back home, the bios has the setting and i changed it to 2048 and started a fresh install, good thing is that it can see the raid 0 again, so somthing has changed for the better.

Thanks a lot!

**UPDATE** Crash "IRQ not less or equal"....?

[Id034]

Id034, I owe you a coctail of your choosing. Used your settings from page 12, after primary vista install, first overclock of an i7, and is working like a champ!

My memory is on the 6 multi as It did not like the 8, probably voltage set too low. No worries, will play with that when I have a chance. 32m pi stable at this point so happy.

This is a build for a friends kid, a rocking gaming rig. Thanks again to eva2000 for all the info on i4memory as well.

It is up crunching numbers while I get some rest, see how it likes running 24/7 @ 100% cpu load on all cores with HT for 8, will check on it after some sleep.

Thanks again guys.

Congrats man, and looking at your temps seems your wc is doing a fabulous work
Sadly my waterchill xternal-apo gtz combo is nowhere near your setup; linx rises my temps to high seventies

before os reinstall definitely try adaptec in pci-e #4

My old good raptor 74gb died a few days ago...but i'll buy something soon to replace it.

Yes I did try without soundcard, did'nt help.

Before or After clean install? All my problems related to pci-e dissapeared when i removed my adaptec card before clean install..but i'm still looking for more evidences

[linflas]

Congrats man, and looking at your temps seems your wc is doing a fabulous work

Not what is in my sig, this is a build for a friends kid. Have a MCR220, and a single Black ice120, on some weak ass fans that came witht he coolermaster Cosmos case, He wanted it quiet so did not switch out to decent CFM fans. An apogee GTZ also, and 2 mcw60's on a 4870x2 in the loop.

Ran WCG for a bit, was not stable for that, so had to make some adjustment, now at 200x19 for 3.8 @ 1.29V, memory @ 1600 Cas8 (gskil 1200 6gb cas 7 kit).

@ 4.2 WCG was heating to 67 or so, @ 3.8 holding @ 58 on all 4 cores @ 100% with HT on. So pretty happy with that. As this is not my rig, and I will not be around to tweak it if things go wrong, settlled on the lower clock to garuntee stability

Thanks again for the help, really gave me a good place to start.

[eva2000]

Not what is in my sig, this is a build for a friends kid. Have a MCR220, and a single Black ice120, on some weak ass fans that came witht he coolermaster Cosmos case, He wanted it quiet so did not switch out to decent CFM fans. An apogee GTZ also, and 2 mcw60's on a 4870x2 in the loop.

Ran WCG for a bit, was not stable for that, so had to make some adjustment, now at 200x19 for 3.8 @ 1.29V, memory @ 1600 Cas8 (gskil 1200 6gb cas 7 kit).

@ 4.2 WCG was heating to 67 or so, @ 3.8 holding @ 58 on all 4 cores @ 100% with HT on. So pretty happy with that. As this is not my rig, and I will not be around to tweak it if things go wrong, settlled on the lower clock to garuntee stability

Thanks again for the help, really gave me a good place to start.

oh so rig was for someone else ? Tempted ya self yet ?

[EJTREM3]

hi everyone, until today i testing my x58ut.
i put the dtek fuzion with my own plate.
i pushed to 3.8ghz and i ghetting estabilize,
my batch ists a 3639

[eva2000]

hi everyone, until today i testing my x58ut.
i put the dtek fuzion with my own plate.
i pushed to 3.8ghz and i ghetting estabilize,
my batch ists a 3639
http://imgplace.com/image_bin/3499/2...5ab64350a8.jpg

Nice but image can't be viewed ?


i7 920 3836A756 #2
Max BClock
Testing 2nd twin batch cpu and definitely this second i7 920 3836A756 is better than my first one in terms of max Bclock and low CPU VTT usage.

i7 920 3836A756 #1

• max memtest86+ bootable bclk = 224 requiring 1.58-1.60v CPU VTT
• max super pi 32m bclk = 222bclk (without IOH/ICH voltage tweaks) & 228bclk (with IOH/ICH voltage tweaks) requiring 1.60-1.61v CPU VTT
• max CPUZ Validation = 230bclk requiring 1.60-1.63v CPU VTT

i7 920 3836A756 #2

• max memtest86+ bootable bclk = 229bclk requiring 1.40-1.48v CPU VTT
• max super pi 32m bclk = 227bclk (without IOH/ICH voltage tweaks) & ???bclk (with IOH/ICH voltage tweaks TBD) requiring 1.48-1.50v CPU VTT
• max CPUZ Validation = 231bclk requiring 1.50v CPU VTT

TBD - Haven't tested 2nd cpu with IOH/ICH tweaks yet

Max Validation @231Bclock

CPU VTT = 1.50v





Max Screenshot @228Bclk with low CPU VTT



Max 32M @225Bclk



Max 32M @227Bclk

[PRJ]

Back home, the bios has the setting and i changed it to 2048 and started a fresh install, good thing is that it can see the raid 0 again, so somthing has changed for the better.

Thanks a lot!

**UPDATE** Crash "IRQ not less or equal"....?

The crash i had was with the Raid driver from DFI, when I used Windows own driver it kept going and installed successfully... With Quad SLi!! So it did work!

[eva2000]

The crash i had was with the Raid driver from DFI, when I used Windows own driver it kept going and installed successfully... With Quad SLi!! So it did work!

raid driver from cd-rom or web site ? tried latest from intel.com ?

[PRJ]

Latest driver from DFI Website ( Vista X64), but raids working fine now on windows own raid driver. Performance of raid seems fine, so i'll avoid a reinstall.

[PRJ]

I seem to be stuck at 19*200 no matter what vcore i use, but if i want to raise the VVT to help out, I'll have to raise the Vdimm aswell (as they need to be ~0,5V apart) My Corsair memmory are rated at 1,65 vdimm and if I'm not to exceed the limit, I'm stuck with a VVT voltage at <~1,21 V

Is it realy important for a 24/7 setup to have a vdimm at max the rated 1,65 V?

Thanks!

[Sam_oslo]

I seem to be stuck at 19*200 no matter what vcore i use, but if i want to raise the VVT to help out, I'll have to raise the Vdimm aswell (as they need to be ~0,5V apart) My Corsair memmory are rated at 1,65 vdimm and if I'm not to exceed the limit, I'm stuck with a VVT voltage at <~1,21 V

Is it realy important for a 24/7 setup to have a vdimm at max the rated 1,65 V?

Thanks!

Nobody knows the longtime effect av Vdimm> 1.65v for 24/7. It is safer to keep it within the Intel's recommended limit for 24/7. But VVT needs to be 0 - 0,5V apart from Vdimm, so you are not stuck at VVT <~1,21 V.

[PRJ]

Nobody knows the longtime effect av Vdimm> 1.65v for 24/7. It is safer to keep it within the Intel's recommended limit for 24/7. But VVT needs to be 0 - 0,5V apart from Vdimm, so you are not stuck at VVT <~1,21 V.

Thanks for your input, I'll try to push the VVT some more, without raising Vdimm.

[diablos]

tuned, probably my next mb

[rusty]

Nobody knows the longtime effect av Vdimm> 1.65v for 24/7. It is safer to keep it within the Intel's recommended limit for 24/7. But VVT needs to be 0 - 0,5V apart from Vdimm, so you are not stuck at VVT <~1,21 V.

But that would be high VTT volts 1.6v wouldn't it ? Or am I missing something ?
If vdim is 1.65 then vtt can't be 1.6v ?

[eva2000]

But that would be high VTT volts 1.6v wouldn't it ? Or am I missing something ?
If vdim is 1.65 then vtt can't be 1.6v ?

VDIMM needs to be within 0.5v of CPU VTT. But the ideal difference between the two depends on number of modules installed as well as the memory themselves and at what memory frequency and timings used.

Samsung HCF0 2x1GB liked 0.24-0.31v difference
Samsung HCF0 3x1GB liked 0.31-0.40v difference
6x1GB mixed HCF0 and Micron D9GTR liked 0.35-0.45v difference
3x2GB Corsair Dominator GT Elpida 50nm based likes 0.20-0.28v difference - hence the general trend to need more cpu vtt than samsung hcf0 but handle CAS7 timings higher than others

[rusty]

VDIMM needs to be within 0.5v of CPU VTT. But the ideal difference between the two depends on number of modules installed as well as the memory themselves and at what memory frequency and timings used.

Samsung HCF0 2x1GB liked 0.24-0.31v difference
Samsung HCF0 3x1GB liked 0.31-0.40v difference
6x1GB mixed HCF0 and Micron D9GTR liked 0.35-0.45v difference
3x2GB Corsair Dominator GT Elpida 50nm based likes 0.20-0.28v difference - hence the general trend to need more cpu vtt than samsung hcf0 but handle CAS7 timings higher than others

I'm a bit lost can you give some examples please?

[Praz]

Samsung HCF0 2x1GB liked 0.24-0.31v difference
Samsung HCF0 3x1GB liked 0.31-0.40v difference
6x1GB mixed HCF0 and Micron D9GTR liked 0.35-0.45v difference
3x2GB Corsair Dominator GT Elpida 50nm based likes 0.20-0.28v difference - hence the general trend to need more cpu vtt than samsung hcf0 but handle CAS7 timings higher than others

I'm a bit lost can you give some examples please?

Those seem to be pretty good examples. The voltages are a balancing act. The difference between VTT and DRAM voltage affects stability. And to a lesser extent the amount of VTT influences the needed VCORE for CPU stability. Once past general guidelines it's trial and error to what each system is happy with.

[EJTREM3]

hehehe, sorry, here the image.


[img=http://img144.imageshack.us/img144/4149/b48b439ae2741dbaf199c29jl0.jpg]
[img=http://img144.imageshack.us/img144/b48b439ae2741dbaf199c29jl0.jpg/1/w1024.png]

i started to try push to 4.0ghz but i have problem to stabilize, i put 1.38 cpu vid and 1.48 cpu vtt, anyone knows how put more volt on nb?
this voltage its a pretty high to 4.0 or its ok, my temps its 33 in cores idle and 54c load. with 1.35vid and 1.42vtt in 3.8ghz

[eva2000]

I'm a bit lost can you give some examples please?

examples in sig already - finding that balance between the 2 voltages will allow your memory to clock higher/more stable. How you find that balance is through trial and error - i start with memtest86+ v2.11, then into windows for stability tests. I like to start with super pi (yeah some folks will say BUT...) but with i7 mem controller on cpu, super pi isn't a bad place to start at. I've experienced an hour blend prime95 stable only to not be able to even pass 2nd loop of super pi 32m before with i7 920/x58. Of course do both tests etc.

Those seem to be pretty good examples. The voltages are a balancing act. The difference between VTT and DRAM voltage affects stability. And to a lesser extent the amount of VTT influences the needed VCORE for CPU stability. Once past general guidelines it's trial and error to what each system is happy with.

Yeah i found the right cpu vtt volts allow me to clock cpu higher with less vcore as well

hehehe, sorry, here the image.

i started to try push to 4.0ghz but i have problem to stabilize, i put 1.38 cpu vid and 1.48 cpu vtt, anyone knows how put more volt on nb?
this voltage its a pretty high to 4.0 or its ok, my temps its 33 in cores idle and 54c load. with 1.35vid and 1.42vtt in 3.8ghz

for cpu clocking NB doesn't play much part in i7 920 anymore. Key volts are vcore (vid), cpu vtt (uncore) and vdimm (memory volts).

might start with some reading at http://i4memory.com/102537-post3.html

[PRJ]

But that would be high VTT volts 1.6v wouldn't it ? Or am I missing something ?
If vdim is 1.65 then vtt can't be 1.6v ?

Needs a difference of maximum 0,50 between VVT and Vdimm - not 0,05

[Sam_oslo]

for cpu clocking NB doesn't play much part in i7 920 anymore. Key volts are vcore (vid), cpu vtt (uncore) and vdimm (memory volts).

In my experience, this is true up until @3.8.GHZ, HT:on, and 4GHZ HT:off, but it seams to need some more tweaking to stabilize @4GHZ, HT:on.

@4GHZ HT:off, i can prime for several houser, and pass memory tests with these values:

@4GHZ HT:off (200x20), 1.31v(load) and 1.26v VVT :



@4GHZ(19x211) HT:off, 1.28v(Load) and 1.29v VVT. This primes with 1.26v too.


But the moment it hits @4GHZ, HT:on, it fails at the moment i start prrime95 with 1.33v(load) and 1.32v VVT. I tried a few other Vcore and VVT too, but still fails/freezes at the first moment of prime95. This doesn't BSOD as usual (when it needs more Vcore), but freezes. It seams to need tweaking on other values. What is your opinion?

Yeah i found the right cpu vtt volts allow me to clock cpu higher with less vcore as well.

You have a big and important point here, and it deserves further investigation. Could this be the reason for less Vcore with my 1.29 VVT?.
@4GHZ HT:off (200x20), 1.26v VVT it needs at least 1.31v(load) to be primestable. But
@4GHZ(19x211) HT:off, 1.29v VVT it is primestable with only 1.28v and even 1.26v(load).

I never figured the reason for this difference, could that "little extra" VVT be the reason?

[linflas]

oh so rig was for someone else ? Tempted ya self yet ?

Very tempted, we shall see, especially if the ATI refresh does show up at the end of the month.