DFI UT X58-T3EH8 and OC Troubles

[devsk]

So, I am having troubles with my DFI X58 board. I am not sure if its dud as yet.

Here are my settings:

3x2GB OCZ gold 8 8 8 24, rated 1600Mhz at 1.65 running at 8 9 9 24 at 1.64V, speed only 1200Mhz. Tried 8 8 8 24 also. 8 9 9 19 is best worked on Asus.

cpu feature: VR enabled, Thermal disabled, CxE C1E, Execute disable Bit enabled, VT enabled, HT enabled, All cores
ram: rated timings. 2:6, so its running slower than rated.
voltages: vcore 1.3 (tried upto 1.35), vid special add enabled, special add Auto, vdroop disabled, dram 1.65, vtt 1.36 (tried upto 1.45), VTT special add Auto, CPU pll 1.8 (tried upto 1.9), IOH/ICH 1.11 (tried upto 1.31), IOH analog voltage 1.1 (tried upto 1.15). Rest default.
PPM enabled.
Turbo enabled.
BCLK 200
PCIE 100
DRAM Freq BCLK*6
UnCore Freq BCLK*12.

Rest default.

I put it under stress using LinX, and the PC shuts off (power off followed by auto restart).

I reduce the BCLK to 190 and take 2 sticks out. Run the stress program and PC shuts off and auto-restarts.

Core temps never go above 85C. With 190 BCLK, they were below 80C.

I have run this CPU on Asus P6T board with 21x206 12hr prime and 3hr Linx stable with full 6GB RAM. vcore used was 1.35, 1.4vtt for 4.3Ghz. I expected it do something similar in this range of vcore and vtt. I had tested RAM on that board with memtest also.

Anybody have any ideas about what could be the problem?

[loonym]

In HW monitor section of bios disable thermal shutdown.

[devsk]

In HW monitor section of bios disable thermal shutdown.

Its set to shutdown at 100C. Why should it engage at 85C of core temp? CPU temp seen by mobo is actually stupidly low 57C.

Tried it anyway but it didn't help.

[loonym]

Its set to shutdown at 100C. Why should it engage at 85C of core temp? CPU temp seen by mobo is actually stupidly low 57C.Did you try disabled?

[devsk]

Did you try disabled?

yeah, it was 100c before. I tried disabled.

[devsk]

My PWM is hot. Its reaching 95C under load. I have no idea why it gets so hot. I have very powerful panaflo at 2700rpm on cpu sink, and a scythe sflex at exhaust at 1500rpm pulling hot air over it.

It seems like the flame freezer is loose. It has an inch of movement behind the case. The ridge on the mounting clip did encapsulate it and I put OCZ freeze on it before putting it in there. Is it supposed to be pushed more? I think its already pressed against the fins. I have no idea why its so loose (the part outside of case moves freely. I haven't attached the covering clip).

[loonym]

You still get the shutdowns even with thermal shutdown disabled?

As far as the little 'flame freezer' gimmick sink, I never used it so I can't speculate.

[devsk]

It seems like something is forcing shutdown under load. The most likely candidate is PWM because that's the only thing so hot. My CPU core temps have been in the 80s in the past. I have run them in 90s also few times.

[devsk]

You still get the shutdowns even with thermal shutdown disabled?

As far a the little 'flame freezer' gimmick sink, I never used it so I can't speculate.

Yeah, all thermal shutdown options are disabled.

The flame freezer seemed to help 3-4C in reviews. So, if that mean 90C instead of 94C, I would like it to work.

[loonym]

I was getting similar shutdowns when trying to max my clocks for benching but disabling thermal shutdown stopped those. If your pwm temps are that high you might want to remove the sink and check the contact pattern.

[devsk]

I was getting similar shutdowns when trying to max my clocks for benching but disabling thermal shutdown stopped those. If your pwm temps are that high you might want to remove the sink and check the contact pattern.which sink are you talking about? The sink on PWM itself? I would hate to break something in there.

Or you meant the flame freezer? How the hell is it supposed to be installed? Why is it loose even after fitting perfectly in the ridge of the mounting clip? Worst design ever and the pics in doc don't help at all. You can hardly see anything in there.

[loonym]

I have no idea about the flame freezer. I immediately decided it looked like cheap junk and I never gave it another glance, it's still in the box in my closet. If my pwm temps were as high as you state, I'd be taking a look under the sink to see why.

[eva2000]

try suggestions at http://i4memory.com/f80/t3eh8-very-high-pwm-temps-18015/ for pwm temps - it can be cpu dependent too i.e. high leakage cpus will have higher pwm temps too. My w3520 pwm temps > 4x i7 920 d0/c0 pwm temps once oc'd with cpu vtt/vcore applied

but still never had DFI UT X58 shutdown on me for any reason... unstable clocks for me usually either freeze, bsod or reboot system.

Some folks don't have much luck on 05/08 or 4/28 beta, so try 2/27 or 2/17 bioses as well.

[devsk]

Do you know what is "CPU" and "IOH" QPI drive strength? What about PWM freqs? Should I try lower or higher?

Also, the fans can't be made quiet. I have set CPU fan to turn fully on at 60C but its fully on when the temp is like 51C.

I have a feeling that overall DFI board is weaker than Asus board in the robustness department. BIOS seems flaky. It threw one random error on me: "non-matching MP processors. The CPU models should match..." Reset and works fine after that.

Many times I clear CMOS, I end up with some weird error or another: most of the times its "checksum error" but many times random stuff like "AHCI BIOS not installed".

[devsk]

Could it be my PSU? Does DFI board require more juice that Asus/Gigabyte boards? I have Antec TP Quattro 850W.

[eva2000]

QPI drive strength may help stablize some cpus, cstkl1 reported his i7 920 d0 liked strong for stability in long runs for linx/prime95 but my 2x i7 920 d0 and 2x i7 920 c0 like normal instead of strong. Haven't tested my W3520 D0 to see which it likes better yet. So as with all settings on any board you use, trial and error and find out what your setup likes ;)

05/08 is a hit or miss with folks more so than 4/28 beta - flash commands can be run manually without autoexec.bat or you edit it yourself - commands here http://i4memory.com/f80/4-28-beta-bios-results-dfi-ut-x58-t3eh8-17224/#post112474 and my 4/28 results http://i4memory.com/f80/4-28-beta-bios-results-dfi-ut-x58-t3eh8-17224/

How did you flash to 05/08 bios using winflash or awdflash with the command lines listed in above link (contained in autoexect.bat) ?

As per 2nd link in list below (http://i4memory.com/f55/intel-core-i7-920-overclocking-introduction-guide-x58-13051/), finding out your settings for memory first before cpu will make things move much more smoothly. You can't translate voltages directly from one board brand to another as they all work differently to some extent.


DFI UT X58-T3EH8 notes, findings & tips (http://i4memory.com/102537-post3.html)
Intel Core i7 920 Overclocking Introduction Guide on X58 (http://i4memory.com/f55/intel-core-i7-920-overclocking-introduction-guide-x58-13051/)
Prime95/LinX/Super Pi 32M stability tweaks (http://i4memory.com/f80/i7-920-d0-dfi-ut-x58-t3eh8-prime95-linx-stability-16617/)
DFI UT X58 - tRRD, RTL, tFAW timing discussion. (http://i4memory.com/f80/dfi-ut-x58-trrd-rtl-tfaw-timing-discussion-16970/)
DFI Lanparty UT X58-T3EH8 LGA1366 Photos, Bios and Overclocking Information Guide (http://i4memory.com/f80/dfi-lanparty-ut-x58-t3eh8-lga1366-photos-bios-overclocking-information-guide-12855/)
DFI LP UT X58-T3EH8 - Bios Template (http://i4memory.com/f80/dfi-lp-ut-x58-t3eh8-bios-template-12919/)
DFI UT X58-T3EH8 - Bios Download / Change Log (http://i4memory.com/f80/dfi-ut-x58-t3eh8-bios-download-change-log-13966/)
DFI LP UT X58-T3EH8 - Bios Screenshots (http://i4memory.com/f80/dfi-lp-ut-x58-t3eh8-bios-screenshots-12949/)
Memtest86+ (v2.11) (http://i4memory.com/f79/memtest86-v2-11-a-6/)
i7 920 3836A756 C0 @4610Mhz AIR - DFI UT X58-T3EH8 (http://i4memory.com/f80/i7-920-3836a756-4610mhz-air-dfi-ut-x58-t3eh8-13007/)
i7 920 C0 @4769Mhz TRUE120 AIR ! (http://i4memory.com/f80/i7-920-4769mhz-true120-air-13746/)
First i7 920 D0 Step = mixed bag of results (http://i4memory.com/f80/first-i7-920-d0-step-mixed-bag-results-17703/)
Second i7 920 D0 Stepping 3851A368 #2 results (http://i4memory.com/f80/second-i7-920-d0-stepping-3851a368-2-results-17969/)
Windows 7 + Turbo mode fixed 12x multi issue [DFI UT X58] - research (http://i4memory.com/f80/windows-7-turbo-mode-fixed-12x-multi-issue-dfi-ut-x58-research-18053/)


I'd try a different set of memory.. could be during clear cmos board default vdimm is 1.56v, if you memory doesn't like such low volts it could chuck a checksum error.

PWM frequencies, 1st link listed above to http://i4memory.com/102537-post3.html tip #11 for memory, different memory ics may like different values so again, you need to find out what your memory likes. Finding the right one can make your memory and thus cpu fly high :)

Also tip #5, 6, and 7 at http://i4memory.com/102537-post3.html make sure you aren't loading old saved profiles on newly updated bioses especially for 01/14, 04/28 and 05/28 bioses which has drastic bios changes.

Also some folks report using mem and uncore multi 8x and 16x respectively are more stable than 6x/12x for mem/uncore.

As to linx/prime95 and hyperpi 8x32m stability look at your ioh/ich and ioh analog voltage relationship, some board/cpu combinations like some additional voltage there. 3rd linked on list http://i4memory.com/f80/i7-920-d0-dfi-ut-x58-t3eh8-prime95-linx-stability-16617/.

As to psu, you could use multimeter to measure your real voltages for cpu vtt/vcore/vdimm etc as per guide at bottom of post here http://i4memory.com/102537-post3.html

http://nx.i4memory.com/reviewimages/motherboards/dfi/UT_X58_T3EH8/dmm/dfi_utx58_dmm.jpg

Different psu, may give you different real voltages from what SG/board/windows app report.

Also share you settings you have tried via this template you can fill in your settings http://i4memory.com/f80/dfi-lp-ut-x58-t3eh8-bios-template-12919/