Z77X-D3H, 3570k and 1600Mhz Memory Woes

chrisgt

chrisgt

chrisgt

Associate
Joined
12 Jul 2010
Posts
743
Hi all,

Original thread here: http://forums.overclockers.co.uk/showthread.php?t=18486081

Basically, I am experiencing random WHEA_UNCORRECTABLE_ERROR BSODs (hal.dll 0x124) on my newly built machine. Specs are:

i5 3570k
Gigabyte Z77X-D3H
Gigabyte 7970
XFX 850w PSU
Kingston 16GB (now 8GB) 1600Mhz RAM
Agility 3 120GB SSD

I put the machine together, installed Windows 8 Pro x64 and things seemed fine. However I noticed in CPU-Z that my DDR Frequency was only 1333Mhz and not 1600Mhz as it should be, so I went into BIOS and enabled XMP Profile1 which corrected this (set the correct timings as per Kingston documentation: 9-9-9-27 1.65v). Note: at this point, the i5 was stock @ 3.4Ghz.

This unfortunately caused BSODs 10sec into a Prime95 Blend Test.

So far I have:
  • Re-installed Windows a number of times
  • Ran MemTest86+ for 12 hours, without a single error.
  • Updated BIOS to latest stable release (F16)
  • Re-built computer; reseated the CPU & CPU Cooler, RAM sticks, all power connections

In order to get Prime95 Blend Test stable I had to load BIOS optimized defaults which reverted the RAM back to 1333Mhz @ 1.5v. This was Blend test stable for 4 hours, so I thought the issue was to due with the Z77 chipset & Ivy bridge not liking >1.5v DRAM voltage, which is what my memory needs to run at to achieve 1600Mhz...or so I thought! The PC seemed fine all day...but then I got another WHEA_UNCORRECTABLE_ERROR BSOD while browsing the net!! This was at the BIOS optimized defaults!

Any clues as to whats wrong? I want to overclock this bad boy, but the fact I can't even get it stable @ stock clocks is worrying me...

Set-up 1: BIOS Optimized Defaults
i5 @ 3.4Ghz - Completely Stock (stock voltages (everything at auto in BIOS))
RAM @ 667Mhz 9-9-9-24 1.5v

Set-up 1 seemed good, was Prime95 blend test stable for 4 hours, but still BSOD'd at night!

Set-up 2: Slight Multiplier OC
i5 @ 4.0Ghz (40x100BLCK with stock (auto) voltages)
RAM @ 667Mhz 9-9-9-24 1.5v

Set-up 2 BSODs 5sec into Blend test.

Please, any help would be appreciated :( I'm about ready to give up...I just need to get it 100% error free @ stock speeds, then I'll focus on OC'ing
 
Last edited:
Current config:
552040321774453841432892.png

43243697780068068479365.jpg

BSOD Details:
400246572633906364672341.jpg

MemTest86+ run:
837222462842606889820644.png



EDIT:
Got 40minutes into the OCCT CPU:OCCT test then got a BSOD WHEA_UNCORRECTABLE_ERROR. This was using BIOS optimized defaults, using test set-up of:
  • CPU:OCCT
  • Test Type: Infinite
  • Duration: 1hour
  • Test Version: 64bit
  • Test Mode: Large Data Set

No Cores exceeded 67C during the test. So it seems I don't even have a system that is stable @ factory settings? I did notice that the CPU Clock never exceeded 3.6Ghz, I thought Turbo was supposed to take that up to 3.8Ghz? I'll try get 30minutes of OCCT:CPU out of it without BSOD'ing, just to see what the Vcore graphs etc look like, will they be useful?

EDIT 2:
Passed 30 minutes of OCCT without error. What graphs, if any, would be useful?
 
Last edited:
MORE SPEED said:
Does it BSOD when you aren't stress testing, such as running a game?
Click to expand...

Yes. This config (everything at default) has BSOD whilst playing Path of Exile, general web browsing, watching a 1080p video in VLC Player.

Oddly enough, 4hours of Prime95 Blend Test went without issue, the 3D Mark 11 bench was OK, but 40min of OCCT crashed it.

Is this fluctuation OK for the Vcore at stock frequencies? This is all 'Auto' in BIOS:

868687235273365353628437.png
 
Last edited:
Yes, WHEA_UNCORRECTABLE_ERROR in hal.dll with a stop code of 0x00000124.

Which unfortunately just means a hardware issue, doesn't give any specifics as to what component. So I have stress tested GPU multiple times (Heaven, 3D Mark 11) and that's been fine. I then ran Memtest86+ on the memory for 12hours and that was fine. P95 FPU test and Blend test appears stable, but the system has still crashed doing mundane things like web browsing/gaming, and has also crashed 40minutes into the OCCT CPU test.
 
So as mentioned above, I've ordered some Crucial Ballistix 1866Mhz RAM @ 9-9-9-27 1.5v which the Crucial website says is 100% compatible with the Z77X-D3H motherboard.

Fingers crossed this sorts the problem, and I can just throw that 16GB Kingston back into my old system, in which it used to run fine!

Be interesting to know if any other forum members have had issue with older ram (needing to run at 1.65v) and the newer Z77 boards, especially the Gigabyte Z77X-D3H!
 
Last edited:
djkav said:
Nope, I said it was an i5 2500k which is a sandy.

I'm just using the XMP profile with 1.65v. What BIOS are you running?
Click to expand...

Apologies, if I had taken 2 secs to read your post and/or your sig I would've known that! :p

F16. Not tried the beta...

I have newer RAM coming tomorrow, which on paper is 100% compatible with this motherboard. If you have a flick through your Gigabyte manual I'm pretty sure it says it only supports DIMMs @ 1.5v. I don't know if that is a motherboard thing, or specific to Ivybridge IMC.
 
I've opened a thread asking 8 Pack to impart some of his wisdom...but in case anyone else is interested...

Just installed my brand new Crucial 1866Mhz RAM, enabled XMP Profile1 to get it to run at advertised speeds. Started a P95 Blend test, 10 seconds in, blue screen WHEA_UNCORRECTABLE_ERROR.

So this is happening with 3 lots of RAM, 2 of which were MemTest86+ tested for 12hours without issue, and 1 was brand new purchased 2 days ago.

This is so frustrating :( Pics as requested:

CPU Voltage
30621189592886062900605.jpg

CPU Frequency
217570683810747160621356.jpg

DRAM Voltage
477104774340370536647780.jpg

Memory settings 1:
76528921442400078385011.jpg

Memory settings 2:
57435318262973983069992.jpg

3D Power settings:
221247177229129113124697.jpg

Again, BIOS is 'optimized defaults' with the only thing touched being the DRAM Profile to Profile1.
 
Cheers for that. How comes in HWMonitor it doesn't show DRAM and CPU VTT under Voltages? That's what I'd be most interested in seeing bud - maybe your version of HWMonitor is out of date?
 
You must log in or register to reply here.
Back
Top Bottom