Who wants a challenge? 2 Z77 pcs, 5 sets of ram, Sandy, Ivy, & failing LinX at stock.

[ZeroOne]

Two separate computers are failing LinX at stock speeds / manually set DDR3 1600 (4 physical sets of DDR3 1600), and I cannot figure out what I'm doing wrong. They WILL pass with a 5th set from another system that is downclocked to 1600 from its native 1866.


First, here is the hardware:

PC 1 (fails LinX):
P8Z77-I Deluxe
3570K (stock clocks)
2x8GB DDR3 1600, 2 sets
CMP16GX3M2A1600C10 Corsair (2 sets, first RMA'ed with errors)
http://www.newegg.com/Product/Produc...82E16820233288
....then....
F3-1600C9D-16GXM GSkill
http://www.newegg.com/Product/Produc...82E16820231568



Computer 2 (runs perfectly):
Asus Maximus V Extreme
3770K (4.8Ghz @ 75C, de-lidded)
2x8GB DDR3 1866 - CMD16GX3M2A1866C9 Corsair
http://www.newegg.com/Product/Produc...82E16820233287



Computer 3 (also fails LinX):
Sabertooth Z77
2700K (stock clocks)
2x8GB DDR3 1600 - CMP16GX3M2A1600C10




The issue is... PC1 and 3 that both have 1600 memory and both fail LinX at stock clocks / 1600 memory speed. On PC1 (haven't been testing PC3 yet), backing the memory down from 1600 to 1333 on PC1 passes 20 runs. Pushing up to 1600 fails after 4-6 runs. This is the same for 2 different sets of Dominator ram and one set of GSkill Ripjaws on this computer. Taking the 1866 memory from PC2 and using that at 1600 in PC1 WILL PASS 20 runs of LinX. The failures have been the case with 2 identical sets of Corsair 2x8GB 1600 Dominator memory, and one set of GSkill 1600. I've been leaving the memory at auto timings, and checking CPU-Z to make sure the timings are automatically set according to the ram. It has been right on, with the only manual setting being the 1600 memory speed in the bios, and verifying 1.5v according to the specs.

PC3, with its OWN (3rd) set of Corsair Dominator 2x8GB 1600, fails LinX around 4-6 runs, all stock clocks, manually set to 1600, timings auto, verified by CPU-Z. Is there something with this 1600 2x8GB memory that I'm missing???

Here's the kicker...... Putting the GSkill (so far. will test Dominator 1600 next.) in the Maximus V Extreme, it PASSES 20 runs at its same 1600 speed (auto timings / manual speed), that it fails with in PC1 P8Z77-I. Going to test the 2 sets of Dominators next, but I suspect they may pass, same as this GSKill.

It's ironic the one computer that is heavily overclocked passes with no issue! WTH am I doing wrong with these other two at STOCK!? Please help! Any comments are appreciated!

[zanzabar]

do you need to up the vccio/pch voltage by any chance on the ones that are failing.

[sin0822]

PSu maybe? OS maybe?

[ZeroOne]

Thanks for the replies.... Win8 Preview on one (PC1), and Win7 Pro on the other (PC3). They are both failing the exact same way, around 4th to 6th LinX run.

Tried 2x4GB 1600 that has the exact same timings as the 8GB sticks last night in PC1, with success. Passed with no isssues. It looks like those 8GB 1600 modules are causing some kind of instability.... thats 4 sets total for PC1, failing, with 2x8GB 1600, but passing at 2x4GB 1600. Time to increase some voltages I think.... There aren't as many options in the Z77 Sabertooth and P8Z77-I as the MVE. Looking for VCCIO, but not an option in BIOS.

[Ozzfest05]

I would update to latest bios and are you using xmp profiles .
I have had issues before and had to use profiles to hit correct speed with stability.

[ZeroOne]

I would update to latest bios and are you using xmp profiles .
I have had issues before and had to use profiles to hit correct speed with stability.

Was using the manual method of entering the 1600, and watching the timings to make sure they were picked up properly. Trying now with the XMP for 1600. Thanks for the suggestion!

[ZeroOne]

Failed with XMP on PC1 at 1600 (GSkill set linked above). Upped VCCSA and VDIMM to 1.15, and 1.60v. Testing again...

[Ozzfest05]

http://support.asus.com/download.asp...Ymwl5uuG2ml3jJ

If you look here there is a bios update as of october 18th
Fixes show improved stability and improve memory compatibility.

I suggest updating bios to latest regardless.

[ZeroOne]

The bios on the P8Z77-I got the update from a few days ago... (PC1) and am testing on that one now... no luck so far. Even with the vdimm @ 1.6 and vccsa @ 1.15. This time, the Asus Anti-Surge was tripped when I came back to the computer this morning, expecting a complete or failed test. Rebooting and testing, it failed in 19 minutes, on run 2. Not doing so well. Might just back down to 1333 even though these are 1600 sticks and call it a day. This is getting ridiculous for something that should be running at advertised speeds.

[ZeroOne]

Had a 2nd test running, with Vdimm @ 1.62 and vccsa @ 1.18 (if i'm remembering correctly). Failed in ONE run, 9min. Going backwards now...

[Leeghoofd]

I don't think this is in the voltages at all, try secondary and tertiary RAM timings... The sabretooth z77, 2600K and 8GB dimms is a problem child, pure bios related. Can you swap any of the Ivies and plug it in the Sabretooth ?

Test the dimms with HCI memtest and if that passes let them be, test IMC with Hyper PI 32M x nr of threads for your CPU, if that passes too you are more than stable enough for daily crunching...

[ZeroOne]

I don't think this is in the voltages at all, try secondary and tertiary RAM timings... The sabretooth z77, 2600K and 8GB dimms is a problem child, pure bios related. Can you swap any of the Ivies and plug it in the Sabretooth ?

Test the dimms with HCI memtest and if that passes let them be, test IMC with Hyper PI 32M x nr of threads for your CPU, if that passes too you are more than stable enough for daily crunching...

Thanks for the advice... Swapping chips would be a pain at the moment...

Currently just testing the p8z77-I and all of the ram. The runs above were with the gskill 8GB. Will look into hci memtest... Tho knowing it will fail linx feels like a bit of a disappointment. It definitely seems like the dimm size is the whole issue. 4GB dimms pass! Same timings (primary, auto).

I'm not familiar with secondary and tertiary timings, and usually buy the memory speed I intend to run, so I'm not 100% comfortable with what to change...

[kiwi]

Are you running Linx 11.0.1.005?

[ZeroOne]

v0.6.0.4 in one part, then below, "Uses Intel Math Kernel Library 10.2" but I downloaded from the latest link in the forum post.... which is supposedly 11.0.1.005

[Raja@ASUS]

Send me UEFI screenshots of the DRAM timing pages with the 8GB modules in place (use a FAT formatted USB drive and then convert the files to JPEG). I'll tell you what to change.

-Raja

[ZeroOne]

Raja, thanks for your reply. I had the 1866 Corsair memory in, and had been testing with that. It was passing. The 1600 Gskill passed in another system. When I swapped back, the strangest thing happened... It started with me forgetting to change the bios on the P8Z77-I back to 1600 before removing the 1866 memory I'd been testing with. In my haste, I accidentally overclocked the Gskill 1600 to 1866. It did not boot, and went into this boot cycle, of power on, 5 full seconds, power off, 2 seconds, and repeat. It kept going with this loop for a while. I left it, thinking, is this the "mem-OK" feature? After checking the manual and reading that it should be around 30 seconds per interval, I cut the power, and reset the bios (like i've done many times before.) This time, with the Gskill 1600 @ 1600, it DID PASS LinX. 2 full tests at 25 runs each. It appears stable. I have no idea what is going on, but it seems to be working...?

[ZeroOne]

This is looking like a motherboard / bios issue.... both the Sabertooth Z77 and P8Z77-I, after the most recent bios updates, seem to be all of a sudden working. The P8Z77-I update was this week. Seems that after the flash, and a bios clear post-flash, 1600 just works. Thanks for all of the suggestions. Probably should have started testing on the Sabertooth, since its bios was out before the -I.

[Ozzfest05]

http://support.asus.com/download.asp...Ymwl5uuG2ml3jJ

If you look here there is a bios update as of october 18th
Fixes show improved stability and improve memory compatibility.

I suggest updating bios to latest regardless.

It smelt like bios, and that doesnt smell good.
Anyway a good idea when updating bios is to copy down your overclock settings you wish to keep
Clear bios. Proceed to flash then after flash clear again.

[Leeghoofd]

Using overclock profiles of previous bios versions usually will not work correctly , not load or mess things up... bad advice mate, if you said take screenshots ( F12 ) on a stick and set the settings manually later on I would agree...

[ZeroOne]

Well, it's back to its old tricks again. Was OC'ing the 3570K in the P8Z77-I. Went back into bios to adjust voltage and noticed randomly, the multiplier was back to "auto." CPU voltage and DDR3-1600 speeds were still manually set. It also loaded back to the easy-bios instead of the advanced one. Something is definitely up with this board I think. Reset CMOS and back to stock speeds / volts, and failing again. Was passing before this weird boot. Just reset everything and set the turbo manuallly in bios x38 and its loading windows and cpu-z reports the 41 multiplier. This is strange.

[ZeroOne]

Maybe it's the base clock? The last time things got stable, was after the off / on / off / on cycle. Tried the auto OC feature and it was doing the same cycle. It ended up at 103 for the base clock. Maybe 100.0 is not stable on this board and needs to be moved from that. at 100.5 now and it is passing LinX again at the 1600 memory speed.

[ZeroOne]

This is getting pretty annoying. Working soley on the P8Z77-I and 3570K.

Completely back to failing stock speeds at 1600 (stock memory speed). Back to 1333 being the only thing passing. Memory passes in another system at 1600. Saved screenshots of the timings. Need to compare, but they look the same. Thinking this is either CPU or motherboard. Need parts to begin the swapping.