768k Skylake Problem/Bug - Page 54

s1riker · 2016-04-12, 15:35

If anyone still cares, my idle crashes were ultimately solved by the latest BIOS update for my gigabyte board. The changelog claims the following: "Improve DDR XMP compatibility". It also comes with MC version 74, so who knows what actually fixed it, but it's fixed.

All I know is that I was right about this platform not at all being ready on launch day. It only took them 9 months to get it stable. Should have trusted my instincts and waited a year before diving in.

pegnose · 2016-04-12, 15:54

Quote:

Originally Posted by s1riker

If anyone still cares, my idle crashes were ultimately solved by the latest BIOS update for my gigabyte board. The changelog claims the following: "Improve DDR XMP compatibility". It also comes with MC version 74, so who knows what actually fixed it, but it's fixed.

All I know is that I was right about this platform not at all being ready on launch day. It only took them 9 months to get it stable. Should have trusted my instincts and waited a year before diving in.

Good to hear that you finally made it. Thumbs up!

Batalov · 2016-07-30, 16:30

The "other", non-768k-FFT-size Skylake stability issue discussion is forked to a thread of its own

BinaryKhaos · 2016-08-13, 07:04

Hi @all,

just as a quick note for those that have not followed by own thread on the forums, the issue is still not fixed with MC 0x74. Upping to MC 0x9e fixed it for me. And both FMA3 and AVX are affected.

Here a link to my post and thread.

If that saves someone else the headache I had, it was at least worth it.

Have a nice weekend,
Matthias

owh · 2016-12-22, 02:29

First time poster, First time Mersenne, Long time IT geek.

I think my iMac is affected by this, but I'm unsure and would like someone to hit me with a clue-bat to correct me please.

I have read the entire thread - and I think I even understood most of it ;-)

I have an Apple iMac 5K which appears to have an i7 6700K CPU. I have added 64 Gb of RAM and VMware crashes under load. This has been happening from day 1, in August when I purchased the machine.

Initially RAM was suspected and after weeks of testing, Apple advised me that the problem was nothing to do with them since I was using non-Apple RAM. VMware advise me that the issue is to do with the bug that is outlined in this thread.

Today I have downloaded the OSX command-line version of prime95 (http://www.mersenne.org/ftp_root/gim...X.noGUI.tar.gz) and started a torture test. (Based on the PC World Article, which in turn pointed me here...)

The initial test failed almost immediately with "Hardware failure detected, consult stress.txt file.", but that was before I read this thread.

I have run the test again and it fails in the same way.

7 of the 8 workers stop with a Hardware failure. I have not waited until (and if) the last one fails.
Both times I've run the test, the same worker was left standing (#6)
I have added CpuSupportsFMA3=0 to local.txt
I did not set FFT2=768K
There is no stress.txt file
According to VMware the microcode in my CPU is ucode level: 0x00000074

As an aside:

Hitting Ctrl-D in the menu of the tool causes mayhem, in terms of a running update of the menu being redrawn as fast as the machine can.
I am unsure how to properly exit the tool.

My questions:

Is this the same problem as described in this thread?
If this is the same problem, how have others been successful in telling their OEM that this is a problem?
Is anyone aware of a firmware update that covers this already?

I look forward to hearing from you.

Kind regards,
Onno

henryzz · 2016-12-22, 22:47

The first thing I would do is make sure I have the latest bios. My pc was unstable until I updated the bios even though the microcode version didn't change when I upgraded the bios.
You will need good cooling with a 6700k. Mine is barely cooled enough and I have a fairly good cooler.

Prime95 · 2016-12-23, 01:21

Quote:

Originally Posted by owh

I have an Apple iMac 5K which appears to have an i7 6700K CPU. I have added 64 Gb of RAM and VMware crashes under load.

Initially RAM was suspected and after weeks of testing, Apple advised me that the problem was nothing to do with them since I was using non-Apple RAM. VMware advise me that the issue is to do with the bug that is outlined in this thread.

Welcome to the forum!

I'd suspect RAM rather than BIOS. Try the following:

1) Run a small FFT torture test. If that works, then RAM is strongly implicated.
2) Replace the RAM with the original RAM. Try the torture test again. If that succeeds, then your new RAM is strongly implicated.

To exit mprime, first start by running "mprime -m". This displays a menu and choice 5 is exit.

owh · 2017-05-19, 10:03

Quote:

Originally Posted by Prime95

Welcome to the forum!

I'd suspect RAM rather than BIOS. Try the following:

1) Run a small FFT torture test. If that works, then RAM is strongly implicated.
2) Replace the RAM with the original RAM. Try the torture test again. If that succeeds, then your new RAM is strongly implicated.

To exit mprime, first start by running "mprime -m". This displays a menu and choice 5 is exit.

This lead me down a long and windy road which ultimately proved to be correct, so thank you very much for your comments.

I returned the 4 x "16GB 1867MHZ DDR3 SO-DIMM PC3-14900" from OWC and installed 4 x "Crucial 16GB PC3L-14900 Unbuffered 204-pin SO-DIMM", which thus far have proven to be stable - touch wood.

Thank you for your help and your time, hope this helps another user.

o

2016-12-22, 02:29	#588
owh Dec 2016 2 Posts	First time poster, First time Mersenne, Long time IT geek. I think my iMac is affected by this, but I'm unsure and would like someone to hit me with a clue-bat to correct me please. I have read the entire thread - and I think I even understood most of it ;-) I have an Apple iMac 5K which appears to have an i7 6700K CPU. I have added 64 Gb of RAM and VMware crashes under load. This has been happening from day 1, in August when I purchased the machine. Initially RAM was suspected and after weeks of testing, Apple advised me that the problem was nothing to do with them since I was using non-Apple RAM. VMware advise me that the issue is to do with the bug that is outlined in this thread. Today I have downloaded the OSX command-line version of prime95 (http://www.mersenne.org/ftp_root/gim...X.noGUI.tar.gz) and started a torture test. (Based on the PC World Article, which in turn pointed me here...) The initial test failed almost immediately with "Hardware failure detected, consult stress.txt file.", but that was before I read this thread. I have run the test again and it fails in the same way. 7 of the 8 workers stop with a Hardware failure. I have not waited until (and if) the last one fails. Both times I've run the test, the same worker was left standing (#6) I have added CpuSupportsFMA3=0 to local.txt I did not set FFT2=768K There is no stress.txt file According to VMware the microcode in my CPU is ucode level: 0x00000074 As an aside: Hitting Ctrl-D in the menu of the tool causes mayhem, in terms of a running update of the menu being redrawn as fast as the machine can. I am unsure how to properly exit the tool. My questions: Is this the same problem as described in this thread? If this is the same problem, how have others been successful in telling their OEM that this is a problem? Is anyone aware of a firmware update that covers this already? I look forward to hearing from you. Kind regards, Onno

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Skylake vs Kabylake	ET_	Hardware	17	2017-05-24 16:19
Skylake and RAM scaling	mackerel	Hardware	34	2016-03-03 19:14
So does skylake-nonXeon actually get us anything?	fivemack	Hardware	36	2015-09-08 01:42
Skylake processor	tha	Hardware	7	2015-03-05 23:49
Skylake AVX-512	clarke	Software	15	2015-03-04 21:48

2016-04-12, 15:35	#584
s1riker Jan 2016 37₈ Posts	If anyone still cares, my idle crashes were ultimately solved by the latest BIOS update for my gigabyte board. The changelog claims the following: "Improve DDR XMP compatibility". It also comes with MC version 74, so who knows what actually fixed it, but it's fixed. All I know is that I was right about this platform not at all being ready on launch day. It only took them 9 months to get it stable. Should have trusted my instincts and waited a year before diving in.

2016-07-30, 16:30	#586
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 2·47·101 Posts	The "other", non-768k-FFT-size Skylake stability issue discussion is forked to a thread of its own

2016-08-13, 07:04	#587
BinaryKhaos Aug 2016 23₁₀ Posts	Hi @all, just as a quick note for those that have not followed by own thread on the forums, the issue is still not fixed with MC 0x74. Upping to MC 0x9e fixed it for me. And both FMA3 and AVX are affected. Here a link to my post and thread. If that saves someone else the headache I had, it was at least worth it. Have a nice weekend, Matthias

2016-12-22, 22:47	#589
henryzz Just call me Henry "David" Sep 2007 Cambridge (GMT/BST) 2×3³×109 Posts	The first thing I would do is make sure I have the latest bios. My pc was unstable until I updated the bios even though the microcode version didn't change when I upgraded the bios. You will need good cooling with a 6700k. Mine is barely cooled enough and I have a fairly good cooler.