Re: [opensuse] Unstable system - culprit identified
The saga is about to conclude. The Gigabyte motherboard GA-MA790FX-DS5 with an AMD Phenom quad-core is unstable under full load, period. I suspect the memory controller setup, but Gigabyte has been unable to reproduce (under Windows). Last week I purchased an MSI board (K9A2 CF) with slightly lower specs (primarily AMD790X instead of -FX chipset) - yesterday I installed it with the same components (cpu, psu, memory, disks etc.). Earlier this afternoon it completed a 24 hour burn-in run with mprime stresstests going at full throttle - showing no problems whatsoever. Right now it's running four copies of mprime plus a Kubuntu installation in Virtualbox, and a Firefox3 beta. (I'm browsing the net whilst waiting for Kubuntu to make some progress). IMO the Gigabyte board is faulty - I'll ask the dealer to take it back tomorrow. /Per Jessen, Zürich -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Wednesday 2008-03-26 at 19:28 +0100, Per Jessen wrote:
The saga is about to conclude.
The Gigabyte motherboard GA-MA790FX-DS5 with an AMD Phenom quad-core is unstable under full load, period. I suspect the memory controller setup, but Gigabyte has been unable to reproduce (under Windows).
Well... at least you can stop worrying.
Last week I purchased an MSI board (K9A2 CF) with slightly lower specs (primarily AMD790X instead of -FX chipset) - yesterday I installed it with the same components (cpu, psu, memory, disks etc.). Earlier this afternoon it completed a 24 hour burn-in run with mprime stresstests going at full throttle - showing no problems whatsoever.
:-)
Right now it's running four copies of mprime plus a Kubuntu installation in Virtualbox, and a Firefox3 beta. (I'm browsing the net whilst waiting for Kubuntu to make some progress).
I'd be surprised if with that load it does "some" progress :-p - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFH6pxYtTMYHG2NR9URAgaBAJwNB+Rt5TdqSwVgw5Wv6d4rTkvx1gCfdWq4 Z8AfO3MFgKUcBoITrZe+aNE= =Hmbx -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Carlos E. R. wrote:
Right now it's running four copies of mprime plus a Kubuntu installation in Virtualbox, and a Firefox3 beta. (I'm browsing the net whilst waiting for Kubuntu to make some progress).
I'd be surprised if with that load it does "some" progress :-p
You're right, it didn't - for some reason it got hung up at 21%, but on the retry it did get a lot further. /Per Jessen, Zürich -- http://www.spamchek.com/ - your spam is our business. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Per Jessen schreef:
The saga is about to conclude.
The Gigabyte motherboard GA-MA790FX-DS5 with an AMD Phenom quad-core is unstable under full load, period. I suspect the memory controller setup, but Gigabyte has been unable to reproduce (under Windows).
Last week I purchased an MSI board (K9A2 CF) with slightly lower specs (primarily AMD790X instead of -FX chipset) - yesterday I installed it with the same components (cpu, psu, memory, disks etc.). Earlier this afternoon it completed a 24 hour burn-in run with mprime stresstests going at full throttle - showing no problems whatsoever. Right now it's running four copies of mprime plus a Kubuntu installation in Virtualbox, and a Firefox3 beta. (I'm browsing the net whilst waiting for Kubuntu to make some progress).
IMO the Gigabyte board is faulty - I'll ask the dealer to take it back tomorrow.
/Per Jessen, Zürich
Well, not to come behind the meal, but i exchanged a gigabyte mobo, few years ago, because it crashed random, with an AMD 64, single core. Memory wasn't the cause. I exchanged against asus K8, same problem, with windows that was. SuSE 64 bit, no problem, but slow, to low on memory, only 512. Added 1GB. Speed was incredible after that. Never completely trusted the board, and i finaly got rid of it, in favor of a AMD 64 Dual Core 3600+ on an elite mobo. ATX, pci e. Like a sun from the start, no problem whatsoever. What i want to say is, that you have done the best you could have done. You would always be paranoid about that pc. As i remember crashing, all work lost, or few hours gaming for nothing. This is no fun at all. From the start you feel better now, same as i did. Your pc is your friend, it has to be reliable, you depend on that. I am glad you finaly cut the nut. -- Enjoy your time around, Oddball (Now or never...) Besturingssysteem: Linux 2.6.25-rc5-git2-5-default x86_64 Current user: oddball@AMD64x2-sfn1 System: openSUSE 11.0 (x86_64) Alpha3 KDE: 4.00.66 (KDE 4.0.66 >= 20080313) "release 6.1" -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
(I apologize Per, you will get two copies, I forgot to change the To: line when I first hit send) Per Jessen wrote:
The saga is about to conclude.
The Gigabyte motherboard GA-MA790FX-DS5 with an AMD Phenom quad-core is unstable under full load, period. I suspect the memory controller setup, but Gigabyte has been unable to reproduce (under Windows).
Last week I purchased an MSI board (K9A2 CF) with slightly lower specs (primarily AMD790X instead of -FX chipset) - yesterday I installed it with the same components (cpu, psu, memory, disks etc.). Earlier this afternoon it completed a 24 hour burn-in run with mprime stresstests going at full throttle - showing no problems whatsoever. Right now it's running four copies of mprime plus a Kubuntu installation in Virtualbox, and a Firefox3 beta. (I'm browsing the net whilst waiting for Kubuntu to make some progress).
IMO the Gigabyte board is faulty - I'll ask the dealer to take it back tomorrow.
/Per Jessen, Zürich
Per, I'm beginning to suspect that it's not just the Gigabyte board that is having the problems with x86_64 and AMD multicore processors. I don't know if you have seen my post asking for your help yet, but I am in the exact same boat with a Tyan Tomcat K8E S2865ANRF Socket939 with Opteron 185 (Ver 2.00.00), (2 Gig OCX Platinum -- certified by OCX in Denmark and returned 2 weeks ago) and a pair of Seagate 500G drives in raid 1. It's not memory, there are no errors and the box will run just fine at idle but crash when you put it under load. (not very functional regardless of the specs) At least in your case with the Gigabyte board and in my case with the Tyan board, there is something in the 10.3 x86_64 setup that is crashing under load and leaving nary a trace as to why? The Tyan board even supports Linux. I don't know what is new in the dual core chipsets on the boards, but somewhere there _is_ something rotten in Zürich (or is that Denmark?) What is helpful and unique about the Tyan board is that is comes with an onboard (literally on the board - you have to take the side cover off) 2 digit lcd display that shows the current bios state and bios codes for the current state of the machine. I have watched and logged all bios codes and watched it run through its internal self tests and everything is shown as 100% OK. No problems from POST to fully loaded in KDE. I have also turned all unneeded peripherals off in the bios to eliminate any irq conflicts and booted with and without noapic and nolapic - makes no difference. I suspect it may be worth a bug report to investigate given that we now have a Gigabyte and Tyan boards experiencing the same thing. I searched bugzilla.novell.com and didn't find any other entries. Did you file a bug report? If not, I'll try and get one in next week. -- David C. Rankin, J.D., P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
David,
I'm beginning to suspect that it's not just the Gigabyte board that is having the problems with x86_64 and AMD multicore processors. I don't know if you have seen my post asking for your help yet, but I am in the exact same boat with a Tyan Tomcat K8E S2865ANRF Socket939 with Opteron 185 (Ver 2.00.00), (2 Gig OCX Platinum -- certified by OCX in Denmark and returned 2 weeks ago) and a pair of Seagate 500G drives in raid 1. It's not memory, there are no errors and the box will run just fine at idle but crash when you put it under load. (not very functional regardless of the specs)
At least in your case with the Gigabyte board and in my case with the Tyan board, there is something in the 10.3 x86_64 setup that is crashing under load and leaving nary a trace as to why? The Tyan board even supports Linux. I don't know what is new in the dual core chipsets on the boards, but somewhere there _is_ something rotten in Zürich (or is that Denmark?)
What is helpful and unique about the Tyan board is that is comes with an onboard (literally on the board - you have to take the side cover off) 2 digit lcd display that shows the current bios state and bios codes for the current state of the machine. I have watched and logged all bios codes and watched it run through its internal self tests and everything is shown as 100% OK. No problems from POST to fully loaded in KDE. I have also turned all unneeded peripherals off in the bios to eliminate any irq conflicts and booted with and without noapic and nolapic - makes no difference.
I suspect it may be worth a bug report to investigate given that we now have a Gigabyte and Tyan boards experiencing the same thing. I searched bugzilla.novell.com and didn't find any other entries. Did you file a bug report? If not, I'll try and get one in next week.
Not sure this is helpful, but I have been running opensuse 10_3 x86_64 without problems on an Epox 9NPA-Ultra motherboard, 4GB of OCZ ram and an AMD Opteron 144. this is a Socket 939 motherboard. I have had zero problems with this setup. this particular setup is the backup server for our small office. The main server runs opensuse 10.2 x86_64 on a Tyan Tomcat n3400 (S2925) This is an AM@ socket, and has 4 GB of A-data Ram DDR2 800 and a dual core Opteron 1218. The only reason reason this has 10.2 instead of 10.3 is that vmware workstation 5.5 did not run on opensuse 10.3. My server runs linux as the file server and then has two virtual machines. One virtual is Windows SBS 2003 (sorry, but our donor management software only runs on windows) and the other virtual machine is SME Server, our linux email and web server. When I get a chance I will upgrade the main server to 10.3, as I have now upgraded vmware workstation to 6.0. In any event I have 10.3 x86_64 running fine on a Socket 939 board, for what that is worth. Presumably you ran Memtest for a few hours to verify your memory. I know your post indicated that OCZ verified the memory, and OCZ is good stuff, but I would certainly run Memtest anyway, if yo haven't already. Mike I -- Michael A. Coan Woodlawn Foundation 524 North Avenue, Suite 203 New Rochelle, NY 10801-3410 Tel 914-632-3778 Fax 914-632-5502 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (5)
-
Carlos E. R. -
David C. Rankin -
Mike Coan -
Oddball -
Per Jessen