Socket 939 Sempron found........

[Keysplayr]

Originally posted by: PetNorth

Originally posted by: keysplayr2003
So, more or less, THG should have just started out with a Intel 945P or 955X chipset. I mean, why wouldn't you?

They installed yesterday-night Asus mobo with Intel 955X (Asus P5WD2 Premium). It has crashed two times in aprox. eighteen hours. You see...

And you think the blame lies where? I really think it's time to scrap this project for now giving the win to AMD for stability until more mature chipsets for the Intel systems emerge. It seems there is nothing that can be done to make the Intel system stable?
Change the memory maybe? PSU? Were the OS's reinstalled properly every time? no idea. Way too many variables. If I had the money, I would conduct this test myself and interact in real time with all of you to make sure every contingency was thought of.

 

[Duvie]

Originally posted by: Sinker

Originally posted by: Duvie
Again charts clearly show 3 (yes three) reboots.....THG cant seemt to count.... those morons

The weekly chart stopped updating days ago, if that's the one you're refering to. Those three reboots were from before the test was restarted. Both the 24 hour charts are working, but that's all.

Mmm, unless you're refering to the idle time that extends off the beginning of the 24 hour chart. Surely there was a reboot at the beginning of that?
That would make 3...

Yep the last thing you said is the chart I am looking at.....3 reboots....I was watching when that first one went down at around 10:00pm pst last night and then upon restart nothing was running for the next several hours.....0% cpu usage on all 4 cores...

 

[Duvie]

Originally posted by: PetNorth

Originally posted by: keysplayr2003
So, more or less, THG should have just started out with a Intel 945P or 955X chipset. I mean, why wouldn't you?

They installed yesterday-night Asus mobo with Intel 955X (Asus P5WD2 Premium). It has crashed two times in aprox. eighteen hours. You see...

Actually 3 and then in the last hour cpu usage all over the place....prime for another crash....

 

[Duvie]

Originally posted by: keysplayr2003

Originally posted by: PetNorth

Originally posted by: keysplayr2003
So, more or less, THG should have just started out with a Intel 945P or 955X chipset. I mean, why wouldn't you?

They installed yesterday-night Asus mobo with Intel 955X (Asus P5WD2 Premium). It has crashed two times in aprox. eighteen hours. You see...

And you think the blame lies where? I really think it's time to scrap this project for now giving the win to AMD for stability until more mature chipsets for the Intel systems emerge. It seems there is nothing that can be done to make the Intel system stable?
Change the memory maybe? PSU? Were the OS's reinstalled properly every time? no idea. Way too many variables. If I had the money, I would conduct this test myself and interact in real time with all of you to make sure every contingency was thought of.

I agree with you....

 

[Sentential]

Jesus.... did anyone see the CPU load temps of that intel system? I wonder if they have the heatsink installed properly. 84C* average is just crazy

 

[PetNorth]

Originally posted by: keysplayr2003

Originally posted by: PetNorth

Originally posted by: keysplayr2003
So, more or less, THG should have just started out with a Intel 945P or 955X chipset. I mean, why wouldn't you?

They installed yesterday-night Asus mobo with Intel 955X (Asus P5WD2 Premium). It has crashed two times in aprox. eighteen hours. You see...

And you think the blame lies where? I really think it's time to scrap this project for now giving the win to AMD for stability until more mature chipsets for the Intel systems emerge. It seems there is nothing that can be done to make the Intel system stable?
Change the memory maybe? PSU? Were the OS's reinstalled properly every time? no idea. Way too many variables. If I had the money, I would conduct this test myself and interact in real time with all of you to make sure every contingency was thought of.

Like I said before in this thread, I guess this is caused by heat, basically. Whole system heat. This is a super stressed test, with full load all the time, with some key components running at the limit. Pentium D is extremely hot by itself and this heat affect to these other components, and to the whole system stability.
Only my two cents.


EDIT: pfff Far Cry chashed in Intel system right now.

 

[Sinker]

Originally posted by: Duvie
Yep the last thing you said is the chart I am looking at.....3 reboots....I was watching when that first one went down at around 10:00pm pst last night and then upon restart nothing was running for the next several hours.....0% cpu usage on all 4 cores...

And have you noticed how the '24 hour' chart only shows 8 hours?!

 

[PetNorth]

Originally posted by: Duvie

Originally posted by: PetNorth

Originally posted by: keysplayr2003
So, more or less, THG should have just started out with a Intel 945P or 955X chipset. I mean, why wouldn't you?

They installed yesterday-night Asus mobo with Intel 955X (Asus P5WD2 Premium). It has crashed two times in aprox. eighteen hours. You see...

Actually 3 and then in the last hour cpu usage all over the place....prime for another crash....

yes... Far cry at this moment is down

 

[PetNorth]

All crashed, idle again.

 

[Duvie]

getting pretty ridiculous now.....

Dammit Keys....Go buy the stuff and do it right!!!!

 

[Duvie]

Originally posted by: PetNorth
All crashed, idle again.

LOL!!! I called that about 4 posts up...i saw the cpu usage bouncing all over the palce and knew it was getting flaky...

OK....So there has been only 2 reboots or shutdowns but 2 times where the system programs have crashed and may not require a reboot....Though honestly a quick reboot upon a hang up or freeze would never even register in the software charting IMO....I think this should count as 4 reboots, and relable them as reboots, crashes, and shutdowns....


Edit: when you are working from home like I do, this is better then a soap opera or a car chase in LA on Foxnews!!!

 

[Zebo]

Originally posted by: Sentential
Jesus.... did anyone see the CPU load temps of that intel system? I wonder if they have the heatsink installed properly. 84C* average is just crazy

This is why I say NEVER believe motherboard temps. Obviously that CPU is not running 90C or it would'nt be running at all.. It has the exact same "heart monitor" throttleing effect it had when running a 65C and 72C on the two other boards. How can you very from 65-90C with same chip. Classical motherboard diode BS which I never look at since I had two NF7-S's which varied by more than 25C with the same chip!!!!!

 

[PetNorth]

Originally posted by: DuvieEdit: when you are working from home like I do, this is better then a soap opera or a car chase in LA on Foxnews!!!

LOL!!

 

[Duvie]

If they think it is the NB chipset then they should actively cool it.....At this point I dont care if they water cool it I would like to get down to the bottom of the source of the instability...immature platform, heat related, etc....

I would also perhaps actively cool the mosfetts as well....if this thing at 84c is wrong like I assume, but other boards showed 71c it may not be too far off....

I think it is more heat related...It acts too similar on all the boards to be the same issue...i ofcourse am assuming they would be smart enough to test the ram and swap them out awhile ago for testing assurances....

 

[Duvie]

Intel machine is offline now officially and picture shows the hSF has been removed from cpu...Maybe in prep for throwing out the window!!!

I hope they are going try a better cooler just to clarify the issue....

 

[PetNorth]

Originally posted by: Duvie
If they think it is the NB chipset then they should actively cool it.....At this point I dont care if they water cool it I would like to get down to the bottom of the source of the instability...immature platform, heat related, etc....

I would also perhaps actively cool the mosfetts as well....if this thing at 84c is wrong like I assume, but other boards showed 71c it may not be too far off....

I think it is more heat related...It acts too similar on all the boards to be the same issue...i ofcourse am assuming they would be smart enough to test the ram and swap them out awhile ago for testing assurances....

agree

now... Offline. haha this is an horror

 

[Duvie]

http://www.tomshardware.com/cpu/20050603/stresstest-01.html

Down where they highlight the Epox manual where it says support for dual core....It says Pentium D 3.0ghz and above....Are they assuming too much being htat this is an EE medel and with 4 cores could draw more current....I mean the 840 EE has a higher power draw under load then the 840 without HT......I wonder if just that extra fried the board...nice to know the board has no headroom for anymore cpus above 3.2ghz 840...however it sounds like this line wont be going up anymore anyways...


< At this point the only solution left was to use a board with an Intel 955X chipset, which, although it is sold only individually, is supposed to run without problems. However, the use of an Intel component involves a clear limitation: the current Nvidia driver deactivates the SLI mode of the graphic card on all Intel boards.

Despite this, in order to continue to guarantee a fair comparison between AMD and Intel systems, we are now determined to use a complete Intel platform based on the 955X chipset. So we established new conditions: all benchmarks will be restarted and the counters simultaneously set back. The SLI mode in the AMD system will also be disabled.

The initial testing of an appropriate motherboard with this chipset had strange results: two of our very own boards, identical ones of the D955XBK type, were unable to work with the processors provided by Intel. The Asus P5WD2 Premium with 955X chipset is now being used.

And after all that, finally, the Intel platform is running again - after the replacement of a total of five motherboards within a single day >

So since the crash 7:00am europe time or 10:00pm pst (last night) and long idle period they have been using the intel 955 board and now it too has bit the dust...It has had 2 reboots on its own and no is offline....

So it has to be the cchips heat or power draw through these boards...Good luck Dell!!!

 

[Duvie]

New system is up!!!!! and now appears to be crashing again....temp ran up to mid 70's and then started coming down...cpu usage at 13min is zero again...

 

[Keysplayr]

Is it me or did they plug the CPU fan into a different fan header this time? RPM's are higher now. Almost 1000 rpm higher. What was it before, about 2150rpm? Now its just over 3000.

 

[Duvie]

Originally posted by: keysplayr2003
Is it me or did they plug the CPU fan into a different fan header this time? RPM's are higher now.

I think it is a different fan...I dont know about the plug becasue they just may have had the cords not draped that way before....I never cache any of the old images....

Did they perhaps hav eit plugged into on eof the power supply 4 pins before??

 

[Keysplayr]

NOW WE ARE TALKING FAN SPEED!!!! (sorry for yelling)

 

[Duvie]

Yep the temps have hit 70c again but the rpm speed is about double....Now lets see if it can hold...The startup looked shaky though...

Edit: the rboots are still wrong...It should be 4 or 3 with a software crash.....Are the Germans slow at counting??? Anybody who can read a chart can ell something happened for 4-5 hours...

 

[Keysplayr]

At this time, I would just like to thank THG for all of this entertainment. At the very least, it sure gives us stuff to talk about. Yes, fan speed is at almost 4000rpm. Thats more like it. And I noticed they must have flashed the bios on the ASUS board because temps are reporting normal (ish) now. Dunno.

 

[Duvie]

I still think it looks differnt (HSF that is)....Why do you think the change in the cpu fan speed...Any theories??

 

[Keysplayr]

Originally posted by: Duvie
I still think it looks differnt (HSF that is)....Why do you think the change in the cpu fan speed...Any theories??

Q-Fan Technology built into the ASUS boards may not have been working correctly. Could be they turned it either on or off, or a BIOS update to actually just make everything report correctly. Could be a different fan because I can no longer easily see the actual HeatSink through the spinning fan blades. Now mostly black indicating larger fan blades indicating a different fan as you say.