THG Intel vs AMD stress test disaster

[saaya]

Thg started a stress test of the latest amd and intel plattform on june 3rd, and so far it has been a disaster. thg hardly documented their methods of testing and the problems they found and how they solved them.
for me its sad to see such a well known site do such a bad job, and have removed the old logs and performence stats making it look like they are trying to hide things.

UPDATE: it looks as if THG is using a BTX heatsink in an ATX case wich results in those ridiculous load temps of 70-90°C!

UPDATE: THG finally posted some more detailed infos about the problems they had, but still not about all the problems that were reported by xs members who checked the results and live feeds of the systems! http://www.tomshardware.com/cpu/20050603/index.html



i hope this gets enough attention to change the quality of their site in future!

here is the article:
http://www.tomshardware.com/cpu/20050603/index.html

here you can see the systems running on a webcam live feed that gets updated every few seconds.
http://www.tomshardware.com/stresstest/index.html

and here you can see the results of the different apps they are stressing both systems with wich gets updated all the time as well.
http://www.tomshardware.com/stresstest/charts.html


amd system:
a64 x2 4800+ with boxed HSF
gigabyte nf4 sli
2 x 512mb ocz pc3200 2-2-2-5
2x msi 6800U 256mb
2x WD 160GB SATA
PCPNC 850W PSU
Case with 2 fans, in front of the hdds and exhaust below the psu.

intel system:
840EE with boxed HSF
gigabyte nf4 intel edition sli
2 x 512mb ocz DDR2 667 4-4-4-5
2x msi 6800U 256mb
2x WD 160GB SATA
PCPNC 850W PSU
Case with 2 fans, in front of the hdds and exhaust below the psu.


thg log:
i made a log of all the events thg posted about in their article
xs log:
i made a log of all the events xs members posted about that thg hasnt mentioned


06-03 test starts
06-04 at 8:45am paciffic time the gigabyte nf4 intel edition board crashes
06-04 at 5pm paciffic time the gigabyte nf4 intel edition board crashes
06-04 bios flashed to increase the chipsets fanspeed, system restarted
06-04 the wattmeter dies
06-06 a bios flash kills the gigabyte nf4 intel edition board
06-06 an epox nf4 intel edition is beeing used
06-06 the epox board suddenly dies as well
06-06 an asus nf4 intel edition is beeing used
06-06 Crucial DDR2 667 is beeing used since they couldnt get the ocz memory to work on the news board
06-06 the reboot counter gets changed from 3 to 1
06-06 the reboot counter gets changed from 1 to 4
06-07 the intel system crashed again
06-07 an intel 955x board is beeing used, amd still running sli
06-07 the 955x board from intel doesnt work with the 840EE and died?
06-07 second 955x board from intel doesnt work with the 840EE either and died?
06-07 05B boxed heatsink gets replaced with a 05A boxed heatsink because of a termal pad??
06-07 enermax 480W psu beeing used instead of pcpnc
06-07 both systems now running with only one card
06-08 an asus 955x board is beeing used
06-08 different boxed HSF with bigger copper base from intel beeing used 05A heatsink removed
06-08 amd is no longer running sli, the crashcounter and the performence stats get reset AGAIN!
06-08 the webcam showed the intel system on the desktop yet the far cry fps of the intel system were increasing and increasing
06-08 the fps in far cry in the intel system dropped and the cd encoding speed jumped up significantly, looking as if the priority of the benchmakrs had been changed
06-08 the amd performence stats show that the amd suddenly encoded cds much slower, only 1 cd in one hour for at least 2 hours
06-08 since the asus 955x board has been used the system has crashed at least twice
06-08 the fanspeed of the intel fan jumped from 2000rpm on the old heatsink to 3600rpm on the new heatsink, so its not just a bigger copper base that makes the difference. the temps of the intel rig went down significantly from 90°C to 67°C and the intel system is suddenly faster in far cry than the amd system in the stats
06-08 the vcore on the intel system is now lower than it used to be in the earlier phase of the stress test
06-10 the fanspeed of the intel heatsink is up to 4300rpm!!!
06-10 the network connecting the test systems seems to be down, yet the intel systems score keeps increasing while the amd systems score froze
06-10 the persformence stats changed drastically all the sudden:
Lame Encoder Intel = 301 AMD = 256
Winrar Intel = 1634 AMD = 2072
Farcry Intel = 27FPS AMD = 35FPS
Divx Intel = 1380 minutes AMD = 80 minutes

About 10 minutes later the following:
Lame Encoder Intel = 303 AMD jumps up to 311
Winrar Intel = 1643 AMD jumps to 2321
FarCry Intel = 27FPS AMD drops to 15FPS
Divx Intel = 1390 AMD jumps to 130
looks like the priorities for the different tasks have been changed manually, no comment from thg...

06-13 ht gets disabled on the intel system?
06-14 final results get posted, intel is now faster in mp3 endocing compared to the previous results, no comment from thg. also no mention of the intel system crashing several times
06-14 intel mainboard changed because the asus board doesnt allow ht disabling?
06-10 ht disabled on the intel plattform to find out why divx is slower on the amd system
06-10 Priorities for Divx get set manually, making the step of disabling ht to find out about windows automated thread prioritizing useless?


at this monent the reboot counter of the intel system shows 3 reboots, wich havent been documented by thg, at least not yet?
further more thg claims the intel system was running steadily under load, wich is not true, not even by their own documentation.

Monday, June 6, 2005: Our stress test has now been going for more than 60 hours. The systems from AMD and Intel are running steadily under full load, and a lot has happened in the Munich THG lab during this time

here is a roundup of what was documented by our forum members:

Ok you guys need to view these with internet explorer since they are mht archive files.

First off we have the inital gigabyte board with already one re-boot

Next Toms switches boards to Asus NF4 and still shows 3 reboots

Can't have that!!! so he re-sets the boot counter back to 1!!!

Crashes 4 more times as shown here before finally upgrading to Asus 955x board. That's at least 7 crashes so far!!!

Meanwhile, as shown here AMD is rock solid 97 hours straight!!! (actually more like 102 but I don't have an archive of that)

Toms does'nt like Intel 7-9 crashes so he resets counter yet again to Zero for the intel system and whipes out AMD rocksolidness restarting all testing


Intel Crashed 3x since last night with the new testing on Asus 955x

Count them up to 12 crashes for intel, Zero for AMD.

and heres a funny link you gotta check out
http://users.pandora.be/compas/stresstest/intel.html


thx a lot to Mr. Tinker and Zebo for their help
im really curious about thgs conclusion of this

heres the discussion about the whole thg disaster on our forums:
http://www.xtremesystems.org/forums/...ad.php?t=64950