Article:Ati Drivers Put To The Test

The_Neon_Cowboy · Oct 18, 2003, 10:17 PM

[SIZE=xx-large]Ati Drivers Put To The Test[/size]
by Michael W. Lady (AKA. The_Neon_Cowboy)

I wanted to see and show other weather the rumors of problems are true or not. I started off with the latest omega drivers based on the 3.8.5's since they were already in my system. I ran good old driver cleaner in between installs. Also all 3 temps were taken @ the exact same point in 3dmark just after test 3 while loading/starting test 4. Nothing on the desk or pc was moved (except my mouse) from start to finish.

Test System Setup...
Built by Ati Radeon 9800 PRO Bios 8004
A7n8x Bios Ver. 1007 (Mwarhead 2.45a REMIX )
Windows XP service pack 2 (Build 1240)
1 GB GeIL PC3500
Ati TV Wonder VE
Sound Blaster Live 5.1

The_Neon_Cowboy · Oct 18, 2003, 10:20 PM

First well start off with the room temperature.

I installed my probe on the back of the PCB with good reasons. Not
Only ease of installation and making sure that I do not disturb the cooling of the GPU.

The_Neon_Cowboy · Oct 18, 2003, 10:21 PM

[SIZE=xx-large]Results[/size]

Temperature just 5 minutes after booting....heat sink feels hot to the touch!

Temperature under the full load of 3dmark 2003 (build 330). heat sink felt like I was goanna burn my hand!!!

The_Neon_Cowboy · Oct 18, 2003, 10:24 PM

[SIZE=xx-large]Results[/size]

Temperature 5 minutes after booting....

Temperature under full load of 3dmark 2003 build 330.

The_Neon_Cowboy · Oct 18, 2003, 10:25 PM

[SIZE=xx-large]Results[/size]

Temperature 5 minutes after booting.... a bit cooler

Temperature under full load of 3dmark 2003 build 330.

[SIZE=xx-large]Memory[/size]

This time the probe was mounted against the side of one of the ram chips and under the side of my added heat sinks.

Temperature 5 minutes after booting....*

Temperature under full load of 3dmark 2003 build 330.*

The_Neon_Cowboy · Oct 18, 2003, 10:26 PM

[SIZE=xx-large]Final Results
[/size]

Omegas based on 3.8.5
GPU
Idle 47.1
Full Load 55.6

Memory*
Idle 40.9
Full Load 52.1

Official Cat. 3.8’s
Idle 40.9
Full Load 46.3

Memory*
Idle 40.9
Full Load 52.1

Official Cat. 3.7’s
Idle 41.5
Full Load 47.6

Memory*
Idle 40.9
Full Load 52.1

[SIZE=xx-large]Summary
[/size]

The drivers based on the dell 3.8.5 leaked betas were the hottest. Closely followed by the cat 3.7’s and finally the cat 3.8’s were the coolest. The results were surprising to me to say the least.

[SIZE=xx-large]Update
[/size]

After some suggestions I went back through and checked the memory temps. This time also with the memory nudged up to 350 MHz* from stock to appease the over-clockers here. I was also surprised to get the exact same readings in each driver.

[SIZE=xx-large]*****Special Update*****[/size]

I’m back once again. It seems I have gotten a lot of requests for test results with older cards. So I looked around and found my old Ati Radeon 64mb VIVO (183/183). My test point on this card since it can’t run all tests were on the first video/sound test right at the start.

[SIZE=xx-large] Test Setup [/size]

I had a little trouble getting good contact but after a few tries I fond a sweet spot.

OMEGA based on 3.8.5

Results

Temperature 5 minutes after booting

Temperature under full load of 3dmark 2003 build 330. A bit toasty for a 7200 core.

Ati Official 3.8's

Results

Temperature 5 minutes after booting....

Temperature under full load of 3dmark 2003 build 330.

Ati Official 3.7's

Temperature 5 minutes after booting.... a bit cooler

Temperature under full load of 3dmark 2003 build 330.

[SIZE=xx-large] Final Results [/size]

Omegas based on 3.8.5

GPU
Idle 33.3
Full Load 37.4

Official Cat. 3.8’s
GPU
Idle 32.9
Full Load 37.4

Official Cat. 3.7’s
GPU
Idle 32.9
Full Load 36.7

[SIZE=xx-large] Summary [/size]

As expected the want much difference on this lower end card. But the results were interesting. The Omega 3.8.5’s was a little warmer on the idle side. Both the 3.8.5 @ 3.8s maxed out at the same temperature. The 3.7 were slightly cooler in this test on both the idle and under a full load..

[SIZE=xx-large]**NEW ASUS DRIVER LEAK**[/size]

Well when I think it all said and done a new leak emerges Asus V6392. So, I stuck the 9800 pro back in. I’m testing only the GPU temps with this driver since the previous test resulted in no difference on the memory side.

Room temp has changed

It’s a lot cooler now…..

Wow that’s getting hot for idle temp.

At this room temp that’s HOT!

Room temp 21.5 C

Asus V6392
GPU
Idle 41.2
Full Load 45.9
[SIZE=xx-large] Summary [/size]

It’s hard to compare because the room temp is so low now. But it look like the Asus driver are as hot if not hotter than the omega 3.8.5’s (based on dells leak)

Vengeance · Oct 18, 2003, 10:55 PM

Good job Cowboy, Thanks a lot. You went to a lot of troble to do this.

The_Neon_Cowboy · Oct 18, 2003, 11:14 PM

Once again, this time with the help of Omega I was able to find a copy of the dell 3.8.5 leak that I had been looking for. I know many of you were probably wondering how temperatures play out on the Dell leak as well. This was done on my 9800 pro card only...

Room temperature ……

[SIZE=xx-large]Dell Leak 3.8.5 [/size]

Temperature just 5 minutes after booting....

Temperature under the full load of 3dmark 2003 (build 330). It flocculated between 48.1 and 48.2 so I’m goanna call it 48.15

[SIZE=xx-large]Omega 3.8.5 [/size]

Temperature just 5 minutes after booting....

Temperature under the full load of 3dmark 2003 (build 330).

[SIZE=xx-large] Final Results [/size]

Room temperature 22.3

Original Dell 3.8.5 Leak
Idle 42.1
Full Load 48.15

Omega's 3.8.5 (dell based)
Idle 42.7
Full Load 48.1

[SIZE=xx-large] Summery [/size]

Once again the dell drivers prove to be hot. It seems that the omegas and the dells are both equal in temperature. The problem appears to be a result of the actual BETA drivers codeing. Also, what could be a factor in all this the fact that those leaked dells were ment for 9800pro XT cards!.

mainman · Oct 19, 2003, 12:09 AM

Good job, so point is that leaked drivers are probably the problem and not the official 3.8??

This all is very confusing,

IronChisel · Oct 19, 2003, 12:16 AM

At least this would explain why I was getting some lockups with the new Omega 3.8.5's. Thanks for the info Cowboy.

Judas · Oct 19, 2003, 12:39 AM

.....*me hands out gold medal*.... er oops... i can do that...

...... most excellent job there neon.... (opps just about said neo... i watch metrix 2 on dvd last night at around 2AM.... man, i love the movie).....

mainman · Oct 19, 2003, 12:52 AM

Quote:

Originally posted by Judas
.....*me hands out gold medal*.... er oops... i can do that... ...... most excellent job there neon.... (opps just about said neo... i watch metrix 2 on dvd last night at around 2AM.... man, i love the movie).....

Yep i agree with you bud, neo did an awsome job i does deserve a medal.

kelvinchi · Oct 19, 2003, 01:35 AM

At least this would explain why I was getting broken-up on my Radeon 9700 Pro with the new Omega 3.8.5's. Thanks for the info Cowboy. now the display card have send to agency to repair..........

tragic · Oct 19, 2003, 01:56 AM

Gonna have to chip in here too. Just to say thanks Neon.

I too was getting a little confused with all the various ideas flying around about the 3.8's.

Outstanding job Neon..........thanks

kinetic · Oct 19, 2003, 02:18 AM

Brilliant work Neon! I haven't slept since all this blew up!

Thanks again!

***Veridian3*** · Oct 19, 2003, 02:49 AM

Neon, as an extra thing could you add a "temperature under overclocking" section? Just to see if anything changes between the 3 drivers when you overclock. Or if the temp order stays the same.

Stu

MaTT01 · Oct 19, 2003, 02:55 AM

Would it be worth testing with and without the VPU feature as well?

Edit- Very nice article you worked up, good job.

**MIG-31** · Oct 19, 2003, 03:02 AM

its good to see some taking the trouble to do stuff like neon just has.good work m8.

Falstaff · Oct 19, 2003, 04:28 AM

Lol...got an old 8500 card with these new drivers and some OC i am running stable. no problems, I dont 3dmark, sisoft sandra, or bryce 3d...just enjoy the drivers and the card..good work neon..thanks for the info..

softman · Oct 19, 2003, 04:30 AM

You tested the drivers with a 9800 pro card !!!!!!!!!

I would like to know, ARE THE TEMP`READINGS THE SAME ON OTHER CARDS ???????

I personaly have a (old) radeon 8500 64 mb (retail)

Can you test with other cards also ???

***Zardon*** · Oct 19, 2003, 04:33 AM

im sure he doesnt have alot of ATI hardware lying around, and the 9800 based series is one of the cards that people say are failing so IMO one of the most important. I would say these readings could safely carry across the board.

Judas · Oct 19, 2003, 05:13 AM

ah yes... if anyone is interested... i've had no problems whatsoever running the standard Cat 3.8's.... and i'm running them across dozens of radeons... anywere from a 7000...to a 9800 pro..... no problems reported..

BWX · Oct 19, 2003, 05:13 AM

Very strange that the ATI3.8's are so much cooler that the Omega 3.8's- I think I need to go buy a thermal probe now to do some experimenting on my 9700pro. Where did you get that one Neon?

Judas · Oct 19, 2003, 05:16 AM

actually, i was debating on gettting a 5 1/2" drive bay full LCD thermal/fan touch screen display... althoug $75-100, i think it's worth it seeing as i can connect it to hardrives/cpu/vid card/northbridge and any other components up to 3 more.... And have it display in bar graphcs or a few numbers....

~~HawK~~ · Oct 19, 2003, 05:18 AM

Great Job Neon!!
was no suprise to me, I have ran the 3.8.5 test too.. after running 3times aquamark3 I had a total system temp. rise of only 3gr. celcius (cpu +6c)
done with a closed normal case, normal cpu cooler, just one fan to push the heat out
room temp 23gr. celcius
Sure it wasend measured on the card it self, but if there was the alleged 10 degree rise
it would surely affect the mobo & Cpu since the extra heat wil go from the agp port to the mobo
& heat from the card itself would warm up the air in the closed case!!

System:
MoBo = Shuttle AK32a Via drivers: 449
PSU = Q-tec 550 watt
mem = 2x 256mb_cl2.5 Hyund. DDR266mhz
Cpu = Amd Athlon 1600xp+
HD = WD- 800JB 80Gb - Seagate Barracuda ATA-IV 80Gb
Vid = Ati9700 driver: Cat 3.8
OS = W2000 sp4 DX9b
Sound = Hercules FortissimoIII 7.1