QuadGT Arrived (56k warning)

Printable View

Show 100 post(s) from this thread on one page

03-06-2007, 04:37 PM
RLM

A temporary solution to this locked at 6 multiplier and quad core issue was posted on the abit forums by beppe. It comes in the form of crystalcpuid. You can manually change the multiplier back to it's correct setting.

The abit thread is here http://forum.abit-usa.com/showthread...=120670&page=3

The download can be found here
http://crystalmark.info/download/index-e.html

I've tested this with some 3D rendering in lightwave and cinebench9.5, and the times are a lot more respectable. Nice one:)

RLM
03-06-2007, 04:37 PM
Ruslan

Hmmm

Quote:

Originally Posted by ziddey

I remember reading that there's a problem with vista possibly overriding c1e/eist settings, so you need to fiddle with power settings possibly. If rmclock works, you could always force it

That makes the most sense. now any idea how to shut the power management in vista off all the way :confused: ???
03-06-2007, 07:49 PM
Revv23

holy crap i cant believe you guys are popping 100c on the PWM, i havent been over 60-70 and i have used 1.75v at 4.1ghz.

on the other hand my cooling is really good...
03-06-2007, 08:18 PM
Gitsum

Your pictures look like ASS, retard.
03-06-2007, 08:40 PM
ziddey

Quote:

Originally Posted by Ruslan

That makes the most sense. now any idea how to shut the power management in vista off all the way :confused: ???

not sure but as mentioned, crystal is a way. also, if rmclock works under vista, you can redisable c1e/eist in it. i've been even able to reenable it on systems i've disabled it in at the bios level
03-06-2007, 09:29 PM
situman

Quote:

Originally Posted by Revv23

holy crap i cant believe you guys are popping 100c on the PWM, i havent been over 60-70 and i have used 1.75v at 4.1ghz.

on the other hand my cooling is really good...

wanna share?
03-06-2007, 09:59 PM
1heone

:) 1heone

AKA

beppe
03-06-2007, 10:01 PM
Revv23

Quote:

Originally Posted by situman

wanna share?

my cooling?

already in this thread somewhere... :p:

EDIT - found em: http://xtremesystems.org/forums/show...&postcount=149

Imagine that in a case with stock cooling on the 1900... otherwise its still like that, i have the extra cooling of the nbsink cooling my pwm, so that explains my low temps.

I want to note, i was wrong, they never climb over 55c.. not 65c like i originally said.
03-06-2007, 10:06 PM
fhpchris

Quote:

Originally Posted by Revv23

holy crap i cant believe you guys are popping 100c on the PWM, i havent been over 60-70 and i have used 1.75v at 4.1ghz.

on the other hand my cooling is really good...

Pretty much every digital PWM board is garbage...

RD600
QuadGT
Foxconn AB
IN9 32x max (not as bad as the others AFIK, but it has other bugs)

Nearly all the boards droop like a Biotch, are impossible to mod, and have that 100+C PWM issue....

I saw mabye 70-80C @ 110A @ 1.7v DMM on my QuadGT with my L628B211 Xeon 3060, but throw in a Quad and it is game over.

EVGA 680i + Xeon3060 @ 1.6v prime = 430 watts measured(kill-a-watt)
EVGA 680i + QX6700 @ 1.4v IDLE = 400 watts measured
EVGA 680i + QX6700 @ 1.4v quad prime = 530 watts measured
03-06-2007, 10:09 PM
Heidfirst

Quote:

Originally Posted by Evanesco

By the way, I just come to think about, that I noticed the PWMs where around 85C and raised well beyon 100C when the CPU was under load. But uGuru didn't give me any warnings, so I didn't bother at the time, had other things to bother about, anyone ???

the default warning on the QuadGT is 125C afaik.
03-07-2007, 05:15 AM
boblemagnifique

Somebody can test this Please :

I have just found some things interesting and I am not only a to have this "bug"

I measure real Vcore with my multimeter to have the true tension :

Vcore : 1.305v bios , 1.257v real under Windows (With 7800GT)
Vcore : 1.305v bios , 1.305v real under Windows (With 8800 GTS 320 & 640)

If I overclock the 8800 GTS, I can go up until 1.34/1.35v real (1.305v
bios)

A friend has to also test, it has similar with a X1900XT and one 8800 GTS
320mo, there is real a difference in tension on the level of the vcore
(With µGuru , he isn't multimeter)

Thank you for the Help ;)

http://boblemagnifique36.free.fr/Har...GT/QuadGT4.jpg

http://boblemagnifique36.free.fr/Har...GT/QuadGT3.jpg
03-07-2007, 05:21 AM
Johnny Bravo

boblemagnifique, just to confirm, the machine had no load on both readings? ie it was running idle? My board is in for an RMA atm but once it's back I'll test with a 7600GS
03-07-2007, 06:05 AM
boblemagnifique

Quote:

Originally Posted by Johnny Bravo

boblemagnifique, just to confirm, the machine had no load on both readings? ie it was running idle? My board is in for an RMA atm but once it's back I'll test with a 7600GS

I had to test the last week end :

-6400 @ 3800mhz at 1.545v Bios
-Leadtek 8800 GTS 320mo stock Fréquency

Under Windaube : 1.580v

Under 3dmark 2006 :

GT1/GT2/GT3/GT4 : 1.57/1.58v approximately
Test Cpu 1 : 1.559/1.561v
Test Cpu 2 : 1.559/1.563v

With 8800 GTS 320Mo @ 650/2040

Under Windaube : 1.595v

GT1/GT2/GT3/GT4 : 1.585/1.595v approximately
Test Cpu 1 : 1.565/1.57v approximately
Test Cpu 2 : 1.565/1.57v approximately
03-07-2007, 06:28 AM
fgw

see my post #252 in this thread regarding vdroop on this board.

will redo the measurements with higher vcore.

also your readoutpoint for ddr2 is not showing the correct values! used the same point before, but measuring on the ddr2 sockets directly showed something different. i'm in a hurry right now, but will show the readout points i use later. ...

if i got you right, you tested the effect on vcore during high gpu loads? have not done this test on the quad so far, but on my aw9d-max i observed something similar. had vcore as low as possible to be able to run my overclock stable. as i started scan for artifacts in ati tool while running orthos the systems hangs or crashed randomly. looking at vcore showed quickly whats going on. vcore dropped from idle to load, thats nothing new, but as soon as i started artifact scanning, vcore dropped even lower and i assume this caused the instabilities during high gpu loads! upped vcore one notch and everything was fine again!

have not yet done this test on the quad, but will look into it tonight or may be tomorrow ...
03-07-2007, 07:27 AM
boblemagnifique

Quote:

Originally Posted by fgw

see my post #252 in this thread regarding vdroop on this board.

will redo the measurements with higher vcore.

also your readoutpoint for ddr2 is not showing the correct values! used the same point before, but measuring on the ddr2 sockets directly showed something different. i'm in a hurry right now, but will show the readout points i use later. ...

if i got you right, you tested the effect on vcore during high gpu loads? have not done this test on the quad so far, but on my aw9d-max i observed something similar. had vcore as low as possible to be able to run my overclock stable. as i started scan for artifacts in ati tool while running orthos the systems hangs or crashed randomly. looking at vcore showed quickly whats going on. vcore dropped from idle to load, thats nothing new, but as soon as i started artifact scanning, vcore dropped even lower and i assume this caused the instabilities during high gpu loads! upped vcore one notch and everything was fine again!

have not yet done this test on the quad, but will look into it tonight or may be tomorrow ...

oki Thank you for your Help

Indeed, I am not sure point of measurement for Vddr, I asked has Shamino to also look at there to confirm :)
I sent this information has Abit Taiwan so that it test also (Cebit soon :( )
I recieved 2 Q6600 from Intel for this week end , I would test compared with the 6400 2mo L2 / 6600 4mo L2

I would test stability more in details this week end with 8800 GTS 640mo stock and Overclock with each Cpus
03-07-2007, 08:23 AM
Johnny Bravo

I agree that the droop on the QuadGT is a big problem but I was not aware that the gfx card was affecting it too! Did you have the aux 4 pin molex connected at the edge of the board when doing these tests boblemagnifique?
03-07-2007, 08:42 AM
Revv23

Quote:

Originally Posted by boblemagnifique

I had to test the last week end :

-6400 @ 3800mhz at 1.545v Bios
-Leadtek 8800 GTS 320mo stock Fréquency

Under Windaube : 1.580v

Under 3dmark 2006 :

GT1/GT2/GT3/GT4 : 1.57/1.58v approximately
Test Cpu 1 : 1.559/1.561v
Test Cpu 2 : 1.559/1.563v

With 8800 GTS 320Mo @ 650/2040

Under Windaube : 1.595v

GT1/GT2/GT3/GT4 : 1.585/1.595v approximately
Test Cpu 1 : 1.565/1.57v approximately
Test Cpu 2 : 1.565/1.57v approximately

IS your 12v line on your psu adequate? man this digital PWM sucks... :(
03-07-2007, 08:45 AM
MADMAX22

Quote:

Pretty much every digital PWM board is garbage...

RD600
QuadGT
Foxconn AB
IN9 32x max (not as bad as the others AFIK, but it has other bugs)

Nearly all the boards droop like a Biotch, are impossible to mod, and have that 100+C PWM issue....

umm what about the am2 board from dfi, the 590 , it is a kick*** board and its got the digital pwm and it overclocks like crazy for the am2 guys
03-07-2007, 09:29 AM
situman

Did the whole stupid ABit office just pick up and go to Cebit? I guess no one needs to work? What they gonna say after they come back? "Oh, we are too tired so new bios next year?"
03-07-2007, 10:05 AM
Millyons

Quote:

Originally Posted by situman

Did the whole stupid ABit office just pick up and go to Cebit? I guess no one needs to work? What they gonna say after they come back? "Oh, we are too tired so new bios next year?"

lol dont be so harsh....mabey fall after the summer vacations and to have the mobo ready for the christmas season :D
03-07-2007, 10:12 AM
Ruslan

Let them have there time.

I prefer a good bios instead of a rushed one. let them get the extra time. the latest beta is fine. better quality than most other companies stable releases.
03-07-2007, 10:46 AM
situman

Quote:

Originally Posted by Revv23

my cooling?

already in this thread somewhere... :p:

EDIT - found em: http://xtremesystems.org/forums/show...&postcount=149

Imagine that in a case with stock cooling on the 1900... otherwise its still like that, i have the extra cooling of the nbsink cooling my pwm, so that explains my low temps.

I want to note, i was wrong, they never climb over 55c.. not 65c like i originally said.

thats cheating man, over the top air cooling lol.
03-07-2007, 11:01 AM
boblemagnifique

Quote:

Originally Posted by Revv23

IS your 12v line on your psu adequate? man this digital PWM sucks... :(

yes I have potentiometer (12.3/12.4v), I could test with my OCZ 1kw but that will make similar I think :)
03-07-2007, 11:14 AM
fhpchris

Quote:

Originally Posted by MADMAX22

umm what about the am2 board from dfi, the 590 , it is a kick*** board and its got the digital pwm and it overclocks like crazy for the am2 guys

The problem is that Kentsfield can pull DOUBLE the power of a conroe B @ 1.6v and most likely THREE times as much as AM2.
03-07-2007, 11:54 AM
boblemagnifique

Quote:

Originally Posted by Johnny Bravo

I agree that the droop on the QuadGT is a big problem but I was not aware that the gfx card was affecting it too! Did you have the aux 4 pin molex connected at the edge of the board when doing these tests boblemagnifique?

Yes that same similar

A 4 pin molex is normally to feed a SLI or Crosffire (not obligatory into individual)

Show 100 post(s) from this thread on one page