260GTX nvidia-drivers-190.42-r3 - powermizer issue - Page 7

cbuchner1 · 05-25-10, 07:38 AM

There was an attempt to fix this bug in the BETA 256.25 drivers.
Powermizer level #1 now has the 600MHz Core, 1000 MHZ Mem Clock setting,
and keeps changing between level #0 and #1 in the Adaptive mode.

But for me it fails to actually apply these clocks, the clock only hits 275 MHz
core, 301 MHz Mem with this driver instead of actually clocking up
(which coincidentially happens to be the setting for levels #2 and #3)

Looks to me like a botched attempt to fix that bug.

Christian

cedric.s · 05-25-10, 09:42 AM

i have exactly same thing on a gtx 285m with beta 256.25 driver.

powermizer show level 1 show 600MHz for Core and 1000 MHZ for Mem Clock.

nvclock show 275 MHz for core and 301 MHz Mem.

there is no gain in performance....

**zander** · 05-25-10, 11:30 AM

The 256.25 BETA driver release does not yet include a fix for the ACPI related Powermizer problems I referred to earlier. However, I hope the next 256.xx BETA driver release will.

gio___ · 05-25-10, 12:49 PM

Quote:

Originally Posted by zander

The 256.25 BETA driver release does not yet include a fix for the ACPI related Powermizer problems I referred to earlier. However, I hope the next 256.xx BETA driver release will.

good words!

cbuchner1 · 05-25-10, 01:15 PM

Quote:

Originally Posted by zander

The 256.25 BETA driver release does not yet include a fix for the ACPI related Powermizer problems I referred to earlier. However, I hope the next 256.xx BETA driver release will.

If it is an ACPI related problem, as you state, would turning off ACPI in the Linux kernel (acpi=off) help?

Of course one would lose all the power saving features and possible the core i7 C3 sleep state that enables the other cores to increase their clock rates. But still we need that GPU power rather sooner than later!

Christian

**zander** · 05-25-10, 02:19 PM

Quote:

Originally Posted by cbuchner1

If it is an ACPI related problem, as you state, would turning off ACPI in the Linux kernel (acpi=off) help?

No, the problem is that the driver needs to retrieve information via the ACPI subsystem, but some driver-internal wiring is incomplete/broken on Linux.

NightOwl · 05-28-10, 02:05 PM

I hope this is fixed soon - or at least make it the default that if no ACPI is available, to run the hardware at full speed, rather than the crippled mode we get now. This problem has been around for months, and it is a crying shame that NVidia won't open source this so someone with the time and expertise can fix this rather than all of us sitting around waiting for manna from NVidia - which doesn't seem likely because there doesn't seem to be the will to support this niche of users.

**danix** · 05-29-10, 11:28 AM

Please try the new 256.29 beta release. Thanks!

cbuchner1 · 05-29-10, 12:04 PM

Quote:

Originally Posted by danix

Please try the new 256.29 beta release. Thanks!

You're working on Saturdays to bring us this new Beta? Awesome.

**AaronP** · 05-29-10, 12:18 PM

He even pushed to have it released before the long weekend, instead of waiting until Tuesday.

gio___ · 05-29-10, 12:21 PM

Quote:

Originally Posted by danix

Please try the new 256.29 beta release. Thanks!

wow... seems fixed!

In the weekend I'll do some test and write about problems that may arise ...
For now it seems to correctly using PowerMizer levels with a huge increase in performance.

Thanks to everyone involved in solving this problem

mrn · 05-29-10, 03:06 PM

I suspect the powermizer fix may have sorted this one out as well:
http://www.nvnews.net/vbulletin/showthread.php?t=142946

05-25-10, 07:38 AM	#73
cbuchner1 Registered User Join Date: May 2010 Posts: 3	Re: 260GTX nvidia-drivers-190.42-r3 - powermizer issue There was an attempt to fix this bug in the BETA 256.25 drivers. Powermizer level #1 now has the 600MHz Core, 1000 MHZ Mem Clock setting, and keeps changing between level #0 and #1 in the Adaptive mode. But for me it fails to actually apply these clocks, the clock only hits 275 MHz core, 301 MHz Mem with this driver instead of actually clocking up (which coincidentially happens to be the setting for levels #2 and #3) Looks to me like a botched attempt to fix that bug. Christian

05-25-10, 09:42 AM	#74
cedric.s Registered User Join Date: Nov 2002 Posts: 10	Re: 260GTX nvidia-drivers-190.42-r3 - powermizer issue i have exactly same thing on a gtx 285m with beta 256.25 driver. powermizer show level 1 show 600MHz for Core and 1000 MHZ for Mem Clock. nvclock show 275 MHz for core and 301 MHz Mem. there is no gain in performance....

05-25-10, 11:30 AM	#75
zander NVIDIA Corporation Join Date: Aug 2002 Posts: 3,740	Re: 260GTX nvidia-drivers-190.42-r3 - powermizer issue The 256.25 BETA driver release does not yet include a fix for the ACPI related Powermizer problems I referred to earlier. However, I hope the next 256.xx BETA driver release will.

05-28-10, 02:05 PM	#79
NightOwl Registered User Join Date: Sep 2006 Location: Toronto, Canada Posts: 23	Re: 260GTX nvidia-drivers-190.42-r3 - powermizer issue I hope this is fixed soon - or at least make it the default that if no ACPI is available, to run the hardware at full speed, rather than the crippled mode we get now. This problem has been around for months, and it is a crying shame that NVidia won't open source this so someone with the time and expertise can fix this rather than all of us sitting around waiting for manna from NVidia - which doesn't seem likely because there doesn't seem to be the will to support this niche of users.

05-29-10, 11:28 AM	#80
danix NVIDIA Corporation Join Date: Feb 2010 Location: Santa Clara, CA Posts: 237	Re: 260GTX nvidia-drivers-190.42-r3 - powermizer issue Please try the new 256.29 beta release. Thanks!

05-29-10, 12:18 PM	#82
AaronP NVIDIA Corporation Join Date: Mar 2005 Posts: 2,487	Re: 260GTX nvidia-drivers-190.42-r3 - powermizer issue He even pushed to have it released before the long weekend, instead of waiting until Tuesday.

05-29-10, 03:06 PM	#84
mrn Registered User Join Date: Dec 2004 Posts: 254	Re: 260GTX nvidia-drivers-190.42-r3 - powermizer issue I suspect the powermizer fix may have sorted this one out as well: http://www.nvnews.net/vbulletin/showthread.php?t=142946