Where are the GeForce GT 240 Linux drivers?

[wvdgaast]

Wow, so apparently this was still an unknown bug with them? Kudos to you for reporting it then. Is their bug tracker open BTW? A few quick Google queries didn't show me anything useful..

I see you have a drop-down box to select PowerMizer modes in your screenshot. I don't have those on my machine, but I'm still stuck with the 190.42 driver; is it a brand new thing?

[ttsiodras]

Originally Posted by wvdgaast

Wow, so apparently this was still an unknown bug with them? Kudos to you for reporting it then. Is their bug tracker open BTW?

No - I got the number from their response mail.

I see you have a drop-down box to select PowerMizer modes in your screenshot. I don't have those on my machine, but I'm still stuck with the 190.42 driver; is it a brand new thing?

No, I don't have such a drop-down box - I wish I did! (we'd all be happy then, since this would solve all the problems reported in this and about a dozen other powermizer-related theads).
The only drop-down existing allows me to select "Adaptive" or "Prefer maximum performance", and with either option I still get the lousy performance of 324MHz (instead of 1.7GHz).

By the way, Phoronix just reviewed a GT240 and published similar speed problems -but Michael apparently didn't notice the powermizer problem that causes this...

[ttsiodras]

Just tried 195.36.03 on my GT240/GDDR5 - powermizer bug remains, memory speed is still 324MHz instead of 1.7GHz, even when running OpenGL/CUDA apps...

$ glxgears & 
$ nvidia-settings -q all 2>&1 | egrep '(Current)?ClockFreqs' | grep '0\.0'
  Attribute 'GPU2DClockFreqs' (home:0.0): 135,135.
  Attribute 'GPU3DClockFreqs' (home:0.0): 550,1700.
  Attribute 'GPUDefault2DClockFreqs' (home:0.0): 135,135.
  Attribute 'GPUDefault3DClockFreqs' (home:0.0): 550,1700.
  Attribute 'GPUCurrentClockFreqs' (home:0.0): 405,324.

Under Windows, when running OpenGL apps, the card bumps up its speed to 1.7GHz (verified by GPU-z) and achieves anywhere between 2-5 times better speeds (depending program tested). The bug is easily verified via any of NVIDIA's own CUDA SDK tools, like bandwidthTest.

Come on, NVIDIA...

[Zoom79]

I replaced a GeForce 7600 with a GT 240-based card in my system yesterday. Normally, I'm using 64-bit Fedora 12 as a MythTV backend and VMware work area. I had been using the nouveau drivers and things seemed to work OK to this point, but I'd been interested in seeing what the NVidia drivers could do with a new card. I'd seed this thread and decided to watch to see how things worked out before switching.

Last night, I realized I had an old installation of UbuntuStudio installed on a test partition with the Nvidia drivers installed so I could check things out. In the categorization of hardware / software / performance, I'd summarize:

Zotac GT 240 / DDR3 / Passive Cooling
64-bit 2.6.31-9-rt kernel / X-server 1.6.4 / Nvidia 185.18.36
PowerMizer runs card at highest perf level (level 2) @ 550 / 790 with glxgears running
Side note: GPU thermal monitor show stable 50 C as I type this

Don't know if this is helpful, but it did seem like another data point.

More Detailed Outputs:

root@mycomputer:~# uname -a
Linux mycomputer 2.6.31-9-rt #152-Ubuntu SMP PREEMPT RT Thu Oct 15 13:22:24 UTC 2009 x86_64 GNU/Linux


NVIDIA Driver Version: 185.18.36
X Server Info:
Number: 11.0
Version: 1.6.4 (10604000)
NV-Control Ver: 1.18

Card: Zotac GeForce GT 240 1GB DDR3 Passive Cooling
http://www.newegg.com/Product/Produc...-138-_-Product

root@mycomputer:~# lspci -v
01:00.0 VGA compatible controller: nVidia Corporation Device 0ca3 (rev a2)
Subsystem: ZOTAC International (MCO) Ltd. Device 2134
Flags: bus master, fast devsel, latency 0, IRQ 16
Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
Memory at c0000000 (64-bit, prefetchable) [size=256M]
Memory at dc000000 (64-bit, prefetchable) [size=32M]
I/O ports at 8c00 [size=128]
[virtual] Expansion ROM at fe780000 [disabled] [size=512K]
Capabilities: [60] Power Management version 3
Capabilities: [68] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
Capabilities: [78] Express Endpoint, MSI 00
Capabilities: [b4] Vendor Specific Information <?>
Capabilities: [100] Virtual Channel <?>
Capabilities: [128] Power Budgeting <?>
Capabilities: [600] Vendor Specific Information <?>
Kernel driver in use: nvidia
Kernel modules: nvidia, nvidiafb


root@mycomputer:~# nvidia-settings -q all | grep GPUPerfModes
Attribute 'GPUPerfModes' (mycomputer:0.0): perf=0, nvclock=135, memclock=135 ; perf=1, nvclock=405, memclock=324 ; perf=2, nvclock=550, memclock=790


steve@mycomputer:~$ glxgears
Running synchronized to the vertical refresh. The framerate should be
approximately the same as the monitor refresh rate.
54150 frames in 5.0 seconds
53373 frames in 5.0 seconds
53914 frames in 5.0 seconds
53927 frames in 5.0 seconds
53943 frames in 5.0 seconds
53566 frames in 5.0 seconds
55400 frames in 5.0 seconds
55522 frames in 5.0 seconds
55470 frames in 5.0 seconds
55212 frames in 5.0 seconds
54692 frames in 5.0 seconds


root@mycomputer:~# nvidia-settings -q all | grep GPUCurrentClockFreqs
Attribute 'GPUCurrentClockFreqs' (mycomputer:0.0): 550,790.
'GPUCurrentClockFreqs' is a packed integer attribute.
'GPUCurrentClockFreqs' is a read-only attribute.
'GPUCurrentClockFreqs' can use the following target types: X Screen, GPU.

[ttsiodras]

Originally Posted by Zoom79

...I'd summarize:

Zotac GT 240 / DDR3 / Passive Cooling
64-bit 2.6.31-9-rt kernel / X-server 1.6.4 / Nvidia 185.18.36
PowerMizer runs card at highest perf level (level 2) @ 550 / 790 with glxgears running

Yep, it seems that this bug affects only GT240 GDDR5, not GDDR3.

Similar thread with reports about GT240/GDDR5 and other chips running at slow speeds due to powermizer (no matter what app is running - i.e even with OpenGL or CUDA running):
http://www.nvnews.net/vbulletin/showthread.php?t=141116

[voidmage]

Sorry to bump an old thread like this, but I've been testing ubuntu lucid and have been getting hard locks again with the WAIT error. I had posted earlier in this thread that this wasn't a problem with the 190 drivers, but they're only packaging the 195 drivers in lucid right now. What can I do in this situation?

[Dizzle7677]

Originally Posted by voidmage

Sorry to bump an old thread like this, but I've been testing ubuntu lucid and have been getting hard locks again with the WAIT error. I had posted earlier in this thread that this wasn't a problem with the 190 drivers, but they're only packaging the 195 drivers in lucid right now. What can I do in this situation?

Set the Powermizer control to 'prefer maximum performance' instead of 'adaptive'. Or if you're using a pre-release of some software like Firefox 3.6.2pre it might cause troubles. But you never really know what the cause is until you debug X. See if that helps.
More info: https://bugs.freedesktop.org/show_bug.cgi?id=15473

[cleanrock]

Still a problem with 195.36.08.
I bought my GT 240 GDDR5 today and i was confident this new driver (its even a stable release) fixed this problem since its been known to nvidia for more than a month now !
I think an nvidia rep should make a statement here so we know when/if our cards will work.

[AaronP]

There was a change that went in just after the current release that may fix the problem. Please try the next beta release.

[dura91]

Originally Posted by AaronP

There was a change that went in just after the current release that may fix the problem. Please try the next beta release.

Fix what? The GT240 with DDR5 stucked at 405/324 instead of 550/1700 or is it about the WAIT error?

[AaronP]

I'm sorry, I mean the PowerMizer issue on certain DDR5-based board. I haven't yet been able to reproduce the WAIT issue.

[Kurtosis]

Thanks Aaron, but when I search Beta drivers at Nvidia's site, I get zero results for the GT 240. And when I try another 200series, like the GTX 295, the most recent beta driver listed is 195.30.

Is there another location to get a more recent beta version that you seem to be referring to? Or is 195.30 the most recent?

Thanks!