CUDA 2.0 Beta 2 GTX support, more Linux distros...

Thanks for the update, Tim.

UPDATED WINDOWS DRIVERS - now works on all CUDA cards
This is a beta driver intended for CUDA development only.

WinXP 177.35 driver
WinXP 64-bit 177.35 driver
Vista 32-bit 177.35 driver
Vista 64-bit 177.35 driver

sorry about the delay for the fix, folks. if there’s ever anything that any of you feel that we need to be aware of ASAP, please feel free to send me a PM.

I installed the fixed 177_35 and now it works with my 8600GTS, yay!

But I got a problem… some examples make a BSoD ( nvlddkkm.sys ).

I ran the alignedTypes.exe x86 (Release) over Vista x64 SP1 with the 177_35 for x64.

edited: The x64 alignedTypes.exe does not cause a BSoD but makes my screen to show “snow”… and the fluidsGL shows a black screen result.

I just installed the Cuda beta 2 and I experience a decrease in device to device bandwidth. The bandwidthTest previously showed 64000 MB/s with Cuda 1.1 and I now only have 57091 MB/s .
Was there any change in the way the bandwidth test works or is it a bug with my config?

Card: Tesla C870
Linux x86-64 Fedora8 driver 177.13 .

I updated motherboard bios to latest version but that didnt change anything. Any Idea what might be the cause ?

Bump for new version of SDK.

CUDA SDK 2.0b2:
UPDATED WINDOWS RELEASE
Fixes glew64.dll errors.

Linux
32-bit WinXP or Vista
64-bit WinXP or Vista

(the Linux release is the same, I’m just including it here for the sake of completeness)

Thanks for the update. Has there been any word on the Vista issues with multiple cards? (i.e. needing a monitor with the desktop extended onto it attached to a card for it to be usable)

Hi,there
174.35 driver CUDA2.0 SDK&Tooolkit beta2 GTX 280
do not run correctly on my machine
device query shows that the device had been found and is running
the bandwidth test past, but matrixmul and transpose and some other examples failed, my programmes that could run correctlu under CUDA2.0 ,GF8800GTS also failed, seems that the device did nothing, the porgramme read back the same data sent to device memory, without changing any value.

When is nVidia going to release new MATLAB plug-in to go with 2.0 toolkit? MATLAB plug-in 1.1 does not recognize --arch or --gpu_name compiler flags that enable double precision computation.

When is nVidia going to release new MATLAB plug-in to go with toolkit 2.0? MATLAB plug-in 1.1 does not recognize --arch and --gpu_name compiler flags that enable double precision computation.

Solution to enabling double precision computations in MATLAB is given here:

[url=“The Official NVIDIA Forums | NVIDIA”]The Official NVIDIA Forums | NVIDIA