Andreas - the patches work, cheers. Both of my tests (which failed
yesterday) work first time, no warning messages are shown either
(yesterday there were heaps of them).
Much obliged,
Ian.
On 3 March 2010 21:58, Andreas Klöckner li...@informa.tiker.net wrote:
On Mittwoch 03 März 2010, Ian
Scratch the last - the same errors occur with the latest master as
listed below. In my haste I didn't remove the already-compiled kernels
(I cleared the wrong cache directory sigh).
The fix is to comment out lines 312 and 457 of pycuda-complex.hpp and
recompile, I still get a ton of warning
Hi,
I had to add the patch in attachment to make work a kernel like
void kernel( float* out, int size){
}
I am not 100% sure it's corret. PTAL
Fabrizio
--
Luck favors the prepared mind. (Pasteur)
0001-added-int-detection.patch
Description: Binary data
On Donnerstag 04 März 2010, Fabrizio Milo aka misto wrote:
Hi,
I had to add the patch in attachment to make work a kernel like
void kernel( float* out, int size){
}
Unless you're using prepared invocation, you have to use Numpy's sized
integers/floats:
Unless you're using prepared invocation, you have to use Numpy's sized
integers/floats:
I see.
Did the patch with the decorators, get trough or got lost yesterday?
Fabrizio
--
Luck favors the prepared mind. (Pasteur)
___
On Donnerstag 04 März 2010, Ian Ozsvald wrote:
Scratch the last - the same errors occur with the latest master as
listed below. In my haste I didn't remove the already-compiled kernels
(I cleared the wrong cache directory sigh).
The fix is to comment out lines 312 and 457 of
Fabrizio Milo aka misto wrote:
b) Imports with side effects are bad. (pycuda.autoinit is the only
module with side-effects in PyCUDA--and the side effect is its only
purpose.)
I agree side effects are bad!! I don't think this is a side effect.
This is how you would use the cuda driver.
any and
On Thursday 04 March 2010 07:59:34 pm Imran Haque wrote:
I disagree. It's important to know if the CUDA libraries are getting
reinitialized; particularly on multi-GPU systems, there can be
significant overhead in calling cuInit.
Agreed,
In particular, if your program uses libraries
that
does anyone has an example of a program where doesn't use the
cuda.autoimport before using any of the pycuda.* ?
I can't think to an example. Also in multi thread, cuInit should be
called only once.
Fabrizio
--
Luck favors the prepared mind. (Pasteur)
Fabrizio Milo aka misto wrote:
does anyone has an example of a program where doesn't use the
cuda.autoimport before using any of the pycuda.* ?
Yes, my library (shameless plug: https://simtk.org/home/siml) manually
handles CUDA initialization, because just getting some context from
autoinit
On Donnerstag 04 März 2010, Imran Haque wrote:
Fabrizio Milo aka misto wrote:
does anyone has an example of a program where doesn't use the
cuda.autoimport before using any of the pycuda.* ?
Yes, my library (shameless plug: https://simtk.org/home/siml)
Cool--I've added that to
On Donnerstag 04 März 2010, Fabrizio Milo aka misto wrote:
The work around should simple be to import pycuda after the fork.
Importing before would be useless, because for sure you can't
initialize cuInit and thus can't use any cu* function..
Or I am missing something ?
Imports might happen
12 matches
Mail list logo