08.07.2011 08:01, Bogdan Opanchuk пишет:
Hello Mikhail,
This program worked without any changes on Tesla C2050. Such error
message usually points to insufficient number of registers on the
device, so try to choose block size<= MAX_REGISTERS_PER_BLOCK (device
attribute) / func.num_regs
Best regards,
Bogdan
On Fri, Jul 8, 2011 at 1:46 PM, Алексей Гурин<[email protected]> wrote:
08 июля 2011, 01:12 от Mikhail Korobko<[email protected]>:
Hello,
I'm trying to use PyCUDA for numerical calculations and I have some problems
with resourses. I want to integrate some function using Sympson
method(scalar multiplication of vector of Sympson coefficients and vector of
function).
When I use simple function everything seems to be ok, but for some reason
more copmlex function rises an error: "pycuda._driver.LaunchError:
cuLaunchGrid failed: launch out of resources"
Traceback (most recent call last):
File "scalar_prod_complex_1_test.py", line 198, in<module>
func(dest,a_gpu,b_gpu,f_gpu,W,S_h_quant, block = (thread_strides, 1, 1),
grid = (block_size,1))
File
"/usr/local/lib/python2.6/dist-packages/pycuda-2011.1-py2.6-linux-x86_64.egg/pycuda/driver.py",
line 187, in function_call
func.launch_grid(*grid)
pycuda._driver.LaunchError: cuLaunchGrid failed: launch out of resources
My program is attached to the mail.
P.S. I'm new in CUDA so it's possible that I've just made some stupid
mistake=)
Best regards,
Mikhail
_______________________________________________
PyCUDA mailing list
[email protected]
http://lists.tiker.net/listinfo/pycuda
I tried your program and it worked without error when i reduced
thread_strides and BLOCK_SIZE to 256
_______________________________________________
PyCUDA mailing list
[email protected]
http://lists.tiker.net/listinfo/pycuda
Thank you very much! Reducing the block size really works good.
_______________________________________________
PyCUDA mailing list
[email protected]
http://lists.tiker.net/listinfo/pycuda
