I've updated the mandelbrot.py demo:
http://wiki.tiker.net/PyCuda/Examples/Mandelbrot
which runs against the latest git, cheers for accepting the patches.
There's a (slightly linkbaity!) post on my blog with timing info for a
9800GT and GTX 480:
22,937* faster Python math using pyCUDA
Hi,
On Thu, Jul 15, 2010 at 4:49 AM, Ian Ozsvald i...@ianozsvald.com wrote:
One question that's raised from my post - how come double precision
CPU math is faster than single precision CPU math? I hadn't expected
that result to drop out of the test and it has been a while since I
did any good
Interesting...we use acos and sin/tan, possibly these are very
optimised for doubles on the CPU? I'll have to check this.
When you say 'amblibm' - did you mean:
http://developer.amd.com/CPU/LIBRARIES/LIBM/Pages/default.aspx
If so - cheers, that's new to me :-)
i.
2010/7/15 Frédéric Bastien
On Thu, Jul 15, 2010 at 10:26 AM, Ian Ozsvald i...@ianozsvald.com wrote:
Interesting...we use acos and sin/tan, possibly these are very
optimised for doubles on the CPU? I'll have to check this.
That could be the case as for exp in my case. Do you use gcc and the
default libm? no mkl or other
I'm on Windows, 32bit, using MS VC 2008. It looks like the lib is 64
bit only so I can't run it here. I'll bear it in mind for the future
though :-)
i.
2010/7/15 Frédéric Bastien no...@nouiz.org:
On Thu, Jul 15, 2010 at 10:26 AM, Ian Ozsvald i...@ianozsvald.com wrote:
Interesting...we use acos
On Tue, 29 Jun 2010 16:44:18 +0100, Ian Ozsvald i...@ianozsvald.com wrote:
Andreas, I'm attaching two patches.
0001 removes the #warning lines in cuda.hpp that make msvc (2008 on WinxP)
fail.
0002 adds GPUArray comparisons for == != = =
Merged, thanks!
Assuming you're cool with the