Hi
I am interested to know: are there are any LLVM targets in the works
for Nvidia's PTX ISA?
Also if anyone knows about Ocelot (a project done by some students at
my school): it does the opposite of what I am trying to do (translates
PTX to LLVM IR to run Cuda kernels on the CPU).
Thanks in advance.
-Puyan
I don't know of any, but that would be a great project if the nvidia toolchain is sufficiently available for it to be useful to open source folks.
-Chris
Hey,
Chris Lattner schrieb:
Hi
I am interested to know: are there are any LLVM targets in the works
for Nvidia's PTX ISA?
Also if anyone knows about Ocelot (a project done by some students at
my school): it does the opposite of what I am trying to do (translates
PTX to LLVM IR to run Cuda kernels on the CPU).
I don't know of any, but that would be a great project if the nvidia toolchain is sufficiently available for it to be useful to open source folks.
Helge Rhodin, a student at Saarland University, has already written a PTX backend for LLVM as part of his bachelor's thesis. We currently use it for a deferred shading system of a rasterizer.
The plan is to reintegrate it into the LLVM trunk, but we have not yet made any steps into this direction.
Best regards,
Ralf