Fix cuda.is_available when cuda libs aren't present
#234
Conversation
|
cc @rwgk for vis |
Going to merge this without the test. This will be replaced by the |
Got it- thanks, @kkraus14 . |
I'd use mocking only as a last resort in this context. I've seen too many ways that can go horribly wrong. (One way I think about it: often creates about as many problems as you're hoping to avoid, if you're lucky.) Systematic testing against real environments is much better, and I think it's feasible (for the path_finder). (Hopefully not famous last words.) |
- Fix Invalid NVVM IR emitted when lowering shfl_sync APIs (NVIDIA#231) - Add Bfloat16 Low++ Bindings (NVIDIA#166) - Fix cuda.jit decorator inline (NVIDIA#181) - Feature: cuda specific make_attribute_wrapper (NVIDIA#193) - return a none tuple if no libdevice path is found (NVIDIA#234)
- Local variable debug info deduplication (NVIDIA#222) - Fix package installation for wheels CI (NVIDIA#238) - Fix Invalid NVVM IR emitted when lowering shfl_sync APIs (NVIDIA#231) - Add Bfloat16 Low++ Bindings (NVIDIA#166) - Fix cuda.jit decorator inline (NVIDIA#181) - Feature: cuda specific make_attribute_wrapper (NVIDIA#193) - return a none tuple if no libdevice path is found (NVIDIA#234)
- Local variable debug info deduplication (#222) - Fix package installation for wheels CI (#238) - Fix Invalid NVVM IR emitted when lowering shfl_sync APIs (#231) - Add Bfloat16 Low++ Bindings (#166) - Fix cuda.jit decorator inline (#181) - Feature: cuda specific make_attribute_wrapper (#193) - return a none tuple if no libdevice path is found (#234)
4 participants
Closes #233.
This could use some kind of test, perhaps one that patches out the finder functions to return None just for the duration of that test.