Skip to content

Remove Cuda toolkit version check #8401

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
esc merged 3 commits into from
Sep 7, 2022
Merged

Remove Cuda toolkit version check #8401

Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
4dcb59e
Remove Cuda toolkit version check
testhound Aug 29, 2022
12e528c
Remove get_version check in habs unit test
testhound Sep 2, 2022
46e13ab
Merge branch 'main' into testhound/remove_get_version_check
testhound Sep 2, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 1 addition & 11 deletions numba/cuda/cudaimpl.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -502,18 +502,8 @@ def ptx_fp16_hneg(context, builder, sig, args):

@lower(stubs.fp16.habs, types.float16)
def ptx_fp16_habs(context, builder, sig, args):
if cuda.runtime.get_version() < (10, 2):
# CUDA < 10.2 does not support abs.f16. For these versions, we mask
# off the sign bit to compute abs instead. We determine whether or
# not to do this based on the runtime version so that our behaviour
# is consistent with the version of NVVM we're using to go from
# NVVM IR -> PTX.
inst = 'and.b16 $0, $1, 0x7FFF;'
else:
inst = 'abs.f16 $0, $1;'

fnty = ir.FunctionType(ir.IntType(16), [ir.IntType(16)])
asm = ir.InlineAsm(fnty, inst, '=h,h')
asm = ir.InlineAsm(fnty, 'abs.f16 $0, $1;', '=h,h')
return builder.call(asm, args)


Expand Down