Skip to content
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

Try to use unsigned types from pytorch #3280

Closed
wants to merge 2 commits into from
Closed

Try to use unsigned types from pytorch #3280

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
42dd180
Try to use unsigned types from pytorch
anmyachev Jan 27, 2025
e6772d8
just try
anmyachev Jan 30, 2025
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
2 changes: 1 addition & 1 deletion lib/Conversion/TritonGPUToLLVM/Utility.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,7 @@
static int __builtin_clz(unsigned x) {
unsigned long r;
_BitScanReverse(&r, x);
return static_cast<int>(r);
return static_cast<int>(r ^ 31);
}

static int __builtin_ctz(unsigned x) {
Expand Down
16 changes: 5 additions & 11 deletions python/triton/_internal_testing.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -110,17 +110,11 @@ def to_triton(x: np.ndarray, device, dst_type=None) -> Union[TensorWrapper, torc
For example: x is of type `float32`, dst_type is `bfloat16`.
If dst_type is None, we infer dst_type from x.
'''
t = x.dtype.name
if t in uint_dtypes:
signed_type_name = t.lstrip('u') # e.g. "uint16" -> "int16"
x_signed = x.astype(getattr(np, signed_type_name))
return reinterpret(torch.tensor(x_signed, device=device), getattr(tl, t))
else:
if dst_type and 'float8' in dst_type:
return reinterpret(torch.tensor(x, device=device), getattr(tl, dst_type))
if t == 'float32' and dst_type == 'bfloat16':
return torch.tensor(x, device=device).bfloat16()
return torch.tensor(x, device=device)
if dst_type and 'float8' in dst_type:
return reinterpret(torch.tensor(x, device=device), getattr(tl, dst_type))
if x.dtype.name == 'float32' and dst_type == 'bfloat16':
return torch.tensor(x, device=device).bfloat16()
return torch.tensor(x, device=device)


def str_to_triton_dtype(x: str) -> tl.dtype:
Expand Down
Loading