Skip to content

Prepare 0.3 release #10

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
merged 26 commits into from
Apr 11, 2025
Merged

Prepare 0.3 release #10

Changes from 1 commit
Commits
Show all changes
26 commits
Select commit Hold shift + click to select a range
5490ea7
Make code compile with HIP on Lumi
stijnh Nov 1, 2024
9385655
Changes to make code compile under HIPRTC
stijnh Nov 18, 2024
de62ad0
Merge branch 'hip-support' into dev
stijnh Nov 18, 2024
4fca9cb
Add approximation functions
stijnh Nov 18, 2024
ae0e6b1
Simplify how policies are implemented internally
stijnh Nov 18, 2024
f89cf98
Rename FP16 primitive names: `__half` to `half_t` and `__nv_bfloat16`…
stijnh Nov 18, 2024
014e32f
Implement approximation for `pow`
stijnh Nov 18, 2024
003ce36
Add `apply_fallback_impl` struct
stijnh Nov 18, 2024
76501fd
Add `approx_*` functions
stijnh Nov 18, 2024
ba7356a
Fix incorrect definition of `KERNEL_FLOAT_POLICY`
stijnh Nov 20, 2024
4231f44
Add `Accuracy` parameter to `zip_common`
stijnh Nov 20, 2024
f5edbc8
Add `add_mul` to `vector`
stijnh Nov 20, 2024
e6c8a7c
Overwrite `fast_policy` for FP16 and BF16
stijnh Nov 20, 2024
5c859b9
`kernel_float::approx::sqrt(0)` now returns 0
stijnh Nov 20, 2024
76c695a
Fix compilation error on HIP due to `KERNEL_FLOAT_FAST_F32_MAP`
stijnh Nov 26, 2024
d8a53a3
Remove call to `__exp2f` since it does not exist
stijnh Nov 26, 2024
a2b08a5
Change github workflow to compile for all architectures
stijnh Nov 26, 2024
846de1f
Fix bug in `approx::exp(bfloat16)` for HIP
stijnh Dec 2, 2024
f94bd10
Fix several issues related to HIP compilation for bfloat16
stijnh Dec 2, 2024
2730789
`vector_ptr` now requires alignment in bytes instead of elements
stijnh Jan 27, 2025
1611258
Remove `apply_fallback_impl`
stijnh Jan 27, 2025
c44c6ed
Fix incorrect type name in `approx.h`
stijnh Jan 27, 2025
212efee
Fix incorrect type name in `binops.h`
stijnh Jan 27, 2025
09dc820
Change `AssignConversionProxy` to also accept rvalues
stijnh Jan 27, 2025
126737c
Change `ops::cast` to get rid of `cast_float_fallback`
stijnh Jan 27, 2025
4d18563
Update `Jimver/cuda-toolkit` workflow action to newer version
stijnh Apr 11, 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
Prev Previous commit
Next Next commit
Fix incorrect type name in approx.h
  • Loading branch information
@stijnh
stijnh committed Jan 27, 2025
commit c44c6ed2b3c8cb8e6040c1e286b7272736c86387
4 changes: 2 additions & 2 deletions include/kernel_float/approx.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -304,13 +304,13 @@ KERNEL_FLOAT_DEVICE bfloat16x2_t normalize_trig_input(bfloat16x2_t x) {
template<int Iter>
KERNEL_FLOAT_DEVICE bfloat16x2_t cos(bfloat16x2_t x) {
bfloat16x2_t xf = normalize_trig_input(x);
return cos_poly<__bfloat16, Iter + 1>::call(__hmul2(xf, xf));
return cos_poly<bfloat16_t, Iter + 1>::call(__hmul2(xf, xf));
}

template<int Iter>
KERNEL_FLOAT_DEVICE bfloat16x2_t sin(bfloat16x2_t x) {
bfloat16x2_t xf = normalize_trig_input(x);
return __hmul2(sin_poly<__bfloat16, Iter>::call(__hmul2(xf, xf)), xf);
return __hmul2(sin_poly<bfloat16_t, Iter>::call(__hmul2(xf, xf)), xf);
}

template<int Iter>
Expand Down