Skip to content

Do only one iteration of newton-raphson refinement for FP16 inv #1185

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 1 commit into from
Oct 3, 2025
Merged

Do only one iteration of newton-raphson refinement for FP16 inv #1185

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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: 0 additions & 2 deletions src/core/NEON/NEMath.inl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -634,15 +634,13 @@ inline float16x4_t vinv_f16(float16x4_t x)
{
float16x4_t recip = vrecpe_f16(x);
recip = vmul_f16(vrecps_f16(x, recip), recip);
recip = vmul_f16(vrecps_f16(x, recip), recip);
return recip;
}

inline float16x8_t vinvq_f16(float16x8_t x)
{
float16x8_t recip = vrecpeq_f16(x);
recip = vmulq_f16(vrecpsq_f16(x, recip), recip);
recip = vmulq_f16(vrecpsq_f16(x, recip), recip);
return recip;
}

Expand Down
1 change: 0 additions & 1 deletion src/core/NEON/SVEMath.inl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,7 +82,6 @@ inline svfloat16_t svinv_f16_z(svbool_t pg, svfloat16_t x)
{
auto recip = svrecpe_f16(x);
recip = svmul_f16_z(pg, svrecps_f16(x, recip), recip);
recip = svmul_f16_z(pg, svrecps_f16(x, recip), recip);
return recip;
}

Expand Down