archive/jdk
1
0
Fork 0
mirror of https://github.com/openjdk/jdk.git synced 2025-09-19 02:24:40 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

8289552: Make intrinsic conversions between bit representations of half precision values and floats

Browse source 
Reviewed-by: kvn, sviswanathan, jbhateja
This commit is contained in:
Smita Kamath 2022-10-11 17:06:19 +00:00 • committed by Sandhya Viswanathan
parent 2586b1a3c1
commit 07946aa49c
19 changed files with 345 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

5
src/hotspot/cpu/x86/assembler_x86.hpp
View file

@ -1156,6 +1156,11 @@ private:
void cvtdq2pd(XMMRegister dst, XMMRegister src);
void vcvtdq2pd(XMMRegister dst, XMMRegister src, int vector_len);
// Convert Halffloat to Single Precision Floating-Point value
void vcvtps2ph(XMMRegister dst, XMMRegister src, int imm8, int vector_len);
void vcvtph2ps(XMMRegister dst, XMMRegister src, int vector_len);
void evcvtps2ph(Address dst, KRegister mask, XMMRegister src, int imm8, int vector_len);
// Convert Packed Signed Doubleword Integers to Packed Single-Precision Floating-Point Value
void cvtdq2ps(XMMRegister dst, XMMRegister src);
void vcvtdq2ps(XMMRegister dst, XMMRegister src, int vector_len);