fp16c

Documentation

Overview

THESE PACKAGES ARE FOR DEMONSTRATION PURPOSES ONLY!

THEY DO NOT NOT CONTAIN WORKING INTRINSICS!

See https://github.com/klauspost/intrinsics

Index

• func CvtphPs(a x86.M128i) (dst x86.M128)
• func CvtpsPh(a x86.M128, rounding int) (dst x86.M128i)
• func M256CvtphPs(a x86.M128i) (dst x86.M256)
• func M256CvtpsPh(a x86.M256, rounding int) (dst x86.M128i)

Functions

func CvtphPs

func CvtphPs(a x86.M128i) (dst x86.M128)

CvtphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'.

FOR j := 0 to 3
	i := j*32
	m := j*16
	dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
ENDFOR
dst[MAX:128] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm_cvtph_ps'. Requires FP16C.

func CvtpsPh

func CvtpsPh(a x86.M128, rounding int) (dst x86.M128i)

CvtpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst'.

Rounding is done according to the 'rounding' parameter, which can be one

of:

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 3
			i := 16*j
			l := 32*j
			dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm_cvtps_ph'. Requires FP16C.

func M256CvtphPs

func M256CvtphPs(a x86.M128i) (dst x86.M256)

M256CvtphPs: Convert packed half-precision (16-bit) floating-point elements in 'a' to packed single-precision (32-bit) floating-point elements, and store the results in 'dst'.

FOR j := 0 to 7
	i := j*32
	m := j*16
	dst[i+31:i] := Convert_FP16_To_FP32(a[m+15:m])
ENDFOR
dst[MAX:256] := 0

Instruction: 'VCVTPH2PS'. Intrinsic: '_mm256_cvtph_ps'. Requires FP16C.

func M256CvtpsPh

func M256CvtpsPh(a x86.M256, rounding int) (dst x86.M128i)

M256CvtpsPh: Convert packed single-precision (32-bit) floating-point elements in 'a' to packed half-precision (16-bit) floating-point elements, and store the results in 'dst'.

Rounding is done according to the 'rounding' parameter, which can be one

of:

    (_MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC) // round to nearest, and suppress exceptions
    (_MM_FROUND_TO_NEG_INF |_MM_FROUND_NO_EXC)     // round down, and suppress exceptions
    (_MM_FROUND_TO_POS_INF |_MM_FROUND_NO_EXC)     // round up, and suppress exceptions
    (_MM_FROUND_TO_ZERO |_MM_FROUND_NO_EXC)        // truncate, and suppress exceptions
    _MM_FROUND_CUR_DIRECTION // use MXCSR.RC; see _MM_SET_ROUNDING_MODE

		FOR j := 0 to 7
			i := 16*j
			l := 32*j
			dst[i+15:i] := Convert_FP32_To_FP16FP(a[l+31:l])
		ENDFOR
		dst[MAX:128] := 0

Instruction: 'VCVTPS2PH'. Intrinsic: '_mm256_cvtps_ph'. Requires FP16C.