MCPcopy Create free account
hub / github.com/ARM-software/astc-encoder / int_as_float

Function int_as_float

Source/astcenc_vecmathlib_avx2_8.h:1013–1016  ·  view source on GitHub ↗

* @brief Return a integer value as a float bit pattern (i.e. no conversion). * * It is a common trick to convert floats into integer bit patterns, perform * some bit hackery based on knowledge they are IEEE 754 layout, and then * convert them back again. This is the second half of that flip. */

Source from the content-addressed store, hash-verified

1011 * convert them back again. This is the second half of that flip.
1012 */
1013ASTCENC_SIMD_INLINE vfloat8 int_as_float(vint8 a)
1014{
1015 return vfloat8(_mm256_castsi256_ps(a.m));
1016}
1017
1018/*
1019 * Table structure for a 16x 8-bit entry table.

Callers

nothing calls this directly

Calls 1

vfloat8Class · 0.70

Tested by

no test coverage detected