MCPcopy Create free account
hub / github.com/ARM-software/ComputeLibrary / exponential_bf16

Function exponential_bf16

src/core/helpers/LUTManager.cpp:69–74  ·  view source on GitHub ↗

Read bf16 value as u16, convert to fp32. Calculate exp in fp32, return as bf16

Source from the content-addressed store, hash-verified

67// Read bf16 value as u16, convert to fp32.
68// Calculate exp in fp32, return as bf16
69inline uint16_t exponential_bf16(uint16_t x, const LUTInfo &info)
70{
71 float fp = bf16_to_float(x);
72 fp = exponential(fp, info);
73 return float_to_bf16(fp);
74}
75
76void init_lut(LookupTable256 &lut, const LUTInfo &info)
77{

Callers 1

init_lutFunction · 0.85

Calls 3

bf16_to_floatFunction · 0.85
exponentialFunction · 0.85
float_to_bf16Function · 0.85

Tested by

no test coverage detected