fnmadd Class — pytorch Architecture
Architecture documentation for the fnmadd class in vec128_bfloat16_neon.h from the pytorch codebase.
Entity Profile
Source Code
aten/src/ATen/cpu/vec/vec128/vec128_bfloat16_neon.h lines 647–661
template <>
Vectorized<c10::BFloat16> inline fnmadd(
const Vectorized<c10::BFloat16>& a,
const Vectorized<c10::BFloat16>& b,
const Vectorized<c10::BFloat16>& c) {
#if BF16_ARITHMETIC_SUPPORTED()
bfloat16x8_t x = a;
bfloat16x8_t y = b;
bfloat16x8_t z = c;
return (-x) * y + z;
#else
// See NOTE [BF16 FMA] above.
return -a * b + c;
#endif
}
Source
Analyze Your Own Codebase
Get architecture documentation, dependency graphs, and domain analysis for your codebase in minutes.
Try Supermodel Free