From here, I can find about Half2 Arithmetic Functions, such as __host__ __device__ __half2__hsub2(const __half2 a, const __half2 b)
.
But in FP8 Intrinsics, I can’t find the function about fp8 vector subtraction calculation.
Is there such a function?
Related topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
Is there any documentation about FP8 arithmetic functions? | 4 | 270 | June 28, 2024 | |
Half precision and int8 computations on cuda cores | 0 | 597 | August 28, 2019 | |
Half2 performance | 4 | 2636 | October 29, 2018 | |
Half2 vs. Half datatype | 3 | 3418 | July 1, 2022 | |
What is the point of the half2 datatype? | 2 | 371 | September 4, 2023 | |
Vector maths on float2, where are the SIMD functions? | 4 | 3227 | July 9, 2018 | |
CUDNN half2 convolution | 1 | 756 | November 11, 2021 | |
Kernel function documentation? | 0 | 941 | July 19, 2021 | |
fp16 <-> fp32 | 2 | 9102 | July 23, 2009 | |
How to cuda half and half functions | 5 | 4043 | January 10, 2019 |