Added bwd data v3r1 v4r1, tweaking v1 (#10)

* Added bwd data v3r1: breaking down compute into a series of load balanced GEMM, and launch in a single kernel
* Added bwd data v4r1: like v3r1, but launch GEMMs in multiple kernels
* Tweaked v1r1  and v1r2 (atomic) on AMD GPU
This commit is contained in:
Chao Liu
2020-01-20 10:20:03 -06:00
committed by GitHub
parent e2b4c5b469
commit c5da0377fb
43 changed files with 2123 additions and 452 deletions

View File

@@ -49,6 +49,12 @@ struct integer_divide_ceiler
}
};
template <class X, class Y>
__host__ __device__ constexpr auto integer_divide_floor(X x, Y y)
{
return x / y;
}
template <class X, class Y>
__host__ __device__ constexpr auto integer_divide_ceil(X x, Y y)
{