std :: ml::device
Struct std::DevGemm
struct DevGemm {
int dtype;
int ctype;
int ta;
int tb;
unsigned long m;
unsigned long n;
unsigned long k;
float alpha;
unsigned long a;
unsigned long lda;
unsigned long sa;
unsigned long b;
unsigned long ldb;
unsigned long sb;
float beta;
unsigned long c;
unsigned long ldc;
unsigned long sc;
unsigned long batch;
}A strided batch of GEMMs: for i < batch, c_i = alpha op(a_i) op(b_i) + beta c_i with x_i = x + i * sx (strides in elements), as gemm(). dtype (DEVICE_FP32 / FP16 / BF16) is a and b's element type; c is fp32 and accumulates in fp32.
Fields
a and b's element type
c's (DEVICE_FP32 / FP16 / BF16)