sgemm_tcore: Support two accumulation reg tiles

This commit is contained in:
Hansung Kim
2024-06-01 01:12:08 -07:00
parent c8d6c56dd9
commit 4e723c4655

View File

@@ -27,10 +27,10 @@
// (BM*BN) / (TM*TN) == threadblock size >= NT * CORES_PER_CLUSTER // (BM*BN) / (TM*TN) == threadblock size >= NT * CORES_PER_CLUSTER
// * Combining BM * BK >= (BM*BN) / (TM*TN) == threadblock yields // * Combining BM * BK >= (BM*BN) / (TM*TN) == threadblock yields
// BM <= BK*TM*TN // BM <= BK*TM*TN
#define BM 8 #define BM 32
#define BN 8 #define BN 32
#define BK 8 #define BK 32
#define WM 8 #define WM 16
#define WN 8 #define WN 8
#define TCM 8 #define TCM 8
#define TCN 8 #define TCN 8
@@ -133,8 +133,12 @@ inline constexpr void map_c(const int tid, int &row, int &col) {
} }
} }
inline void vx_wmma() { inline void vx_wmma(const int dest_reg) {
asm volatile (".insn r %0, 0, 0, x0, x0, x0" :: "i"(RISCV_CUSTOM3)); if (dest_reg == 0) {
asm volatile (".insn r %0, 0, 0, x0, x0, x0" :: "i"(RISCV_CUSTOM3));
} else {
asm volatile (".insn r %0, 0, 0, x1, x0, x0" :: "i"(RISCV_CUSTOM3));
}
} }
// `local_k` is assumed to be multiple of TCK // `local_k` is assumed to be multiple of TCK
@@ -196,23 +200,35 @@ inline void vx_wmma_load(volatile float *smem_A, volatile float *smem_B, const i
asm volatile("flw f15, %0" ::"m"(smem_B[((local_k + 7) * smem_B_cols) + (WN * warp_col + TCN * wn_iter) + col])); asm volatile("flw f15, %0" ::"m"(smem_B[((local_k + 7) * smem_B_cols) + (WN * warp_col + TCN * wn_iter) + col]));
} }
inline void initialize_C() { inline void initialize_C(const int dest_reg) {
// initialize C to zeros // initialize C to zeros
asm volatile("fmv.w.x f16, x0"); if (dest_reg == 0) {
asm volatile("fmv.w.x f17, x0"); asm volatile("fmv.w.x f16, x0");
asm volatile("fmv.w.x f18, x0"); asm volatile("fmv.w.x f17, x0");
asm volatile("fmv.w.x f19, x0"); asm volatile("fmv.w.x f18, x0");
asm volatile("fmv.w.x f20, x0"); asm volatile("fmv.w.x f19, x0");
asm volatile("fmv.w.x f21, x0"); asm volatile("fmv.w.x f20, x0");
asm volatile("fmv.w.x f22, x0"); asm volatile("fmv.w.x f21, x0");
asm volatile("fmv.w.x f23, x0"); asm volatile("fmv.w.x f22, x0");
asm volatile("fmv.w.x f23, x0");
} else {
asm volatile("fmv.w.x f24, x0");
asm volatile("fmv.w.x f25, x0");
asm volatile("fmv.w.x f26, x0");
asm volatile("fmv.w.x f27, x0");
asm volatile("fmv.w.x f28, x0");
asm volatile("fmv.w.x f29, x0");
asm volatile("fmv.w.x f30, x0");
asm volatile("fmv.w.x f31, x0");
}
} }
inline void write_results(volatile float *local_warp_results, inline void write_results(volatile float *local_warp_results,
int thread_in_warp, int warp_col, int warp_row, const int thread_in_warp, const int warp_col,
int wn_iter, int wm_iter, int dim_m, int dim_n, const int warp_row, const int wn_iter,
float *C, int threadblock_id_x, const int wm_iter, const int dim_m, const int dim_n,
int threadblock_id_y) { float *C, const int threadblock_id_x,
const int threadblock_id_y) {
int tid = thread_in_warp; int tid = thread_in_warp;
int tg = tid / 4; int tg = tid / 4;
@@ -229,14 +245,25 @@ inline void write_results(volatile float *local_warp_results,
BN * threadblock_id_x; BN * threadblock_id_x;
// @perf: this likely causes a lot of gmem bank conflicts // @perf: this likely causes a lot of gmem bank conflicts
asm volatile ("fsw f16, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 0)])); if (wm_iter == 0) {
asm volatile ("fsw f17, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 1)])); asm volatile ("fsw f16, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 0)]));
asm volatile ("fsw f18, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 0)])); asm volatile ("fsw f17, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 1)]));
asm volatile ("fsw f19, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 1)])); asm volatile ("fsw f18, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 0)]));
asm volatile ("fsw f20, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 4)])); asm volatile ("fsw f19, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 1)]));
asm volatile ("fsw f21, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 5)])); asm volatile ("fsw f20, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 4)]));
asm volatile ("fsw f22, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 4)])); asm volatile ("fsw f21, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 5)]));
asm volatile ("fsw f23, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 5)])); asm volatile ("fsw f22, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 4)]));
asm volatile ("fsw f23, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 5)]));
} else {
asm volatile ("fsw f24, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 0)]));
asm volatile ("fsw f25, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 1)]));
asm volatile ("fsw f26, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 0)]));
asm volatile ("fsw f27, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 1)]));
asm volatile ("fsw f28, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 4)]));
asm volatile ("fsw f29, %0" :: "m"(global_offset_C[dim_n * (local_row + 0) + (local_col + 5)]));
asm volatile ("fsw f30, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 4)]));
asm volatile ("fsw f31, %0" :: "m"(global_offset_C[dim_n * (local_row + 2) + (local_col + 5)]));
}
} }
inline void threadblock_barrier(unsigned int tid_in_threadblock, inline void threadblock_barrier(unsigned int tid_in_threadblock,
@@ -349,7 +376,8 @@ void thread_block_gemm(kernel_arg_t *__UNIFORM__ arg,
local_b + local_b_elems + (warp_in_threadblock * TCM * TCN); local_b + local_b_elems + (warp_in_threadblock * TCM * TCN);
// clear out C // clear out C
initialize_C(); initialize_C(0);
initialize_C(1);
#pragma GCC unroll 1 #pragma GCC unroll 1
for (uint32_t k = 0; k < dim_k; k += BK) { for (uint32_t k = 0; k < dim_k; k += BK) {
@@ -394,7 +422,7 @@ void thread_block_gemm(kernel_arg_t *__UNIFORM__ arg,
vx_wmma_load(local_a, local_b, local_k, warp_col, warp_row, vx_wmma_load(local_a, local_b, local_k, warp_col, warp_row,
wn_iter, wm_iter, tid_in_warp); wn_iter, wm_iter, tid_in_warp);
// compute // compute
vx_wmma(); vx_wmma(wm_iter);
#if TC_SINGLE_WARP #if TC_SINGLE_WARP
} }
#endif #endif