p35189024
/
InfiniTensor
forked from jiuyuan/InfiniTensor
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

clean up

This commit is contained in:
xiaonans 2024-03-21 10:17:06 +08:00
parent fc3d38f80e
commit 0740d26f43
1 changed files with 39 additions and 27 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

66
src/kernels/cuda/attention_kvcache.cu
View File

@ -20,7 +20,7 @@ __global__ void _attention_kvcache_kernel_128_1(T* input_k_cache,
if(blockIdx.y >= stride)
return;
int lane_id_x4 = threadIdx.x % WARP_SIZE * 4;
int lane_id_x2 = threadIdx.x % WARP_SIZE * 2;
int group_id = threadIdx.x / WARP_SIZE;
int parallel_idx = blockIdx.x * (blockDim.x / WARP_SIZE) + group_id;
int idx_seq = blockIdx.y * SEQ_UNIT;
@ -38,64 +38,75 @@ __global__ void _attention_kvcache_kernel_128_1(T* input_k_cache,
float temp[4];
bool is_fp16 = sizeof(T) == 2 ? true : false;
int idx_qkv = lane_id_x2 + parallel_idx * compMeta.stride[2];
// readin Q
if(!is_fp16){
(float4 &)temp[0] = (float4 &)input_q[lane_id_x4 + (parallel_idx * 128)];
for(int i = 0; i < 4; i += 2){
(float2 &)temp[i] = (float2 &)input_q[idx_qkv + i*WARP_SIZE];
*((half2*)(&ptr_Q[i])) = __float22half2_rn(*((float2*)(&temp[i])));
}
}
else{
for(int i = 0; i < 4; i += 2){
(half2 &)ptr_Q[i] = (half2 &)input_q[lane_id_x4 + i + (parallel_idx * 128)];
(half2 &)ptr_Q[i] = (half2 &)input_q[idx_qkv + i*WARP_SIZE];
}
}
int common_idx = lane_id_x4 + (parallel_idx * compMeta.stride[1]);
int idx_kv = lane_id_x4 + parallel_idx * compMeta.stride[2];
int common_idx = lane_id_x2 + (parallel_idx * compMeta.stride[1]);
// Q*K
#pragma unroll
for (int idx_SEQ_UNIT = 0; idx_SEQ_UNIT < SEQ_UNIT && idx_SEQ_UNIT + idx_seq < seq_length; idx_SEQ_UNIT ++) {
int idx_kvcache = common_idx + ((idx_SEQ_UNIT + idx_seq) * compMeta.stride[2]);
if(idx_SEQ_UNIT + idx_seq < seq_length - 1){
*((int2*)(&ptr_K[0])) = *((int2*)(&((half*)input_k_cache)[idx_kvcache]));
if(idx_SEQ_UNIT + idx_seq < seq_length - 1){
for(int i = 0; i < 4; i += 2){
*((half2*)(&ptr_K[i])) = *((half2*)(&((half*)input_k_cache)[idx_kvcache + i*WARP_SIZE]));
}
}
else{
if(!is_fp16){
(float4 &)temp[0] = (float4 &) input_k[idx_kv];
for(int i = 0; i < 4; i += 2)
for(int i = 0; i < 4; i += 2){
(float2 &)temp[i] = (float2 &) input_k[idx_qkv + i*WARP_SIZE];
*((half2*)(&ptr_K[i])) = __float22half2_rn(*((float2*)(&temp[i])));
}
}
else{
for(int i = 0; i < 4; i += 2){
(half2 &)ptr_K[i] = (half2 &)input_k[lane_id_x4 + i + (parallel_idx * 128)];
(half2 &)ptr_K[i] = (half2 &)input_k[idx_qkv + i*WARP_SIZE];
}
}
*((int2*)(&((half*)input_k_cache)[idx_kvcache])) = *((int2*)(&ptr_K[0]));
for(int i = 0; i < 4; i += 2){
*((half2*)(&((half*)input_k_cache)[idx_kvcache + i*WARP_SIZE])) = *((half2*)(&ptr_K[i]));
}
}
// * V
if(idx_SEQ_UNIT + idx_seq < seq_length - 1){
*((int2*)(&ptr_V[0])) = *((int2*)(&((half*)input_v_cache)[idx_kvcache]));
if(idx_SEQ_UNIT + idx_seq < seq_length - 1){
for(int i = 0; i < 4; i += 2){
*((half2*)(&ptr_V[i])) = *((half2*)(&((half*)input_v_cache)[idx_kvcache + i*WARP_SIZE]));
}
}
else{
if(!is_fp16){
(float4 &)temp[0] = (float4 &) input_v[idx_kv];
for(int i = 0; i < 4; i += 2)
for(int i = 0; i < 4; i += 2){
(float2 &)temp[i] = (float2 &) input_v[idx_qkv + i*WARP_SIZE];
*((half2*)(&ptr_V[i])) = __float22half2_rn(*((float2*)(&temp[i])));
}
}
else{
for(int i = 0; i < 4; i += 2){
(half2 &)ptr_V[i] = (half2 &)input_v[lane_id_x4 + i + (parallel_idx * 128)];
(half2 &)ptr_V[i] = (half2 &)input_v[idx_qkv + i*WARP_SIZE];
}
}
*((int2*)(&((half*)input_v_cache)[idx_kvcache])) = *((int2*)(&ptr_V[0]));
for(int i = 0; i < 4; i += 2){
*((half2*)(&((half*)input_v_cache)[idx_kvcache + i*WARP_SIZE])) = *((half2*)(&ptr_V[i]));
}
}
#pragma unroll
for (int i = 0; i < 4; i ++){
ptr_K[i] = ptr_Q[i] * ptr_K[i];
#pragma unroll
for (int offset = 16; offset > 0; offset /= 2) {
for (int offset = WARP_SIZE/2; offset > 0; offset /= 2) {
ptr_K[i] += __shfl_down_sync(0xffffffff, ptr_K[i], offset);
}
ptr_P[idx_SEQ_UNIT] += __half2float(ptr_K[i]);
@ -119,8 +130,8 @@ __global__ void _attention_kvcache_kernel_128_1(T* input_k_cache,
ptr_O[i] /= ptr_sum[0];
for(int i = 0; i < 4; i += 2)
(half2 &)output_O_temp[(lane_id_x4 + i) + (blockIdx.y * compMeta.dimSize[3]) + (parallel_idx * compMeta.dimSize[3] * stride)] = __float22half2_rn((float2 &)ptr_O[i]);
if(lane_id_x4 == 0){
(half2 &)output_O_temp[(lane_id_x2 + i*WARP_SIZE) + (blockIdx.y * compMeta.dimSize[3]) + (parallel_idx * compMeta.dimSize[3] * stride)] = __float22half2_rn((float2 &)ptr_O[i]);
if(lane_id_x2 == 0){
output_sum_temp[blockIdx.y + parallel_idx * stride] = ptr_sum[0];
}
@ -148,7 +159,7 @@ __global__ void _attention_kvcache_kernel_128_2(int* position_id,
for(int i = 0; i < size; i ++){
for(int j = 0; j < 4; j += 2)
(float2 &)ptr_O[j]
= __half22float2((half2 &)output_O_temp[(lane_id_x2 + j*32) + (i * compMeta.dimSize[3]) + parallel_idx * compMeta.dimSize[3] * size]);
= __half22float2((half2 &)output_O_temp[(lane_id_x2 + j*WARP_SIZE) + (i * compMeta.dimSize[3]) + parallel_idx * compMeta.dimSize[3] * size]);
ptr_sum_temp = output_sum_temp[i + parallel_idx * size];
#pragma unroll
@ -163,23 +174,23 @@ __global__ void _attention_kvcache_kernel_128_2(int* position_id,
if(!is_fp16){
for(int j = 0; j < 4; j += 2)
(float2 &)output_matmul[lane_id_x2 + j*32 + (parallel_idx * compMeta.dimSize[3])] = (float2 &)ptr_O_sum[j];
(float2 &)output_matmul[lane_id_x2 + j*WARP_SIZE + (parallel_idx * compMeta.dimSize[3])] = (float2 &)ptr_O_sum[j];
}
else{
for(int j = 0; j < 4; j += 2)
(half2 &)output_matmul[lane_id_x2 + j*32 + (parallel_idx * compMeta.dimSize[3])] = __float22half2_rn((float2 &)ptr_O_sum[j]);
(half2 &)output_matmul[lane_id_x2 + j*WARP_SIZE + (parallel_idx * compMeta.dimSize[3])] = __float22half2_rn((float2 &)ptr_O_sum[j]);
}
}
else{
if(!is_fp16){
for(int i = 0; i < 4; i += 2)
(float2 &)output_matmul[(lane_id_x2 + i*32) + (parallel_idx * compMeta.dimSize[3])]
= __half22float2((half2 &)output_O_temp[(lane_id_x2 + i*32) + parallel_idx * compMeta.dimSize[3]]);
(float2 &)output_matmul[(lane_id_x2 + i*WARP_SIZE) + (parallel_idx * compMeta.dimSize[3])]
= __half22float2((half2 &)output_O_temp[(lane_id_x2 + i*WARP_SIZE) + parallel_idx * compMeta.dimSize[3]]);
}
else{
for(int i = 0; i < 4; i += 2)
(half2 &)output_matmul[(lane_id_x2 + i*32) + (parallel_idx * compMeta.dimSize[3])]
= (half2 &)output_O_temp[(lane_id_x2 + i*32) + parallel_idx * compMeta.dimSize[3]];
(half2 &)output_matmul[(lane_id_x2 + i*WARP_SIZE) + (parallel_idx * compMeta.dimSize[3])]
= (half2 &)output_O_temp[(lane_id_x2 + i*WARP_SIZE) + parallel_idx * compMeta.dimSize[3]];
}
}
}
@ -219,6 +230,7 @@ void attention_kvcache_kernel(int dType, void *input_k_cache, void *input_v_cach
0, CUDAStream::getCurrentStream()>>>
(position_id, (half*)output_matmul, compMeta, (half*)output_O_temp, output_sum_temp);
}
}
} // namespace infini