Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

CUDA: Implemented row flattening for non-glm RoPE #2468

Merged
merged 1 commit into from
Jul 31, 2023
Merged

CUDA: Implemented row flattening for non-glm RoPE #2468

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 15 additions & 8 deletions ggml-cuda.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2790,7 +2790,8 @@ static __global__ void cpy_f32_f16(const char * cx, char * cdst, const int ne,
}

// rope == RoPE == rotary positional embedding
static __global__ void rope_f32(const float * x, float * dst, const int ncols, const float p, const float theta_scale) {
static __global__ void rope_f32(const float * x, float * dst, const int ncols, const float p0,
const float p_delta, const int p_delta_rows, const float theta_scale) {
const int col = 2*(blockDim.x*blockIdx.x + threadIdx.x);

if (col >= ncols) {
Expand All @@ -2800,7 +2801,7 @@ static __global__ void rope_f32(const float * x, float * dst, const int ncols, c
const int row = blockDim.y*blockIdx.y + threadIdx.y;
const int i = row*ncols + col;

const float theta = p*powf(theta_scale, col/2);
const float theta = (p0 + p_delta * (row/p_delta_rows))*powf(theta_scale, col/2);
const float sin_theta = sinf(theta);
const float cos_theta = cosf(theta);

Expand Down Expand Up @@ -3404,12 +3405,13 @@ static void scale_f32_cuda(const float * x, float * dst, const float scale, cons
scale_f32<<<num_blocks, CUDA_SCALE_BLOCK_SIZE, 0, stream>>>(x, dst, scale, k);
}

static void rope_f32_cuda(const float * x, float * dst, const int ncols, const int nrows, const float p, const float theta_scale, cudaStream_t stream) {
static void rope_f32_cuda(const float * x, float * dst, const int ncols, const int nrows, const float p0,
const float p_delta, const int p_delta_rows, const float theta_scale, cudaStream_t stream) {
GGML_ASSERT(nrows % 2 == 0);
const dim3 block_dims(2*CUDA_ROPE_BLOCK_SIZE, 1, 1);
const int num_blocks_x = (ncols + 2*CUDA_ROPE_BLOCK_SIZE - 1) / (2*CUDA_ROPE_BLOCK_SIZE);
const dim3 block_nums(num_blocks_x, nrows, 1);
rope_f32<<<block_nums, block_dims, 0, stream>>>(x, dst, ncols, p, theta_scale);
rope_f32<<<block_nums, block_dims, 0, stream>>>(x, dst, ncols, p0, p_delta, p_delta_rows, theta_scale);
}

static void rope_glm_f32_cuda(const float * x, float * dst, const int ncols, const int nrows, const float p, const float block_p, const float theta_scale, cudaStream_t stream) {
Expand Down Expand Up @@ -4105,6 +4107,7 @@ inline void ggml_cuda_op_rope(
GGML_ASSERT(dst_ddf_i != nullptr);

const int64_t ne00 = src0->ne[0];
const int64_t ne01 = src0->ne[1];
const int64_t i01_diff = i01_high - i01_low;

const int n_past = ((int32_t *) dst->op_params)[0];
Expand All @@ -4118,17 +4121,18 @@ inline void ggml_cuda_op_rope(
memcpy(&freq_scale, (int32_t *) dst->op_params + 5, sizeof(float));

const float theta_scale = powf(freq_base, -2.0f/n_dims);
const float p = (((mode & 1) == 0 ? n_past + i02 : i02)) * freq_scale;

bool is_glm = mode & 4;
const bool is_glm = mode & 4;

// compute
if (is_glm) {
const float p = (((mode & 1) == 0 ? n_past + i02 : i02)) * freq_scale;
const float id_p = min(p, n_ctx - 2.f);
const float block_p = max(p - (n_ctx - 2.f), 0.f);
rope_glm_f32_cuda(src0_ddf_i, dst_ddf_i, ne00, i01_diff, id_p, block_p, theta_scale, cudaStream_main);
} else {
rope_f32_cuda(src0_ddf_i, dst_ddf_i, ne00, i01_diff, p, theta_scale, cudaStream_main);
const float p0 = (((mode & 1) == 0 ? n_past : 0)) * freq_scale;
rope_f32_cuda(src0_ddf_i, dst_ddf_i, ne00, i01_diff, p0, freq_scale, ne01, theta_scale, cudaStream_main);
}

(void) src1;
Expand Down Expand Up @@ -4743,7 +4747,10 @@ void ggml_cuda_soft_max(const ggml_tensor * src0, const ggml_tensor * src1, ggml

void ggml_cuda_rope(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) {
GGML_ASSERT(src0->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32);
ggml_cuda_op(src0, src1, dst, ggml_cuda_op_rope, true, false); // FIXME flatten changes results

const int mode = ((int32_t *) dst->op_params)[2];
const bool is_glm = mode & 4;
ggml_cuda_op(src0, src1, dst, ggml_cuda_op_rope, true, !is_glm); // flatten support not implemented for glm
}

void ggml_cuda_nop(const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) {
Expand Down