happyz
/
gemma.cpp
mirror of https://github.com/google/gemma.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Internal changes 

PiperOrigin-RevId: 836654012
This commit is contained in:
Krzysztof Rymski 2025-11-25 07:05:19 -08:00 committed by Copybara-Service
parent c153d5255b
commit b31e8f98e8
3 changed files with 36 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
gemma/flash_attention.cc
View File

@ -17,7 +17,9 @@
#include <stdint.h>
#include <algorithm>
#include <array>
#include <cmath>
#include <cstdlib>
#include <limits>
#include "compression/types.h" // GEMMA_DISABLED_TARGETS

65
gemma/flash_attention.h
View File

@ -29,38 +29,39 @@
namespace gcpp {
// Passed to HWY_VISIT_TARGETS; declares for one target.
#define GEMMA_DECL_FLASH_ATTENTION(TARGET, NAMESPACE) \
namespace NAMESPACE { \
void RMSNormAndPositionalEncoding( \
size_t num_tokens, const QBatch& qbatch, MatPtrT<float>& q, \
const MatPtr& query_norm_scale, size_t layer_idx, \
const AttentionActivationsPtrs& activations, ThreadingContext& ctx); \
\
void SingleFlashAttention(size_t start_pos, size_t last_pos, \
const float* HWY_RESTRICT q, \
const MatPtrT<KV_t>& k, const MatPtrT<KV_t>& v, \
size_t layer_idx, \
const AttentionActivationsPtrs& activations, \
float* HWY_RESTRICT att_out, \
ThreadingContext& ctx, size_t worker); \
\
Tile4FlashState TileFlashAttention4( \
const MatPtrT<BF16>& q, const uint32_t* HWY_RESTRICT q_offsets, \
const MatPtrT<KV_t>& k, size_t start_pos, \
const uint32_t* HWY_RESTRICT last_pos, size_t min_last_pos, \
size_t max_last_pos, const MatPtrT<KV_t>& v, size_t layer_idx, \
const LayerWeightsPtrs& layer, const AttentionActivations& activations, \
MatPtrT<float>& att_out, const uint32_t* HWY_RESTRICT out_offsets, \
ThreadingContext& ctx, const size_t worker); \
\
size_t GetVTileSize(size_t kNF, size_t num_head_groups, size_t num_tokens, \
size_t total_tasks, size_t target_parallelism); \
\
void FlashAttention(size_t num_tokens, size_t target_parallelism, \
size_t layer_idx, const MatPtr& query_norm_scale, \
AttentionActivationsPtrs& activations, QBatch& qbatch, \
ThreadingContext& ctx); \
/* NOLINTNEXTLINE(google-readability-namespace-comments) */ \
#define GEMMA_DECL_FLASH_ATTENTION(TARGET, NAMESPACE) \
namespace NAMESPACE { \
void RMSNormAndPositionalEncoding( \
size_t num_tokens, const QBatch& qbatch, MatPtrT<float>& q, \
const MatPtr& query_norm_scale, size_t layer_idx, \
const AttentionActivationsPtrs& activations, ThreadingContext& ctx); \
\
void SingleFlashAttention(size_t start_pos, size_t last_pos, \
const BF16* HWY_RESTRICT q, \
const MatPtrT<KV_t>& k, const MatPtrT<KV_t>& v, \
size_t layer_idx, \
const AttentionActivationsPtrs& activations, \
float* HWY_RESTRICT att_out, \
ThreadingContext& ctx, size_t worker); \
\
Tile4FlashState TileFlashAttention4( \
const MatPtrT<BF16>& q, const uint32_t* HWY_RESTRICT q_offsets, \
const MatPtrT<KV_t>& k, size_t start_pos, \
const uint32_t* HWY_RESTRICT last_pos, size_t min_last_pos, \
size_t max_last_pos, const MatPtrT<KV_t>& v, size_t layer_idx, \
const LayerWeightsPtrs& layer, const AttentionActivations& activations, \
MatPtrT<float>& att_out, const uint32_t* HWY_RESTRICT out_offsets, \
ThreadingContext& ctx, const size_t worker); \
\
size_t GetVTileSize(size_t kNF, size_t num_head_groups, size_t num_tokens, \
size_t total_tasks, size_t target_parallelism); \
\
void FlashAttention(size_t num_tokens, size_t target_parallelism, \
size_t layer_idx, const MatPtr& query_norm_scale, \
AttentionActivationsPtrs& activations, QBatch& qbatch, \
ThreadingContext& ctx); \
\
/* NOLINTNEXTLINE(google-readability-namespace-comments) */ \
} // namespace NAMESPACE
// Function declarations for each SIMD target. Allows direct call from the

1
ops/ops-inl.h
View File

@ -25,6 +25,7 @@
#include <cstdint>
#include <random>
#include <type_traits> // std::enable_if_t
#include <utility>
#include <vector>
#include "ops/matmul.h"