tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-05 02:23:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

batch : optional requirement for sequential sequence ids

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-06-25 17:02:38 +03:00
parent 6663128448
commit 5eb1a88dc0
6 changed files with 11 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
src/llama-batch.cpp
View File

@ -457,7 +457,7 @@ llama_ubatch llama_batch_allocr::split_simple(uint32_t n_ubatch) {
return ubatch_add(idxs, idxs.size(), false);
}
llama_ubatch llama_batch_allocr::split_equal(uint32_t n_ubatch) {
llama_ubatch llama_batch_allocr::split_equal(uint32_t n_ubatch, bool sequential) {
std::vector<seq_set_t> cur_seq_set;
llama_seq_id last_seq_id = -1;
@ -479,7 +479,9 @@ llama_ubatch llama_batch_allocr::split_equal(uint32_t n_ubatch) {
}
// accept only increasing sequence ids
add = add && (cur_seq_set.empty() || batch.seq_id[i][0] == last_seq_id + 1);
if (sequential) {
add = add && (cur_seq_set.empty() || batch.seq_id[i][0] == last_seq_id + 1);
}
if (add) {
cur_seq_set.push_back(seq_set[i]);

3
src/llama-batch.h
View File

@ -69,7 +69,8 @@ public:
llama_ubatch split_simple(uint32_t n_ubatch);
// make ubatches of equal-length sequences sets
llama_ubatch split_equal(uint32_t n_ubatch);
// if sequential == true, the tokens in the ubatch will have increasing sequential sequence ids
llama_ubatch split_equal(uint32_t n_ubatch, bool sequential);
// sequence-set-wise split - each ubatch contains a single sequence-set
llama_ubatch split_seq(uint32_t n_ubatch);

4
src/llama-kv-cache-unified-iswa.cpp
View File

@ -102,7 +102,7 @@ llama_memory_context_ptr llama_kv_cache_unified_iswa::init_batch(llama_batch_all
// first try simple split
do {
if (n_seq_virt > 1) {
// requires equal splits
// requires equal splits, so we skip the simple split
break;
}
@ -141,7 +141,7 @@ llama_memory_context_ptr llama_kv_cache_unified_iswa::init_batch(llama_batch_all
std::vector<llama_ubatch> ubatches;
while (true) {
auto ubatch = balloc.split_equal(n_ubatch);
auto ubatch = balloc.split_equal(n_ubatch, n_seq_virt > 1);
if (ubatch.n_tokens == 0) {
break;

2
src/llama-kv-cache-unified.cpp
View File

@ -418,7 +418,7 @@ llama_memory_context_ptr llama_kv_cache_unified::init_batch(
std::vector<llama_ubatch> ubatches;
while (true) {
auto ubatch = n_seq_virt == 1 ? balloc.split_simple(n_ubatch) : balloc.split_equal(n_ubatch);
auto ubatch = n_seq_virt == 1 ? balloc.split_simple(n_ubatch) : balloc.split_equal(n_ubatch, true);
if (ubatch.n_tokens == 0) {
break;

2
src/llama-memory-hybrid.cpp
View File

@ -71,7 +71,7 @@ llama_memory_context_ptr llama_memory_hybrid::init_batch(llama_batch_allocr & ba
// if all tokens are output, split by sequence
ubatch = balloc.split_seq(n_ubatch);
} else {
ubatch = balloc.split_equal(n_ubatch);
ubatch = balloc.split_equal(n_ubatch, false);
}
if (ubatch.n_tokens == 0) {

2
src/llama-memory-recurrent.cpp
View File

@ -372,7 +372,7 @@ llama_memory_context_ptr llama_memory_recurrent::init_batch(llama_batch_allocr &
// if all tokens are output, split by sequence
ubatch = balloc.split_seq(n_ubatch);
} else {
ubatch = balloc.split_equal(n_ubatch);
ubatch = balloc.split_equal(n_ubatch, false);
}
if (ubatch.n_tokens == 0) {