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

Revert "fix: replace soft rollback with proper failure in recurrent seq_rm" 

This reverts commit 9a04ac4e10.
This commit is contained in:
eauchs 2026-03-05 11:11:38 +01:00
parent 9a04ac4e10
commit 8fe0e0353e
1 changed files with 4 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
src/llama-memory-recurrent.cpp
View File

@ -175,9 +175,9 @@ bool llama_memory_recurrent::seq_rm(llama_seq_id seq_id, llama_pos p0, llama_pos
if (best_cell >= 0) {
tail_id = best_cell;
} else {
// no checkpoint found at p0-1: the SSM tensor state cannot be rolled back
// without re-evaluating the sequence. Signal failure to the caller.
return false;
// if no checkpoint found, we still move the position back (soft rollback)
// only if it's the current sequence's tail
cells[tail_id].pos = p0 - 1;
}
}
// invalidate tails which will be cleared
@ -648,17 +648,10 @@ bool llama_memory_recurrent::find_slot(const llama_ubatch & ubatch) {
// Copy state data
copy_cell(seq_meta.tail, next_empty_cell);
// Keep history of previous states for rollback (up to 8 cells per sequence).
// The 0.9 threshold prevents the checkpoint history from filling the cache.
// When the cache is too full to keep checkpoints, speculative decoding rollback
// will fail (seq_rm returns false) and the caller must re-evaluate.
// Keep history of previous states for rollback (up to 8 cells per sequence)
if (get_cell_count(seq_id) < 8 && used < size * 0.9) {
// Do not erase seq_id from orig_cell to keep it as a checkpoint
} else {
if (used >= size * 0.9) {
LLAMA_LOG_DEBUG("%s: cache too full (used=%u/%u) to keep checkpoint for seq %d; speculative rollback will require re-evaluation\n",
__func__, used, size, seq_id);
}
// Erase oldest history point for this sequence
int32_t oldest_cell = -1;
llama_pos min_pos = std::numeric_limits<llama_pos>::max();