ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

kv-cache : fix shift and defrag logic (#14081)

Browse source 
* kv-cache : fix shift

ggml-ci

* cont : reset shift[i]

ggml-ci

* cont : fix defrag erasing cells that didn't move

ggml-ci
This commit is contained in:
Georgi Gerganov 2025-06-09 23:04:35 +03:00 • committed by GitHub
parent 7f4fbe5183
commit 40cbf571c9
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 12 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

15
src/llama-kv-cells.h
View file

@ -80,6 +80,9 @@ public:
assert(isrc < pos.size());
assert(idst < pos.size());
assert(pos[idst] == -1);
assert(pos[isrc] != -1);
pos [idst] = pos [isrc];
shift[idst] = shift[isrc];
seq [idst] = seq [isrc];
@ -144,9 +147,10 @@ public:
assert(pos[i] != -1);
seq_pos_rm(i);
seq[i].reset();
pos[i] = -1;
seq[i].reset();
shift[i] = 0;
used.erase(i);
}
@ -164,6 +168,7 @@ public:
if (seq[i].none()) {
pos[i] = -1;
shift[i] = 0;
used.erase(i);
@ -192,6 +197,7 @@ public:
seq[i].reset();
pos[i] = -1;
shift[i] = 0;
used.erase(i);
@ -317,21 +323,20 @@ public:
pos[i] += d;
shift[i] += d;
seq_pos_add(i);
has_shift = true;
if (pos[i] < 0) {
seq_pos_rm(i);
seq[i].reset();
pos[i] = -1;
shift[i] = 0;
used.erase(i);
return true;
}
seq_pos_add(i);
return false;
}