Skip to content

kleidiai: generalize compute_forward_kv_cache to compute_forward_fp16 #15817

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

Merged
merged 1 commit into from
Sep 6, 2025
Merged

kleidiai: generalize compute_forward_kv_cache to compute_forward_fp16 #15817

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
13 changes: 4 additions & 9 deletions ggml/src/ggml-cpu/kleidiai/kleidiai.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -154,7 +154,7 @@ class tensor_traits : public ggml::cpu::tensor_traits {
if (dst->src[0]->type == GGML_TYPE_Q4_0) {
return compute_forward_q4_0(params, dst);
} else if (dst->src[0]->type == GGML_TYPE_F16) {
return compute_forward_kv_cache(params, dst);
return compute_forward_fp16(params, dst);
}
} else if (dst->op == GGML_OP_GET_ROWS) {
if (dst->src[0]->type == GGML_TYPE_Q4_0) {
Expand All @@ -164,7 +164,7 @@ class tensor_traits : public ggml::cpu::tensor_traits {
return false;
}

bool compute_forward_kv_cache(ggml_compute_params * params, struct ggml_tensor * dst) {
bool compute_forward_fp16(ggml_compute_params * params, struct ggml_tensor * dst) {
static std::atomic_flag first_to_arrive = ATOMIC_FLAG_INIT;

const ggml_tensor * src0 = dst->src[0];
Expand Down Expand Up @@ -534,13 +534,8 @@ class extra_buffer_type : ggml::cpu::extra_buffer_type {
if (op->src[0]->buffer && op->src[0]->buffer->buft == ggml_backend_cpu_kleidiai_buffer_type()) {
return (ggml::cpu::tensor_traits *) op->src[0]->extra;
}
else if (ggml_kleidiai_select_kernels(ctx.features, op) &&
op->src[0]->op == GGML_OP_VIEW &&
(op->src[1]->op == GGML_OP_PERMUTE || op->src[1]->op == GGML_OP_SOFT_MAX) &&
op->src[1]->ne[1] > 1) {
if ((op->src[0]->nb[0] != 2) ||
(op->src[1]->nb[0] != 4) ||
(op->src[0]->nb[1] * op->src[0]->ne[1] != op->src[0]->nb[2]) ||
else if (ggml_kleidiai_select_kernels(ctx.features, op) && op->src[1]->ne[1] > 1) {
if ((op->src[0]->nb[1] * op->src[0]->ne[1] != op->src[0]->nb[2]) ||
(op->src[1]->nb[1] * op->src[1]->ne[1] != op->src[1]->nb[2])) {
return nullptr;
}
Expand Down
Loading