Skip to content

llama: Enable K-shift for quantized KV cache for cuda #760

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 5, 2025
Merged

llama: Enable K-shift for quantized KV cache for cuda #760

Show file tree
Hide file tree
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
3 changes: 3 additions & 0 deletions ggml/src/ggml-cuda.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4119,6 +4119,9 @@ GGML_CALL static bool ggml_backend_cuda_supports_op(ggml_backend_t backend, cons
if (src0_type == GGML_TYPE_F32 && src1_type == GGML_TYPE_Q8_0) {
return true;
}
if (src0_type == GGML_TYPE_Q8_0 && src1_type == GGML_TYPE_F32) {
return true;
}
if (src0_type == GGML_TYPE_F32 && src1_type == GGML_TYPE_Q4_0) {
return true;
}
Expand Down
35 changes: 27 additions & 8 deletions src/llama.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8367,17 +8367,36 @@ struct llm_build_context {
const int64_t n_head_kv = hparams.n_head_kv(il);
const int64_t n_embd_k_gqa = hparams.n_embd_k_gqa(il);
struct ggml_tensor * rope_factors = build_rope_factors(il);
struct ggml_tensor * tmp =
struct ggml_tensor * k =
ggml_view_3d(ctx0, kv_self.k_l[il],
n_embd_head_k, n_head_kv, n_ctx,
ggml_row_size(kv_self.k_l[il]->type, n_embd_head_k),
ggml_row_size(kv_self.k_l[il]->type, n_embd_k_gqa),
0);

struct ggml_tensor * tmp;
if (ggml_is_quantized(k->type)) {
// dequantize to f32 -> RoPE -> quantize back
tmp = ggml_cast(ctx0, k, GGML_TYPE_F32);
cb(tmp, "K_f32", il);
for (auto * backend : lctx.backends) {
// Figure out which backend KV cache belongs to
if (ggml_backend_supports_buft(backend, lctx.model.buft_layer[il].buft)) {
ggml_backend_sched_set_tensor_backend(lctx.sched, tmp, backend);
break;
}
}
tmp = ggml_rope_ext_inplace(ctx0, tmp,
lctx.inp_K_shift, rope_factors, n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
ext_factor, attn_factor, beta_fast, beta_slow);
cb(tmp, "K_shifted_f32", il);
tmp = ggml_cpy(ctx0, tmp, k);
} else {
// we rotate only the first n_rot dimensions
ggml_rope_ext_inplace(ctx0,
ggml_view_3d(ctx0, kv_self.k_l[il],
n_embd_head_k, n_head_kv, n_ctx,
ggml_row_size(kv_self.k_l[il]->type, n_embd_head_k),
ggml_row_size(kv_self.k_l[il]->type, n_embd_k_gqa),
0),
tmp = ggml_rope_ext_inplace(ctx0, k,
lctx.inp_K_shift, rope_factors, n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
ext_factor, attn_factor, beta_fast, beta_slow);

}
cb(tmp, "K_shifted", il);
ggml_build_forward_expand(gf, tmp);
}
Expand Down