Skip to content

fix: re-feed persistent prefill inputs on cached-graph recompute in QwenCausalDecodeRuntime #331

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
0xShug0 merged 1 commit into from
Aug 29, 2026
+48 −12
Merged

fix: re-feed persistent prefill inputs on cached-graph recompute in QwenCausalDecodeRuntime #331

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
60 changes: 48 additions & 12 deletions src/framework/modules/transformers/qwen_causal_decode_runtime.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -747,18 +747,18 @@ class QwenCausalDecodeRuntime::Impl {
!ggml_gallocr_alloc_graph(prefill_gallocr_, prefill_graph_)) {
throw std::runtime_error("failed to allocate QwenCausalDecodeRuntime prefill graph");
}
const auto positions_values = qwen_position_ids(steps);
prefill_positions_values_ = qwen_position_ids(steps);
ggml_backend_tensor_set(
prefill_positions_,
positions_values.data(),
prefill_positions_values_.data(),
0,
positions_values.size() * sizeof(int32_t));
const auto mask = prefill_attention_mask_values(config_, 1, steps);
prefill_positions_values_.size() * sizeof(int32_t));
prefill_attention_mask_values_ = prefill_attention_mask_values(config_, 1, steps);
ggml_backend_tensor_set(
prefill_attention_mask_,
mask.data(),
prefill_attention_mask_values_.data(),
0,
mask.size() * sizeof(ggml_fp16_t));
prefill_attention_mask_values_.size() * sizeof(ggml_fp16_t));
if (prefill_logits_readback_token_ids_ != nullptr) {
upload_logits_readback_token_ids(prefill_logits_readback_token_ids_, config_);
}
Expand All @@ -771,6 +771,19 @@ class QwenCausalDecodeRuntime::Impl {
}

QwenCausalPrefillResult run_prefill() {
// The gallocr considers the persistent position/mask inputs dead after
// their last read inside a compute and may hand their memory to other
// tensors, so a cached prefill graph must be re-fed before recompute.
ggml_backend_tensor_set(
prefill_positions_,
prefill_positions_values_.data(),
0,
prefill_positions_values_.size() * sizeof(int32_t));
ggml_backend_tensor_set(
prefill_attention_mask_,
prefill_attention_mask_values_.data(),
0,
prefill_attention_mask_values_.size() * sizeof(ggml_fp16_t));
core::set_backend_threads(backend_, threads_);
const ggml_status status = core::compute_backend_graph(backend_, prefill_graph_);
ggml_backend_synchronize(backend_);
Expand Down Expand Up @@ -922,18 +935,18 @@ class QwenCausalDecodeRuntime::Impl {
!ggml_gallocr_alloc_graph(batched_prefill_gallocr_, batched_prefill_graph_)) {
throw std::runtime_error("failed to allocate QwenCausalDecodeRuntime batched prefill graph");
}
const auto positions_values = qwen_position_ids(steps);
batched_prefill_positions_values_ = qwen_position_ids(steps);
ggml_backend_tensor_set(
batched_prefill_positions_,
positions_values.data(),
batched_prefill_positions_values_.data(),
0,
positions_values.size() * sizeof(int32_t));
const auto mask = prefill_attention_mask_values(config_, batch_size, steps);
batched_prefill_positions_values_.size() * sizeof(int32_t));
batched_prefill_attention_mask_values_ = prefill_attention_mask_values(config_, batch_size, steps);
ggml_backend_tensor_set(
batched_prefill_attention_mask_,
mask.data(),
batched_prefill_attention_mask_values_.data(),
0,
mask.size() * sizeof(ggml_fp16_t));
batched_prefill_attention_mask_values_.size() * sizeof(ggml_fp16_t));
if (batched_prefill_logits_readback_token_ids_ != nullptr) {
upload_logits_readback_token_ids(batched_prefill_logits_readback_token_ids_, config_);
}
Expand All @@ -946,6 +959,17 @@ class QwenCausalDecodeRuntime::Impl {
}

QwenCausalBatchedPrefillResult run_batched_prefill() {
// Re-feed persistent inputs before recompute (see run_prefill note).
ggml_backend_tensor_set(
batched_prefill_positions_,
batched_prefill_positions_values_.data(),
0,
batched_prefill_positions_values_.size() * sizeof(int32_t));
ggml_backend_tensor_set(
batched_prefill_attention_mask_,
batched_prefill_attention_mask_values_.data(),
0,
batched_prefill_attention_mask_values_.size() * sizeof(ggml_fp16_t));
core::set_backend_threads(backend_, threads_);
const ggml_status status = core::compute_backend_graph(backend_, batched_prefill_graph_);
ggml_backend_synchronize(backend_);
Expand Down Expand Up @@ -1347,6 +1371,8 @@ class QwenCausalDecodeRuntime::Impl {
prefill_keys_.clear();
prefill_values_.clear();
prefill_graph_ = nullptr;
prefill_positions_values_.clear();
prefill_attention_mask_values_.clear();
prefill_steps_ = 0;
prefill_input_kind_ = InputKind::None;
}
Expand All @@ -1370,6 +1396,8 @@ class QwenCausalDecodeRuntime::Impl {
batched_prefill_keys_.clear();
batched_prefill_values_.clear();
batched_prefill_graph_ = nullptr;
batched_prefill_positions_values_.clear();
batched_prefill_attention_mask_values_.clear();
batched_prefill_batch_size_ = 0;
batched_prefill_steps_ = 0;
batched_prefill_input_kind_ = InputKind::None;
Expand Down Expand Up @@ -1442,6 +1470,11 @@ class QwenCausalDecodeRuntime::Impl {
std::vector<ggml_tensor *> prefill_values_;
ggml_cgraph * prefill_graph_ = nullptr;
ggml_gallocr_t prefill_gallocr_ = nullptr;
// Host-side copies of the persistent graph inputs. The gallocr may reuse
// their memory after the last read within one compute, so they must be
// re-uploaded before every recompute of a cached prefill graph.
std::vector<int32_t> prefill_positions_values_;
std::vector<ggml_fp16_t> prefill_attention_mask_values_;
int64_t prefill_steps_ = 0;
InputKind prefill_input_kind_ = InputKind::None;

Expand All @@ -1456,6 +1489,9 @@ class QwenCausalDecodeRuntime::Impl {
std::vector<ggml_tensor *> batched_prefill_values_;
ggml_cgraph * batched_prefill_graph_ = nullptr;
ggml_gallocr_t batched_prefill_gallocr_ = nullptr;
// Host-side copies, re-uploaded before every recompute (see prefill note).
std::vector<int32_t> batched_prefill_positions_values_;
std::vector<ggml_fp16_t> batched_prefill_attention_mask_values_;
int64_t batched_prefill_batch_size_ = 0;
int64_t batched_prefill_steps_ = 0;
InputKind batched_prefill_input_kind_ = InputKind::None;
Expand Down
Loading