Skip to content

[JAX SC] Fix unique ID count propagation in minibatching splits and update test assertion. #523

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
Oct 22, 2025
+9 −8
Merged

[JAX SC] Fix unique ID count propagation in minibatching splits and update test assertion. #523

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
15 changes: 7 additions & 8 deletions jax_tpu_embedding/sparsecore/lib/core/input_preprocessing_test.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -963,14 +963,13 @@ void ValidateMinibatchOrSparseCoreSlice(
int ids_count = 0;
absl::flat_hash_set<int> unique_ids;
for (int j = start_index; j < end_index; ++j) {
if (embedding_ids_slice(j) != INT_MAX) {
ValidateCooId(embedding_ids_slice(j), sample_ids_slice(j),
table_shard_size, batch_size_per_sc);
ids_count++;
unique_ids.insert(embedding_ids_slice(j));
// ASSERT_LE(ids_count, max_ids_per_partition);
// ASSERT_LE(unique_ids.size(), max_unique_ids_per_partition);
}
ASSERT_NE(embedding_ids_slice(j), INT_MAX);
ValidateCooId(embedding_ids_slice(j), sample_ids_slice(j),
table_shard_size, batch_size_per_sc);
ids_count++;
unique_ids.insert(embedding_ids_slice(j));
ASSERT_LE(ids_count, max_ids_per_partition);
ASSERT_LE(unique_ids.size(), max_unique_ids_per_partition);
}
start_index = xla::RoundUpTo(end_index, TPU_VECTOR_REGISTER_ALIGNMENT_SIZE);
}
Expand Down
2 changes: 2 additions & 0 deletions jax_tpu_embedding/sparsecore/lib/core/minibatching_splits_impl.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@
#ifndef JAX_TPU_EMBEDDING_SPARSECORE_LIB_CORE_MINIBATCHING_SPLITS_IMPL_H_
#define JAX_TPU_EMBEDDING_SPARSECORE_LIB_CORE_MINIBATCHING_SPLITS_IMPL_H_

#include <algorithm>
#include <bitset>
#include <cstddef>
#include <cstdint>
Expand Down Expand Up @@ -72,6 +73,7 @@ std::bitset<N - 1> ComputeMinibatchingSplit(
const int val_right = unique_ids_per_bucket[i + subtree_size / 2];
if (val_left + val_right > max_unique_ids_per_partition) {
split.set(split_index);
unique_ids_per_bucket[i] = std::max(val_left, val_right);
} else {
unique_ids_per_bucket[i] += val_right;
}
Expand Down