Skip to content

Commit

Permalink
Fix an issue in PartitionGatherTrivialSlicedOperandDimensions when …
Browse files Browse the repository at this point in the history 
…handling out-of-bound indices.

A gather operation will clamp the fetched indices such that we always retrieve the corresponding entries in the operand. However, the result of `PartitionGatherTrivialSlicedOperandDimensions` will do not handle these indices. Namely, if the indices is out of bound, we do not retrieve the entries from the operand and the result is 0.

This is a execution bug in SPMD partitioner in both GSPMD and Shardy. The compilation succeeds. This issue does not exist in scatter since scatter does not need to clamp the indices.

This change fixes this issue by clamping the indices at the very beginning of `PartitionGatherTrivialSlicedOperandDimensions`.

PiperOrigin-RevId: 704842079
  • Loading branch information
@ZixuanJiang @Google-ML-Automation
ZixuanJiang authored and Google-ML-Automation committed Dec 12, 2024
1 parent fb8e7d5 commit dfc2f90
Show file tree
Hide file tree
Showing 2 changed files with 61 additions and 15 deletions.
50 changes: 45 additions & 5 deletions xla/service/spmd/gather_scatter_handler.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -193,6 +193,44 @@ std::vector<int64_t> GatherOutputDimsByPriority(
return priority_dims_for_output;
}

PartitionedHlo ClampGatherIndices(const PartitionedHlo& indices,
const Shape& operand_base_shape,
absl::Span<const int64_t> start_index_map,
int64_t index_vector_dim, SpmdBuilder* b) {
const PrimitiveType indices_type = indices.hlo()->shape().element_type();

HloInstruction* max_indices;
if (index_vector_dim < indices.rank()) {
std::vector<int32_t> max_indices_values;
max_indices_values.reserve(start_index_map.size());
for (int64_t operand_dim : start_index_map) {
max_indices_values.push_back(operand_base_shape.dimensions(operand_dim) -
1);
}
max_indices = b->AddInstruction(HloInstruction::CreateConstant(
LiteralUtil::CreateR1<int32_t>(max_indices_values)));
max_indices = b->AddInstruction(HloInstruction::CreateBroadcast(
indices.hlo()->shape(), max_indices, {index_vector_dim}));
} else {
CHECK_EQ(start_index_map.size(), 1);
max_indices = CreateR0WithType<int32_t>(
indices_type, operand_base_shape.dimensions(start_index_map[0]) - 1, b);
max_indices = b->AddInstruction(HloInstruction::CreateBroadcast(
indices.hlo()->shape(), max_indices, {}));
}

HloInstruction* constant_zero = CreateR0WithType<int32_t>(indices_type, 0, b);
HloInstruction* min_indices =
b->AddInstruction(HloInstruction::CreateBroadcast(indices.hlo()->shape(),
constant_zero, {}));

HloInstruction* clamped_indices = b->AddInstruction(
HloInstruction::CreateTernary(indices.hlo()->shape(), HloOpcode::kClamp,
min_indices, indices.hlo(), max_indices));
clamped_indices->set_sharding(indices.sharding());
return PartitionedHlo(clamped_indices, indices.base_shape(), indices.state());
}

// Returns the min and max for the indices in a scatter/gather which has the
// operand partitioned on trivial slice dimensions (slice size 1).
std::pair<HloInstruction*, HloInstruction*>
Expand Down Expand Up @@ -451,11 +489,9 @@ absl::StatusOr<HloInstruction*> PartitionGatherTrivialSlicedOperandDimensions(

SpmdBuilder* b = visitor->builder();
const GatherDimensionNumbers& dnums = gather->gather_dimension_numbers();
std::vector<int64_t> start_index_map(dnums.start_index_map().begin(),
dnums.start_index_map().end());
if (std::optional<std::vector<int64_t>> trivial_slice_dims =
GatherScatterOperandPartitionedOnTrivialSliceDims(
operand, start_index_map, slice_sizes)) {
operand, dnums.start_index_map(), slice_sizes)) {
const HloSharding original_operand_sharding = operand.sharding();
const int64_t num_groups = operand.sharding().NumTiles(*trivial_slice_dims);
const int64_t num_tiles = operand.sharding().TotalNumTiles();
Expand Down Expand Up @@ -504,6 +540,9 @@ absl::StatusOr<HloInstruction*> PartitionGatherTrivialSlicedOperandDimensions(
// Reshard indices to its intended sharding before clamping and adjusting.
indices =
indices.Reshard(hlo_sharding_util::UngroupSharding(indices_grouped));
indices = ClampGatherIndices(indices, operand.base_shape(),
dnums.start_index_map(),
dnums.index_vector_dim(), b);
// Now the operand is partitioned in trivial slice dimensions, and the
// indices are replicated. We execute a gather on partitioned operand,
// with full number of indices, where out-of-bounds indices are clamped,
Expand All @@ -514,8 +553,9 @@ absl::StatusOr<HloInstruction*> PartitionGatherTrivialSlicedOperandDimensions(
HloInstruction* indices_max;
std::tie(indices_min, indices_max) =
IndexBoundsForGatherScatterOperandPartitionedOnTrivialSliceDims(
operand, indices, operand.state().partition_id, start_index_map,
*trivial_slice_dims, dnums.index_vector_dim(), b);
operand, indices, operand.state().partition_id,
dnums.start_index_map(), *trivial_slice_dims,
dnums.index_vector_dim(), b);
// Clamp the indices.
auto adjusted_indices = b->AddInstruction(
HloInstruction::CreateTernary(indices.hlo()->shape(), HloOpcode::kClamp,
Expand Down
26 changes: 16 additions & 10 deletions xla/service/spmd/spmd_partitioner_test.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7926,10 +7926,13 @@ ENTRY entry {
auto min = AllOf(op::Broadcast(offset), op::Shape("s32[2,3]"));
auto max = AllOf(op::Broadcast(op::Add(offset, op::Constant())),
op::Shape("s32[2,3]"));
auto clamp = op::Clamp(min, op::Parameter(1), max);
auto clamped_indices =
op::Clamp(op::Broadcast(op::Constant()), op::Parameter(1),
op::Broadcast(op::Constant()));
auto clamp = op::Clamp(min, clamped_indices, max);
auto gather = op::Gather(op::Parameter(0), op::Subtract(clamp, min));
auto mask =
op::Or(op::Lt(op::Parameter(1), min), op::Gt(op::Parameter(1), max));
op::Or(op::Lt(clamped_indices, min), op::Gt(clamped_indices, max));
auto masked =
op::Select(op::Broadcast(mask), op::Broadcast(op::Constant()), gather);
HloInstruction* root = module->entry_computation()->root_instruction();
Expand All @@ -7952,15 +7955,18 @@ ENTRY entry {
TF_ASSERT_OK_AND_ASSIGN(auto module,
PartitionComputation(hlo_string, /*num_devices=*/4));
VLOG(1) << module->ToString();
auto clamped_indices =
op::Clamp(op::Broadcast(op::Constant()), op::Parameter(1),
op::Broadcast(op::Constant()));
auto offset =
op::Reshape(op::DynamicSlice(op::Constant(), op::PartitionId()));
auto min = AllOf(op::Broadcast(offset), op::Shape("s32[2,3]"));
auto max = AllOf(op::Broadcast(op::Add(offset, op::Constant())),
op::Shape("s32[2,3]"));
auto clamp = op::Clamp(min, op::Parameter(1), max);
auto clamp = op::Clamp(min, clamped_indices, max);
auto gather = op::Gather(op::Parameter(0), op::Subtract(clamp, min));
auto mask =
op::Or(op::Lt(op::Parameter(1), min), op::Gt(op::Parameter(1), max));
op::Or(op::Lt(clamped_indices, min), op::Gt(clamped_indices, max));
auto masked =
op::Select(op::Broadcast(mask), op::Broadcast(op::Constant()), gather);
HloInstruction* root = module->entry_computation()->root_instruction();
Expand Down Expand Up @@ -11919,11 +11925,10 @@ ENTRY entry {
VLOG(1) << module->ToString();
HloInstruction* root = module->entry_computation()->root_instruction();
EXPECT_THAT(root, op::AllReduce(op::Select(_, _, op::Gather(_, _))));
EXPECT_THAT(root->operand(0)->operand(2)->operand(1),
op::Subtract(op::Clamp(_, op::Parameter(1), _), _));
EXPECT_THAT(
root->operand(0)->operand(2)->operand(1),
op::Subtract(op::Clamp(_, op::Clamp(_, op::Parameter(1), _), _), _));

auto clamp = FindInstruction(module.get(), HloOpcode::kClamp);
EXPECT_THAT(clamp->operand(1), op::Parameter(1));
auto dynamic_slice = FindInstruction(module.get(), HloOpcode::kDynamicSlice);
EXPECT_THAT(dynamic_slice->operand(1), op::PartitionId());
auto collective_permute =
Expand Down Expand Up @@ -11955,8 +11960,9 @@ ENTRY entry {
_, op::AllReduce(op::Select(_, _, op::Gather(op::AllReduce(_), _))),
_, _, _)));
auto gather = FindInstruction(module.get(), HloOpcode::kGather);
EXPECT_THAT(gather->operand(1),
op::Subtract(op::Clamp(_, op::Parameter(1), _), _));
EXPECT_THAT(
gather->operand(1),
op::Subtract(op::Clamp(_, op::Clamp(_, op::Parameter(1), _), _), _));
auto collective_permute =
FindInstruction(module.get(), HloOpcode::kCollectivePermute);
EXPECT_NE(collective_permute, nullptr);
Expand Down

0 comments on commit dfc2f90

Please sign in to comment.