Skip to content

Forward policy hub from dispatch_streaming_arg_reduce_t to reduce::dispatch #7805

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
bernhardmgruber merged 3 commits into from
Feb 26, 2026
Merged

Forward policy hub from dispatch_streaming_arg_reduce_t to reduce::dispatch #7805

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
d67b045
Forward policy hub from dispatch_streaming_arg_reduce_t to reduce::di…
bernhardmgruber Feb 26, 2026
a7dfbfa
fix
bernhardmgruber Feb 26, 2026
b189da8
Fix
bernhardmgruber Feb 26, 2026
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
1 change: 1 addition & 0 deletions cub/benchmarks/bench/reduce/arg_extrema.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@
// %RANGE% TUNE_ITEMS_PER_VEC_LOAD_POW2 ipv 1:2:1

#if !TUNE_BASE
# error "Cannot tune until https://github.com/NVIDIA/cccl/pull/7807 is merged"
# define TUNE_ITEMS_PER_VEC_LOAD (1 << TUNE_ITEMS_PER_VEC_LOAD_POW2)
template <typename AccumT, typename OffsetT>
struct policy_hub_t
Expand Down
12 changes: 7 additions & 5 deletions cub/cub/device/dispatch/dispatch_streaming_reduce.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -141,16 +141,16 @@ struct local_to_global_op
// @tparam InitT
// Initial value type
//
// @tparam PolicyChainT
// The policy chain passed to the DispatchReduce template specialization
// @tparam PolicySelector
// Selects the tuning policy
template <typename InputIteratorT,
typename OutputIteratorT,
typename PerPartitionOffsetT,
typename GlobalOffsetT,
typename ReductionOpT,
typename InitT,
typename PolicyChainT =
detail::reduce::policy_hub<KeyValuePair<PerPartitionOffsetT, InitT>, PerPartitionOffsetT, ReductionOpT>>
typename PolicySelector = detail::reduce::
policy_selector_from_types<KeyValuePair<PerPartitionOffsetT, InitT>, PerPartitionOffsetT, ReductionOpT>>
struct dispatch_streaming_arg_reduce_t
{
// Internal dispatch routine for computing a device-wide argument extremum, like `ArgMin` and `ArgMax`
Expand Down Expand Up @@ -254,7 +254,9 @@ struct dispatch_streaming_arg_reduce_t
static_cast<PerPartitionOffsetT>(largest_partition_size),
reduce_op,
initial_value,
stream);
stream,
::cuda::std::identity{},
PolicySelector{});

// Alias the temporary allocations from the single storage blob (or compute the necessary size
// of the blob)
Expand Down
2 changes: 1 addition & 1 deletion cub/cub/device/dispatch/tuning/tuning_reduce.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -323,7 +323,7 @@ struct policy_hub
using ReducePolicy =
decltype(select_agent_policy<sm100_tuning<AccumT,
OffsetT,
classify_op<ReductionOpT>(),
classify_op<ReductionOpT>,
classify_offset_size<OffsetT>(),
classify_accum_size<AccumT>()>>(0));

Expand Down
Loading