Skip to content
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

Remove caching on some CUB types temporarily #216

Merged
merged 1 commit into from
Jul 15, 2022
Merged

Remove caching on some CUB types temporarily #216

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
78 changes: 41 additions & 37 deletions include/matx_cub.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1314,20 +1314,21 @@ void sort(OutputTensor &a_out, const InputOperator &a,

detail::SortParams_t p{dir};

// Don't cache until we have a good plan for hashing parameters here
// Get cache or new Sort plan if it doesn't exist
auto ret = detail::cub_cache.Lookup(params);
if (ret == std::nullopt) {
// auto ret = detail::cub_cache.Lookup(params);
// if (ret == std::nullopt) {
auto tmp = new detail::matxCubPlan_t<OutputTensor, InputOperator, detail::CUB_OP_RADIX_SORT, decltype(p)>{
a_out, a, p, stream};
detail::cub_cache.Insert(params, static_cast<void *>(tmp));
tmp->ExecSort(a_out, a, stream, dir);
}
else {
auto sort_type =
static_cast<detail::matxCubPlan_t<OutputTensor, InputOperator, detail::CUB_OP_RADIX_SORT, decltype(p)> *>(
ret.value());
sort_type->ExecSort(a_out, a, stream, dir);
}
// }
// else {
// auto sort_type =
// static_cast<detail::matxCubPlan_t<OutputTensor, InputOperator, detail::CUB_OP_RADIX_SORT, decltype(p)> *>(
// ret.value());
// sort_type->ExecSort(a_out, a, stream, dir);
// }
#endif
}

Expand Down Expand Up @@ -1411,9 +1412,10 @@ void hist(OutputTensor &a_out, const InputOperator &a,
detail::matxCubPlan_t<OutputTensor, InputOperator, detail::CUB_OP_HIST_EVEN>::GetCubParams(a_out, a);
params.stream = stream;

// Don't cache until we have a good plan for hashing parameters here
// Get cache or new Sort plan if it doesn't exist
auto ret = detail::cub_cache.Lookup(params);
if (ret == std::nullopt) {
// auto ret = detail::cub_cache.Lookup(params);
// if (ret == std::nullopt) {
detail::HistEvenParams_t<typename InputOperator::scalar_type> hp{lower, upper};
auto tmp = new detail::matxCubPlan_t< OutputTensor,
InputOperator,
Expand All @@ -1422,14 +1424,14 @@ void hist(OutputTensor &a_out, const InputOperator &a,
a_out, a, detail::HistEvenParams_t<typename InputOperator::scalar_type>{hp}, stream};
detail::cub_cache.Insert(params, static_cast<void *>(tmp));
tmp->ExecHistEven(a_out, a, lower, upper, stream);
}
else {
auto sort_type =
static_cast<detail::matxCubPlan_t<OutputTensor, InputOperator,
detail::CUB_OP_HIST_EVEN, detail::HistEvenParams_t<typename InputOperator::scalar_type>> *>(
ret.value());
sort_type->ExecHistEven(a_out, a, lower, upper, stream);
}
// }
// else {
// auto sort_type =
// static_cast<detail::matxCubPlan_t<OutputTensor, InputOperator,
// detail::CUB_OP_HIST_EVEN, detail::HistEvenParams_t<typename InputOperator::scalar_type>> *>(
// ret.value());
// sort_type->ExecHistEven(a_out, a, lower, upper, stream);
// }
#endif
}

Expand Down Expand Up @@ -1542,23 +1544,24 @@ void find(OutputTensor &a_out, CountTensor &num_found, const InputOperator &a, S
params.stream = stream;

// Get cache or new Sort plan if it doesn't exist
auto ret = detail::cub_cache.Lookup(params);
//auto ret = detail::cub_cache.Lookup(params);
auto cparams = detail::SelectParams_t<SelectType, CountTensor>{sel, num_found};

if (ret == std::nullopt) {
// Don't cache until we have a good plan for hashing parameters here
//if (ret == std::nullopt) {
auto tmp = new detail::matxCubPlan_t< OutputTensor,
InputOperator,
detail::CUB_OP_SELECT,
decltype(cparams)>{a_out, a, cparams, stream};
detail::cub_cache.Insert(params, static_cast<void *>(tmp));
tmp->ExecSelect(a_out, a, stream);
}
else {
auto sort_type =
static_cast<detail::matxCubPlan_t<OutputTensor, InputOperator, detail::CUB_OP_SELECT, decltype(cparams)> *>(
ret.value());
sort_type->ExecSelect(a_out, a, stream);
}
// }
// else {
// auto sort_type =
// static_cast<detail::matxCubPlan_t<OutputTensor, InputOperator, detail::CUB_OP_SELECT, decltype(cparams)> *>(
// ret.value());
// sort_type->ExecSelect(a_out, a, stream);
// }
#endif
}

Expand Down Expand Up @@ -1602,23 +1605,24 @@ void find_idx(OutputTensor &a_out, CountTensor &num_found, const InputOperator &
params.stream = stream;

// Get cache or new Sort plan if it doesn't exist
auto ret = detail::cub_cache.Lookup(params);
//auto ret = detail::cub_cache.Lookup(params);
auto cparams = detail::SelectParams_t<SelectType, CountTensor>{sel, num_found};

if (ret == std::nullopt) {
// Don't cache until we have a good plan for hashing parameters here
//if (ret == std::nullopt) {
auto tmp = new detail::matxCubPlan_t< OutputTensor,
InputOperator,
detail::CUB_OP_SELECT_IDX,
decltype(cparams)>{a_out, a, cparams, stream};
detail::cub_cache.Insert(params, static_cast<void *>(tmp));
tmp->ExecSelectIndex(a_out, a, stream);
}
else {
auto sort_type =
static_cast<detail::matxCubPlan_t<OutputTensor, InputOperator, detail::CUB_OP_SELECT_IDX, decltype(cparams)> *>(
ret.value());
sort_type->ExecSelectIndex(a_out, a, stream);
}
// }
// else {
// auto sort_type =
// static_cast<detail::matxCubPlan_t<OutputTensor, InputOperator, detail::CUB_OP_SELECT_IDX, decltype(cparams)> *>(
// ret.value());
// sort_type->ExecSelectIndex(a_out, a, stream);
// }
#endif
}

Expand Down