From 41873b6ede77ce6126b94d4b307d9b1d5be31426 Mon Sep 17 00:00:00 2001
From: Michael Buehlmann <buehlmann.michi@gmail.com>
Date: Thu, 7 Sep 2023 15:31:27 -0500
Subject: [PATCH] remove rebalance_sourceranks function
(was added 2 years ago and did not turn out to be useful.
Also causing problems on some compilers, eg on theta)
---
GenericIO.cxx | 33 ---------------------------------
GenericIO.h | 3 ---
python/genericio.cpp | 13 -------------
3 files changed, 49 deletions(-)
diff --git a/GenericIO.cxx b/GenericIO.cxx
index ac00f34..bcae435 100644
--- a/GenericIO.cxx
+++ b/GenericIO.cxx
@@ -1793,39 +1793,6 @@ void GenericIO::readData(int EffRank, size_t RowOffset, int Rank,
}
}
-#ifndef GENERICIO_NO_MPI
-void GenericIO::rebalanceSourceRanks() {
- if(Redistributing) {
- int NRanks, Rank;
- MPI_Comm_rank(Comm, &Rank);
- MPI_Comm_size(Comm, &NRanks);
-
- std::vector<std::pair<int, size_t>> rank_sizes;
- std::vector<std::tuple<int, size_t, std::vector<int>>> new_source_ranks;
- for(int i=0; i<NRanks; ++i) {
- new_source_ranks.emplace_back(std::make_tuple(i, 0ul, std::vector<int>()));
- }
- for(int i=0; i<readNRanks(); ++i) {
- rank_sizes.emplace_back(std::make_pair(i, readNumElems(i)));
- }
- std::sort(rank_sizes.begin(), rank_sizes.end(), [](const auto& p1, const auto& p2){ return p1.second > p2.second; });
- // Distribute ranks
- for(size_t i=0; i<rank_sizes.size(); ++i) {
- // Assign to first rank
- std::get<2>(new_source_ranks[0]).push_back(rank_sizes[i].first);
- std::get<1>(new_source_ranks[0]) += rank_sizes[i].second;
- // Reorder ranks (could be optimized since array already sorted)
- std::stable_sort(new_source_ranks.begin(), new_source_ranks.end(), [](const auto& s1, const auto& s2){ return std::get<1>(s1) < std::get<1>(s2); });
- }
- // copy own array
- SourceRanks.resize(0);
- std::copy(std::get<2>(new_source_ranks[Rank]).begin(), std::get<2>(new_source_ranks[Rank]).end(), std::back_inserter(SourceRanks));
- } else {
- std::cerr << "rebalancing source ranks has no effect when Redistributing==false" << std::endl;
- }
-}
-#endif
-
void GenericIO::getVariableInfo(vector<VariableInfo> &VI) {
if (FH.isBigEndian())
getVariableInfo<true>(VI);
diff --git a/GenericIO.h b/GenericIO.h
index 02bf685..3241415 100644
--- a/GenericIO.h
+++ b/GenericIO.h
@@ -439,9 +439,6 @@ public:
int readGlobalRankNumber(int EffRank = -1);
void readData(int EffRank = -1, bool PrintStats = true, bool CollStats = true);
-#ifndef GENERICIO_NO_MPI
- void rebalanceSourceRanks();
-#endif
void getSourceRanks(std::vector<int> &SR);
void close() {
diff --git a/python/genericio.cpp b/python/genericio.cpp
index a29d478..64bc989 100644
--- a/python/genericio.cpp
+++ b/python/genericio.cpp
@@ -204,14 +204,9 @@ std::map<std::string, py::array> read_genericio(
PyGenericIO::MismatchBehavior redistribute=PyGenericIO::MismatchBehavior::MismatchRedistribute,
bool print_stats=true,
bool collective_stats=true,
- bool rebalance_source_ranks=false,
int eff_rank=-1
) {
PyGenericIO reader(filename, method, redistribute, eff_rank);
-#ifndef GENERICIO_NO_MPI
- if(rebalance_source_ranks)
- reader.rebalanceSourceRanks();
-#endif
return reader.read(var_names, print_stats, collective_stats, eff_rank);
}
@@ -383,7 +378,6 @@ PYBIND11_MODULE(pygio, m) {
.def_static("setNaturalDefaultPartition", &PyGenericIO::setNaturalDefaultPartition)
.def_static("setDefaultFileIOType", &PyGenericIO::setDefaultFileIOType)
#ifndef GENERICIO_NO_MPI
- .def("rebalance_source_ranks", &PyGenericIO::rebalanceSourceRanks)
.def_static("setCollectiveMPIIOThreshold", &PyGenericIO::setCollectiveMPIIOThreshold)
#endif
;
@@ -423,12 +417,6 @@ PYBIND11_MODULE(pygio, m) {
collective_stats: bool
if ``True``, aggregate statistics among reading ranks (if using MPI)
- rebalance_sourceranks: bool
- if ``True``, the code will re-assign the file ranks to the reading
- MPI ranks to equalize the data size each rank is reading. Only
- relevant if using MPI and more ranks were used to write the file
- than reading.
-
Returns
-------
data: Mapping[str, np.ndarray]
@@ -441,7 +429,6 @@ PYBIND11_MODULE(pygio, m) {
py::arg("redistribute")=PyGenericIO::MismatchBehavior::MismatchRedistribute,
py::arg("print_stats")=true,
py::arg("collective_stats")=true,
- py::arg("rebalance_sourceranks")=false,
py::arg("eff_rank")=-1,
py::return_value_policy::move);
--
GitLab