Files
Grid/Grid/parallelIO/BinaryIO.cc
T
Peter Boyle 41f5a02204 IO updates for AllToAllV aggregation of contiguous chunks.
May finally clean up the poor MPI2 IO performance issue that has been persistent.
2026-08-11 13:33:04 -04:00

9 lines
372 B
C++

#include <Grid/GridCore.h>
int Grid::BinaryIO::latticeWriteMaxRetry = -1;
Grid::BinaryIO::IoPerf Grid::BinaryIO::lastPerf;
// Target size of a single contiguous file extent under BINARYIO_AGGREGATE.
// 4MB is around the knee for Lustre; exposed so it can be swept at runtime.
uint64_t Grid::BinaryIO::aggregateTargetBytes = 4*1024*1024;