namespace Gambit::PostProcessor
[No description available]
Classes
Name | |
---|---|
class | Gambit::PostProcessor::PPDriver Driver class to handle the actual postprocessing tasks. |
struct | Gambit::PostProcessor::PPOptions |
Functions
Name | |
---|---|
ChunkSet | get_done_points(Gambit::Printers::BaseBaseReader & filebase) Read through resume data files and reconstruct which chunks of points have already been processed. |
bool | point_done(const ChunkSet done_chunks, size_t index) Helper functions for performing resume related tasks. |
Chunk | get_effective_chunk(const std::size_t total_length, const unsigned int rank, const unsigned int numtasks) |
Chunk | get_my_chunk(const std::size_t dset_length, const ChunkSet & done_chunks, const int rank, const int numtasks) Compute start/end indices for a given rank process, given previous “done_chunk” data. |
ChunkSet | get_done_points(const std::string & filebase) Read through resume data files and reconstruct which chunks of points have already been processed. |
ChunkSet | merge_chunks(const ChunkSet & input_chunks) Simplify a ChunkSet by merging chunks which overlap. |
void | record_done_points(const ChunkSet & done_chunks, const Chunk & mydone, const std::string & filebase, unsigned int rank, unsigned int size) |
Attributes
Name | |
---|---|
const Chunk | stopchunk |
Functions Documentation
function get_done_points
ChunkSet get_done_points(
Gambit::Printers::BaseBaseReader & filebase
)
Read through resume data files and reconstruct which chunks of points have already been processed.
Read through any pre-existing output and reconstruct which chunks of points have already been processed.
function point_done
bool point_done(
const ChunkSet done_chunks,
size_t index
)
Helper functions for performing resume related tasks.
Answer queries as to whether a given dataset index has been postprocessed in a previous run or not
function get_effective_chunk
Chunk get_effective_chunk(
const std::size_t total_length,
const unsigned int rank,
const unsigned int numtasks
)
Get ’effective’ start and end positions for a processing batch i.e. simply divides up an integer into the most even parts possible over a given number of processes
function get_my_chunk
Chunk get_my_chunk(
const std::size_t dset_length,
const ChunkSet & done_chunks,
const int rank,
const int numtasks
)
Compute start/end indices for a given rank process, given previous “done_chunk” data.
First compute number of points left to process
First compute number of points left to process
function get_done_points
ChunkSet get_done_points(
const std::string & filebase
)
Read through resume data files and reconstruct which chunks of points have already been processed.
function merge_chunks
ChunkSet merge_chunks(
const ChunkSet & input_chunks
)
Simplify a ChunkSet by merging chunks which overlap.
Simplify a ChunkSet by merging chunks which overlap (or are directly adjacent).
function record_done_points
void record_done_points(
const ChunkSet & done_chunks,
const Chunk & mydone,
const std::string & filebase,
unsigned int rank,
unsigned int size
)
Write resume data files These specify which chunks of points have been processed during this run
Attributes Documentation
variable stopchunk
const Chunk stopchunk = Chunk(0,0);
Updated on 2024-07-18 at 13:53:31 +0000