mirror of
https://github.com/Unidata/netcdf-c.git
synced 2024-12-03 08:01:25 +08:00
166 lines
5.3 KiB
C
166 lines
5.3 KiB
C
/*
|
|
Copyright 2007, UCAR/Unidata
|
|
See COPYRIGHT file for copying and redistribution conditions.
|
|
|
|
This program benchmarks the write and read of reasonable large files,
|
|
with different chunking parameters set.
|
|
|
|
$Id: bm_chunking.c,v 1.4 2009/08/06 17:29:32 ed Exp $
|
|
*/
|
|
#include <config.h>
|
|
#include <nc_tests.h>
|
|
#include <netcdf.h>
|
|
#include <stdio.h>
|
|
#include <string.h>
|
|
#include <time.h>
|
|
#include <sys/time.h> /* Extra high precision time info. */
|
|
#include <math.h>
|
|
|
|
/* This is one megabyte (2^20), in decimal. */
|
|
#define MEGABYTE 1048576
|
|
|
|
/* Define this to debug. */
|
|
#define DIM_LEN (MEGABYTE*32)
|
|
/*#define DIM_LEN (1024*8)*/
|
|
|
|
/* Size, in bytes, of a double. */
|
|
#define DOUBLE_SIZE 8
|
|
|
|
/* We will create this file. */
|
|
#define FILE_NAME "bm_chunking.nc"
|
|
|
|
/* Subtract the `struct timeval' values X and Y, storing the result in
|
|
RESULT. Return 1 if the difference is negative, otherwise 0. This
|
|
function from the GNU documentation. */
|
|
int
|
|
timeval_subtract (result, x, y)
|
|
struct timeval *result, *x, *y;
|
|
{
|
|
/* Perform the carry for the later subtraction by updating Y. */
|
|
if (x->tv_usec < y->tv_usec) {
|
|
int nsec = (y->tv_usec - x->tv_usec) / 1000000 + 1;
|
|
y->tv_usec -= 1000000 * nsec;
|
|
y->tv_sec += nsec;
|
|
}
|
|
if (x->tv_usec - y->tv_usec > 1000000) {
|
|
int nsec = (x->tv_usec - y->tv_usec) / 1000000;
|
|
y->tv_usec += 1000000 * nsec;
|
|
y->tv_sec -= nsec;
|
|
}
|
|
|
|
/* Compute the time remaining to wait.
|
|
`tv_usec' is certainly positive. */
|
|
result->tv_sec = x->tv_sec - y->tv_sec;
|
|
result->tv_usec = x->tv_usec - y->tv_usec;
|
|
|
|
/* Return 1 if result is negative. */
|
|
return x->tv_sec < y->tv_sec;
|
|
}
|
|
|
|
|
|
int
|
|
main(int argc, char **argv)
|
|
{
|
|
struct timeval start_time, end_time, diff_time;
|
|
printf("\n*** Benchmarking chunking.\n");
|
|
|
|
printf("\n*** Trying different chunksizes in 4 var, 1D file with "
|
|
"large dimension (%d)...\n", DIM_LEN);
|
|
{
|
|
#define DIM_NAME "Some_really_long_dimension"
|
|
#define NUMDIMS 1
|
|
#define NUMVARS 4
|
|
#define NUM_ATTEMPTS 10
|
|
#define NUM_VALUES_TO_WRITE (DIM_LEN/1024)
|
|
|
|
int ncid, dimids[NUMDIMS], varid[NUMVARS], chunksize[NUMDIMS];
|
|
char var_name[NUMVARS][NC_MAX_NAME + 1] = {"England", "Ireland", "Scotland", "Wales"};
|
|
size_t start[NUMDIMS], count[NUMDIMS] = {NUM_VALUES_TO_WRITE};
|
|
int ndims, nvars, natts, unlimdimid;
|
|
nc_type xtype;
|
|
char name_in[NC_MAX_NAME + 1];
|
|
size_t len;
|
|
double pi = 3.1459;
|
|
double data[NUM_VALUES_TO_WRITE], data_in[NUM_VALUES_TO_WRITE];
|
|
int a, i, j;
|
|
|
|
/* Generate phoney data. */
|
|
for (i = 0; i < NUM_VALUES_TO_WRITE; i++)
|
|
data[i] = pi * i;
|
|
|
|
/* Try NUM_ATTEMPTS different values of chunksize, timing each one. */
|
|
for (a = 0; a < NUM_ATTEMPTS; a++)
|
|
{
|
|
/* Create a netCDF netCDF-4/HDF5 format file, with 4 vars. */
|
|
if (nc_create(FILE_NAME, NC_NETCDF4, &ncid)) ERR;
|
|
if (nc_set_fill(ncid, NC_NOFILL, NULL)) ERR;
|
|
if (nc_def_dim(ncid, DIM_NAME, DIM_LEN, dimids)) ERR;
|
|
|
|
/* According to Quncey, a square chunk around 1 MB in size is a
|
|
* reasonable idea. */
|
|
chunksize[0] = DIM_LEN/pow(2, a+6);
|
|
for (i = 0; i < NUMVARS; i++)
|
|
{
|
|
if (nc_def_var(ncid, var_name[i], NC_DOUBLE, NUMDIMS,
|
|
dimids, &varid[i])) ERR;
|
|
if (nc_def_var_chunking(ncid, i, NC_CHUNKED, chunksize)) ERR;
|
|
}
|
|
if (nc_enddef(ncid)) ERR;
|
|
|
|
/* Write the data and time it. */
|
|
if (gettimeofday(&start_time, NULL))
|
|
ERR;
|
|
for (start[0] = 0; start[0] < DIM_LEN; start[0] += NUM_VALUES_TO_WRITE)
|
|
for (i = 0; i < NUMVARS; i++)
|
|
if (nc_put_vara_double(ncid, i, start, count, data)) ERR;
|
|
|
|
if (nc_close(ncid)) ERR;
|
|
if (gettimeofday(&end_time, NULL)) ERR;
|
|
if (timeval_subtract(&diff_time, &end_time, &start_time)) ERR;
|
|
printf("chunksize %d write time %d sec %d usec\n", chunksize[0],
|
|
(int)diff_time.tv_sec, (int)diff_time.tv_usec);
|
|
|
|
/* Reopen and check the file metadata. */
|
|
if (nc_open(FILE_NAME, 0, &ncid)) ERR;
|
|
if (nc_inq(ncid, &ndims, &nvars, &natts, &unlimdimid)) ERR;
|
|
if (ndims != NUMDIMS || nvars != NUMVARS || natts != 0 || unlimdimid != -1) ERR;
|
|
if (nc_inq_dimids(ncid, &ndims, dimids, 1)) ERR;
|
|
if (ndims != 1 || dimids[0] != 0) ERR;
|
|
if (nc_inq_dim(ncid, 0, name_in, &len)) ERR;
|
|
if (strcmp(name_in, DIM_NAME) || len != DIM_LEN) ERR;
|
|
for (i = 0; i < NUMVARS; i++)
|
|
{
|
|
if (nc_inq_var(ncid, i, name_in, &xtype, &ndims, dimids, &natts)) ERR;
|
|
if (strcmp(name_in, var_name[i]) || xtype != NC_DOUBLE || ndims != 1 ||
|
|
dimids[0] != 0 || natts != 0) ERR;
|
|
}
|
|
|
|
/* Read the data and time it. */
|
|
if (gettimeofday(&start_time, NULL)) ERR;
|
|
for (start[0] = 0; start[0] < DIM_LEN; start[0] += NUM_VALUES_TO_WRITE)
|
|
for (i = 0; i < NUMVARS; i++)
|
|
if (nc_get_vara_double(ncid, i, start, count, data_in)) ERR;
|
|
if (gettimeofday(&end_time, NULL)) ERR;
|
|
if (timeval_subtract(&diff_time, &end_time, &start_time)) ERR;
|
|
printf("chunksize %d read time %d sec %d usec\n", chunksize[0],
|
|
(int)diff_time.tv_sec, (int)diff_time.tv_usec);
|
|
|
|
/* Reread, and check the data (but don't time this operation). */
|
|
for (start[0] = DIM_LEN - NUM_VALUES_TO_WRITE; start[0] > 0; start[0] -= NUM_VALUES_TO_WRITE)
|
|
for (i = 0; i < NUMVARS; i++)
|
|
{
|
|
if (nc_get_vara_double(ncid, i, start, count, data_in)) ERR;
|
|
for (j = 0; j < NUM_VALUES_TO_WRITE; j++)
|
|
if (data[j] != data_in[j]) ERR;
|
|
}
|
|
|
|
|
|
if (nc_close(ncid)) ERR;
|
|
}
|
|
}
|
|
|
|
SUMMARIZE_ERR;
|
|
FINAL_RESULTS;
|
|
}
|
|
|