mirror/hdf5
2
0
Fork 0
mirror of https://github.com/HDFGroup/hdf5.git synced 2024-11-21 01:04:10 +08:00
Code Issues Packages Projects Releases Wiki Activity
c8645048e2
hdf5/perform/pio_engine.c
Quincey Koziol 4e23c80758 [svn-r7181] Purpose: 
Version update

Description:
    Removed 1.4 compatibility code in the library.

Platforms tested:
    FreeBSD 4.8 (sleipnir)
    h5committest
2003-07-07 14:02:46 -05:00

2142 lines
74 KiB
C
Raw Blame History

/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* Copyright by the Board of Trustees of the University of Illinois. *
* All rights reserved. *
* *
* This file is part of HDF5. The full HDF5 copyright notice, including *
* terms governing use, modification, and redistribution, is contained in *
* the files COPYING and Copyright.html. COPYING can be found at the root *
* of the source code distribution tree; Copyright.html can be found at the *
* root level of an installed copy of the electronic HDF5 document set and *
* is linked from the top-level documents page. It can also be found at *
* http://hdf.ncsa.uiuc.edu/HDF5/doc/Copyright.html. If you do not have *
* access to either file, you may request a copy from hdfhelp@ncsa.uiuc.edu. *
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
/*
* Author: Albert Cheng of NCSA, Oct 24, 2001.
*/
#include <sys/types.h>
#include <sys/stat.h>
#include <stdio.h>
#include <stdlib.h>
#include <fcntl.h>
#include <unistd.h>
#include <errno.h>
#include "hdf5.h"
#ifdef H5_HAVE_PARALLEL
#include <mpi.h>
#ifndef MPI_FILE_NULL /*MPIO may be defined in mpi.h already */
# include <mpio.h>
#endif /* !MPI_FILE_NULL */
#ifdef H5_HAVE_GPFS
# include <gpfs_fcntl.h>
#endif /* H5_HAVE_GPFS */
#include "pio_perf.h"
#include "pio_timer.h"
/* Macro definitions */
/* sizes of various items. these sizes won't change during program execution */
/* The following three must have the same type */
#define ELMT_SIZE (sizeof(unsigned char)) /* we're doing bytes */
#define ELMT_MPI_TYPE MPI_BYTE
#define ELMT_H5_TYPE H5T_NATIVE_UCHAR
#define GOTOERROR(errcode) { ret_code = errcode; goto done; }
#define GOTODONE { goto done; }
#define ERRMSG(mesg) { \
fprintf(stderr, "Proc %d: ", pio_mpi_rank_g); \
fprintf(stderr, "*** Assertion failed (%s) at line %4d in %s\n", \
mesg, (int)__LINE__, __FILE__); \
}
#define MSG(mesg) { \
fprintf(stderr, "Proc %d: ", pio_mpi_rank_g); \
fprintf(stderr, "(%s) at line %4d in %s\n", \
mesg, (int)__LINE__, __FILE__); \
}
/* verify: if val is false (0), print mesg. */
#define VRFY(val, mesg) do { \
if (!val) { \
ERRMSG(mesg); \
GOTOERROR(FAIL); \
} \
} while(0)
/* POSIX I/O macros */
#define POSIXCREATE(fn) HDopen(fn, O_CREAT|O_TRUNC|O_RDWR, 0600)
#define POSIXOPEN(fn, F) HDopen(fn, F, 0600)
#define POSIXCLOSE(F) HDclose(F)
#define POSIXSEEK(F,L) HDlseek(F, L, SEEK_SET)
#define POSIXWRITE(F,B,S) HDwrite(F,B,S)
#define POSIXREAD(F,B,S) HDread(F,B,S)
enum {
PIO_CREATE = 1,
PIO_WRITE = 2,
PIO_READ = 4
};
/* Global variables */
static int clean_file_g = -1; /*whether to cleanup temporary test */
/*files. -1 is not defined; */
/*0 is no cleanup; 1 is do cleanup */
/*
* In a parallel machine, the filesystem suitable for compiling is
* unlikely a parallel file system that is suitable for parallel I/O.
* There is no standard pathname for the parallel file system. /tmp
* is about the best guess.
*/
#ifndef HDF5_PARAPREFIX
# ifdef __PUMAGON__
/* For the PFS of TFLOPS */
# define HDF5_PARAPREFIX "pfs:/pfs_grande/multi/tmp_1"
# else
# define HDF5_PARAPREFIX ""
# endif /* __PUMAGON__ */
#endif /* !HDF5_PARAPREFIX */
#ifndef MIN
# define MIN(a,b) ((a) < (b) ? (a) : (b))
#endif /* !MIN */
/* the different types of file descriptors we can expect */
typedef union _file_descr {
int posixfd; /* POSIX file handle*/
MPI_File mpifd; /* MPI file */
hid_t h5fd; /* HDF5 file */
} file_descr;
/* local functions */
static char *pio_create_filename(iotype iot, const char *base_name,
char *fullname, size_t size);
static herr_t do_write(results *res, file_descr *fd, parameters *parms,
long ndsets, off_t nelmts, size_t buf_size, void *buffer);
static herr_t do_read(results *res, file_descr *fd, parameters *parms,
long ndsets, off_t nelmts, size_t buf_size, void *buffer /*out*/);
static herr_t do_fopen(parameters *param, char *fname, file_descr *fd /*out*/,
int flags);
static herr_t do_fclose(iotype iot, file_descr *fd);
static void do_cleanupfile(iotype iot, char *fname);
/* GPFS-specific functions */
#ifdef H5_HAVE_GPFS
static void gpfs_access_range(int handle, off_t start, off_t length, int is_write);
static void gpfs_free_range(int handle, off_t start, off_t length);
static void gpfs_clear_file_cache(int handle);
static void gpfs_cancel_hints(int handle);
static void gpfs_start_data_shipping(int handle, int num_insts);
static void gpfs_start_data_ship_map(int handle, int partition_size,
int agent_count, int *agent_node_num);
static void gpfs_stop_data_shipping(int handle);
static void gpfs_invalidate_file_cache(const char *filename);
#endif /* H5_HAVE_GPFS */
/*
* Function: do_pio
* Purpose: PIO Engine where Parallel IO are executed.
* Return: results
* Programmer: Albert Cheng, Bill Wendling 2001/12/12
* Modifications:
*/
results
do_pio(parameters param)
{
/* return codes */
herr_t ret_code = 0; /*return code */
results res;
file_descr fd;
iotype iot;
char fname[FILENAME_MAX];
long nf;
long ndsets;
off_t nbytes; /* Number of bytes per dataset */
char *buffer = NULL; /*data buffer pointer */
size_t buf_size; /*data buffer size in bytes */
size_t blk_size; /*data block size in bytes */
/* HDF5 variables */
herr_t hrc; /*HDF5 return code */
/* Sanity check parameters */
/* IO type */
iot = param.io_type;
switch (iot) {
case MPIO:
fd.mpifd = MPI_FILE_NULL;
res.timers = pio_time_new(MPI_TIMER);
break;
case POSIXIO:
fd.posixfd = -1;
res.timers = pio_time_new(MPI_TIMER);
break;
case PHDF5:
fd.h5fd = -1;
res.timers = pio_time_new(MPI_TIMER);
break;
default:
/* unknown request */
fprintf(stderr, "Unknown IO type request (%d)\n", iot);
GOTOERROR(FAIL);
}
ndsets = param.num_dsets; /* number of datasets per file */
nbytes = param.num_bytes; /* number of bytes per dataset */
buf_size = param.buf_size;
blk_size = param.blk_size;
if (param.num_files < 0 ) {
fprintf(stderr,
"number of files must be >= 0 (%ld)\n",
param.num_files);
GOTOERROR(FAIL);
}
if (ndsets < 0 ) {
fprintf(stderr,
"number of datasets per file must be >= 0 (%ld)\n",
ndsets);
GOTOERROR(FAIL);
}
if (param.num_procs <= 0 ) {
fprintf(stderr,
"maximum number of process to use must be > 0 (%d)\n",
param.num_procs);
GOTOERROR(FAIL);
}
/* Validate transfer buffer size & block size*/
if(blk_size<=0) {
HDfprintf(stderr,
"Transfer block size (%Hd) must be > 0\n", (long_long)blk_size);
GOTOERROR(FAIL);
}
if(buf_size<=0) {
HDfprintf(stderr,
"Transfer buffer size (%Hd) must be > 0\n", (long_long)buf_size);
GOTOERROR(FAIL);
}
if ((buf_size % blk_size) != 0){
HDfprintf(stderr,
"Transfer buffer size (%Hd) must be a multiple of the "
"interleaved I/O block size (%Hd)\n",
(long_long)buf_size, (long_long)blk_size);
GOTOERROR(FAIL);
}
if((nbytes%pio_mpi_nprocs_g)!=0) {
HDfprintf(stderr,
"Dataset size (%Hd) must be a multiple of the "
"number of processes (%d)\n",
(long_long)nbytes, pio_mpi_nprocs_g);
GOTOERROR(FAIL);
}
if(((nbytes/pio_mpi_nprocs_g)%buf_size)!=0) {
HDfprintf(stderr,
"Dataset size/process (%Hd) must be a multiple of the "
"trasfer buffer size (%Hd)\n",
(long_long)(nbytes/pio_mpi_nprocs_g), (long_long)buf_size);
GOTOERROR(FAIL);
}
/* Allocate transfer buffer */
if ((buffer = malloc(buf_size)) == NULL){
HDfprintf(stderr, "malloc for transfer buffer size (%Hd) failed\n",
(long_long)(buf_size));
GOTOERROR(FAIL);
}
if (pio_debug_level >= 4) {
int myrank;
MPI_Comm_rank(pio_comm_g, &myrank);
/* output all of the times for all iterations */
if (myrank == 0)
fprintf(output, "Timer details:\n");
}
for (nf = 1; nf <= param.num_files; nf++) {
/*
* Write performance measurement
*/
/* Open file for write */
char base_name[256];
sprintf(base_name, "#pio_tmp_%lu", nf);
pio_create_filename(iot, base_name, fname, sizeof(fname));
if (pio_debug_level > 0)
HDfprintf(output, "rank %d: data filename=%s\n",
pio_mpi_rank_g, fname);
/* Need barrier to make sure everyone starts at the same time */
MPI_Barrier(pio_comm_g);
set_time(res.timers, HDF5_GROSS_WRITE_FIXED_DIMS, START);
hrc = do_fopen(&param, fname, &fd, PIO_CREATE | PIO_WRITE);
VRFY((hrc == SUCCESS), "do_fopen failed");
set_time(res.timers, HDF5_FINE_WRITE_FIXED_DIMS, START);
hrc = do_write(&res, &fd, &param, ndsets, nbytes, buf_size, buffer);
set_time(res.timers, HDF5_FINE_WRITE_FIXED_DIMS, STOP);
VRFY((hrc == SUCCESS), "do_write failed");
/* Close file for write */
hrc = do_fclose(iot, &fd);
set_time(res.timers, HDF5_GROSS_WRITE_FIXED_DIMS, STOP);
VRFY((hrc == SUCCESS), "do_fclose failed");
if (!param.h5_write_only) {
/*
* Read performance measurement
*/
/* Need barrier to make sure everyone is done writing and has
* closed the file. Also to make sure everyone starts reading
* at the same time.
*/
MPI_Barrier(pio_comm_g);
/* Open file for read */
set_time(res.timers, HDF5_GROSS_READ_FIXED_DIMS, START);
hrc = do_fopen(&param, fname, &fd, PIO_READ);
VRFY((hrc == SUCCESS), "do_fopen failed");
set_time(res.timers, HDF5_FINE_READ_FIXED_DIMS, START);
hrc = do_read(&res, &fd, &param, ndsets, nbytes, buf_size, buffer);
set_time(res.timers, HDF5_FINE_READ_FIXED_DIMS, STOP);
VRFY((hrc == SUCCESS), "do_read failed");
/* Close file for read */
hrc = do_fclose(iot, &fd);
set_time(res.timers, HDF5_GROSS_READ_FIXED_DIMS, STOP);
VRFY((hrc == SUCCESS), "do_fclose failed");
}
/* Need barrier to make sure everyone is done with the file */
/* before it may be removed by do_cleanupfile */
MPI_Barrier(pio_comm_g);
do_cleanupfile(iot, fname);
}
done:
/* clean up */
/* release HDF5 objects */
/* close any opened files */
/* no remove(fname) because that should have happened normally. */
switch (iot) {
case POSIXIO:
if (fd.posixfd != -1)
hrc = do_fclose(iot, &fd);
break;
case MPIO:
if (fd.mpifd != MPI_FILE_NULL)
hrc = do_fclose(iot, &fd);
break;
case PHDF5:
if (fd.h5fd != -1)
hrc = do_fclose(iot, &fd);
break;
}
/* release generic resources */
if(buffer)
free(buffer);
res.ret_code = ret_code;
return res;
}
/*
* Function: pio_create_filename
* Purpose: Create a new filename to write to. Determine the correct
* suffix to append to the filename by the type of I/O we're
* doing. Also, place in the /tmp/{$USER,$LOGIN} directory if
* USER or LOGIN are specified in the environment.
* Return: Pointer to filename or NULL
* Programmer: Bill Wendling, 21. November 2001
* Modifications:
*/
static char *
pio_create_filename(iotype iot, const char *base_name, char *fullname, size_t size)
{
const char *prefix, *suffix="";
char *ptr, last = '\0';
size_t i, j;
if (!base_name || !fullname || size < 1)
return NULL;
memset(fullname, 0, size);
switch (iot) {
case POSIXIO:
suffix = ".posix";
break;
case MPIO:
suffix = ".mpio";
break;
case PHDF5:
suffix = ".h5";
break;
}
/* First use the environment variable and then try the constant */
prefix = getenv("HDF5_PARAPREFIX");
#ifdef HDF5_PARAPREFIX
if (!prefix)
prefix = HDF5_PARAPREFIX;
#endif /* HDF5_PARAPREFIX */
/* Prepend the prefix value to the base name */
if (prefix && *prefix) {
/* If the prefix specifies the HDF5_PARAPREFIX directory, then
* default to using the "/tmp/$USER" or "/tmp/$LOGIN"
* directory instead. */
register char *user, *login, *subdir;
user = getenv("USER");
login = getenv("LOGIN");
subdir = (user ? user : login);
if (subdir) {
for (i = 0; i < size && prefix[i]; i++)
fullname[i] = prefix[i];
fullname[i++] = '/';
for (j = 0; i < size && subdir[j]; i++, j++)
fullname[i] = subdir[j];
} else {
/* We didn't append the prefix yet */
strncpy(fullname, prefix, MIN(strlen(prefix), size));
}
if ((strlen(fullname) + strlen(base_name) + 1) < size) {
/* Append the base_name with a slash first. Multiple slashes are
* handled below. */
h5_stat_t buf;
if (HDstat(fullname, &buf) < 0)
/* The directory doesn't exist just yet */
if (mkdir(fullname, (mode_t)0755) < 0 && errno != EEXIST) {
/* We couldn't make the "/tmp/${USER,LOGIN}" subdirectory.
* Default to PREFIX's original prefix value. */
strcpy(fullname, prefix);
}
strcat(fullname, "/");
strcat(fullname, base_name);
} else {
/* Buffer is too small */
return NULL;
}
} else if (strlen(base_name) >= size) {
/* Buffer is too small */
return NULL;
} else {
strcpy(fullname, base_name);
}
/* Append a suffix */
if (suffix) {
if (strlen(fullname) + strlen(suffix) >= size)
return NULL;
strcat(fullname, suffix);
}
/* Remove any double slashes in the filename */
for (ptr = fullname, i = j = 0; ptr && i < size; i++, ptr++) {
if (*ptr != '/' || last != '/')
fullname[j++] = *ptr;
last = *ptr;
}
return fullname;
}
/*
* Function: do_write
* Purpose: Write the required amount of data to the file.
* Return: SUCCESS or FAIL
* Programmer: Albert Cheng, Bill Wendling, 2001/12/13
* Modifications:
*/
static herr_t
do_write(results *res, file_descr *fd, parameters *parms, long ndsets,
off_t nbytes, size_t buf_size, void *buffer)
{
int ret_code = SUCCESS;
int rc; /*routine return code */
long ndset;
size_t blk_size; /* The block size to subdivide the xfer buffer into */
off_t nbytes_xfer; /* Total number of bytes transferred so far */
size_t nbytes_toxfer; /* Number of bytes to transfer a particular time */
char dname[64];
off_t dset_offset=0; /*dataset offset in a file */
off_t bytes_begin; /*first elmt this process transfer */
off_t bytes_count; /*number of elmts this process transfer */
unsigned char *buf_p; /* Current buffer pointer */
/* POSIX variables */
off_t file_offset; /* File offset of the next transfer */
off_t posix_file_offset; /* Base file offset of the next transfer */
/* MPI variables */
MPI_Offset mpi_file_offset;/* Base file offset of the next transfer*/
MPI_Offset mpi_offset; /* Offset in MPI file */
MPI_Datatype mpi_file_type; /* MPI derived type for file */
MPI_Datatype mpi_blk_type; /* MPI derived type for buffer */
MPI_Status mpi_status;
int mrc; /* MPI return code */
/* HDF5 variables */
herr_t hrc; /*HDF5 return code */
hsize_t h5dims[1]; /*dataset dim sizes */
hid_t h5dset_space_id = -1; /*dataset space ID */
hid_t h5mem_space_id = -1; /*memory dataspace ID */
hid_t h5ds_id = -1; /*dataset handle */
hsize_t h5block[1]; /*dataspace selection */
hsize_t h5stride[1];
hsize_t h5count[1];
hssize_t h5start[1];
hssize_t h5offset[1]; /* Selection offset within dataspace */
hid_t h5dcpl = -1; /* Dataset creation property list */
hid_t h5dxpl = -1; /* Dataset transfer property list */
/* Get the parameters from the parameter block */
blk_size=parms->blk_size;
/* Prepare buffer for verifying data */
if (parms->verify)
memset(buffer,pio_mpi_rank_g,buf_size);
/* There are two kinds of transfer patterns, contiguous and interleaved.
* Let 0,1,2,...,n be data accessed by process 0,1,2,...,n
* where n is rank of the last process.
* In contiguous pattern, data are accessed as
* 000...111...222...nnn...
* In interleaved pattern, data are accessed as
* 012...n012...n...
* These are all in the scope of one dataset.
*/
if (parms->interleaved==0) {
/* Contiguous Pattern: */
bytes_begin = (off_t)(((double)nbytes*pio_mpi_rank_g)/pio_mpi_nprocs_g);
} /* end if */
else {
/* Interleaved Pattern: */
bytes_begin = (off_t)(blk_size*pio_mpi_rank_g);
} /* end else */
/* Calculate the total number of bytes (bytes_count) to be
* transferred by this process. It may be different for different
* transfer pattern due to rounding to integral values.
*/
/*
* Calculate the beginning bytes of this process and the next.
* bytes_count is the difference between these two beginnings.
* This way, it eliminates any rounding errors.
* (This is tricky, don't mess with the formula, rounding errors
* can easily get introduced) */
bytes_count = (off_t)(((double)nbytes*(pio_mpi_rank_g+1)) / pio_mpi_nprocs_g)
- (off_t)(((double)nbytes*pio_mpi_rank_g) / pio_mpi_nprocs_g);
/* debug */
if (pio_debug_level >= 4) {
HDprint_rank(output);
HDfprintf(output, "Debug(do_write): "
"buf_size=%Hd, bytes_begin=%Hd, bytes_count=%Hd\n",
(long_long)buf_size, (long_long)bytes_begin,
(long_long)bytes_count);
}
/* I/O Access specific setup */
switch (parms->io_type) {
case POSIXIO:
/* No extra setup */
break;
case MPIO: /* MPI-I/O setup */
/* Build block's derived type */
mrc = MPI_Type_contiguous((int)blk_size,
MPI_BYTE, &mpi_blk_type);
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_CREATE");
/* Build file's derived type */
mrc = MPI_Type_vector((int)(buf_size/blk_size), (int)1,
(int)pio_mpi_nprocs_g, mpi_blk_type, &mpi_file_type);
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_CREATE");
/* Commit file type */
mrc = MPI_Type_commit( &mpi_file_type );
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_COMMIT");
/* Commit buffer type */
mrc = MPI_Type_commit( &mpi_blk_type );
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_COMMIT");
break;
case PHDF5: /* HDF5 setup */
if(nbytes>0) {
/* define a contiquous dataset of nbytes native bytes */
h5dims[0] = nbytes;
h5dset_space_id = H5Screate_simple(1, h5dims, NULL);
VRFY((h5dset_space_id >= 0), "H5Screate_simple");
/* Set up the file dset space id to select the pattern to access */
if (parms->interleaved==0){
/* Contiguous pattern */
h5start[0] = bytes_begin;
h5stride[0] = h5block[0] = blk_size;
h5count[0] = buf_size/blk_size;
} /* end if */
else {
/* Interleaved access pattern */
/* Skip offset over blocks of other processes */
h5start[0] = bytes_begin;
h5stride[0] = blk_size*pio_mpi_nprocs_g;
h5block[0] = blk_size;
h5count[0] = buf_size/blk_size;
} /* end else */
hrc = H5Sselect_hyperslab(h5dset_space_id, H5S_SELECT_SET,
h5start, h5stride, h5count, h5block);
VRFY((hrc >= 0), "H5Sselect_hyperslab");
} /* end if */
else {
h5dset_space_id = H5Screate(H5S_SCALAR);
VRFY((h5dset_space_id >= 0), "H5Screate");
} /* end else */
/* Create the memory dataspace that corresponds to the xfer buffer */
if(buf_size>0) {
h5dims[0] = buf_size;
h5mem_space_id = H5Screate_simple(1, h5dims, NULL);
VRFY((h5mem_space_id >= 0), "H5Screate_simple");
} /* end if */
else {
h5mem_space_id = H5Screate(H5S_SCALAR);
VRFY((h5mem_space_id >= 0), "H5Screate");
} /* end else */
/* Create the dataset transfer property list */
h5dxpl = H5Pcreate(H5P_DATASET_XFER);
if (h5dxpl < 0) {
fprintf(stderr, "HDF5 Property List Create failed\n");
GOTOERROR(FAIL);
}
/* Change to collective I/O, if asked */
if(parms->collective) {
hrc = H5Pset_dxpl_mpio(h5dxpl, H5FD_MPIO_COLLECTIVE);
if (hrc < 0) {
fprintf(stderr, "HDF5 Property List Set failed\n");
GOTOERROR(FAIL);
} /* end if */
} /* end if */
break;
} /* end switch */
for (ndset = 1; ndset <= ndsets; ++ndset) {
/* Calculate dataset offset within a file */
/* create dataset */
switch (parms->io_type) {
case POSIXIO:
case MPIO:
/* both posix and mpi io just need dataset offset in file*/
dset_offset = (ndset - 1) * nbytes;
break;
case PHDF5:
h5dcpl = H5Pcreate(H5P_DATASET_CREATE);
if (h5dcpl < 0) {
fprintf(stderr, "HDF5 Property List Create failed\n");
GOTOERROR(FAIL);
}
/* Make the dataset chunked if asked */
if(parms->h5_use_chunks) {
/* Set the chunk size to be the same as the buffer size */
h5dims[0] = buf_size;
hrc = H5Pset_chunk(h5dcpl, 1, h5dims);
if (hrc < 0) {
fprintf(stderr, "HDF5 Property List Set failed\n");
GOTOERROR(FAIL);
} /* end if */
} /* end if */
sprintf(dname, "Dataset_%ld", ndset);
h5ds_id = H5Dcreate(fd->h5fd, dname, ELMT_H5_TYPE,
h5dset_space_id, h5dcpl);
if (h5ds_id < 0) {
fprintf(stderr, "HDF5 Dataset Create failed\n");
GOTOERROR(FAIL);
}
hrc = H5Pclose(h5dcpl);
/* verifying the close of the dcpl */
if (hrc < 0) {
fprintf(stderr, "HDF5 Property List Close failed\n");
GOTOERROR(FAIL);
}
break;
}
/* The task is to transfer bytes_count bytes, starting at
* bytes_begin position, using transfer buffer of buf_size bytes.
* If interleaved, select buf_size at a time, in round robin
* fashion, according to number of process. Otherwise, select
* all bytes_count in contiguous.
*/
nbytes_xfer = 0 ;
/* Set base file offset for all I/O patterns and POSIX access */
posix_file_offset = dset_offset + bytes_begin;
/* Set base file offset for all I/O patterns and MPI access */
mpi_file_offset = (MPI_Offset)(dset_offset + bytes_begin);
/* Start "raw data" write timer */
set_time(res->timers, HDF5_RAW_WRITE_FIXED_DIMS, START);
while (nbytes_xfer < bytes_count){
/* Write */
/* Calculate offset of write within a dataset/file */
switch (parms->io_type) {
case POSIXIO:
/* Contiguous pattern */
if (parms->interleaved==0) {
/* Compute file offset */
file_offset = posix_file_offset + (off_t)nbytes_xfer;
/* only care if seek returns error */
rc = POSIXSEEK(fd->posixfd, file_offset) < 0 ? -1 : 0;
VRFY((rc==0), "POSIXSEEK");
/* check if all bytes are written */
rc = ((ssize_t)buf_size ==
POSIXWRITE(fd->posixfd, buffer, buf_size));
VRFY((rc != 0), "POSIXWRITE");
/* Advance global offset in dataset */
nbytes_xfer+=buf_size;
} /* end if */
/* Interleaved access pattern */
else {
/* Set the base of user's buffer */
buf_p=(unsigned char *)buffer;
/* Set the number of bytes to transfer this time */
nbytes_toxfer = buf_size;
/* Loop over the buffers to write */
while(nbytes_toxfer>0) {
/* Skip offset over blocks of other processes */
file_offset = posix_file_offset +
(off_t)(nbytes_xfer*pio_mpi_nprocs_g);
/* only care if seek returns error */
rc = POSIXSEEK(fd->posixfd, file_offset) < 0 ? -1 : 0;
VRFY((rc==0), "POSIXSEEK");
/* check if all bytes are written */
rc = ((ssize_t)blk_size ==
POSIXWRITE(fd->posixfd, buf_p, blk_size));
VRFY((rc != 0), "POSIXWRITE");
/* Advance location in buffer */
buf_p+=blk_size;
/* Advance global offset in dataset */
nbytes_xfer+=blk_size;
/* Decrement number of bytes left this time */
nbytes_toxfer-=blk_size;
} /* end while */
} /* end else */
break;
case MPIO:
/* Independent file access */
if(parms->collective==0) {
/* Contiguous pattern */
if (parms->interleaved==0){
/* Compute offset in file */
mpi_offset = mpi_file_offset +
nbytes_xfer;
/* Perform independent write */
mrc = MPI_File_write_at(fd->mpifd, mpi_offset, buffer,
(int)(buf_size/blk_size), mpi_blk_type,
&mpi_status);
VRFY((mrc==MPI_SUCCESS), "MPIO_WRITE");
/* Advance global offset in dataset */
nbytes_xfer+=buf_size;
} /* end if */
/* Interleaved access pattern */
else {
/* Set the base of user's buffer */
buf_p=(unsigned char *)buffer;
/* Set the number of bytes to transfer this time */
nbytes_toxfer = buf_size;
/* Loop over the buffers to write */
while(nbytes_toxfer>0) {
/* Skip offset over blocks of other processes */
mpi_offset = mpi_file_offset +
(nbytes_xfer*pio_mpi_nprocs_g);
/* Perform independent write */
mrc = MPI_File_write_at(fd->mpifd, mpi_offset, buf_p,
(int)1, mpi_blk_type, &mpi_status);
VRFY((mrc==MPI_SUCCESS), "MPIO_WRITE");
/* Advance location in buffer */
buf_p+=blk_size;
/* Advance global offset in dataset */
nbytes_xfer+=blk_size;
/* Decrement number of bytes left this time */
nbytes_toxfer-=blk_size;
} /* end while */
} /* end else */
} /* end if */
/* Collective file access */
else {
/* Contiguous access pattern */
if (parms->interleaved==0){
/* Compute offset in file */
mpi_offset = mpi_file_offset +
nbytes_xfer;
/* Perform independent write */
mrc = MPI_File_write_at_all(fd->mpifd, mpi_offset, buffer,
(int)(buf_size/blk_size), mpi_blk_type, &mpi_status);
VRFY((mrc==MPI_SUCCESS), "MPIO_WRITE");
/* Advance global offset in dataset */
nbytes_xfer+=buf_size;
} /* end if */
/* Interleaved access pattern */
else {
/* Compute offset in file */
mpi_offset = mpi_file_offset +
(nbytes_xfer*pio_mpi_nprocs_g);
/* Set the file view */
mrc = MPI_File_set_view(fd->mpifd, mpi_offset, mpi_blk_type,
mpi_file_type, (char*)"native", h5_io_info_g);
VRFY((mrc==MPI_SUCCESS), "MPIO_VIEW");
/* Perform write */
mrc = MPI_File_write_at_all(fd->mpifd, 0, buffer,
(int)(buf_size/blk_size), mpi_blk_type, &mpi_status);
VRFY((mrc==MPI_SUCCESS), "MPIO_WRITE");
/* Advance global offset in dataset */
nbytes_xfer+=buf_size;
} /* end else */
} /* end else */
break;
case PHDF5:
/* Set up the file dset space id to move the selection to process */
if (parms->interleaved==0){
/* Contiguous pattern */
h5offset[0] = nbytes_xfer;
} /* end if */
else {
/* Interleaved access pattern */
/* Skip offset over blocks of other processes */
h5offset[0] = (nbytes_xfer*pio_mpi_nprocs_g);
} /* end else */
hrc = H5Soffset_simple(h5dset_space_id, h5offset);
VRFY((hrc >= 0), "H5Soffset_simple");
/* Write the buffer out */
hrc = H5Dwrite(h5ds_id, ELMT_H5_TYPE, h5mem_space_id,
h5dset_space_id, h5dxpl, buffer);
VRFY((hrc >= 0), "H5Dwrite");
/* Increment number of bytes transferred */
nbytes_xfer += buf_size;
break;
} /* switch (parms->io_type) */
} /* end while */
/* Stop "raw data" write timer */
set_time(res->timers, HDF5_RAW_WRITE_FIXED_DIMS, STOP);
/* Calculate write time */
/* Close dataset. Only HDF5 needs to do an explicit close. */
if (parms->io_type == PHDF5) {
hrc = H5Dclose(h5ds_id);
if (hrc < 0) {
fprintf(stderr, "HDF5 Dataset Close failed\n");
GOTOERROR(FAIL);
}
h5ds_id = -1;
} /* end if */
} /* end for */
done:
/* release MPI-I/O objects */
if (parms->io_type == MPIO) {
/* Free file type */
mrc = MPI_Type_free( &mpi_file_type );
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_FREE");
/* Free buffer type */
mrc = MPI_Type_free( &mpi_blk_type );
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_FREE");
} /* end if */
/* release HDF5 objects */
if (h5dset_space_id != -1) {
hrc = H5Sclose(h5dset_space_id);
if (hrc < 0){
fprintf(stderr, "HDF5 Dataset Space Close failed\n");
ret_code = FAIL;
} else {
h5dset_space_id = -1;
}
}
if (h5mem_space_id != -1) {
hrc = H5Sclose(h5mem_space_id);
if (hrc < 0) {
fprintf(stderr, "HDF5 Memory Space Close failed\n");
ret_code = FAIL;
} else {
h5mem_space_id = -1;
}
}
if (h5dxpl != -1) {
hrc = H5Pclose(h5dxpl);
if (hrc < 0) {
fprintf(stderr, "HDF5 Dataset Transfer Property List Close failed\n");
ret_code = FAIL;
} else {
h5dxpl = -1;
}
}
return ret_code;
}
/*
* Function: do_read
* Purpose: read the required amount of data from the file.
* Return: SUCCESS or FAIL
* Programmer: Albert Cheng 2001/12/13
* Modifications:
*/
static herr_t
do_read(results *res, file_descr *fd, parameters *parms, long ndsets,
off_t nbytes, size_t buf_size, void *buffer /*out*/)
{
int ret_code = SUCCESS;
int rc; /*routine return code */
long ndset;
size_t blk_size; /* The block size to subdivide the xfer buffer into */
off_t nbytes_xfer; /* Total number of bytes transferred so far */
size_t nbytes_toxfer; /* Number of bytes to transfer a particular time */
char dname[64];
off_t dset_offset=0; /*dataset offset in a file */
off_t bytes_begin; /*first elmt this process transfer */
off_t bytes_count; /*number of elmts this process transfer */
unsigned char *buf_p; /* Current buffer pointer */
/* POSIX variables */
off_t file_offset; /* File offset of the next transfer */
off_t posix_file_offset; /* Base file offset of the next transfer */
/* MPI variables */
MPI_Offset mpi_file_offset;/* Base file offset of the next transfer*/
MPI_Offset mpi_offset; /* Offset in MPI file */
MPI_Datatype mpi_file_type; /* MPI derived type for file */
MPI_Datatype mpi_blk_type; /* MPI derived type for buffer */
MPI_Status mpi_status;
int mrc; /* MPI return code */
/* HDF5 variables */
herr_t hrc; /*HDF5 return code */
hsize_t h5dims[1]; /*dataset dim sizes */
hid_t h5dset_space_id = -1; /*dataset space ID */
hid_t h5mem_space_id = -1; /*memory dataspace ID */
hid_t h5ds_id = -1; /*dataset handle */
hsize_t h5block[1]; /*dataspace selection */
hsize_t h5stride[1];
hsize_t h5count[1];
hssize_t h5start[1];
hssize_t h5offset[1]; /* Selection offset within dataspace */
hid_t h5dxpl = -1; /* Dataset transfer property list */
/* Get the parameters from the parameter block */
blk_size=parms->blk_size;
/* There are two kinds of transfer patterns, contiguous and interleaved.
* Let 0,1,2,...,n be data accessed by process 0,1,2,...,n
* where n is rank of the last process.
* In contiguous pattern, data are accessed as
* 000...111...222...nnn...
* In interleaved pattern, data are accessed as
* 012...n012...n...
* These are all in the scope of one dataset.
*/
if (parms->interleaved==0) {
/* Contiguous Pattern: */
bytes_begin = (off_t)(((double)nbytes*pio_mpi_rank_g)/pio_mpi_nprocs_g);
} /* end if */
else {
/* Interleaved Pattern: */
bytes_begin = (off_t)(blk_size*pio_mpi_rank_g);
} /* end else */
/* Calculate the total number of bytes (bytes_count) to be
* transferred by this process. It may be different for different
* transfer pattern due to rounding to integral values.
*/
/*
* Calculate the beginning bytes of this process and the next.
* bytes_count is the difference between these two beginnings.
* This way, it eliminates any rounding errors.
* (This is tricky, don't mess with the formula, rounding errors
* can easily get introduced) */
bytes_count = (off_t)(((double)nbytes*(pio_mpi_rank_g+1)) / pio_mpi_nprocs_g)
- (off_t)(((double)nbytes*pio_mpi_rank_g) / pio_mpi_nprocs_g);
/* debug */
if (pio_debug_level >= 4) {
HDprint_rank(output);
HDfprintf(output, "Debug(do_read): "
"buf_size=%Hd, bytes_begin=%Hd, bytes_count=%Hd\n",
(long_long)buf_size, (long_long)bytes_begin,
(long_long)bytes_count);
}
/* I/O Access specific setup */
switch (parms->io_type) {
case POSIXIO:
/* No extra setup */
break;
case MPIO: /* MPI-I/O setup */
/* Build block's derived type */
mrc = MPI_Type_contiguous((int)blk_size,
MPI_BYTE, &mpi_blk_type);
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_CREATE");
/* Build file's derived type */
mrc = MPI_Type_vector((int)(buf_size/blk_size), (int)1,
(int)pio_mpi_nprocs_g, mpi_blk_type, &mpi_file_type);
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_CREATE");
/* Commit file type */
mrc = MPI_Type_commit( &mpi_file_type );
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_COMMIT");
/* Commit buffer type */
mrc = MPI_Type_commit( &mpi_blk_type );
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_COMMIT");
break;
case PHDF5: /* HDF5 setup */
if(nbytes>0) {
/* define a contiquous dataset of nbytes native bytes */
h5dims[0] = nbytes;
h5dset_space_id = H5Screate_simple(1, h5dims, NULL);
VRFY((h5dset_space_id >= 0), "H5Screate_simple");
/* Set up the file dset space id to select the pattern to access */
if (parms->interleaved==0){
/* Contiguous pattern */
h5start[0] = bytes_begin;
h5stride[0] = h5block[0] = blk_size;
h5count[0] = buf_size/blk_size;
} /* end if */
else {
/* Interleaved access pattern */
/* Skip offset over blocks of other processes */
h5start[0] = bytes_begin;
h5stride[0] = blk_size*pio_mpi_nprocs_g;
h5block[0] = blk_size;
h5count[0] = buf_size/blk_size;
} /* end else */
hrc = H5Sselect_hyperslab(h5dset_space_id, H5S_SELECT_SET,
h5start, h5stride, h5count, h5block);
VRFY((hrc >= 0), "H5Sselect_hyperslab");
} /* end if */
else {
h5dset_space_id = H5Screate(H5S_SCALAR);
VRFY((h5dset_space_id >= 0), "H5Screate");
} /* end else */
/* Create the memory dataspace that corresponds to the xfer buffer */
if(buf_size>0) {
h5dims[0] = buf_size;
h5mem_space_id = H5Screate_simple(1, h5dims, NULL);
VRFY((h5mem_space_id >= 0), "H5Screate_simple");
} /* end if */
else {
h5mem_space_id = H5Screate(H5S_SCALAR);
VRFY((h5mem_space_id >= 0), "H5Screate");
} /* end else */
/* Create the dataset transfer property list */
h5dxpl = H5Pcreate(H5P_DATASET_XFER);
if (h5dxpl < 0) {
fprintf(stderr, "HDF5 Property List Create failed\n");
GOTOERROR(FAIL);
}
/* Change to collective I/O, if asked */
if(parms->collective) {
hrc = H5Pset_dxpl_mpio(h5dxpl, H5FD_MPIO_COLLECTIVE);
if (hrc < 0) {
fprintf(stderr, "HDF5 Property List Set failed\n");
GOTOERROR(FAIL);
} /* end if */
} /* end if */
break;
} /* end switch */
for (ndset = 1; ndset <= ndsets; ++ndset) {
/* Calculate dataset offset within a file */
/* create dataset */
switch (parms->io_type) {
case POSIXIO:
case MPIO:
/* both posix and mpi io just need dataset offset in file*/
dset_offset = (ndset - 1) * nbytes;
break;
case PHDF5:
sprintf(dname, "Dataset_%ld", ndset);
h5ds_id = H5Dopen(fd->h5fd, dname);
if (h5ds_id < 0) {
fprintf(stderr, "HDF5 Dataset open failed\n");
GOTOERROR(FAIL);
}
break;
}
/* The task is to transfer bytes_count bytes, starting at
* bytes_begin position, using transfer buffer of buf_size bytes.
* If interleaved, select buf_size at a time, in round robin
* fashion, according to number of process. Otherwise, select
* all bytes_count in contiguous.
*/
nbytes_xfer = 0 ;
/* Set base file offset for all I/O patterns and POSIX access */
posix_file_offset = dset_offset + bytes_begin;
/* Set base file offset for all I/O patterns and MPI access */
mpi_file_offset = (MPI_Offset)(dset_offset + bytes_begin);
/* Start "raw data" read timer */
set_time(res->timers, HDF5_RAW_READ_FIXED_DIMS, START);
while (nbytes_xfer < bytes_count){
/* Read */
/* Calculate offset of read within a dataset/file */
switch (parms->io_type) {
case POSIXIO:
/* Contiguous pattern */
if (parms->interleaved==0) {
/* Compute file offset */
file_offset = posix_file_offset + (off_t)nbytes_xfer;
/* only care if seek returns error */
rc = POSIXSEEK(fd->posixfd, file_offset) < 0 ? -1 : 0;
VRFY((rc==0), "POSIXSEEK");
/* check if all bytes are written */
rc = ((ssize_t)buf_size ==
POSIXREAD(fd->posixfd, buffer, buf_size));
VRFY((rc != 0), "POSIXREAD");
/* Advance global offset in dataset */
nbytes_xfer+=buf_size;
} /* end if */
/* Interleaved access pattern */
else {
/* Set the base of user's buffer */
buf_p=(unsigned char *)buffer;
/* Set the number of bytes to transfer this time */
nbytes_toxfer = buf_size;
/* Loop over the buffers to read */
while(nbytes_toxfer>0) {
/* Skip offset over blocks of other processes */
file_offset = posix_file_offset +
(off_t)(nbytes_xfer*pio_mpi_nprocs_g);
/* only care if seek returns error */
rc = POSIXSEEK(fd->posixfd, file_offset) < 0 ? -1 : 0;
VRFY((rc==0), "POSIXSEEK");
/* check if all bytes are written */
rc = ((ssize_t)blk_size ==
POSIXREAD(fd->posixfd, buf_p, blk_size));
VRFY((rc != 0), "POSIXREAD");
/* Advance location in buffer */
buf_p+=blk_size;
/* Advance global offset in dataset */
nbytes_xfer+=blk_size;
/* Decrement number of bytes left this time */
nbytes_toxfer-=blk_size;
} /* end while */
} /* end else */
break;
case MPIO:
/* Independent file access */
if(parms->collective==0) {
/* Contiguous pattern */
if (parms->interleaved==0){
/* Compute offset in file */
mpi_offset = mpi_file_offset +
nbytes_xfer;
/* Perform independent read */
mrc = MPI_File_read_at(fd->mpifd, mpi_offset, buffer,
(int)(buf_size/blk_size), mpi_blk_type,
&mpi_status);
VRFY((mrc==MPI_SUCCESS), "MPIO_READ");
/* Advance global offset in dataset */
nbytes_xfer+=buf_size;
} /* end if */
/* Interleaved access pattern */
else {
/* Set the base of user's buffer */
buf_p=(unsigned char *)buffer;
/* Set the number of bytes to transfer this time */
nbytes_toxfer = buf_size;
/* Loop over the buffers to read */
while(nbytes_toxfer>0) {
/* Skip offset over blocks of other processes */
mpi_offset = mpi_file_offset +
(nbytes_xfer*pio_mpi_nprocs_g);
/* Perform independent read */
mrc = MPI_File_read_at(fd->mpifd, mpi_offset, buf_p,
(int)1, mpi_blk_type, &mpi_status);
VRFY((mrc==MPI_SUCCESS), "MPIO_READ");
/* Advance location in buffer */
buf_p+=blk_size;
/* Advance global offset in dataset */
nbytes_xfer+=blk_size;
/* Decrement number of bytes left this time */
nbytes_toxfer-=blk_size;
} /* end while */
} /* end else */
} /* end if */
/* Collective file access */
else {
/* Contiguous access pattern */
if (parms->interleaved==0){
/* Compute offset in file */
mpi_offset = mpi_file_offset +
nbytes_xfer;
/* Perform collective read */
mrc = MPI_File_read_at_all(fd->mpifd, mpi_offset, buffer,
(int)(buf_size/blk_size), mpi_blk_type, &mpi_status);
VRFY((mrc==MPI_SUCCESS), "MPIO_READ");
/* Advance global offset in dataset */
nbytes_xfer+=buf_size;
} /* end if */
/* Interleaved access pattern */
else {
/* Compute offset in file */
mpi_offset = mpi_file_offset +
(nbytes_xfer*pio_mpi_nprocs_g);
/* Set the file view */
mrc = MPI_File_set_view(fd->mpifd, mpi_offset, mpi_blk_type,
mpi_file_type, (char*)"native", h5_io_info_g);
VRFY((mrc==MPI_SUCCESS), "MPIO_VIEW");
/* Perform collective read */
mrc = MPI_File_read_at_all(fd->mpifd, 0, buffer,
(int)(buf_size/blk_size), mpi_blk_type, &mpi_status);
VRFY((mrc==MPI_SUCCESS), "MPIO_READ");
/* Advance global offset in dataset */
nbytes_xfer+=buf_size;
} /* end else */
} /* end else */
break;
case PHDF5:
/* Set up the file dset space id to move the selection to process */
if (parms->interleaved==0){
/* Contiguous pattern */
h5offset[0] = nbytes_xfer;
} /* end if */
else {
/* Interleaved access pattern */
/* Skip offset over blocks of other processes */
h5offset[0] = (nbytes_xfer*pio_mpi_nprocs_g);
} /* end else */
hrc = H5Soffset_simple(h5dset_space_id, h5offset);
VRFY((hrc >= 0), "H5Soffset_simple");
/* Read the buffer in */
hrc = H5Dread(h5ds_id, ELMT_H5_TYPE, h5mem_space_id,
h5dset_space_id, h5dxpl, buffer);
VRFY((hrc >= 0), "H5Dread");
/* Increment number of bytes transferred */
nbytes_xfer += buf_size;
break;
} /* switch (parms->io_type) */
/* Verify raw data, if asked */
if (parms->verify) {
/* Verify data read */
unsigned char *ucharptr = (unsigned char *)buffer;
size_t i;
int nerror=0;
for (i = 0; i < buf_size; ++i){
if (*ucharptr++ != pio_mpi_rank_g) {
if (++nerror < 20){
/* report at most 20 errors */
HDprint_rank(output);
HDfprintf(output, "read data error, expected (%Hd), "
"got (%Hd)\n",
(long_long)pio_mpi_rank_g,
(long_long)*(ucharptr-1));
} /* end if */
} /* end if */
} /* end for */
if (nerror >= 20) {
HDprint_rank(output);
HDfprintf(output, "...");
HDfprintf(output, "total read data errors=%d\n",
nerror);
} /* end if */
} /* if (parms->verify) */
} /* end while */
/* Stop "raw data" read timer */
set_time(res->timers, HDF5_RAW_READ_FIXED_DIMS, STOP);
/* Calculate read time */
/* Close dataset. Only HDF5 needs to do an explicit close. */
if (parms->io_type == PHDF5) {
hrc = H5Dclose(h5ds_id);
if (hrc < 0) {
fprintf(stderr, "HDF5 Dataset Close failed\n");
GOTOERROR(FAIL);
}
h5ds_id = -1;
} /* end if */
} /* end for */
done:
/* release MPI-I/O objects */
if (parms->io_type == MPIO) {
/* Free file type */
mrc = MPI_Type_free( &mpi_file_type );
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_FREE");
/* Free buffer type */
mrc = MPI_Type_free( &mpi_blk_type );
VRFY((mrc==MPI_SUCCESS), "MPIO_TYPE_FREE");
} /* end if */
/* release HDF5 objects */
if (h5dset_space_id != -1) {
hrc = H5Sclose(h5dset_space_id);
if (hrc < 0){
fprintf(stderr, "HDF5 Dataset Space Close failed\n");
ret_code = FAIL;
} else {
h5dset_space_id = -1;
}
}
if (h5mem_space_id != -1) {
hrc = H5Sclose(h5mem_space_id);
if (hrc < 0) {
fprintf(stderr, "HDF5 Memory Space Close failed\n");
ret_code = FAIL;
} else {
h5mem_space_id = -1;
}
}
if (h5dxpl != -1) {
hrc = H5Pclose(h5dxpl);
if (hrc < 0) {
fprintf(stderr, "HDF5 Dataset Transfer Property List Close failed\n");
ret_code = FAIL;
} else {
h5dxpl = -1;
}
}
return ret_code;
}
/*
* Function: do_fopen
* Purpose: Open the specified file.
* Return: SUCCESS or FAIL
* Programmer: Albert Cheng, Bill Wendling, 2001/12/13
* Modifications:
*/
static herr_t
do_fopen(parameters *param, char *fname, file_descr *fd /*out*/, int flags)
{
int ret_code = SUCCESS, mrc;
herr_t hrc;
hid_t acc_tpl = -1; /* file access templates */
hbool_t use_gpfs = FALSE; /* use GPFS hints */
switch (param->io_type) {
case POSIXIO:
if (flags & (PIO_CREATE | PIO_WRITE))
fd->posixfd = POSIXCREATE(fname);
else
fd->posixfd = POSIXOPEN(fname, O_RDONLY);
if (fd->posixfd < 0 ) {
fprintf(stderr, "POSIX File Open failed(%s)\n", fname);
GOTOERROR(FAIL);
}
/* The perils of POSIX I/O in a parallel environment. The problem is:
*
* - Process n opens a file with truncation and then starts
* writing to the file.
* - Process m also opens the file with truncation, but after
* process n has already started to write to the file. Thus,
* all of the stuff process n wrote is now lost.
*/
MPI_Barrier(pio_comm_g);
break;
case MPIO:
if (flags & (PIO_CREATE | PIO_WRITE)) {
MPI_File_delete(fname, h5_io_info_g);
mrc = MPI_File_open(pio_comm_g, fname, MPI_MODE_CREATE | MPI_MODE_RDWR,
h5_io_info_g, &fd->mpifd);
if (mrc != MPI_SUCCESS) {
fprintf(stderr, "MPI File Open failed(%s)\n", fname);
GOTOERROR(FAIL);
}
/*since MPI_File_open with MPI_MODE_CREATE does not truncate */
/*filesize , set size to 0 explicitedly. */
mrc = MPI_File_set_size(fd->mpifd, (MPI_Offset)0);
if (mrc != MPI_SUCCESS) {
fprintf(stderr, "MPI_File_set_size failed\n");
GOTOERROR(FAIL);
}
} else {
mrc = MPI_File_open(pio_comm_g, fname, MPI_MODE_RDONLY,
h5_io_info_g, &fd->mpifd);
if (mrc != MPI_SUCCESS) {
fprintf(stderr, "MPI File Open failed(%s)\n", fname);
GOTOERROR(FAIL);
}
}
break;
case PHDF5:
acc_tpl = H5Pcreate(H5P_FILE_ACCESS);
if (acc_tpl < 0) {
fprintf(stderr, "HDF5 Property List Create failed\n");
GOTOERROR(FAIL);
}
/* Use the appropriate VFL driver */
if(param->h5_use_mpi_posix) {
/* Set the file driver to the MPI-posix driver */
hrc = H5Pset_fapl_mpiposix(acc_tpl, pio_comm_g, use_gpfs);
if (hrc < 0) {
fprintf(stderr, "HDF5 Property List Set failed\n");
GOTOERROR(FAIL);
}
} /* end if */
else {
/* Set the file driver to the MPI-I/O driver */
hrc = H5Pset_fapl_mpio(acc_tpl, pio_comm_g, h5_io_info_g);
if (hrc < 0) {
fprintf(stderr, "HDF5 Property List Set failed\n");
GOTOERROR(FAIL);
}
} /* end else */
/* Set the alignment of objects in HDF5 file */
hrc = H5Pset_alignment(acc_tpl, param->h5_thresh, param->h5_align);
if (hrc < 0) {
fprintf(stderr, "HDF5 Property List Set failed\n");
GOTOERROR(FAIL);
}
/* create the parallel file */
if (flags & (PIO_CREATE | PIO_WRITE)) {
fd->h5fd = H5Fcreate(fname, H5F_ACC_TRUNC, H5P_DEFAULT, acc_tpl);
} else {
fd->h5fd = H5Fopen(fname, H5F_ACC_RDONLY, acc_tpl);
}
hrc = H5Pclose(acc_tpl);
if (fd->h5fd < 0) {
fprintf(stderr, "HDF5 File Create failed(%s)\n", fname);
GOTOERROR(FAIL);
}
/* verifying the close of the acc_tpl */
if (hrc < 0) {
fprintf(stderr, "HDF5 Property List Close failed\n");
GOTOERROR(FAIL);
}
break;
}
done:
return ret_code;
}
/*
* Function: do_fclose
* Purpose: Close the specified file descriptor.
* Return: SUCCESS or FAIL
* Programmer: Albert Cheng, Bill Wendling, 2001/12/13
* Modifications:
*/
static herr_t
do_fclose(iotype iot, file_descr *fd /*out*/)
{
herr_t ret_code = SUCCESS, hrc;
int mrc = 0, rc = 0;
switch (iot) {
case POSIXIO:
rc = POSIXCLOSE(fd->posixfd);
if (rc != 0){
fprintf(stderr, "POSIX File Close failed\n");
GOTOERROR(FAIL);
}
fd->posixfd = -1;
break;
case MPIO:
mrc = MPI_File_close(&fd->mpifd);
if (mrc != MPI_SUCCESS){
fprintf(stderr, "MPI File close failed\n");
GOTOERROR(FAIL);
}
fd->mpifd = MPI_FILE_NULL;
break;
case PHDF5:
hrc = H5Fclose(fd->h5fd);
if (hrc < 0) {
fprintf(stderr, "HDF5 File Close failed\n");
GOTOERROR(FAIL);
}
fd->h5fd = -1;
break;
}
done:
return ret_code;
}
/*
* Function: do_fclose
* Purpose: Cleanup temporary file unless HDF5_NOCLEANUP is set.
* Only Proc 0 of the PIO communicator will do the cleanup.
* Other processes just return.
* Return: void
* Programmer: Albert Cheng 2001/12/12
* Modifications:
*/
static void
do_cleanupfile(iotype iot, char *fname)
{
if (pio_mpi_rank_g != 0)
return;
if (clean_file_g == -1)
clean_file_g = (getenv("HDF5_NOCLEANUP")==NULL) ? 1 : 0;
if (clean_file_g){
switch (iot){
case POSIXIO:
remove(fname);
break;
case MPIO:
case PHDF5:
MPI_File_delete(fname, h5_io_info_g);
break;
}
}
}
#ifdef H5_HAVE_GPFS
/* Descriptions here come from the IBM GPFS Manual */
/*
* Function: gpfs_access_range
* Purpose: Declares an access range within a file for an
* application.
*
* The application will access file offsets within the given
* range, and will not access offsets outside the range.
* Violating this hint may produce worse performance than if
* no hint was specified.
*
* This hint is useful in situations where a file is
* partitioned coarsely among several nodes. If the ranges
* do not overlap, each node can specify which range of the
* file it will access, with a performance improvement in
* some cases, such as for sequential writing within a
* range.
*
* Subsequent GPFS_ACCESS_RANGE hints will replace a hint
* passed earlier.
*
* START - The start of the access range offset, in
* bytes, from the beginning of the file
* LENGTH - Length of the access range. 0 indicates to
* the end of the file
* IS_WRITE - 0 indicates READ access, 1 indicates WRITE access
* Return: Nothing
* Programmer: Bill Wendling, 03. June 2002
* Modifications:
*/
static void
gpfs_access_range(int handle, off_t start, off_t length, int is_write)
{
struct {
gpfsFcntlHeader_t hdr;
gpfsAccessRange_t access;
} access_range;
access_range.hdr.totalLength = sizeof(access_range);
access_range.hdr.fcntlVersion = GPFS_FCNTL_CURRENT_VERSION;
access_range.hdr.fcntlReserved = 0;
access_range.access.structLen = sizeof(gpfsAccessRange_t);
access_range.access.structType = GPFS_ACCESS_RANGE;
access_range.access.start = start;
access_range.access.length = length;
access_range.access.isWrite = is_write;
if (gpfs_fcntl(handle, &access_range) != 0) {
fprintf(stderr,
"gpfs_fcntl DS start directive failed. errno=%d errorOffset=%d\n",
errno, access_range.hdr.errorOffset);
exit(EXIT_FAILURE);
}
}
/*
* Function: gpfs_free_range
* Purpose: Undeclares an access range within a file for an
* application.
*
* The application will no longer access file offsets within
* the given range. GPFS flushes the data at the file
* offsets and removes it from the cache.
*
* Multi-node applications that have finished one phase of
* their computation may wish to use this hint before the
* file is accessed in a conflicting mode from another node
* in a later phase. The potential performance benefit is
* that GPFS can avoid later synchronous cache consistency
* operations.
*
* START - The start of the access range offset, in
* bytes from the beginning of the file.
* LENGTH - Length of the access range. 0 indicates to
* the end of the file.
* Return: Nothing
* Programmer: Bill Wendling, 03. June 2002
* Modifications:
*/
static void
gpfs_free_range(int handle, off_t start, off_t length)
{
struct {
gpfsFcntlHeader_t hdr;
gpfsFreeRange_t range;
} free_range;
/* Issue the invalidate hint */
free_range.hdr.totalLength = sizeof(free_range);
free_range.hdr.fcntlVersion = GPFS_FCNTL_CURRENT_VERSION;
free_range.hdr.fcntlReserved = 0;
free_range.range.structLen = sizeof(gpfsFreeRange_t);
free_range.range.structType = GPFS_FREE_RANGE;
free_range.range.start = start;
free_range.range.length = length;
if (gpfs_fcntl(handle, &free_range) != 0) {
fprintf(stderr,
"gpfs_fcntl free range failed for range %d:%d. errno=%d errorOffset=%d\n",
start, length, errno, free_range.hdr.errorOffset);
exit(EXIT_FAILURE);
}
}
/*
* Function: gpfs_clear_file_cache
* Purpose: Indicates file access in the near future is not expected.
*
* The application does not expect to make any further
* accesses to the file in the near future, so GPFS removes
* any data or metadata pertaining to the file from its
* cache.
*
* Multi-node applications that have finished one phase of
* their computation may wish to use this hint before the
* file is accessed in a conflicting mode from another node
* in a later phase. The potential performance benefit is
* that GPFS can avoid later synchronous cache consistency
* operations.
* Return: Nothing
* Programmer: Bill Wendling, 03. June 2002
* Modifications:
*/
static void
gpfs_clear_file_cache(int handle)
{
struct {
gpfsFcntlHeader_t hdr;
gpfsClearFileCache_t clear;
} clear_cache;
clear_cache.hdr.totalLength = sizeof(clear_cache);
clear_cache.hdr.fcntlVersion = GPFS_FCNTL_CURRENT_VERSION;
clear_cache.hdr.fcntlReserved = 0;
clear_cache.clear.structLen = sizeof(gpfsClearFileCache_t);
clear_cache.clear.structType = GPFS_CLEAR_FILE_CACHE;
if (gpfs_fcntl(handle, &clear_cache) != 0) {
fprintf(stderr,
"gpfs_fcntl clear file cache directive failed. errno=%d errorOffset=%d\n",
errno, clear_cache.hdr.errorOffset);
exit(EXIT_FAILURE);
}
}
/*
* Function: gpfs_cancel_hints
* Purpose: Indicates to remove any hints against the open file
* handle.
*
* GPFS removes any hints that may have been issued against
* this open file handle:
*
* - The hint status of the file is restored ot what it
* would have been immediately after being opened, but
* does not affect the contents of the GPFS file
* cache. Cancelling an earlier hint that resulted in
* data being removed from the GPFS file cache does
* not bring that data back int othe cache; data
* re-enters the cache only pon access by the
* application or by user-driven or automatic
* prefetching.
* - Only the GPFS_MULTIPLE_ACCESS_RANGE hint has a
* state that might be removed by the
* GPFS_CANCEL_HINTS directive.
* Return: Nothing
* Programmer: Bill Wendling, 03. June 2002
* Modifications:
*/
static void
gpfs_cancel_hints(int handle)
{
struct {
gpfsFcntlHeader_t hdr;
gpfsCancelHints_t cancel;
} cancel_hints;
cancel_hints.hdr.totalLength = sizeof(cancel_hints);
cancel_hints.hdr.fcntlVersion = GPFS_FCNTL_CURRENT_VERSION;
cancel_hints.hdr.fcntlReserved = 0;
cancel_hints.cancel.structLen = sizeof(gpfsCancelHints_t);
cancel_hints.cancel.structType = GPFS_CANCEL_HINTS;
if (gpfs_fcntl(handle, &cancel_hints) != 0) {
fprintf(stderr,
"gpfs_fcntl cancel hints directive failed. errno=%d errorOffset=%d\n",
errno, cancel_hints.hdr.errorOffset);
exit(EXIT_FAILURE);
}
}
/*
* Function: gpfs_start_data_shipping
* Purpose: Initiates data shipping mode.
*
* Once all participating threads have issued this directive
* for a file, GPFS enters a mode where it logically
* partitions the blocks of the file among a group of agent
* nodes. The agents are those nodes on which one or more
* threads have issued the GPFS_DATA_SHIP_START directive.
* Each thread that has issued a GPFS_DATA_SHIP_START
* directive and the associated agent nodes are referred to
* as the data shipping collective.
*
* The second parameter is the total number of open
* instances on all nodes that will be operating on the
* file. Must be called for every such instance with the
* same value of NUM_INSTS.
*
* NUM_INSTS - The number of open file instances, on all
* nodes, collaborating to operate on the file
* Return: Nothing
* Programmer: Bill Wendling, 28. May 2002
* Modifications:
*/
static void
gpfs_start_data_shipping(int handle, int num_insts)
{
struct {
gpfsFcntlHeader_t hdr;
gpfsDataShipStart_t start;
} ds_start;
ds_start.hdr.totalLength = sizeof(ds_start);
ds_start.hdr.fcntlVersion = GPFS_FCNTL_CURRENT_VERSION;
ds_start.hdr.fcntlReserved = 0;
ds_start.start.structLen = sizeof(gpfsDataShipStart_t);
ds_start.start.structType = GPFS_DATA_SHIP_START;
ds_start.start.numInstances = num_insts;
ds_start.start.reserved = 0;
if (gpfs_fcntl(handle, &ds_start) != 0) {
fprintf(stderr,
"gpfs_fcntl DS start directive failed. errno=%d errorOffset=%d\n",
errno, ds_start.hdr.errorOffset);
exit(EXIT_FAILURE);
}
}
/*
* Function: gpfs_start_data_ship_map
* Purpose: Indicates which agent nodes are to be used for data
* shipping. GPFS recognizes which agent nodes to use for
* data shipping.
*
* PARTITION_SIZE - The number of contiguous bytes per
* server. This value must be a
* multiple of the number of bytes in a
* single file system block
* AGENT_COUNT - The number of entries in the
* agentNodeNumber array
* AGENT_NODE_NUM - The data ship agent node numbers as
* listed in the SDT or the global ODM
*
* Return: Nothing
* Programmer: Bill Wendling, 10. Jul 2002
* Modifications:
*/
static void
gpfs_start_data_ship_map(int handle, int partition_size, int agent_count,
int *agent_node_num)
{
int i;
struct {
gpfsFcntlHeader_t hdr;
gpfsDataShipMap_t map;
} ds_map;
ds_map.hdr.totalLength = sizeof(ds_map);
ds_map.hdr.fcntlVersion = GPFS_FCNTL_CURRENT_VERSION;
ds_map.hdr.fcntlReserved = 0;
ds_map.map.structLen = sizeof(gpfsDataShipMap_t);
ds_map.map.structType = GPFS_DATA_SHIP_MAP;
ds_map.map.partitionSize = partition_size;
ds_map.map.agentCount = agent_count;
for (i = 0; i < agent_count; ++i)
ds_map.map.agentNodeNumber[i] = agent_node_num[i];
if (gpfs_fcntl(handle, &ds_map) != 0) {
fprintf(stderr,
"gpfs_fcntl DS map directive failed. errno=%d errorOffset=%d\n",
errno, ds_map.hdr.errorOffset);
exit(EXIT_FAILURE);
}
}
/*
* Function: gpfs_stop_data_shipping
* Purpose: Takes a file out of the data shipping mode.
*
* - GPFS waits for all threads that issued the
* GPFS_DATA_SHIP_START directive to issue this directive,
* then flushes the dirty file data to disk.
*
* - While a gpfs_cntl() call is blocked for other threads,
* the call can be interrupted by any signal. If a signal
* is delivered to any of the waiting calls, all waiting
* calls on every node will be interrupted and will return
* EINTR. GPFS will not cancel data shipping mode if such
* a signal occurs. It is the responsibility of the
* application to mask off any signals that might normally
* occur while waiting for another node in the data
* shipping collective. Several libraries use SIGALRM; the
* thread that makes the gpfs_fcntl() call should use
* sigthreadmask to mask off delivery of this signal while
* inside the call.
* Return: Nothing
* Programmer: Bill Wendling, 28. May 2002
* Modifications:
*/
static void
gpfs_stop_data_shipping(int handle)
{
struct {
gpfsFcntlHeader_t hdr;
gpfsDataShipStop_t stop;
} ds_stop;
ds_stop.hdr.totalLength = sizeof(ds_stop);
ds_stop.hdr.fcntlVersion = GPFS_FCNTL_CURRENT_VERSION;
ds_stop.hdr.fcntlReserved = 0;
ds_stop.stop.structLen = sizeof(ds_stop.stop);
ds_stop.stop.structType = GPFS_DATA_SHIP_STOP;
if (gpfs_fcntl(handle, &ds_stop) != 0)
fprintf(stderr,
"gpfs_fcntl DS stop directive failed. errno=%d errorOffset=%d\n",
errno, ds_stop.hdr.errorOffset);
}
/*
* Function: gpfs_invalidate_file_cache
* Purpose: Invalidate all cached data held on behalf of a file on
* this node.
* Return: Nothing
* Programmer: Bill Wendling, 03. June 2002
* Modifications:
*/
static void
gpfs_invalidate_file_cache(const char *filename)
{
int handle;
struct {
gpfsFcntlHeader_t hdr;
gpfsClearFileCache_t inv;
} inv_cache_hint;
/* Open the file. If the open fails, the file cannot be cached. */
handle = open(filename, O_RDONLY, 0);
if (handle == -1)
return;
/* Issue the invalidate hint */
inv_cache_hint.hdr.totalLength = sizeof(inv_cache_hint);
inv_cache_hint.hdr.fcntlVersion = GPFS_FCNTL_CURRENT_VERSION;
inv_cache_hint.hdr.fcntlReserved = 0;
inv_cache_hint.inv.structLen = sizeof(gpfsClearFileCache_t);
inv_cache_hint.inv.structType = GPFS_CLEAR_FILE_CACHE;
if (gpfs_fcntl(handle, &inv_cache_hint) != 0) {
fprintf(stderr,
"gpfs_fcntl clear cache hint failed for file '%s'.",
filename);
fprintf(stderr, " errno=%d errorOffset=%d\n",
errno, inv_cache_hint.hdr.errorOffset);
exit(1);
}
/* Close the file */
if (close(handle) == -1) {
fprintf(stderr,
"could not close file '%s' after flushing file cache, ",
filename);
fprintf(stderr, "errno=%d\n", errno);
exit(1);
}
}
#else
/* turn the stubs off since some compilers are warning they are not used */
#if 0
/* H5_HAVE_GPFS isn't defined...stub functions */
static void
gpfs_access_range(int UNUSED handle, off_t UNUSED start, off_t UNUSED length,
int UNUSED is_write)
{
return;
}
static void
gpfs_free_range(int UNUSED handle, off_t UNUSED start, off_t UNUSED length)
{
return;
}
static void
gpfs_clear_file_cache(int UNUSED handle)
{
return;
}
static void
gpfs_cancel_hints(int UNUSED handle)
{
return;
}
static void
gpfs_start_data_shipping(int UNUSED handle, int UNUSED num_insts)
{
return;
}
static void
gpfs_stop_data_shipping(int UNUSED handle)
{
return;
}
static void
gpfs_start_data_ship_map(int UNUSED handle, int UNUSED partition_size,
int UNUSED agent_count, int UNUSED *agent_node_num)
{
return;
}
static void
gpfs_invalidate_file_cache(const char UNUSED *filename)
{
return;
}
#endif /* 0 */
#endif /* H5_HAVE_GPFS */
#ifdef TIME_MPI
/* instrument the MPI_File_wrirte_xxx and read_xxx calls to measure
* pure time spent in MPI_File code.
*/
int MPI_File_read_at(MPI_File fh, MPI_Offset offset, void *buf,
int count, MPI_Datatype datatype, MPI_Status *status)
{
int err;
set_time(timer_g, HDF5_MPI_READ, START);
err=PMPI_File_read_at(fh, offset, buf, count, datatype, status);
set_time(timer_g, HDF5_MPI_READ, STOP);
return err;
}
int MPI_File_read_at_all(MPI_File fh, MPI_Offset offset, void *buf,
int count, MPI_Datatype datatype, MPI_Status *status)
{
int err;
set_time(timer_g, HDF5_MPI_READ, START);
err=PMPI_File_read_at_all(fh, offset, buf, count, datatype, status);
set_time(timer_g, HDF5_MPI_READ, STOP);
return err;
}
int MPI_File_write_at(MPI_File fh, MPI_Offset offset, void *buf,
int count, MPI_Datatype datatype, MPI_Status *status)
{
int err;
set_time(timer_g, HDF5_MPI_WRITE, START);
err=PMPI_File_write_at(fh, offset, buf, count, datatype, status);
set_time(timer_g, HDF5_MPI_WRITE, STOP);
return err;
}
int MPI_File_write_at_all(MPI_File fh, MPI_Offset offset, void *buf,
int count, MPI_Datatype datatype, MPI_Status *status)
{
int err;
set_time(timer_g, HDF5_MPI_WRITE, START);
err=PMPI_File_write_at_all(fh, offset, buf, count, datatype, status);
set_time(timer_g, HDF5_MPI_WRITE, STOP);
return err;
}
#endif /* TIME_MPI */
#endif /* H5_HAVE_PARALLEL */
Reference in New Issue View Git Blame Copy Permalink