/* * Copyright (c) 1999-2002 NCSA * All rights reserved. * * Programmer: Robb Matzke * Thursday, July 29, 1999 * * Purpose: This is the MPI-2 I/O driver. * * Limitations: * H5FD_mpio_read & H5FD_mpio_write * Eventually these should choose collective or independent i/o * based on a parameter that is passed down to it from H5Dwrite, * rather than the access_parms (which are fixed at the open). * * H5FD_mpio_read * One implementation of MPI/MPI-IO causes MPI_Get_count * to return (incorrectly) a negative count. I (who?) added code * to detect this, and a kludge to pretend that the number of * bytes read is always equal to the number requested. This * kluge is activated by #ifdef MPI_KLUGE0202. */ #include "H5private.h" /*library functions */ #include "H5Eprivate.h" /*error handling */ #include "H5Fprivate.h" /*files */ #include "H5FDprivate.h" /*file driver */ #include "H5FDmpio.h" /* MPI I/O file driver */ #include "H5Iprivate.h" /*object IDs */ #include "H5MMprivate.h" /* Memory allocation */ #include "H5Pprivate.h" /*property lists */ /* * The driver identification number, initialized at runtime if H5_HAVE_PARALLEL * is defined. This allows applications to still have the H5FD_MPIO * "constants" in their source code (it also makes this file strictly ANSI * compliant when H5_HAVE_PARALLEL isn't defined) */ static hid_t H5FD_MPIO_g = 0; #ifdef H5_HAVE_PARALLEL /* * The description of a file belonging to this driver. * The EOF value is only used just after the file is opened in order for the * library to determine whether the file is empty, truncated, or okay. The MPIO * driver doesn't bother to keep it updated since it's an expensive operation. */ typedef struct H5FD_mpio_t { H5FD_t pub; /*public stuff, must be first */ MPI_File f; /*MPIO file handle */ MPI_Comm comm; /*communicator */ MPI_Info info; /*file information */ int mpi_rank; /* This process's rank */ int mpi_size; /* Total number of processes */ int mpi_round; /* Current round robin process (for metadata I/O) */ haddr_t eof; /*end-of-file marker */ haddr_t eoa; /*end-of-address marker */ haddr_t last_eoa; /* Last known end-of-address marker */ unsigned old_use_view; /*remember value of use_view */ } H5FD_mpio_t; /* Prototypes */ static haddr_t H5FD_mpio_MPIOff_to_haddr(MPI_Offset mpi_off); static herr_t H5FD_mpio_haddr_to_MPIOff(haddr_t addr, MPI_Offset *mpi_off/*out*/); /* Callbacks */ static void *H5FD_mpio_fapl_get(H5FD_t *_file); static H5FD_t *H5FD_mpio_open(const char *name, unsigned flags, hid_t fapl_id, haddr_t maxaddr); static herr_t H5FD_mpio_close(H5FD_t *_file); static herr_t H5FD_mpio_query(const H5FD_t *_f1, unsigned long *flags); static haddr_t H5FD_mpio_get_eoa(H5FD_t *_file); static herr_t H5FD_mpio_set_eoa(H5FD_t *_file, haddr_t addr); static haddr_t H5FD_mpio_get_eof(H5FD_t *_file); static herr_t H5FD_mpio_read(H5FD_t *_file, H5FD_mem_t type, hid_t fapl_id, haddr_t addr, size_t size, void *buf); static herr_t H5FD_mpio_write(H5FD_t *_file, H5FD_mem_t type, hid_t fapl_id, haddr_t addr, size_t size, const void *buf); static herr_t H5FD_mpio_flush(H5FD_t *_file, unsigned closing); /* MPIO-specific file access properties */ typedef struct H5FD_mpio_fapl_t { MPI_Comm comm; /*communicator */ MPI_Info info; /*file information */ } H5FD_mpio_fapl_t; /* The MPIO file driver information */ static const H5FD_class_t H5FD_mpio_g = { "mpio", /*name */ HADDR_MAX, /*maxaddr */ H5F_CLOSE_SEMI, /* fc_degree */ NULL, /*sb_size */ NULL, /*sb_encode */ NULL, /*sb_decode */ sizeof(H5FD_mpio_fapl_t), /*fapl_size */ H5FD_mpio_fapl_get, /*fapl_get */ NULL, /*fapl_copy */ NULL, /*fapl_free */ sizeof(H5FD_mpio_dxpl_t), /*dxpl_size */ NULL, /*dxpl_copy */ NULL, /*dxpl_free */ H5FD_mpio_open, /*open */ H5FD_mpio_close, /*close */ NULL, /*cmp */ H5FD_mpio_query, /*query */ NULL, /*alloc */ NULL, /*free */ H5FD_mpio_get_eoa, /*get_eoa */ H5FD_mpio_set_eoa, /*set_eoa */ H5FD_mpio_get_eof, /*get_eof */ H5FD_mpio_read, /*read */ H5FD_mpio_write, /*write */ H5FD_mpio_flush, /*flush */ H5FD_FLMAP_SINGLE, /*fl_map */ }; #ifdef H5FDmpio_DEBUG /* Flags to control debug actions in H5Fmpio. * Meant to be indexed by characters. * * 'c' show result of MPI_Get_count after read * 'r' show read offset and size * 't' trace function entry and exit * 'w' show write offset and size */ static int H5FD_mpio_Debug[256] = { 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 }; #endif /* Global var to allow elimination of redundant metadata writes * to be controlled by the value of an environment variable. */ /* Use the elimination by default unless this is the Intel Red machine */ #ifndef __PUMAGON__ hbool_t H5_mpi_1_metawrite_g = TRUE; #else hbool_t H5_mpi_1_metawrite_g = FALSE; #endif /* Interface initialization */ #define PABLO_MASK H5FD_mpio_mask #define INTERFACE_INIT H5FD_mpio_init static int interface_initialize_g = 0; /* ======== Temporary, Local data transfer properties ======== */ /* Definitions for memory MPI type property */ #define H5FD_MPIO_XFER_MEM_MPI_TYPE_NAME "H5FD_mpio_mem_mpi_type" #define H5FD_MPIO_XFER_MEM_MPI_TYPE_SIZE sizeof(MPI_Datatype) /* Definitions for file MPI type property */ #define H5FD_MPIO_XFER_FILE_MPI_TYPE_NAME "H5FD_mpio_file_mpi_type" #define H5FD_MPIO_XFER_FILE_MPI_TYPE_SIZE sizeof(MPI_Datatype) /* Definitions for whether to use MPI types property */ #define H5FD_MPIO_XFER_USE_VIEW_NAME "H5FD_mpio_use_view" #define H5FD_MPIO_XFER_USE_VIEW_SIZE sizeof(unsigned) /*------------------------------------------------------------------------- * Function: H5FD_mpio_init * * Purpose: Initialize this driver by registering the driver with the * library. * * Return: Success: The driver ID for the mpio driver. * * Failure: Negative. * * Programmer: Robb Matzke * Thursday, August 5, 1999 * * Modifications: * *------------------------------------------------------------------------- */ hid_t H5FD_mpio_init(void) { FUNC_ENTER_NOAPI(H5FD_mpio_init, FAIL); if (H5I_VFL!=H5Iget_type(H5FD_MPIO_g)) H5FD_MPIO_g = H5FDregister(&H5FD_mpio_g); FUNC_LEAVE(H5FD_MPIO_g); } /*------------------------------------------------------------------------- * Function: H5Pset_fapl_mpio * * Purpose: Store the user supplied MPIO communicator COMM and INFO in * the file access property list FAPL_ID which can then be used * to create and/or open the file. This function is available * only in the parallel HDF5 library and is not collective. * * COMM is the MPI communicator to be used for file open as * defined in MPI_FILE_OPEN of MPI-2. This function does not * make a duplicated communicator. Any modification to COMM * after this function call returns may have undetermined effect * on the access property list. Users should not modify the * communicator while it is defined in a property list. * * INFO is the MPI info object to be used for file open as * defined in MPI_FILE_OPEN of MPI-2. This function does not * make a duplicated info. Any modification to info after this * function call returns may have undetermined effect on the * access property list. Users should not modify the info while * it is defined in a property list. * * Return: Success: Non-negative * * Failure: Negative * * Programmer: Albert Cheng * Feb 3, 1998 * * Modifications: * Robb Matzke, 1998-02-18 * Check all arguments before the property list is updated so we * don't leave the property list in a bad state if something * goes wrong. Also, the property list data type changed to * allow more generality so all the mpi-related stuff is in the * `u.mpi' member. The `access_mode' will contain only * mpi-related flags defined in H5Fpublic.h. * * Albert Cheng, 1998-04-16 * Removed the ACCESS_MODE argument. The access mode is changed * to be controlled by data transfer property list during data * read/write calls. * * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. * * Raymond Lu * Tuesday, Oct 23, 2001 * Changed the file access list to the new generic property * list. * *------------------------------------------------------------------------- */ herr_t H5Pset_fapl_mpio(hid_t fapl_id, MPI_Comm comm, MPI_Info info) { H5FD_mpio_fapl_t fa; H5P_genplist_t *plist; /* Property list pointer */ herr_t ret_value=FAIL; FUNC_ENTER_API(H5Pset_fapl_mpio, FAIL); H5TRACE3("e","iMcMi",fapl_id,comm,info); /* Check arguments */ if(TRUE!=H5P_isa_class(fapl_id,H5P_FILE_ACCESS) || NULL == (plist = H5I_object(fapl_id))) HRETURN_ERROR(H5E_PLIST, H5E_BADTYPE, FAIL, "not a file access list"); #ifdef LATER #warning "We need to verify that COMM and INFO contain sensible information." #endif /* Initialize driver specific properties */ fa.comm = comm; fa.info = info; ret_value= H5P_set_driver(plist, H5FD_MPIO, &fa); FUNC_LEAVE(ret_value); } /*------------------------------------------------------------------------- * Function: H5Pget_fapl_mpio * * Purpose: If the file access property list is set to the H5FD_MPIO * driver then this function returns the MPI communicator and * information through the COMM and INFO pointers. * * Return: Success: Non-negative with the communicator and * information returned through the COMM and * INFO arguments if non-null. Neither piece of * information is copied and they are therefore * valid only until the file access property * list is modified or closed. * * Failure: Negative * * Programmer: Robb Matzke * Thursday, February 26, 1998 * * Modifications: * * Albert Cheng, Apr 16, 1998 * Removed the access_mode argument. The access_mode is changed * to be controlled by data transfer property list during data * read/write calls. * * Raymond Lu * Tuesday, Oct 23, 2001 * Changed the file access list to the new generic property * list. * *------------------------------------------------------------------------- */ herr_t H5Pget_fapl_mpio(hid_t fapl_id, MPI_Comm *comm/*out*/, MPI_Info *info/*out*/) { H5FD_mpio_fapl_t *fa; H5P_genplist_t *plist; /* Property list pointer */ FUNC_ENTER_API(H5Pget_fapl_mpio, FAIL); H5TRACE3("e","ixx",fapl_id,comm,info); if(TRUE!=H5P_isa_class(fapl_id,H5P_FILE_ACCESS) || NULL == (plist = H5I_object(fapl_id))) HRETURN_ERROR(H5E_PLIST, H5E_BADTYPE, FAIL, "not a file access list"); if (H5FD_MPIO!=H5P_get_driver(plist)) HRETURN_ERROR(H5E_PLIST, H5E_BADVALUE, FAIL, "incorrect VFL driver"); if (NULL==(fa=H5P_get_driver_info(plist))) HRETURN_ERROR(H5E_PLIST, H5E_BADVALUE, FAIL, "bad VFL driver info"); if (comm) *comm = fa->comm; if (info) *info = fa->info; FUNC_LEAVE(SUCCEED); } /*------------------------------------------------------------------------- * Function: H5Pset_dxpl_mpio * * Purpose: Set the data transfer property list DXPL_ID to use transfer * mode XFER_MODE. The property list can then be used to control * the I/O transfer mode during data I/O operations. The valid * transfer modes are: * * H5FD_MPIO_INDEPENDENT: * Use independent I/O access (the default). * * H5FD_MPIO_COLLECTIVE: * Use collective I/O access. * * Return: Success: Non-negative * * Failure: Negative * * Programmer: Albert Cheng * April 2, 1998 * * Modifications: * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. *------------------------------------------------------------------------- */ herr_t H5Pset_dxpl_mpio(hid_t dxpl_id, H5FD_mpio_xfer_t xfer_mode) { H5FD_mpio_dxpl_t dx; H5P_genplist_t *plist; /* Property list pointer */ herr_t ret_value=FAIL; FUNC_ENTER_API(H5Pset_dxpl_mpio, FAIL); H5TRACE2("e","iDt",dxpl_id,xfer_mode); /* Check arguments */ if(TRUE!=H5P_isa_class(dxpl_id,H5P_DATASET_XFER) || NULL == (plist = H5I_object(dxpl_id))) HRETURN_ERROR(H5E_PLIST, H5E_BADTYPE, FAIL, "not a dxpl"); if (H5FD_MPIO_INDEPENDENT!=xfer_mode && H5FD_MPIO_COLLECTIVE!=xfer_mode) HRETURN_ERROR(H5E_ARGS, H5E_BADVALUE, FAIL, "incorrect xfer_mode"); /* Initialize driver-specific properties */ dx.xfer_mode = xfer_mode; ret_value= H5P_set_driver(plist, H5FD_MPIO, &dx); FUNC_LEAVE(ret_value); } /*------------------------------------------------------------------------- * Function: H5Pget_dxpl_mpio * * Purpose: Queries the transfer mode current set in the data transfer * property list DXPL_ID. This is not collective. * * Return: Success: Non-negative, with the transfer mode returned * through the XFER_MODE argument if it is * non-null. * * Failure: Negative * * Programmer: Albert Cheng * April 2, 1998 * * Modifications: * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. *------------------------------------------------------------------------- */ herr_t H5Pget_dxpl_mpio(hid_t dxpl_id, H5FD_mpio_xfer_t *xfer_mode/*out*/) { H5FD_mpio_dxpl_t *dx; H5P_genplist_t *plist; /* Property list pointer */ FUNC_ENTER_API(H5Pget_dxpl_mpio, FAIL); H5TRACE2("e","ix",dxpl_id,xfer_mode); if(TRUE!=H5P_isa_class(dxpl_id,H5P_DATASET_XFER) || NULL == (plist = H5I_object(dxpl_id))) HRETURN_ERROR(H5E_PLIST, H5E_BADTYPE, FAIL, "not a dxpl"); if (H5FD_MPIO!=H5P_get_driver(plist)) HRETURN_ERROR(H5E_PLIST, H5E_BADVALUE, FAIL, "incorrect VFL driver"); if (NULL==(dx=H5P_get_driver_info(plist))) HRETURN_ERROR(H5E_PLIST, H5E_BADVALUE, FAIL, "bad VFL driver info"); if (xfer_mode) *xfer_mode = dx->xfer_mode; FUNC_LEAVE(SUCCEED); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_communicator * * Purpose: Returns the MPI communicator for the file. * * Return: Success: The communicator * * Failure: NULL * * Programmer: Robb Matzke * Monday, August 9, 1999 * * Modifications: * *------------------------------------------------------------------------- */ MPI_Comm H5FD_mpio_communicator(H5FD_t *_file) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; FUNC_ENTER_NOAPI(H5FD_mpio_communicator, MPI_COMM_NULL); assert(file); assert(H5FD_MPIO==file->pub.driver_id); FUNC_LEAVE(file->comm); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_mpi_rank * * Purpose: Returns the MPI rank for a process * * Return: Success: non-negative * Failure: negative * * Programmer: Quincey Koziol * Thursday, May 16, 2002 * * Modifications: * *------------------------------------------------------------------------- */ int H5FD_mpio_mpi_rank(H5FD_t *_file) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; FUNC_ENTER_NOAPI(H5FD_mpio_mpi_rank, FAIL); assert(file); assert(H5FD_MPIO==file->pub.driver_id); FUNC_LEAVE(file->mpi_rank); } /* end H5FD_mpio_mpi_rank() */ /*------------------------------------------------------------------------- * Function: H5FD_mpio_mpi_size * * Purpose: Returns the number of MPI processes * * Return: Success: non-negative * Failure: negative * * Programmer: Quincey Koziol * Thursday, May 16, 2002 * * Modifications: * *------------------------------------------------------------------------- */ int H5FD_mpio_mpi_size(H5FD_t *_file) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; FUNC_ENTER_NOAPI(H5FD_mpio_mpi_rank, FAIL); assert(file); assert(H5FD_MPIO==file->pub.driver_id); FUNC_LEAVE(file->mpi_size); } /* end H5FD_mpio_mpi_size() */ /*------------------------------------------------------------------------- * Function: H5FD_mpio_setup * * Purpose: Set the buffer type BTYPE, file type FTYPE for a data * transfer. Also request a MPI type transfer. * * Return: Success: 0 * Failure: -1 * * Programmer: Robb Matzke * Monday, August 9, 1999 * * Modifications: * * Quincey Koziol - 2002/06/17 * Removed 'disp' parameter, read & write routines will use * the address of the dataset in MPI_File_set_view() calls, as * necessary. * * Quincey Koziol - 2002/06/17 * Changed to set temporary properties in a dxpl, instead of * flags in the file struct, which will make this more threadsafe. * *------------------------------------------------------------------------- */ herr_t H5FD_mpio_setup(hid_t dxpl_id, MPI_Datatype btype, MPI_Datatype ftype, unsigned use_view) { H5P_genplist_t *plist; /* Property list pointer */ FUNC_ENTER_NOAPI(H5FD_mpio_setup, FAIL); /* Check arguments */ if(TRUE!=H5P_isa_class(dxpl_id,H5P_DATASET_XFER) || NULL == (plist = H5I_object(dxpl_id))) HRETURN_ERROR(H5E_PLIST, H5E_BADTYPE, FAIL, "not a dataset transfer list"); /* Set buffer MPI type */ if(H5P_insert(plist,H5FD_MPIO_XFER_MEM_MPI_TYPE_NAME,H5FD_MPIO_XFER_MEM_MPI_TYPE_SIZE,&btype,NULL,NULL,NULL,NULL,NULL)<0) HRETURN_ERROR(H5E_PLIST, H5E_CANTSET, FAIL, "can't insert MPI-I/O property"); /* Set file MPI type */ if(H5P_insert(plist,H5FD_MPIO_XFER_FILE_MPI_TYPE_NAME,H5FD_MPIO_XFER_FILE_MPI_TYPE_SIZE,&ftype,NULL,NULL,NULL,NULL,NULL)<0) HRETURN_ERROR(H5E_PLIST, H5E_CANTSET, FAIL, "can't insert MPI-I/O property"); /* Set 'use view' property */ if(H5P_insert(plist,H5FD_MPIO_XFER_USE_VIEW_NAME,H5FD_MPIO_XFER_USE_VIEW_SIZE,&use_view,NULL,NULL,NULL,NULL,NULL)<0) HRETURN_ERROR(H5E_PLIST, H5E_CANTSET, FAIL, "can't insert MPI-I/O property"); FUNC_LEAVE(SUCCEED); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_teardown * * Purpose: Remove the temporary MPI-I/O properties from dxpl. * * Return: Success: Non-negative * Failure: Negative * * Programmer: Quincey Koziol * Monday, June 17, 2002 * * Modifications: * *------------------------------------------------------------------------- */ herr_t H5FD_mpio_teardown(hid_t dxpl_id) { H5P_genplist_t *plist; /* Property list pointer */ FUNC_ENTER_NOAPI(H5FD_mpio_teardown, FAIL); /* Check arguments */ if(TRUE!=H5P_isa_class(dxpl_id,H5P_DATASET_XFER) || NULL == (plist = H5I_object(dxpl_id))) HRETURN_ERROR(H5E_PLIST, H5E_BADTYPE, FAIL, "not a dataset transfer list"); /* Remove buffer MPI type */ if(H5P_remove(dxpl_id,plist,H5FD_MPIO_XFER_MEM_MPI_TYPE_NAME)<0) HRETURN_ERROR(H5E_PLIST, H5E_CANTDELETE, FAIL, "can't remove MPI-I/O property"); /* Remove file MPI type */ if(H5P_remove(dxpl_id,plist,H5FD_MPIO_XFER_FILE_MPI_TYPE_NAME)<0) HRETURN_ERROR(H5E_PLIST, H5E_CANTDELETE, FAIL, "can't remove MPI-I/O property"); /* Remove 'use view' property */ if(H5P_remove(dxpl_id,plist,H5FD_MPIO_XFER_USE_VIEW_NAME)<0) HRETURN_ERROR(H5E_PLIST, H5E_CANTDELETE, FAIL, "can't remove MPI-I/O property"); FUNC_LEAVE(SUCCEED); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_wait_for_left_neighbor * * Purpose: Blocks until (empty) msg is received from immediately * lower-rank neighbor. In conjunction with * H5FD_mpio_signal_right_neighbor, useful for enforcing * 1-process-at-at-time access to critical regions to avoid race * conditions (though it is overkill to require that the * processes be allowed to proceed strictly in order of their * rank). * * Note: This routine doesn't read or write any file, just performs * interprocess coordination. It really should reside in a * separate package of such routines. * * Return: Success: 0 * Failure: -1 * * Programmer: rky * 19981207 * * Modifications: * Robb Matzke, 1999-08-09 * Modified to work with the virtual file layer. *------------------------------------------------------------------------- */ herr_t H5FD_mpio_wait_for_left_neighbor(H5FD_t *_file) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; char msgbuf[1]; MPI_Status rcvstat; FUNC_ENTER_NOAPI(H5FD_mpio_wait_for_left_neighbor, FAIL); assert(file); assert(H5FD_MPIO==file->pub.driver_id); /* p0 has no left neighbor; all other procs wait for msg */ if (file->mpi_rank != 0) { if (MPI_SUCCESS!= MPI_Recv( &msgbuf, 1, MPI_CHAR, file->mpi_rank-1, MPI_ANY_TAG, file->comm, &rcvstat )) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_Recv failed"); } FUNC_LEAVE(SUCCEED); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_signal_right_neighbor * * Purpose: Blocks until (empty) msg is received from immediately * lower-rank neighbor. In conjunction with * H5FD_mpio_wait_for_left_neighbor, useful for enforcing * 1-process-at-at-time access to critical regions to avoid race * conditions (though it is overkill to require that the * processes be allowed to proceed strictly in order of their * rank). * * Note: This routine doesn't read or write any file, just performs * interprocess coordination. It really should reside in a * separate package of such routines. * * Return: Success: 0 * Failure: -1 * * Programmer: rky * 19981207 * * Modifications: * Robb Matzke, 1999-08-09 * Modified to work with the virtual file layer. *------------------------------------------------------------------------- */ herr_t H5FD_mpio_signal_right_neighbor(H5FD_t *_file) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; char msgbuf[1]; FUNC_ENTER_NOAPI(H5FD_mpio_signal_right_neighbor, FAIL); assert(file); assert(H5FD_MPIO==file->pub.driver_id); if (file->mpi_rank != (file->mpi_size-1)) { if (MPI_SUCCESS!= MPI_Send(&msgbuf, 0/*empty msg*/, MPI_CHAR, file->mpi_rank+1, 0, file->comm)) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_Send failed"); } FUNC_LEAVE(SUCCEED); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_fapl_get * * Purpose: Returns a file access property list which could be used to * create another file the same as this one. * * Return: Success: Ptr to new file access property list with all * fields copied from the file pointer. * * Failure: NULL * * Programmer: Robb Matzke * Friday, August 13, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static void * H5FD_mpio_fapl_get(H5FD_t *_file) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; H5FD_mpio_fapl_t *fa = NULL; FUNC_ENTER_NOAPI(H5FD_mpio_fapl_get, NULL); assert(file); assert(H5FD_MPIO==file->pub.driver_id); if (NULL==(fa=H5MM_calloc(sizeof(H5FD_mpio_fapl_t)))) HRETURN_ERROR(H5E_RESOURCE, H5E_NOSPACE, NULL, "memory allocation failed"); /* These should both be copied. --rpm, 1999-08-13 */ fa->comm = file->comm; fa->info = file->info; FUNC_LEAVE(fa); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_open * * Purpose: Opens a file with name NAME. The FLAGS are a bit field with * purpose similar to the second argument of open(2) and which * are defined in H5Fpublic.h. The file access property list * FAPL_ID contains the properties driver properties and MAXADDR * is the largest address which this file will be expected to * access. This is collective. * * Return: Success: A new file pointer. * * Failure: NULL * * Programmer: * January 30, 1998 * * Modifications: * Robb Matzke, 1998-02-18 * Added the ACCESS_PARMS argument. Moved some error checking * here from elsewhere. * * rky, 1998-01-11 * Added H5FD_mpio_Debug debug flags controlled by MPI_Info. * * rky, 1998-08-28 * Init flag controlling redundant metadata writes to disk. * * rky, 1998-12-07 * Added barrier after MPI_File_set_size to prevent race * condition -- subsequent writes were being truncated, causing * holes in file. * * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. * * rky & ppw, 1999-11-07 * Modified "H5FD_mpio_open" so that file-truncation is * avoided for brand-new files (with zero filesize). *------------------------------------------------------------------------- */ static H5FD_t * H5FD_mpio_open(const char *name, unsigned flags, hid_t fapl_id, haddr_t maxaddr/*unused*/) { H5FD_mpio_t *file=NULL; MPI_File fh; int mpi_amode; int mpi_rank; /* MPI rank of this process */ int mpi_size; /* Total number of MPI processes */ int mpi_code; /* mpi return code */ MPI_Offset size; const H5FD_mpio_fapl_t *fa=NULL; H5FD_mpio_fapl_t _fa; H5P_genplist_t *plist; /* Property list pointer */ FUNC_ENTER_NOAPI(H5FD_mpio_open, NULL); #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) { fprintf(stdout, "Entering H5FD_mpio_open(name=\"%s\", flags=0x%x, " "fapl_id=%d, maxaddr=%lu)\n", name, flags, (int)fapl_id, (unsigned long)maxaddr); } #endif /* Obtain a pointer to mpio-specific file access properties */ if(TRUE!=H5P_isa_class(fapl_id,H5P_FILE_ACCESS) || NULL == (plist = H5I_object(fapl_id))) HRETURN_ERROR(H5E_ARGS, H5E_BADTYPE, NULL, "not a file access property list"); if (H5P_DEFAULT==fapl_id || H5FD_MPIO!=H5P_get_driver(plist)) { _fa.comm = MPI_COMM_SELF; /*default*/ _fa.info = MPI_INFO_NULL; /*default*/ fa = &_fa; } else { fa = H5P_get_driver_info(plist); assert(fa); } /* convert HDF5 flags to MPI-IO flags */ /* some combinations are illegal; let MPI-IO figure it out */ mpi_amode = (flags&H5F_ACC_RDWR) ? MPI_MODE_RDWR : MPI_MODE_RDONLY; if (flags&H5F_ACC_CREAT) mpi_amode |= MPI_MODE_CREATE; if (flags&H5F_ACC_EXCL) mpi_amode |= MPI_MODE_EXCL; #ifdef H5FDmpio_DEBUG { /* set debug mask */ /* Should this be done in H5F global initialization instead of here? */ const char *s = HDgetenv ("H5FD_mpio_Debug"); if (s) { while (*s){ H5FD_mpio_Debug[(int)*s]++; s++; } } } /* Check for debug commands in the info parameter */ #if 0 /* Temporary KLUGE rky 2000-06-29, because fa->info is invalid (-1)*/ { char debug_str[128]; int infoerr, flag, i; if (fa->info) { infoerr = MPI_Info_get(fa->info, H5F_MPIO_DEBUG_KEY, 127, debug_str, &flag); if (flag) { fprintf(stdout, "H5FD_mpio debug flags=%s\n", debug_str ); for (i=0; debug_str[i]/*end of string*/ && i<128/*just in case*/; ++i) { H5FD_mpio_Debug[(int)debug_str[i]] = 1; } } } } /* END Temporary KLUGE rky 2000-06-29, because fa->info is invalid (-1) */ #endif #endif /*OKAY: CAST DISCARDS CONST*/ if (MPI_SUCCESS != (mpi_code=MPI_File_open(fa->comm, (char*)name, mpi_amode, fa->info, &fh))) HMPI_RETURN_ERROR(NULL, "MPI_File_open failed", mpi_code); /* Get the MPI rank of this process and the total number of processes */ if (MPI_SUCCESS != MPI_Comm_rank (fa->comm, &mpi_rank)) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, NULL, "MPI_Comm_rank failed"); if (MPI_SUCCESS != MPI_Comm_size (fa->comm, &mpi_size)) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, NULL, "MPI_Comm_size failed"); /* Following changes in handling file-truncation made be rkyates and ppweidhaas, sep 99 */ /* Only processor p0 will get the filesize and broadcast it. */ if (mpi_rank == 0) { /* Get current file size */ if (MPI_SUCCESS != MPI_File_get_size(fh, &size)) { MPI_File_close(&fh); HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, NULL, "MPI_File_get_size failed"); } } /* Broadcast file-size */ if (MPI_SUCCESS != MPI_Bcast(&size, sizeof(MPI_Offset), MPI_BYTE, 0, fa->comm)) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, NULL, "MPI_Bcast failed"); /* Only if size > 0, truncate the file - if requested */ if (size && (flags & H5F_ACC_TRUNC)) { if (MPI_SUCCESS != MPI_File_set_size(fh, (MPI_Offset)0)) { MPI_File_close(&fh); HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, NULL, "MPI_File_set_size failed"); } /* Don't let any proc return until all have truncated the file. */ if (MPI_SUCCESS!= MPI_Barrier(fa->comm)) { MPI_File_close(&fh); HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, NULL, "MPI_Barrier failed"); } size = 0; } /* Build the return value and initialize it */ if (NULL==(file=H5MM_calloc(sizeof(H5FD_mpio_t)))) HRETURN_ERROR(H5E_RESOURCE, H5E_NOSPACE, NULL, "memory allocation failed"); file->f = fh; file->comm = fa->comm; file->info = fa->info; file->mpi_rank = mpi_rank; file->mpi_size = mpi_size; file->eof = H5FD_mpio_MPIOff_to_haddr(size); #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) { fprintf(stdout, "Leaving H5FD_mpio_open\n" ); } #endif FUNC_LEAVE((H5FD_t*)file); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_close * * Purpose: Closes a file. This is collective. * * Return: Success: Non-negative * * Failure: Negative * * Programmer: Unknown * January 30, 1998 * * Modifications: * Robb Matzke, 1998-02-18 * Added the ACCESS_PARMS argument. * * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. *------------------------------------------------------------------------- */ static herr_t H5FD_mpio_close(H5FD_t *_file) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; FUNC_ENTER_NOAPI(H5FD_mpio_close, FAIL); #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "Entering H5FD_mpio_close\n"); #endif assert(file); assert(H5FD_MPIO==file->pub.driver_id); /* MPI_File_close sets argument to MPI_FILE_NULL */ if (MPI_SUCCESS != MPI_File_close(&(file->f)/*in,out*/)) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_close failed"); /* Clean up other stuff */ H5MM_xfree(file); #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "Leaving H5FD_mpio_close\n"); #endif FUNC_LEAVE(SUCCEED); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_query * * Purpose: Set the flags that this VFL driver is capable of supporting. * (listed in H5FDpublic.h) * * Return: Success: non-negative * * Failure: negative * * Programmer: Quincey Koziol * Friday, August 25, 2000 * * Modifications: * *------------------------------------------------------------------------- */ static herr_t H5FD_mpio_query(const H5FD_t *_file, unsigned long *flags /* out */) { #ifndef NDEBUG const H5FD_mpio_t *file = (const H5FD_mpio_t*)_file; #endif /* NDEBUG */ herr_t ret_value=SUCCEED; FUNC_ENTER_NOAPI(H5FD_mpio_query, FAIL); assert(file); assert(H5FD_MPIO==file->pub.driver_id); /* Set the VFL feature flags that this driver supports */ if(flags) { *flags=0; *flags|=H5FD_FEAT_AGGREGATE_METADATA; /* OK to aggregate metadata allocations */ /* Distinguish between updating the metadata accumulator on writes and * reads. This is particularly (perhaps only, even) important for MPI-I/O * where we guarantee that writes are collective, but reads may not be. * If we were to allow the metadata accumulator to be written during a * read operation, the application would hang. */ *flags|=H5FD_FEAT_ACCUMULATE_METADATA_WRITE; /* OK to accumulate metadata for faster writes */ *flags|=H5FD_FEAT_AGGREGATE_SMALLDATA; /* OK to aggregate "small" raw data allocations */ } /* end if */ FUNC_LEAVE(ret_value); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_get_eoa * * Purpose: Gets the end-of-address marker for the file. The EOA marker * is the first address past the last byte allocated in the * format address space. * * Return: Success: The end-of-address marker. * * Failure: HADDR_UNDEF * * Programmer: Robb Matzke * Friday, August 6, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static haddr_t H5FD_mpio_get_eoa(H5FD_t *_file) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; FUNC_ENTER_NOAPI(H5FD_mpio_get_eoa, HADDR_UNDEF); assert(file); assert(H5FD_MPIO==file->pub.driver_id); FUNC_LEAVE(file->eoa); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_set_eoa * * Purpose: Set the end-of-address marker for the file. This function is * called shortly after an existing HDF5 file is opened in order * to tell the driver where the end of the HDF5 data is located. * * Return: Success: 0 * * Failure: -1 * * Programmer: Robb Matzke * Friday, August 6, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static herr_t H5FD_mpio_set_eoa(H5FD_t *_file, haddr_t addr) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; FUNC_ENTER_NOAPI(H5FD_mpio_set_eoa, FAIL); assert(file); assert(H5FD_MPIO==file->pub.driver_id); file->eoa = addr; FUNC_LEAVE(SUCCEED); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_get_eof * * Purpose: Gets the end-of-file marker for the file. The EOF marker * is the real size of the file. * * The MPIO driver doesn't bother keeping this field updated * since that's a relatively expensive operation. Fortunately * the library only needs the EOF just after the file is opened * in order to determine whether the file is empty, truncated, * or okay. Therefore, any MPIO I/O function will set its value * to HADDR_UNDEF which is the error return value of this * function. * * Return: Success: The end-of-address marker. * * Failure: HADDR_UNDEF * * Programmer: Robb Matzke * Friday, August 6, 1999 * * Modifications: * *------------------------------------------------------------------------- */ static haddr_t H5FD_mpio_get_eof(H5FD_t *_file) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; FUNC_ENTER_NOAPI(H5FD_mpio_get_eof, HADDR_UNDEF); assert(file); assert(H5FD_MPIO==file->pub.driver_id); FUNC_LEAVE(file->eof); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_read * * Purpose: Reads SIZE bytes of data from FILE beginning at address ADDR * into buffer BUF according to data transfer properties in * DXPL_ID using potentially complex file and buffer types to * effect the transfer. * * Reading past the end of the MPI file returns zeros instead of * failing. MPI is able to coalesce requests from different * processes (collective or independent). * * Return: Success: Zero. Result is stored in caller-supplied * buffer BUF. * * Failure: -1, Contents of buffer BUF are undefined. * * Programmer: rky, 1998-01-30 * * Modifications: * Robb Matzke, 1998-02-18 * Added the ACCESS_PARMS argument. * * rky, 1998-04-10 * Call independent or collective MPI read, based on * ACCESS_PARMS. * * Albert Cheng, 1998-06-01 * Added XFER_MODE to control independent or collective MPI * read. * * rky, 1998-08-16 * Use BTYPE, FTYPE, and DISP from access parms. The guts of * H5FD_mpio_read and H5FD_mpio_write should be replaced by a * single dual-purpose routine. * * Robb Matzke, 1999-04-21 * Changed XFER_MODE to XFER_PARMS for all H5F_*_read() * callbacks. * * Robb Matzke, 1999-07-28 * The ADDR argument is passed by value. * * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. * * Quincey Koziol, 2002-05-14 * Only call MPI_Get_count if we can use MPI_BYTE for the MPI type * for the I/O transfer. Someday we might include code to decode * the MPI type used for more complicated transfers and call * MPI_Get_count all the time. * * Quincey Koziol - 2002/06/17 * Removed 'disp' parameter from H5FD_mpio_setup routine and use * the address of the dataset in MPI_File_set_view() calls, as * necessary. * *------------------------------------------------------------------------- */ static herr_t H5FD_mpio_read(H5FD_t *_file, H5FD_mem_t UNUSED type, hid_t dxpl_id, haddr_t addr, size_t size, void *buf/*out*/) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; const H5FD_mpio_dxpl_t *dx=NULL; H5FD_mpio_dxpl_t _dx; MPI_Offset mpi_off, mpi_disp; MPI_Status mpi_stat; MPI_Datatype buf_type, file_type; int size_i, bytes_read, n; unsigned use_view_this_time=0, used_view_last_time; H5P_genplist_t *plist; /* Property list pointer */ herr_t ret_value=SUCCEED; FUNC_ENTER_NOAPI(H5FD_mpio_read, FAIL); #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "Entering H5FD_mpio_read\n" ); #endif assert(file); assert(H5FD_MPIO==file->pub.driver_id); /* Make certain we have the correct type of property list */ assert(H5I_GENPROP_LST==H5I_get_type(dxpl_id)); assert(TRUE==H5P_isa_class(dxpl_id,H5P_DATASET_XFER)); /* Portably initialize MPI status variable */ HDmemset(&mpi_stat,0,sizeof(MPI_Status)); /* some numeric conversions */ if (H5FD_mpio_haddr_to_MPIOff(addr, &mpi_off/*out*/)<0) HGOTO_ERROR(H5E_INTERNAL, H5E_BADRANGE, FAIL, "can't convert from haddr to MPI off"); size_i = (int)size; if ((hsize_t)size_i != size) HGOTO_ERROR(H5E_INTERNAL, H5E_BADRANGE, FAIL, "can't convert from size to size_i"); #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'r']) fprintf(stdout, "in H5FD_mpio_read mpi_off=%ld size_i=%d\n", (long)mpi_off, size_i ); #endif /* Obtain the data transfer properties */ if(NULL == (plist = H5I_object(dxpl_id))) HGOTO_ERROR(H5E_ARGS, H5E_BADTYPE, FAIL, "not a file access property list"); if (H5FD_MPIO!=H5P_get_driver(plist)) { _dx.xfer_mode = H5FD_MPIO_INDEPENDENT; /*the default*/ dx = &_dx; } else { dx = H5P_get_driver_info(plist); assert(dx); } /* * Set up for a fancy xfer using complex types, or single byte block. We * wouldn't need to rely on the use_view field if MPI semantics allowed * us to test that btype=ftype=MPI_BYTE (or even MPI_TYPE_NULL, which * could mean "use MPI_BYTE" by convention). */ if(H5P_exist_plist(plist,H5FD_MPIO_XFER_USE_VIEW_NAME)>0) if(H5P_get(plist,H5FD_MPIO_XFER_USE_VIEW_NAME,&use_view_this_time)<0) HGOTO_ERROR(H5E_PLIST, H5E_CANTGET, FAIL, "can't get MPI-I/O type property"); if (use_view_this_time) { /* prepare for a full-blown xfer using btype, ftype, and disp */ if(H5P_get(plist,H5FD_MPIO_XFER_MEM_MPI_TYPE_NAME,&buf_type)<0) HGOTO_ERROR(H5E_PLIST, H5E_CANTGET, FAIL, "can't get MPI-I/O type property"); if(H5P_get(plist,H5FD_MPIO_XFER_FILE_MPI_TYPE_NAME,&file_type)<0) HGOTO_ERROR(H5E_PLIST, H5E_CANTGET, FAIL, "can't get MPI-I/O type property"); /* When using types, use the address as the displacement for * MPI_File_set_view and reset the address for the read to zero */ mpi_disp=mpi_off; mpi_off=0; } else { /* * Prepare for a simple xfer of a contiguous block of bytes. The * btype, ftype, and disp fields are not used. */ buf_type = MPI_BYTE; file_type = MPI_BYTE; mpi_disp = 0; /* mpi_off is alread set */ } /* * Don't bother to reset the view if we're not using the types this time, * and did we didn't use them last time either. */ used_view_last_time = file->old_use_view; if (used_view_last_time || /* change to new ftype or MPI_BYTE */ use_view_this_time) { /* almost certainly a different ftype */ /*OKAY: CAST DISCARDS CONST QUALIFIER*/ if (MPI_SUCCESS != MPI_File_set_view(file->f, mpi_disp, MPI_BYTE, file_type, (char*)"native", file->info)) HGOTO_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_set_view failed"); } /* Keep the 'use view' flag around for the next I/O */ file->old_use_view = use_view_this_time; /* Read the data. */ assert(H5FD_MPIO_INDEPENDENT==dx->xfer_mode || H5FD_MPIO_COLLECTIVE==dx->xfer_mode); if (H5FD_MPIO_INDEPENDENT==dx->xfer_mode) { if (MPI_SUCCESS!= MPI_File_read_at(file->f, mpi_off, buf, size_i, buf_type, &mpi_stat)) HGOTO_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_read_at failed"); } else { #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "H5FD_mpio_read: using MPIO collective mode\n"); #endif if (MPI_SUCCESS!= MPI_File_read_at_all(file->f, mpi_off, buf, size_i, buf_type, &mpi_stat )) HGOTO_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_read_at_all failed"); } /* KLUDGE, Robb Matzke, 2000-12-29 * The LAM implementation of MPI_Get_count() says * MPI_Get_count: invalid argument (rank 0, MPI_COMM_WORLD) * So I'm commenting this out until it can be investigated. The * returned `bytes_written' isn't used anyway because of Kim's * kludge to avoid bytes_written<0. Likewise in H5FD_mpio_write(). */ #ifdef H5_HAVE_MPI_GET_COUNT /* Bill and Albert's kludge*/ /* Yet Another KLUDGE, Albert Cheng & Bill Wendling, 2001-05-11. * Many systems don't support MPI_Get_count so we need to do a * configure thingy to fix this. */ /* Calling MPI_Get_count with "MPI_BYTE" is only valid when we actually * had the 'buf_type' set to MPI_BYTE -QAK */ if(use_view_this_time) { /* Figure out the mapping from the MPI 'buf_type' to bytes, someday... * If this gets fixed (and MPI_Get_count() is reliable), the * kludge below where the 'bytes_read' value from MPI_Get_count() is * overwritten with the 'size_i' parameter can be removed. -QAK */ } /* end if */ else { /* How many bytes were actually read? */ if (MPI_SUCCESS != MPI_Get_count(&mpi_stat, MPI_BYTE, &bytes_read)) HGOTO_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_Get_count failed"); } /* end else */ #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'c']) fprintf(stdout, "In H5FD_mpio_read after Get_count size_i=%d bytes_read=%d\n", size_i, bytes_read ); #endif #endif /* H5_HAVE_MPI_GET_COUNT */ /* * KLUGE rky 1998-02-02 * MPI_Get_count incorrectly returns negative count; fake a complete * read. */ bytes_read = size_i; /* Check for read failure */ if (bytes_read<0 || bytes_read>size_i) HGOTO_ERROR(H5E_IO, H5E_READERROR, FAIL, "file read failed"); /* * This gives us zeroes beyond end of physical MPI file. What about * reading past logical end of HDF5 file??? */ if ((n=(size_i-bytes_read)) > 0) { if (use_view_this_time) { /* * INCOMPLETE rky 1998-09-18 * Haven't implemented reading zeros beyond EOF. What to do??? */ HGOTO_ERROR(H5E_IO, H5E_READERROR, FAIL, "eof file read failed"); } else { memset((char*)buf+bytes_read, 0, (size_t)n); } } done: #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "Leaving H5FD_mpio_read\n" ); #endif FUNC_LEAVE(ret_value); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_write * * Purpose: Writes SIZE bytes of data to FILE beginning at address ADDR * from buffer BUF according to data transfer properties in * DXPL_ID using potentially complex file and buffer types to * effect the transfer. * * MPI is able to coalesce requests from different processes * (collective and independent). * * Return: Success: Zero. USE_TYPES and OLD_USE_TYPES in the * access params are altered. * * Failure: -1, USE_TYPES and OLD_USE_TYPES in the * access params may be altered. * * Programmer: Unknown * January 30, 1998 * * Modifications: * rky, 1998-08-28 * If the file->allsame flag is set, we assume that all the * procs in the relevant MPI communicator will write identical * data at identical offsets in the file, so only proc 0 will * write, and all other procs will wait for p0 to finish. This * is useful for writing metadata, for example. Note that we * don't _check_ that the data is identical. Also, the mechanism * we use to eliminate the redundant writes is by requiring a * call to H5FD_mpio_tas_allsame before the write, which is * rather klugey. Would it be better to pass a parameter to * low-level writes like H5F_block_write and H5F_low_write, * instead? Or...??? Also, when I created this mechanism I * wanted to minimize the difference in behavior between the old * way of doing things (i.e., all procs write) and the new way, * so the writes are eliminated at the very lowest level, here * in H5FD_mpio_write. It may be better to rethink that, and * short-circuit the writes at a higher level (e.g., at the * points in the code where H5FD_mpio_tas_allsame is called). * * * Robb Matzke, 1998-02-18 * Added the ACCESS_PARMS argument. * * rky, 1998-04-10 * Call independent or collective MPI write, based on * ACCESS_PARMS. * * rky, 1998-04-24 * Removed redundant write from H5FD_mpio_write. * * Albert Cheng, 1998-06-01 * Added XFER_MODE to control independent or collective MPI * write. * * rky, 1998-08-16 * Use BTYPE, FTYPE, and DISP from access parms. The guts of * H5FD_mpio_read and H5FD_mpio_write should be replaced by a * single dual-purpose routine. * * rky, 1998-08-28 * Added ALLSAME parameter to make all but proc 0 skip the * actual write. * * Robb Matzke, 1999-04-21 * Changed XFER_MODE to XFER_PARMS for all H5FD_*_write() * callbacks. * * Robb Matzke, 1999-07-28 * The ADDR argument is passed by value. * * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. * * Albert Cheng, 1999-12-19 * When only-p0-write-allsame-data, p0 Bcasts the * ret_value to other processes. This prevents * a racing condition (that other processes try to * read the file before p0 finishes writing) and also * allows all processes to report the same ret_value. * * Kim Yates, Pat Weidhaas, 2000-09-26 * Move block of coding where only p0 writes after the * MPI_File_set_view call. * * Quincey Koziol, 2002-05-10 * Instead of always writing metadata from process 0, spread the * burden among all the processes by using a round-robin rotation * scheme. * * Quincey Koziol, 2002-05-10 * Removed allsame code, keying off the type parameter instead. * * Quincey Koziol, 2002-05-14 * Only call MPI_Get_count if we can use MPI_BYTE for the MPI type * for the I/O transfer. Someday we might include code to decode * the MPI type used for more complicated transfers and call * MPI_Get_count all the time. * * Quincey Koziol - 2002/06/17 * Removed 'disp' parameter from H5FD_mpio_setup routine and use * the address of the dataset in MPI_File_set_view() calls, as * necessary. * *------------------------------------------------------------------------- */ static herr_t H5FD_mpio_write(H5FD_t *_file, H5FD_mem_t type, hid_t dxpl_id, haddr_t addr, size_t size, const void *buf) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; const H5FD_mpio_dxpl_t *dx=NULL; H5FD_mpio_dxpl_t _dx; MPI_Offset mpi_off, mpi_disp; MPI_Status mpi_stat; MPI_Datatype buf_type, file_type; int size_i, bytes_written; unsigned use_view_this_time=0, used_view_last_time; H5P_genplist_t *plist; /* Property list pointer */ herr_t ret_value=SUCCEED; FUNC_ENTER_NOAPI(H5FD_mpio_write, FAIL); #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "Entering H5FD_mpio_write\n" ); #endif assert(file); assert(H5FD_MPIO==file->pub.driver_id); /* Make certain we have the correct type of property list */ assert(H5I_GENPROP_LST==H5I_get_type(dxpl_id)); assert(TRUE==H5P_isa_class(dxpl_id,H5P_DATASET_XFER)); /* Portably initialize MPI status variable */ HDmemset(&mpi_stat,0,sizeof(MPI_Status)); /* some numeric conversions */ if (H5FD_mpio_haddr_to_MPIOff(addr, &mpi_off)<0) HGOTO_ERROR(H5E_INTERNAL, H5E_BADRANGE, FAIL, "can't convert from haddr to MPI off"); size_i = (int)size; if ((hsize_t)size_i != size) HGOTO_ERROR(H5E_INTERNAL, H5E_BADRANGE, FAIL, "can't convert from size to size_i"); #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'w']) fprintf(stdout, "in H5FD_mpio_write mpi_off=%ld size_i=%d\n", (long)mpi_off, size_i); #endif /* Obtain the data transfer properties */ if(NULL == (plist = H5I_object(dxpl_id))) HGOTO_ERROR(H5E_ARGS, H5E_BADTYPE, FAIL, "not a file access property list"); if (H5FD_MPIO!=H5P_get_driver(plist)) { _dx.xfer_mode = H5FD_MPIO_INDEPENDENT; /*the default*/ dx = &_dx; } else { dx = H5P_get_driver_info(plist); assert(dx); } /* * Set up for a fancy xfer using complex types, or single byte block. We * wouldn't need to rely on the use_view field if MPI semantics allowed * us to test that btype=ftype=MPI_BYTE (or even MPI_TYPE_NULL, which * could mean "use MPI_BYTE" by convention). */ if(H5P_exist_plist(plist,H5FD_MPIO_XFER_USE_VIEW_NAME)>0) if(H5P_get(plist,H5FD_MPIO_XFER_USE_VIEW_NAME,&use_view_this_time)<0) HGOTO_ERROR(H5E_PLIST, H5E_CANTGET, FAIL, "can't get MPI-I/O type property"); if (use_view_this_time) { /* prepare for a full-blown xfer using btype, ftype, and disp */ if(H5P_get(plist,H5FD_MPIO_XFER_MEM_MPI_TYPE_NAME,&buf_type)<0) HGOTO_ERROR(H5E_PLIST, H5E_CANTGET, FAIL, "can't get MPI-I/O type property"); if(H5P_get(plist,H5FD_MPIO_XFER_FILE_MPI_TYPE_NAME,&file_type)<0) HGOTO_ERROR(H5E_PLIST, H5E_CANTGET, FAIL, "can't get MPI-I/O type property"); /* When using types, use the address as the displacement for * MPI_File_set_view and reset the address for the read to zero */ mpi_disp=mpi_off; mpi_off=0; } else { /* * Prepare for a simple xfer of a contiguous block of bytes. * The btype, ftype, and disp fields are not used. */ buf_type = MPI_BYTE; file_type = MPI_BYTE; mpi_disp = 0; /* mpi_off is already set */ } /* * Don't bother to reset the view if we're not using the types this time, * and did we didn't use them last time either. */ used_view_last_time = file->old_use_view; if (used_view_last_time || /* change to new ftype or MPI_BYTE */ use_view_this_time) { /* almost certainly a different ftype */ /*OKAY: CAST DISCARDS CONST QUALIFIER*/ if (MPI_SUCCESS != MPI_File_set_view(file->f, mpi_disp, MPI_BYTE, file_type, (char*)"native", file->info)) HGOTO_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_set_view failed"); } /* * We always set the use_view flag to 0 because the default is not to * use types next time, unless someone explicitly requests it by setting * this flag to !=0. */ file->old_use_view = use_view_this_time; /* Only p will do the actual write if all procs in comm write same data */ if ((type!=H5FD_MEM_DRAW) && H5_mpi_1_metawrite_g) { if (file->mpi_rank != file->mpi_round) { #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'w']) { fprintf(stdout, " proc %d: in H5FD_mpio_write (write omitted)\n", file->mpi_rank ); } #endif HGOTO_DONE(SUCCEED) /* skip the actual write */ } } /* Write the data. */ assert(H5FD_MPIO_INDEPENDENT==dx->xfer_mode || H5FD_MPIO_COLLECTIVE==dx->xfer_mode); if (H5FD_MPIO_INDEPENDENT==dx->xfer_mode) { /*OKAY: CAST DISCARDS CONST QUALIFIER*/ if (MPI_SUCCESS != MPI_File_write_at(file->f, mpi_off, (void*)buf, size_i, buf_type, &mpi_stat)) HGOTO_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_write_at failed"); } else { #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "H5FD_mpio_write: using MPIO collective mode\n"); #endif /*OKAY: CAST DISCARDS CONST QUALIFIER*/ if (MPI_SUCCESS != MPI_File_write_at_all(file->f, mpi_off, (void*)buf, size_i, buf_type, &mpi_stat)) HGOTO_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_write_at_all failed"); } /* KLUDGE, Robb Matzke, 2000-12-29 * The LAM implementation of MPI_Get_count() says * MPI_Get_count: invalid argument (rank 0, MPI_COMM_WORLD) * So I'm commenting this out until it can be investigated. The * returned `bytes_written' isn't used anyway because of Kim's * kludge to avoid bytes_written<0. Likewise in H5FD_mpio_read(). */ #ifdef H5_HAVE_MPI_GET_COUNT /* Bill and Albert's kludge*/ /* Yet Another KLUDGE, Albert Cheng & Bill Wendling, 2001-05-11. * Many systems don't support MPI_Get_count so we need to do a * configure thingy to fix this. */ /* Calling MPI_Get_count with "MPI_BYTE" is only valid when we actually * had the 'buf_type' set to MPI_BYTE -QAK */ if(use_view_this_time) { /* Figure out the mapping from the MPI 'buf_type' to bytes, someday... * If this gets fixed (and MPI_Get_count() is reliable), the * kludge below where the 'bytes_written' value from MPI_Get_count() is * overwritten with the 'size_i' parameter can be removed. -QAK */ } /* end if */ else { /* How many bytes were actually written? */ if (MPI_SUCCESS!= MPI_Get_count(&mpi_stat, MPI_BYTE, &bytes_written)) HGOTO_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_Get_count failed"); } /* end else */ #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'c']) fprintf(stdout, "In H5FD_mpio_write after Get_count size_i=%d bytes_written=%d\n", size_i, bytes_written ); #endif #endif /* H5_HAVE_MPI_GET_COUNT */ /* * KLUGE rky, 1998-02-02 * MPI_Get_count incorrectly returns negative count; fake a complete * write. */ bytes_written = size_i; /* Check for write failure */ if (bytes_written<0 || bytes_written>size_i) HGOTO_ERROR(H5E_IO, H5E_READERROR, FAIL, "file write failed"); /* Forget the EOF value (see H5FD_mpio_get_eof()) --rpm 1999-08-06 */ file->eof = HADDR_UNDEF; done: /* Guard against getting into metadate broadcast in failure cases */ if(ret_value!=FAIL) { /* if only p writes, need to broadcast the ret_value to other processes */ if ((type!=H5FD_MEM_DRAW) && H5_mpi_1_metawrite_g) { if (MPI_SUCCESS != MPI_Bcast(&ret_value, sizeof(ret_value), MPI_BYTE, file->mpi_round, file->comm)) HGOTO_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_Bcast failed"); /* Round-robin rotate to the next process */ file->mpi_round = (++file->mpi_round)%file->mpi_size; #ifdef QAK { int max,min; MPI_Allreduce(&file->mpi_round, &max, 1, MPI_INT, MPI_MAX, file->comm); MPI_Allreduce(&file->mpi_round, &min, 1, MPI_INT, MPI_MIN, file->comm); if(max!=file->mpi_round) printf("%s: rank=%d, round=%d, max=%d\n",FUNC,file->mpi_rank,file->mpi_round,max); if(min!=file->mpi_round) printf("%s: rank=%d, round=%d, min=%d\n",FUNC,file->mpi_rank,file->mpi_round,min); } #endif /* QAK */ } /* end if */ } /* end if */ #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "proc %d: Leaving H5FD_mpio_write with ret_value=%d\n", file->mpi_rank, ret_value ); #endif FUNC_LEAVE(ret_value); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_flush * * Purpose: Makes sure that all data is on disk. This is collective. * * Return: Success: Non-negative * * Failure: Negative * * Programmer: Unknown * January 30, 1998 * * Modifications: * Robb Matzke, 1998-02-18 * Added the ACCESS_PARMS argument. * * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. * * Robb Matzke, 2000-12-29 * Make sure file size is at least as large as the last * allocated byte. * * Quincey Koziol, 2002-06-?? * Changed file extension method to use MPI_File_set_size instead * read->write method. * *------------------------------------------------------------------------- */ static herr_t H5FD_mpio_flush(H5FD_t *_file, unsigned closing) { H5FD_mpio_t *file = (H5FD_mpio_t*)_file; int mpi_code; /* mpi return code */ MPI_Offset mpi_off; herr_t ret_value=SUCCEED; #ifdef OLD_WAY uint8_t byte=0; MPI_Status mpi_stat; #endif /* OLD_WAY */ FUNC_ENTER_NOAPI(H5FD_mpio_flush, FAIL); #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "Entering H5FD_mpio_flush\n" ); #endif assert(file); assert(H5FD_MPIO==file->pub.driver_id); #ifdef OLD_WAY /* Portably initialize MPI status variable */ HDmemset(&mpi_stat,0,sizeof(MPI_Status)); #endif /* OLD_WAY */ /* Extend the file to make sure it's large enough, then sync. * Unfortunately, keeping track of EOF is an expensive operation, so * we can't just check whether EOFeoa>file->last_eoa) { #ifdef OLD_WAY if (0==file->mpi_rank) { if (H5FD_mpio_haddr_to_MPIOff(file->eoa-1, &mpi_off)<0) HRETURN_ERROR(H5E_INTERNAL, H5E_BADRANGE, FAIL, "cannot convert from haddr_t to MPI_Offset"); if (MPI_SUCCESS != MPI_File_read_at(file->f, mpi_off, &byte, 1, MPI_BYTE, &mpi_stat)) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_read_at() failed"); if (MPI_SUCCESS != MPI_File_write_at(file->f, mpi_off, &byte, 1, MPI_BYTE, &mpi_stat)) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_write_at() failed"); } /* end if */ #else /* OLD_WAY */ if (H5FD_mpio_haddr_to_MPIOff(file->eoa, &mpi_off)<0) HRETURN_ERROR(H5E_INTERNAL, H5E_BADRANGE, FAIL, "cannot convert from haddr_t to MPI_Offset"); /* Extend the file's size */ if (MPI_SUCCESS != MPI_File_set_size(file->f, mpi_off)) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_File_set_size failed"); /* Don't let any proc return until all have extended the file. * (Prevents race condition where some processes go ahead and write * more data to the file before all the processes have finished making * it the shorter length, potentially truncating the file and dropping * the new data written) */ if (MPI_SUCCESS!= MPI_Barrier(file->comm)) HRETURN_ERROR(H5E_INTERNAL, H5E_MPI, FAIL, "MPI_Barrier failed"); #endif /* OLD_WAY */ /* Update the 'last' eoa value */ file->last_eoa=file->eoa; } /* end if */ /* Only sync the file if we are not going to immediately close it */ if(!closing) { if (MPI_SUCCESS != (mpi_code=MPI_File_sync(file->f))) HMPI_GOTO_ERROR(FAIL, "MPI_File_sync failed", mpi_code); } /* end if */ done: #ifdef H5FDmpio_DEBUG if (H5FD_mpio_Debug[(int)'t']) fprintf(stdout, "Leaving H5FD_mpio_flush\n" ); #endif FUNC_LEAVE(ret_value); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_MPIOff_to_haddr * * Purpose: Convert an MPI_Offset value to haddr_t. * * Return: Success: The haddr_t equivalent of the MPI_OFF * argument. * * Failure: HADDR_UNDEF * * Programmer: Unknown * January 30, 1998 * * Modifications: * Robb Matzke, 1999-04-23 * An error is reported for address overflows. The ADDR output * argument is optional. * * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. *------------------------------------------------------------------------- */ static haddr_t H5FD_mpio_MPIOff_to_haddr(MPI_Offset mpi_off) { haddr_t ret_value=HADDR_UNDEF; FUNC_ENTER_NOINIT(H5FD_mpio_MPIOff_to_haddr); if (mpi_off != (MPI_Offset)(haddr_t)mpi_off) ret_value=HADDR_UNDEF; else ret_value=(haddr_t)mpi_off; FUNC_LEAVE(ret_value); } /*------------------------------------------------------------------------- * Function: H5FD_mpio_haddr_to_MPIOff * * Purpose: Convert an haddr_t value to MPI_Offset. * * Return: Success: Non-negative, the MPI_OFF argument contains * the converted value. * * Failure: Negative, MPI_OFF is undefined. * * Programmer: Unknown * January 30, 1998 * * Modifications: * Robb Matzke, 1999-04-23 * An error is reported for address overflows. The ADDR output * argument is optional. * * Robb Matzke, 1999-07-28 * The ADDR argument is passed by value. * * Robb Matzke, 1999-08-06 * Modified to work with the virtual file layer. *------------------------------------------------------------------------- */ static herr_t H5FD_mpio_haddr_to_MPIOff(haddr_t addr, MPI_Offset *mpi_off/*out*/) { herr_t ret_value=FAIL; FUNC_ENTER_NOINIT(H5FD_mpio_haddr_to_MPIOff); if (mpi_off) *mpi_off = (MPI_Offset)addr; if (addr != (haddr_t)(MPI_Offset)addr) ret_value=FAIL; else ret_value=SUCCEED; FUNC_LEAVE(ret_value); } #endif /*H5_HAVE_PARALLEL*/