1
1
Граф коммитов

6378 Коммитов

Автор SHA1 Сообщение Дата
Jeff Squyres
d55893349f Convert MPI_Offset to be a typedef instead of a #define. See the
following for the rationale:

http://www.open-mpi.org/community/lists/users/2005/11/0391.php

Note that mpi.h only receives AC_DEFINE's (not AC_SUBST's), so we
still have to AC_DEFINE to get the value in there -- so we AC_DEFINE
something that we give to a typedef.

This commit was SVN r8326.
2005-11-30 03:16:24 +00:00
George Bosilca
b47aa5c142 If pending_length is not zero the datatype engine believe that a partial conversion is still in the buffer attached to the convertor. We have to force it to zero for all new convertors. This bug is still half fixed until I figure out a correct way to do it when we clone a convertor ... Anyway, this future correction will not affect open mpi.
This commit was SVN r8324.
2005-11-30 00:16:58 +00:00
Jeff Squyres
f12bd6f442 Sync with v1.0 NEWS
This commit was SVN r8323.
2005-11-29 23:44:56 +00:00
Jeff Squyres
20eebeaa14 MPI_TESTANY is described in MPI-1 p46. The last 2 full paragraphs of
the page address three situations:

1. if the array of requests are all REQUEST_NULL, set flag=true and
   index=MPI_UNDEFINED
2. if the array of requests are not all REQUEST_NULL, if any of them
   finished, set flag=true and index=whatever_the_index_was
3. if the array of requests are not all REQUEST_NULL, if none of them
   finished, set flag=false and index=MPI_UNDEFINED

With regards to the index value, we are currently doing 1 and 2, but
not 3.  More specifically, index should be set to MPI_UNDEFINED if no
requests are found to have completed (regardless of whether they are
REQUEST_NULL or not).

This patch fixes this problem.

This commit was SVN r8319.
2005-11-29 23:41:04 +00:00
Brian Barrett
13b8e166ba * documentation updates for memory hooks. Reordered some declarations, but there
were no interface changes in this commit.

This commit was SVN r8317.
2005-11-29 19:42:07 +00:00
Jeff Squyres
4220a5cd95 Add bullet about Fortran arguments
This commit was SVN r8315.
2005-11-29 18:43:27 +00:00
Brian Barrett
a46d0d3801 Add support to Open MPI for MPI_Register_datarep, as ROMIO now has support for the
functionality.  This is a global registration, so all components have to be made
aware of the registration -- therefore, do things at the component level, not the
module level.

This commit was SVN r8314.
2005-11-29 18:25:35 +00:00
Jeff Squyres
c62ba797a0 Add some symbols from ROMIO to ignore
This commit was SVN r8313.
2005-11-29 17:57:25 +00:00
Tim Woodall
5db38b38f5 corrections for latency issue
- don't do additional select until non-blocking read fails 
- don't do an additional read for 0 byte message

This commit was SVN r8312.
2005-11-29 17:33:01 +00:00
Brian Barrett
6bd76c8f91 * prefix the newest of the ROMIO MPI interface functions
This commit was SVN r8311.
2005-11-29 15:45:00 +00:00
Brian Barrett
f35d5d67a8 * PGI compilers found this little bugglet in ROMIO - check the datatype, not
the count when using the DATATYPE check macro....

This commit was SVN r8310.
2005-11-29 14:56:30 +00:00
Rainer Keller
b528a4dbcd - Add string.h for --enable-picky build.
This commit was SVN r8307.
2005-11-29 09:17:47 +00:00
George Bosilca
906d19a73e A simpler version for the C function who is supposed to print the value of TRUE from Fortran. We support
only the fortran compilers who define the LOGICAL as being at most of the same size as an int.

This commit was SVN r8306.
2005-11-29 06:35:46 +00:00
George Bosilca
cae4b603f4 First cut of the checksum capabilities for the convertor. Right now it's always set to zero, but it will change shortly.
This commit was SVN r8305.
2005-11-29 06:33:21 +00:00
Brian Barrett
79bf8843d2 * update memory hooks interface to allow for callbacks on both allocations
and dealllocations, per request from Galen and Tim

This commit was SVN r8303.
2005-11-29 04:46:14 +00:00
Brian Barrett
5e0c779974 * add note to NEWS about updated ROMIO release
This commit was SVN r8301.
2005-11-29 02:07:34 +00:00
Brian Barrett
8b9a539986 * update ROMIO with the version from MPICH 1.2.7p1, marked as version
2005-06-09

This commit was SVN r8300.
2005-11-29 01:41:53 +00:00
Brian Barrett
682615dafe * make sure to distribute crc.h in dist tarballs
This commit was SVN r8298.
2005-11-29 01:07:30 +00:00
Jeff Squyres
8c31ab1dee Update README and NEWS about Fortran .TRUE. handling.
This commit was SVN r8297.
2005-11-29 01:06:44 +00:00
Tim Woodall
cf53d3e48f missing include
This commit was SVN r8295.
2005-11-28 23:13:36 +00:00
Galen Shipman
6e64e8a144 bproc fixes, these exist in the release 1.0 branch.
This commit was SVN r8292.
2005-11-28 21:10:02 +00:00
Galen Shipman
ff1a820384 set the lib path correctly, build system doesn't like white space..
This commit was SVN r8291.
2005-11-28 21:08:39 +00:00
Galen Shipman
209bc9ca93 fix..
This commit was SVN r8288.
2005-11-28 20:44:53 +00:00
George Bosilca
9d990af4a5 Remove 2 useless functions. They have been replaced by the mca_base version few commits ago.
This commit was SVN r8287.
2005-11-28 20:14:23 +00:00
Jeff Squyres
63a7ea98e2 Added bullet about MX startup scalability
This commit was SVN r8286.
2005-11-28 20:09:46 +00:00
Tim Woodall
a6e91439b5 corrected copyrights
This commit was SVN r8284.
2005-11-28 20:09:21 +00:00
Jeff Squyres
09325ebb92 Be a bit more precise on the MPI_SCATTERV bullet.
This commit was SVN r8281.
2005-11-28 19:39:07 +00:00
Jeff Squyres
170bff7f7e Add bullets for MPI_SCATTERV and iof fixes
This commit was SVN r8280.
2005-11-28 19:36:59 +00:00
Tim Woodall
7667cfb104 checksum/crc routines from la-mpi
This commit was SVN r8272.
2005-11-28 17:20:21 +00:00
George Bosilca
c9c205302d Correct the TRUE detection code to be more Fortran friendly and not based on some compiler
features. Now it works with gfortran as well.

The problem was that the cast between LOGICAL and INTEGER is not a Fortran language feature. The easiest solution around this problem is to have a C function that print the value. We set the value in fortran and then call the C function to print it out.

This commit was SVN r8266.
2005-11-28 16:26:00 +00:00
Tim Woodall
943e6f0cd5 corrections for stdin
- when eof is reached at orterun, send a 0 byte message to peer indicating eof
- on receipt of zero byte message - close corresponding file descriptor associated with the endpoint
- require setup ptys for stdin and stdout so that stdin can be closed independently of stdout

This commit was SVN r8264.
2005-11-28 14:58:53 +00:00
Galen Shipman
55a9fbefd8 fix misc compiler warnings..
This commit was SVN r8263.
2005-11-27 22:53:30 +00:00
George Bosilca
16ca6e4c88 error seems to be a reserved keyword for some compilers ...
This commit was SVN r8262.
2005-11-26 21:18:47 +00:00
George Bosilca
bfa4a40983 Cast it the a well known type to remove a warning.
This commit was SVN r8261.
2005-11-26 21:17:15 +00:00
George Bosilca
b9a739e2b6 Remove 2 useless assignments (they are done at the end before the return).
This commit was SVN r8260.
2005-11-26 21:16:30 +00:00
George Bosilca
011e46017f Match the latest changes on the TEG PML (increase ompi_request_completed when a request is
finished).

This commit was SVN r8259.
2005-11-26 21:06:37 +00:00
Rainer Keller
bf0892bb32 - Implement correct Fortran Logical-handling in f77/f90 interface in
case of:
    sizeof(MPI_Flogical) != sizeof (int)
  and
    Fortran value of .TRUE. != 1
  as is often the case.
- Check in configure the value of .TRUE., the C-type coresponding to
  logical and check, that fortran compiler does not do something strange
  with arrays of logicals
- Convert all occurrences of logicals in the fortran wrappers, only
  in case it is needed.
  *Please note* Implementation of MPI_Cart_sub needed special treatment.
- Output these value in ompi_info -a
- Clean up the prototypes_mpi.h to just have a single definition and
  thereby deleting the necessity for prototypes_pmpi.h

- configured, compiled and tested with F90-program, which uses
  MPI_Cart_create and MPI_Cart_get:
  linux ia32, gcc (no testing, as no f90)
  linux ia32, gcc --disable-mpi-f77 --disable-mpi-f90 (had a bug there)
  linux ia32, icc-8.1
  linux opteron, gcc-3.3.5, pgcc, pathccx/pathf90 (tested just
pgi-compiler)
  linux em64t, gcc, icc-8.1 (tested just icc)

This commit was SVN r8254.
2005-11-24 16:52:35 +00:00
George Bosilca
00c10a6372 Make the MX BTL startup scalable. When the number of processes involved in the MPI application
increase the previous connection code was broken. It can take as much as 60 seconds to connect
64 processes. Now we do not create the connections when we add the procs but only when we send
them the first message. Now it take only 1.6 seconds to setup a 64 procs MPI job over MX (doing a 2 steps barrier in order to insure that we create all the connections).

This commit was SVN r8252.
2005-11-23 23:48:56 +00:00
Edgar Gabriel
83cef7f8ac in scatterv we tested unfortunatly the wrong datatype for the
displacement (for both the inter and intra-communicator version). The
displacements in scatterv are given in multiples of the sendtype.
This fix should probably make to v1.0.1 as well?

This commit was SVN r8251.
2005-11-23 21:42:45 +00:00
Brian Barrett
c6fb3217f8 always include ltdl.h with the full opal path, so that we always grab our
version instead of the (possible) system installed version.

This commit was SVN r8248.
2005-11-23 19:30:12 +00:00
Jeff Squyres
6b823e3663 Remove a little extra memory that is no longer needed
This commit was SVN r8246.
2005-11-23 11:16:59 +00:00
Edgar Gabriel
c96e4d43c7 fixes for other routines which take an MPI-2 byte displacement or an address as
an argument.

This commit was SVN r8244.
2005-11-23 03:52:19 +00:00
Edgar Gabriel
a1b00e0a49 fixing the fortran interface for MPI_Get_address. Since the address
argument is defined to be 

integer (KIND=MPI_ADDRESS_KIND)

we have to map that to an MPI_Aint in the f->c interface, else the
typecast back from the C to the Fortran interface will be wrong as well.
The same holds (at least) for all other new MPI-2 datatype functions
which take a byte-displacement or an address as an argument.

This commit was SVN r8243.
2005-11-23 03:22:50 +00:00
Edgar Gabriel
b6ddb73f38 fixing the condition under which the values of the c-functions are
passed back to the fortran routines. Should probably go to v1.0.1, but
is unfortunatly just half of the overall fix. 

This commit was SVN r8242.
2005-11-23 03:12:40 +00:00
George Bosilca
05a5fb4b05 One less dynamically alllocated array. We can just use vfprint.
This commit was SVN r8239.
2005-11-22 21:53:39 +00:00
George Bosilca
58ba8f033d Handle the case when there is no F77 support available for us.
This commit was SVN r8238.
2005-11-22 21:52:14 +00:00
George Bosilca
7c73095440 Update the correct sended size.
This commit was SVN r8237.
2005-11-22 21:51:04 +00:00
Tim Woodall
b06335abe2 start of a pml for data reliability
This commit was SVN r8236.
2005-11-22 17:24:47 +00:00
Jeff Squyres
949fbcabc7 A pox on Linux and the convoluted configure tests that it makes me
do.  Really fix the MCA info params this time -- ensure to not use
#define's that may be #undef'ed.

This commit was SVN r8235.
2005-11-22 17:18:14 +00:00
Jeff Squyres
2dec4df0b5 Update NEWS
This commit was SVN r8233.
2005-11-22 15:33:04 +00:00