1
1
Граф коммитов

110 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
8f16780ee0 Add a test for datatypes composed by multiple predefined
elements that can be merged into a larger UINT1 type.

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
(cherry picked from commit 82d632278a)
2019-09-03 15:09:34 -04:00
Jeff Squyres
8b3fd5682f Update OPAL DDT variable names
These variables were renamed in
904276bb44caec207638247f23139bc21bc6a09e; update them to use the new
names.

Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
(cherry picked from commit 2ab8109be1)
2019-08-27 13:08:16 -07:00
George Bosilca
0a00b02e48
Small improvements on the test.
Rework the to_self test to be able to be used as a benchmark.

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
2019-08-05 09:39:02 -04:00
George Bosilca
8b794235b8
Update the datatype dump to match the actual types.
Update the comments to better reflect what is going on.
Minor indentations.

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
2019-08-05 09:37:47 -04:00
George Bosilca
9aff3a9a54
Add more details about what is going on.
Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
2019-03-05 19:43:44 -05:00
George Bosilca
e4aae6b5c8
Add a test for very large data.
Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
2019-03-05 19:43:31 -05:00
Howard Pritchard
8552d0e608
Merge pull request #6330 from ggouaillardet/topic/v4.0.x/ompi_datatype_set_args
ompi/datatype: fix how we compute the space needed for the args
2019-02-08 14:44:08 -07:00
George Bosilca
8acdc53892 Provide a better fix for #6285.
The issue was a little complicated due to the internal stack used in the
convertor. The main issue was that in the case where we run out of iov
space to save the raw description of the data while hanbdling a
repetition (loop), instead of saving the current position and bailing out
directly we reading of the next predefined type element. It worked in
most cases, except the one identified by the HDF5 test. However, the
biggest issue here was the drop in performance for all ensuing calls to
the convertor pack/unpack, as instead of handling contiguous loops as a
whole (and minimizing the number of memory copies) we copied data
description by data description.

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>

(back-ported from commit open-mpi/ompi@5a82c4fd07)
2019-02-01 09:28:52 +09:00
Gilles Gouaillardet
f7327735a0 opal/datatype: fix opal_convertor_raw
correctly handle the case in which iovec is full and the
last accessed element of the datatype is the beginning of a loop

Refs. open-mpi/ompi#6285

Thanks Axel Huebl for reporting this

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>

(back-ported from commit open-mpi/ompi@0832ab5acc)
2019-02-01 09:26:30 +09:00
Gilles Gouaillardet
fd157a960a ompi/datatype: fix how we compute the space needed for the args
Refs. open-mpi/ompi#6275

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>

(cherry picked from commit open-mpi/ompi@45fb69b2b9)
2019-01-30 11:01:11 +09:00
Nathan Hjelm
74563d22a0 test/ddt_lib: remove UB/LB tests
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2018-05-31 09:44:19 -06:00
luz.paz
06b121eb70 Misc. trivial typos
Found via `codespell -q 3`

Signed-off-by: luz paz <luzpaz@users.noreply.github.com>
2018-04-09 11:45:58 -04:00
bosilca
cbf03b3113 Topic/datatype (#3441)
* Don't overflow the internal datatype count.
Change the type of the count to be a size_t (it does not alter the total
size of the internal structures, so has no impact on the ABI).

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>

* Optimize the datatype creation.
The internal array of counts of predefined types is now only created
when needed, which is either in a heterogeneous environment, or when
one call get_elements. It saves space and makes the convertor creation a
little faster in some cases.

Rearrange the fields in the datatype description structs.

The macro OPAL_DATATYPE_INIT_PTYPES_ARRAY had a bug, and the
static array was only partially created. All predefined types should
have the ptypes array created and initialized.

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>

* Fix the boundary computation.

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>

* test/datatype: add test for short unpack on heteregeneous cluster

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
Signed-off-by: George Bosilca <bosilca@icl.utk.edu>

* Trying to reduce the cost of creating a convertor.

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>

* Respect the unpack boundaries.
As Gilles suggested on #2535 the opal_unpack_general_function was
unpacking based on the requested count and not on the amount of packed
data provided.
Fixes #2535.

Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
2017-05-09 09:31:40 -04:00
Gilles Gouaillardet
fa5cd0dbe5 use ptrdiff_t instead of OPAL_PTRDIFF_TYPE
since Open MPI now requires a C99, and ptrdiff_t type is part of C99,
there is no more need for the abstract OPAL_PTRDIFF_TYPE type.

Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
2017-04-19 13:41:56 +09:00
Ralph Castain
1e2019ce2a Revert "Update to sync with OMPI master and cleanup to build"
This reverts commit cb55c88a8b.
2016-11-22 15:03:20 -08:00
Ralph Castain
cb55c88a8b Update to sync with OMPI master and cleanup to build
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
2016-11-22 14:24:54 -08:00
Joshua Hursey
f6f24a4f67 build: Custom libmpi(_FOO) name option in configure
* Add a configure time option to rename libmpi(_FOO).*
   - `--with-libmpi-name=STRING`
 * This commit only impacts the installed libraries.
   Internal, temporary libraries have not been renamed to limit the
   scope of the patch to only what is needed.

For example:
```shell
shell$ ./configure --with-libmpi-name=wookie
...
shell$ find . -name "libmpi*"
shell$ find . -name "libwookie*"
./lib/libwookie.so.0.0.0
./lib/libwookie.so.0
./lib/libwookie.so
./lib/libwookie.la
./lib/libwookie_mpifh.so.0.0.0
./lib/libwookie_mpifh.so.0
./lib/libwookie_mpifh.so
./lib/libwookie_mpifh.la
./lib/libwookie_usempi.so.0.0.0
./lib/libwookie_usempi.so.0
./lib/libwookie_usempi.so
./lib/libwookie_usempi.la
shell$
```
2016-09-29 21:47:24 -05:00
Gilles Gouaillardet
5932287cef datatype/[un]pack_external[_size]: move subroutines down to ompi/datatype
so it can be directly used by test/datatype/external32
2016-03-30 13:01:33 +09:00
Gilles Gouaillardet
63eec552b2 datatype/external32: fix test 2016-03-30 11:40:18 +09:00
Gilles Gouaillardet
0f9ddbe02e test/datatype/external32: make it a non MPI application 2016-03-30 11:40:18 +09:00
George Bosilca
383b78d8a6 Validate the packed data
Use htonl and htons to check that the packed data is
indeed the correct data.
2016-03-30 11:40:18 +09:00
George Bosilca
8da67f5ef8 Add a test to validate the support for external32. 2016-03-30 11:40:18 +09:00
Ralph Castain
06f0322f7b Enable the ability to make tarballs on the Mac tarball by ensuring that "make distclean" completely cleans up the test directories
Add one more thing to be cleaned
2015-11-24 21:14:38 -08:00
Jeff Squyres
447eb4238a unpack_ooo: fix empty initializer 2015-08-29 04:57:02 -07:00
Jeff Squyres
ea935df632 test Makefile.am: fix flags
WRAPPER_EXTRA_LDFLAGS is both no longer used, and even if we fixed to
use the new names (i.e., <PACKAGE>_WRAPPER_EXTRA_LDFLAGS), it contains
a token that will not be understandable in Makefile.ams: @{libdir}.

Hence, update all instances to use <PACKAGE>_PKG_CONFIG_LDFLAGS, which
does not contain the @{libdir} token.  See comment in
opal_setup_wrappers.m4 for an explanation.
2015-08-25 14:04:51 -07:00
Jeff Squyres
42b9a966d6 Makefile.am's: if calling OPAL functions, must link to it
On some OSs (e.g., Ubuntu 14.04.2 LTS), the linker is configured such
that the symbols of library dependencies are not available to the
application.  Hence, we need to explicitly list such dependencies when
creating the executable.

For this commit, these tests are use OPAL function calls, so we must
explicitly link in libopen-pal.so.
2015-08-14 07:51:55 -07:00
Nathan Hjelm
ee36d813dc Merge pull request #657 from hjelmn/c99
more c99 updates
2015-06-25 11:21:09 -06:00
Nathan Hjelm
4d92c9989e more c99 updates
This commit does two things. It removes checks for C99 required
headers (stdlib.h, string.h, signal.h, etc). Additionally it removes
definitions for required C99 types (intptr_t, int64_t, int32_t, etc).

Signed-off-by: Nathan Hjelm <hjelmn@me.com>
2015-06-25 10:14:13 -06:00
rhc54
1a767ed47c Merge pull request #654 from rhc54/topic/config
Remove internal bool type definitions
2015-06-25 09:10:21 -07:00
Ralph Castain
57b7c44aec Cleanup the rest of the no-longer-required bool configury 2015-06-24 08:25:49 -07:00
Ralph Castain
869041f770 Purge whitespace from the repo 2015-06-23 20:59:57 -07:00
Gilles Gouaillardet
89882f11bd rewrite ddt_pack test
so make check works without make install
2015-05-07 18:14:32 +09:00
Jeff Squyres
98943d3c8b test/datatype: move ddt_pack to MPI_CHECKS
This still leaves the test disabled (per
b25a7bfe14), but it does still build the
test.
2015-05-02 03:57:19 -07:00
Jeff Squyres
b25a7bfe14 test/datatype: temporarily disable ddt_pack test
This test needs to be tweaked to not use MPI_Init (and only use
opal_init_util(), per #568).  @ggouaillardet can't get to this until
next week sometime, so we're just temporarily disabling it for now.

Fixes #568
2015-05-02 03:53:08 -07:00
George Bosilca
53e9543be4 This test should pass. 2015-04-30 22:51:27 -04:00
George Bosilca
27c3631466 Correctly free the temporary datatypes 2015-04-30 22:51:01 -04:00
George Bosilca
267fdfe435 Add a corner case for struct with multiple predefined types.
Cleanup the datatypes once they are not useful.
Other minor fixes.
2015-04-30 22:06:49 -04:00
Gilles Gouaillardet
06ee9f9ddf enhance test/datatype/ddt_pack
* add a simple test with MPI_Type_vector
 * add some extra checks : (true)lower_bound and (true)extent
   of original and unpacked ddt should match
2015-04-30 15:04:53 +09:00
Gilles Gouaillardet
ed0c48de74 test/datatype/ddt_pack fix
unpacking a packed MPI_INT returns MPI_INT32T instead of MPI_INT
2015-04-30 15:04:53 +09:00
George Bosilca
98cca1b6b3 Remove all uninitialized common symbols in the test directory. 2015-04-25 13:54:45 -04:00
Jeff Squyres
4f1139590f datatypes: trivial spelling fix
s/commited/committed/gi
2015-02-27 18:16:01 -08:00
George Bosilca
82c02b471e Take in acount the lower bound of the data. 2014-12-20 21:28:58 -05:00
George Bosilca
4f9a3bdbab Correctly compute the size of the needed memory for the datatype tests.
Fixes open-mpi/ompi#294.
2014-12-20 01:30:37 -05:00
George Bosilca
cf3ff3fe58 This was not supposed to be part of the 1895f29 commit. 2014-12-19 09:55:18 -05:00
George Bosilca
1895f29537 Remove all warnings from the datatype tests. 2014-12-18 02:58:21 -05:00
Ralph Castain
91bec7e9dd Fix some type declarations so make check works for SPARC. Thanks to Paul Hargrove for the report and correction 2014-12-15 06:44:51 -08:00
George Bosilca
8ee501350b Implement strict validation for the packing/unpacking of the data. 2014-12-02 16:22:18 +09:00
Ralph Castain
53af0f1594 Temporarily turn off a specific datatype test that is preventing the nightly tarball from running 2014-12-01 19:52:37 -08:00
George Bosilca
59b739ee90 Add a resized datatype to the test. Implement generic
data corectness for most of the conversion functions.
2014-11-29 19:47:25 -05:00
George Bosilca
ee3d1ed5fd Add tests for vector type. 2014-11-24 01:52:49 -05:00