ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
5 194 коммитов 3 Ветки 0 Теги 130 MiB
2e0b649362
Граф коммитов

32 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
ea0cf322f3 Taking care of the source pointer. 
This commit was SVN r6691.
 2005-08-01 17:09:46 +00:00
George Bosilca
8e1f169ffd When we copy a convertor we should copy it's flags too. Otherwise the informations inside will 
not match the flags (uninitialized).

This commit was SVN r6681.
 2005-07-29 00:17:02 +00:00
George Bosilca
10d95144b6 Add a global local architecture. It is required by the proc.c file as all new procs by default 
have the same architecture as the current one.

This commit was SVN r6678.
 2005-07-29 00:12:33 +00:00
George Bosilca
4304115e0a Activate the assert as now the ddt engine handle correctly the ref count on the datatypes. 
This commit was SVN r6502.
 2005-07-14 22:02:43 +00:00
George Bosilca
1dfbd0e296 For contiguous datatype set the type in the stack and compute the displacement depending 
on the basic type where we have to start.

This commit was SVN r6497.
 2005-07-14 19:27:58 +00:00
Tim Woodall
76e287a29a George - the change you made breaks the btls - as they 
think the data is contiguous - and the convertor routine
you've changed this to doesn't support returning the correct
offset into the user buffer when a NULL address is provided
in the iovec array

This commit was SVN r6496.
 2005-07-14 19:21:07 +00:00
George Bosilca
269a939a09 Do not allow the macro to execute if there is nothing to do. 
Correctly compute the next displacement, if we stop in the middle of a conversion.

This commit was SVN r6495.
 2005-07-14 19:16:25 +00:00
George Bosilca
77e4e5d1d4 The new pack function should get intensively tested. It should be the default one use 
all the time. There is a performance problem (it's a lot slower than the optimized versions)
but otherwise it will never get tested intensively.

This commit was SVN r6467.
 2005-07-13 21:29:14 +00:00
George Bosilca
b01c7ee66b The flags should be set to zero when we allocate the convertor. 
This commit was SVN r6466.
 2005-07-13 21:27:54 +00:00
George Bosilca
8641a794f8 They get freed in just one place. It's useless to have a macro for that. 
This commit was SVN r6465.
 2005-07-13 21:27:12 +00:00
Galen Shipman
dcbda13a72 Various bug fixes.. 
This commit was SVN r6464.
 2005-07-13 21:13:30 +00:00
George Bosilca
7ca76f83aa pedantic ... 
This commit was SVN r6451.
 2005-07-12 21:53:47 +00:00
George Bosilca
02828d3c97 Finally I get the time to implement a correct reference count for datatypes. The problem 
was that some datatypes can be used in order to create additional datatypes. In such cases
they should not get destroyed otherwise the user will not be able to retrieve how a 
datatype was created. So I decide to never increase the reference count for any predefined
datatypes (as we already know they will never get destroyed, except on finalize). For
the others, every time a datatype is user by another one I increase the reference count.
When I destroy a datatype, I parse (the internal structure args) and release one ref count
on all not predefined datatypes used by this one. Thus the datatypes get cleaned. The main
problem with this approach is the recursivity as this function can trigger another call
to itself (but I dont think it will be an issue).

This commit was SVN r6450.
 2005-07-12 21:52:30 +00:00
George Bosilca
59d886e6c7 pedantic ... 
This commit was SVN r6433.
 2005-07-12 05:31:30 +00:00
George Bosilca
3551f97329 Give a change to all datatype to get an optimized version. Even the contiguous ones ... 
This commit was SVN r6432.
 2005-07-12 05:31:12 +00:00
George Bosilca
a4548c9e89 This optimization require that the true extent of the data is equal to the size of the data. Not the extent. 
This commit was SVN r6431.
 2005-07-12 05:30:35 +00:00
George Bosilca
07ad024ed3 OMPI_ARCH_LONGLONGISxx is a mask so we can get the length of a long long. 
This commit was SVN r6430.
 2005-07-12 05:29:33 +00:00
George Bosilca
717fa6c34f The ompi_ddt_optimize_short function is local to the datatype engine. Do not export it ! 
This commit was SVN r6429.
 2005-07-12 05:28:39 +00:00
George Bosilca
b226af1176 Tolerate the datatype with ZERO elements. They can have an optimized expression (sic!) ... Anyway there is a 
warning message when we try to pack suck kind of datatypes.

This commit was SVN r6388.
 2005-07-08 07:00:18 +00:00
George Bosilca
d38c8b6895 Be nice and remember the real type of the last data. Optimizing the optimization leads to strange behaviours ... 
This commit was SVN r6387.
 2005-07-08 06:50:43 +00:00
George Bosilca
6c6b18eb20 Quit when there is no more space in the iovec. 
This commit was SVN r6386.
 2005-07-08 06:49:07 +00:00
George Bosilca
9a448ece15 A more cleaner version of the pack function. It does not (yet!) have all the optimizations 
from the old version but they will get inside quickly. 

This commit was SVN r6381.
 2005-07-07 23:07:43 +00:00
George Bosilca
e3e49ae42f Force the usage of the new pack function. 
This commit was SVN r6380.
 2005-07-07 23:06:39 +00:00
George Bosilca
c502b6b38c Declare the global "local" convertor. This one can be used for all pack/unpack 
direct calls.

This commit was SVN r6379.
 2005-07-07 23:06:04 +00:00
George Bosilca
d0df88316e Declare and initialize the 2 predefined convertors. 
This commit was SVN r6378.
 2005-07-07 23:03:59 +00:00
George Bosilca
d796186911 If there is no datatype attached to the convertor there is nothing to release. This bug 
happens only with the 2 predefined convertors ...

This commit was SVN r6377.
 2005-07-07 23:03:19 +00:00
George Bosilca
8dea323088 Remember the last type when we do the datatype optimization. 
This commit was SVN r6376.
 2005-07-07 23:01:51 +00:00
Brian Barrett
a13166b500 * rename ompi_output to opal_output 
This commit was SVN r6329.
 2005-07-03 23:31:27 +00:00
Brian Barrett
9f0c969bb4 * rename ompi_hash_table opal_hash_table 
This commit was SVN r6324.
 2005-07-03 16:52:32 +00:00
Brian Barrett
499e4de1e7 * rename ompi_object and ompi_class to opal_object and opal_class 
This commit was SVN r6321.
 2005-07-03 16:06:07 +00:00
Jeff Squyres
aa056f7bfd First cut of OMPI Makefile.am's, plus a few more catchup updates in orte 
This commit was SVN r6286.
 2005-07-02 15:06:47 +00:00
Jeff Squyres
4ab17f019b Rename src -> ompi 
This commit was SVN r6269.
 2005-07-02 13:43:57 +00:00