Statistiques
| Révision :

root / Pi @ 186

# Date Auteur Commentaire
186 07/10/2018 17:12 Emmanuel Quemener

Add dynamic allocation of each inside exploration.
185 07/10/2018 16:59 Emmanuel Quemener

Add metrology inside code. Add dynamic allocation on inside array.
184 07/10/2018 16:10 Emmanuel Quemener

Include metrology inside code. Correct mrproper in Makefile
183 07/10/2018 12:11 Emmanuel Quemener

Add OpenACC with PGI compiler for Nvidia GPUs
182 28/06/2018 14:22 Emmanuel Quemener

Correct to compile.
181 11/06/2018 04:35 Emmanuel Quemener

Add -k option to perform classical test based on if.
177 01/04/2018 12:02 Emmanuel Quemener

Correct CUDA call with direct ccbin compiler call (clang compiles but gcc over 5 not...)
174 21/03/2018 14:29 Emmanuel Quemener

Add comments on top of source files.
159 03/02/2018 11:07 Emmanuel Quemener

Tiny changes.
158 03/02/2018 05:15 Emmanuel Quemener

Wrong cast in FP64 subroutine.
157 25/01/2018 11:03 Emmanuel Quemener

Replace *PU by xPU to avoid prolems on files.
156 21/01/2018 06:02 Emmanuel Quemener

Add INT32, INT64, FP64 possibilities.
154 11/11/2017 03:47 Emmanuel Quemener

Add granularity on variables and Marsaglia RNG.
153 13/10/2017 11:39 Emmanuel Quemener

Correct tiny bug on metrology estimation.
152 27/09/2017 23:25 Emmanuel Quemener

Add C/OpenCL implementation.
143 17/10/2016 22:42 Emmanuel Quemener

Cleaning process.
138 29/09/2016 11:35 Emmanuel Quemener

Modify the "device.type" in order not to crash with POCL implementation.
131 09/05/2016 15:45 Emmanuel Quemener

Change default name of output.
130 09/05/2016 14:30 Emmanuel Quemener

Add Pi estimation.
129 14/03/2016 22:27 Emmanuel Quemener

Correct some errors on CUDA implementation.
127 03/03/2016 13:22 Emmanuel Quemener

Transform from Python 2 to Python 3
126 03/03/2016 10:12 Emmanuel Quemener

Suppress stdout messages and retreive computing time on master.
125 02/03/2016 15:28 Emmanuel Quemener

Suppress stdout messages on slaves and import elapsed on master.
124 05/02/2016 11:58 Emmanuel Quemener

Add asynchrone MPI communications under Hybrid MPI/OpenMP implementation.
123 21/01/2016 09:38 Emmanuel Quemener

Identation for XPU parsing.
122 06/01/2016 17:26 Emmanuel Quemener

Correct CUDA implementation.
108 18/11/2015 16:03 Emmanuel Quemener

Improve MPI distribution.
107 09/11/2015 13:28 Emmanuel Quemener

Add MPI and pThreads version to use multiple OpenCL devices.
106 07/11/2015 21:48 Emmanuel Quemener

Add MPI version which support multi GPU.
105 06/11/2015 22:38 Emmanuel Quemener

Change structure of call for OpenCL Metropolis Routine
104 05/11/2015 22:08 Emmanuel Quemener

Internal modifications on statistics.
103 31/10/2015 04:07 Emmanuel Quemener

Changes on micmac global variables.
102 31/10/2015 03:14 Emmanuel Quemener

Minor change on doc.
101 22/10/2015 06:37 Emmanuel Quemener

Major evolution (statistics, etc).
100 21/10/2015 06:37 Emmanuel Quemener

First revision. Crash on hybrid.
99 20/10/2015 06:22 Emmanuel Quemener

Add hybrid version.
87 27/03/2015 17:16 Emmanuel Quemener

Correct bug with NP=1
86 12/03/2015 15:55 Emmanuel Quemener

Licence definition. Change default type and RNG.
85 12/03/2015 11:20 Emmanuel Quemener

Minor changes on Maximum thread values.
84 09/03/2015 13:58 Emmanuel Quemener

Add granularity on variables types and Marsaglia RNG versions. Add licence
83 09/03/2015 13:51 Emmanuel Quemener

Add Cecill v2 licence on source code
82 04/03/2015 09:52 Emmanuel Quemener

Add granularity on variable types and Marsaglia RNG generators.
81 04/03/2015 09:51 Emmanuel Quemener

Minor change about check.
80 04/03/2015 09:40 Emmanuel Quemener

Extend granularity on size and Marsaglia RNG generators. Add both asynchrone and synchrone MPI calls.
79 03/03/2015 14:53 Emmanuel Quemener

Add granularity choice on type of counters and type of Marsaglia generator.
78 06/02/2015 18:41 Emmanuel Quemener

Support for Intel Xeon Phi
77 05/12/2014 18:23 Emmanuel Quemener

Replace synchrone to asynchrone MPI calls as in Hybrid version.
76 05/12/2014 18:22 Emmanuel Quemener

Replace synchrone to asynchrone MPI Send/Receive. At the beginning only to avoid distribution of tasks
but it was a problem on OpenIB (mlx4_core.log_mtts_per_seg=5 to add in GRUB)
75 26/11/2014 18:04 Emmanuel Quemener

Modify output to provide rates.
74 23/11/2014 17:33 Emmanuel Quemener

Convert CUDA implementation as OpenCL one.
73 23/11/2014 16:16 Emmanuel Quemener

Split MainLoop* by calls on one MainLoop
72 22/11/2014 00:28 Emmanuel Quemener

Add different Marsaglia RNG.
71 17/11/2014 14:56 Emmanuel Quemener

Minor modifications.
70 09/10/2014 17:22 Emmanuel Quemener

Add vendor print and strip output on device name.
65 24/06/2014 17:42 Emmanuel Quemener

Minor changes
64 22/06/2014 00:35 Emmanuel Quemener

Add Hybrid MPI/OpenMP version
59 15/05/2014 16:57 Emmanuel Quemener

Add comment.
56 06/05/2014 11:06 Emmanuel Quemener

Add hostname print. Correct bugs.
55 26/03/2014 17:47 Emmanuel Quemener

Exception add.
50 26/02/2014 11:07 Emmanuel Quemener

Add Simple/Double precision in Kernels. Modify output to simplify import in CSV and GNUplot
49 10/02/2014 16:09 Emmanuel Quemener

Change Pi estimation from global division to atomic division.
48 10/02/2014 11:30 Emmanuel Quemener

Add Xeon Phi support for ACCELERATOR type
47 04/12/2013 11:25 Emmanuel Quemener

Improved version.
46 15/10/2013 10:10 Emmanuel Quemener

Changes on Device selection and metrology statements.
45 15/10/2013 07:23 Emmanuel Quemener

Minor changes on output filename.
42 30/09/2013 11:29 Emmanuel Quemener

Minor changes.
41 30/09/2013 09:14 Emmanuel Quemener

Change size of iterations variables (in functions)...
40 30/09/2013 09:12 Emmanuel Quemener

Change size of Iterations variable.
39 30/09/2013 07:27 Emmanuel Quemener

Minor changes on default values for bench.
38 30/09/2013 05:54 Emmanuel Quemener

Change minor bug.
37 30/09/2013 05:53 Emmanuel Quemener

Change minor bug.
36 30/09/2013 05:52 Emmanuel Quemener

Change optimization to O3.
35 30/09/2013 05:46 Emmanuel Quemener

Change to O3 optimization.
34 30/09/2013 05:45 Emmanuel Quemener

Change to O3 optimization.
33 30/09/2013 05:43 Emmanuel Quemener

Change name of executable.
32 30/09/2013 05:43 Emmanuel Quemener

Change name of executable.
31 29/09/2013 20:53 Emmanuel Quemener

Change atol to atoll to support long long
30 29/09/2013 20:53 Emmanuel Quemener

Minor changes.
29 29/09/2013 20:47 Emmanuel Quemener

Add architecture information. Add better support on 32bits architecture
28 29/09/2013 20:35 Emmanuel Quemener

Minor changes on output information about architecture.
27 23/09/2013 00:00 Emmanuel Quemener

Minor changes.
26 22/09/2013 11:58 Emmanuel Quemener

Change unsigned long to long long for 32 bits compatibility.
25 21/09/2013 14:06 Emmanuel Quemener

Add support for 32 bits and large number of iterations (use of long long int)
24 18/09/2013 00:22 Emmanuel Quemener

Delete executable
23 18/09/2013 00:21 Emmanuel Quemener

Changes to support long (over 2^32 iterations)
22 04/09/2013 06:18 Emmanuel Quemener

Add script to reduce GProf output to show source code lines.
21 04/09/2013 06:15 Emmanuel Quemener

Merge INT and LOG elements.
Add sqrt and classical test on quadrant.
20 04/09/2013 06:14 Emmanuel Quemener
17 05/06/2013 14:36 Emmanuel Quemener

Improve choice of GPU/CPU.
16 08/04/2013 12:02 Emmanuel Quemener

Modify randint call domain for 32 bits machines.
15 08/04/2013 11:10 Emmanuel Quemener

Add jobs mention in output process.
14 03/04/2013 17:02 Emmanuel Quemener

Delete executables.
13 03/04/2013 17:01 Emmanuel Quemener

Modify bench to specify OMP_NUM_THREADS, needed on ARM architecture
12 01/04/2013 21:21 Emmanuel Quemener

Add Pthreads implementation.
11 01/04/2013 21:06 Emmanuel Quemener

Add script for bench on each implementation.
10 01/04/2013 21:05 Emmanuel Quemener

Tiny modifications for input/output
9 31/03/2013 07:46 Emmanuel Quemener

Add MPI simple version
8 31/03/2013 07:45 Emmanuel Quemener

Add OpenMP version.
7 30/03/2013 17:50 Emmanuel Quemener

Add Pi test for Metrology tests.