-
Notifications
You must be signed in to change notification settings - Fork 0
/
CHANGES
250 lines (226 loc) · 8.31 KB
/
CHANGES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
OSU Micro Benchmarks v5.2 (02/05/16)
* New Features & Enhancements
- Support for CUDA-Aware Managed memory
* osu_bibw
* osu_bw
* osu_latency
* osu_allgather
* osu_allgatherv
* osu_allreduce
* osu_alltoall
* osu_alltoallv
* osu_bcast
* osu_gather
* osu_gatherv
* osu_reduce
* osu_reduce_scatter
* osu_scatter
* osu_scatterv
- Add ability to specify minimum message size in addition to maximum
message size for all collective benchmarks
OSU Micro Benchmarks v5.1 (11/10/15)
* New Features & Enhancements
- Introduce non-blocking collective v-variants as well as ialltoallw
* osu_iallgatherv
* osu_ialltoallv
* osu_igatherv
* osu_iscatterv
* osu_ialltoallw
- Add support for benchmarking GPU-Aware non-blocking collectives. Overlap
can be computed using either CPU or GPU kernels
* osu_iallgather
* osu_iallgatherv
* osu_ialltoall
* osu_ialltoallv
* osu_ialltoallw
* osu_ibcast
* osu_igather
* osu_igatherv
* osu_iscatter
* osu_iscatterv
- Allow users the ability to specify zero warmup iterations
* Bug Fixes
- fix openacc pragma
OSU Micro Benchmarks v5.0 (08/17/15)
* New Features & Enhancements
- Support for a set of non-blocking collectives. The benchmarks can display
both the amount of time spent in the collectives and the amount of
overlap achievable
* osu_iallgather
* osu_ialltoall
* osu_ibarrier
* osu_ibcast
* osu_igather
* osu_iscatter
- Add startup benchmarks to facilitate the ability to measure the amount of
time it takes for an MPI library to complete MPI_Init
* osu_init
* osu_hello
- Allocate and align data dynamically
- Thanks to Devendar Bureddy from Mellanox for the suggestion
- Add options for number of warmup iterations [-x] and number of iterations
used per message size [-i] to MPI benchmarks
- Thanks to Devendar Bureddy from Mellanox for the suggestion
* Bug Fixes
- Do not truncate user specified max memory limits
- Thanks to Devendar Bureddy from Mellanox for the report and patch
OSU Micro Benchmarks v4.4.1 (10/30/14)
* Bug Fixes
- adding missing MPI3 guard for WIN_ALLOCATE
- capture getopt return value in an int instead of char
OSU Micro Benchmarks v4.4 (8/23/14)
* New Features & Enhancements
- Support for MPI-3 RMA (one-sided) and atomic operations using GPU buffers
* osu_acc_latency
* osu_cas_latency
* osu_fop_latency
* osu_get_bw
* osu_get_latency
* osu_put_bibw
* osu_put_bw
* osu_put_latency
* Bug Fixes
- remove use of AC_FUNC_MALLOC to avoid undefined rpl_malloc reference
- add missing upc benchmarks for make dist rule
OSU Micro Benchmarks v4.3.1 (6/20/14)
* Bug Fixes
- Fix typo in MPI collective benchmark help message
- Explicitly mention that -m and -M parameters are specified in bytes
OSU Micro Benchmarks v4.3 (3/24/14)
* New Features & Enhancements
- This new suite includes several new (or updated) benchmarks to measure
performance of MPI-3 RMA communication operations with options to select
different window creation (WIN_CREATE, WIN_DYNAMIC, and WIN_ALLOCATE) and
synchronization functions (LOCK, PSCW, FENCE, FLUSH, FLUSH_LOCAL, and
LOCK_ALL) in each benchmark
* osu_acc_latency
* osu_cas_latency
* osu_fop_latency
* osu_get_acc_latency
* osu_get_bw
* osu_get_latency
* osu_put_bibw
* osu_put_bw
* osu_put_latency
- New UPC Collective Benchmarks
* osu_upc_all_barrier
* osu_upc_all_broadcast
* osu_upc_all_exchange
* osu_upc_all_gather
* osu_upc_all_gather_all
* osu_upc_all_reduce
* osu_upc_all_scatter
- Build MPI3 benchmarks when MPI library support is detected
* Bug Fixes
- Add shmem_quiet() in OpenSHMEM Message Rate benchmark to ensure all
previously issued operations are completed
- Allocate pWrk from symmetric heap in OpenSHMEM Reduce benchmark
OSU Micro Benchmarks v4.2 (11/08/13)
* New Features & Enhancements
- New OpenSHMEM benchmarks
* osu_oshm_fcollect
- Enable handling of GPU device buffers in all MPI collective benchmarks
- Add device binding for OpenACC benchmarks
* Bug Fixes
- Add upc_fence after memput in osu_upc_memput benchmark
- Correct CUDA configuration example in README
- Fix several warnings
OSU Micro Benchmarks v4.1 (8/24/13)
* New Features & Enhancements
- New OpenSHMEM benchmarks
* osu_oshm_barrier
* osu_oshm_broadcast
* osu_oshm_collect
* osu_oshm_reduce
- New MPI-3 RMA Atomics benchmarks
* osu_cas_flush
* osu_fop_flush
OSU Micro Benchmarks v4.0.1 (5/06/13)
* Bug Fixes
- Fix several warnings
OSU Micro Benchmarks v4.0 (4/16/13)
* New Features & Enhancements
- Support buffer allocation using OpenACC and CUDA in osu_alltoall,
osu_gather, and osu_scatter benchmarks
- Limit amount of memory allocated by collective benchmarks dynamically
based on number of processes
- Memory limit can also be explicitly set by the user through the -m
option
- Support for 64-bit atomic operations in osu_oshm_atomics
* Bug Fixes
- Fix numerical overflow error with reporting bandwidth in osu_mbw_mr
OSU Micro Benchmarks v3.9 (2/28/13)
* New Features & Enhancements
- Support buffer allocation using OpenACC in GPU benchmarks
- Use average time instead of max time for calculating the bandwidth and
message rate in osu_mbw_mr
- Thanks to Alex Mikheev from Mellanox for the patch
* Bug Fixes
- Properly initialize host buffers for DH and HD transfers in GPU
benchmarks
OSU Micro Benchmarks v3.8 (11/07/12)
* New Features & Enhancements
- New UPC benchmarks
* osu_upc_memput
* osu_upc_memget
OSU Micro Benchmarks v3.7 (9/07/12)
* New Features & Enhancements
- New OpenSHMEM benchmarks
* osu_oshm_get
* osu_oshm_put_mr
* osu_oshm_atomics
* osu_oshm_put
- Organize installation directory according to benchmark type
* Bug Fixes
- Destroy cuda context before exiting
OSU Micro Benchmarks v3.6 (4/30/12)
* New Features & Enhancements
- New collective benchmarks
* osu_allgather
* osu_allgatherv
* osu_allreduce
* osu_alltoall
* osu_alltoallv
* osu_barrier
* osu_bcast
* osu_gather
* osu_gatherv
* osu_reduce
* osu_reduce_scatter
* osu_scatter
* osu_scatterv
* Bug Fixes
- Fix GPU binding issue when running with HH mode
OSU Micro Benchmarks v3.5.2 (3/22/12)
* Bug Fixes
- Fix typo which led to use of incorrect buffers
OSU Micro Benchmarks v3.5.1 (2/02/12)
* New Features & Enhancements
- Provide script to set GPU affinity for MPI processes
* Bug Fixes
- Removed GPU binding after MPI_Init to avoid switching context
OSU Micro Benchmarks v3.5 (11/09/11)
* New Features & Enhancements
- Extension of osu_latency, osu_bw, and osu_bibw benchmarks to evaluate the
performance of MPI_Send/MPI_Recv operation with NVIDIA GPU device and
CUDA support
- This functionality is exposed when configured with --enable-cuda
option
- Flexibility for using buffers in NVIDIA GPU device (D) and host memory (H)
- Flexibility for selecting data movement between D->D, D->H and H->D
OSU Micro Benchmarks v3.4 (09/13/11)
* New Features & Enhancements
- Add passive one-sided communication benchmarks
- Update one-sided communication benchmarks to provide shared memory hint
in MPI_Alloc_mem calls
- Update one-sided communication benchmarks to use MPI_Alloc_mem for buffer
allocation
- Give default values to configure definitions (can now build directly with
mpicc)
- Update latency benchmarks to begin from 0 byte message
* Bug Fixes
- Remove memory leaks in one-sided communication benchmarks
- Update benchmarks to touch buffers before using them for communication
- Fix osu_get_bw test to use different buffers for concurrent communication
operations
- Fix compilation warnings