RELEASE_NOTES 7.14 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14
0.5.1 (June 09, 2016)

network:
==========

corrected link latency calculation in dragonfly model

printf argument mismatch in dragonfly model

refactors to the torus bandwidth calculation to mirror that of the other
  networks (no functional change)

more robust type conversions in dragonfly (int sizes -> uint64_t)

Jonathan Jenkins's avatar
Jonathan Jenkins committed
15 16 17
removed the redundant and obsolete MPI replay simulator
  (modelnet-mpi-wrklds.c). The proper version to use is modelnet-mpi-replay.c

18 19 20
----------

0.5.0 (May 24, 2016)
21 22 23 24 25 26 27

general:
==========

codes-base and codes-net have been combined into a single project (now at
https://xgitlab.cels.anl.gov/codes/codes).

Jonathan Jenkins's avatar
Jonathan Jenkins committed
28
updated to ROSS revision d9cef53.
29 30 31 32 33 34 35 36

fixed a large number of warnings across the codebase.

networks:
==========

addition of the SlimFly network topology, corresponding to the Wolfe et al.
  paper "Modeling a Million-node Slim Fly Network using Parallel Discrete-event
37
  Simulation", at SIGSIM-PADS'16. See README.slimfly.txt 
38
  (src/networks/model-net/doc).
39 40 41

modelnet now supports sampling at regular intervals. Dragonfly LPs can
  currently make use of this - others can be added based on demand. See
42
  (src/network-workloads/README_synthetic.txt).
43

44 45 46
dragonfly and torus network models credit-based flow control has been updated.
    Dragonfly model's adaptive routing algorithms have been updated. For details,
    see paper "Enabling Parallel Simulation of Large-Scale HPC Network Systems",
47
    M.Mubarak et al., at IEEE Trans. on Parallel and Distributed Systems (TPDS).
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
allow 0-byte messages in model-net.

enable "local" model-net messages (for LPs sharing the same model-net endpoint -
  approximates a zero-copy ownership pass of the payload)

the model_net_event family of functions now return a token value which must
  be passed to model_net_event_rc2.

workloads:
==========

concurrent workload support added to workload generators and the MPI
  simulation layer. See codes-jobmap.h and the modified codes-workload.h.
  Thanks to Xu Yang for the partial contributions. Not all workload generators
  support concurrent workloads at this time.

scripts for generating job allocations specific to the torus and dragonfly
  topologies. See scripts/allocation_gen. Thanks to Xu Yang for the
  contributions.

multiple fixes to the MPI simulation layer and DUMPI workload generator.

concurrent workload support and more flexible rank mappings in the MPI
  simulation layer. Thanks to Xu Yang for the initial code.

removed scalatrace workload generator, which never made it to a usable state.

a new checkpoint IO workload generator has been added, based on the Daly paper
  "A higher order estimate of the optimum checkpoint interval for restart
78 79
  dumps" at Future Generation Computing Systems 2004. See README.codes-workload
  (src/workload).
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

utilities:
==========

fixes to rc-stack - memory no longer leaks in sequential mode, and optimistic
  debug mode is now supported.

added a more performance-sensitive function, codes_mapping_get_lp_info2, to
  codes-mapping (passes const pointers around instead of copying strings)

added a "mapping context" API for better controlling implicit LP-LP mappings,
  including modelnet, local storage model, and resource. See
  codes/codes-mapping-context.h and added functions in the mentioned LPs for
  details.

formalized a callback mechanism for CODES, replacing previous ad-hoc
  methods of passing control from LPs to arbitrary other LPs. The request and
  local storage model LP APIs have been changed to use this mechanism, and
  model-net has additional APIs to use user-provided mapping contexts. See
  codes/codes-callback.h for the API and tests/resource-test.c for advanced
  usage.

deprecations:
==========

model_net_event_rc (use model_net_event_rc2, which will eventually be renamed
  to model_net_event_rc)

codes_event_new (define your own bounds-checking macro if need be)

110
----------
111

112 113 114 115 116 117 118 119 120 121
0.4.1 (September 30, 2015)

general:
==========

fix compatibility with recent ROSS releases

----------

0.4.0 (May 6, 2015)
122

123 124
codes-base

125 126
general:
==========
127

128
significant source reorganization / refactoring
129

130
refactor some private headers out of the public eye
131

132 133 134 135
dead code removal

documentation:
==========
136

137
improved example_heterogeneous example program
138

Jonathan Jenkins's avatar
Jonathan Jenkins committed
139
added configuration to example_heterogeneous showing two torus networks
140

141 142
reorganized files to prevent name collisions on OSX. Top-level docs other than
  copyright now in doc directory
143

144 145 146 147
additions to best practice document

configurator:
==========
148

149
more stable file format for configurator output
150

151
ignore unrelated parameters passed into filter_configs
152

153 154 155 156
handle empty cfields in configurator

workloads:
==========
157

158
combined network and IO workload APIs into a single one
159

160
adding dumpi workload support in codes-workload-dump utility
161

162
workload dump utility option cleanup
163

164
renamed "bgp" workload generator to "iolang", significant cleanups
165

166
put network workload ops in workload dump util
167

168
removing one of the dumpi libraries from the build. It was generating some unwanted dumpi files.
169

170 171 172 173
network workload API more fleshed out

utilities:
==========
174

Jonathan Jenkins's avatar
Jonathan Jenkins committed
175
configuration bug fixes for larger LP type counts
176

177
resource LP annotation mapping hooks
178

179
local storage model API switch to use annotations
180

181
better configuration error handling
182

183
hedge against precision loss in codes_local_latency (see codes.h)
184

185 186 187
use a different RNG than default for codes_local_latency
- prevents addition/removal of codes_local_latency calls from poisoning RNG
  stream of calling model
188

189 190
added simple GVT-aware stack with garbage collection (see rc-stack.h)

191 192
codes-net

Jonathan Jenkins's avatar
Jonathan Jenkins committed
193 194
general:
==========
195

Jonathan Jenkins's avatar
Jonathan Jenkins committed
196
cleanup of much of the code base
197

Jonathan Jenkins's avatar
Jonathan Jenkins committed
198
more informative error for failure to find modelnet lps
199

Jonathan Jenkins's avatar
Jonathan Jenkins committed
200
removed redundant include directory on install (was 'install/codes/codes/*.h')
Jonathan Jenkins's avatar
Jonathan Jenkins committed
201 202 203

documentation:
==========
204

Jonathan Jenkins's avatar
Jonathan Jenkins committed
205 206
reorganized files to prevent name collisions on OSX. Top-level docs other than
  copyright now in doc directory
207

Jonathan Jenkins's avatar
Jonathan Jenkins committed
208
updated code documentation
209

Jonathan Jenkins's avatar
Jonathan Jenkins committed
210
fix linker error in certain cases with codes-base
211

Jonathan Jenkins's avatar
Jonathan Jenkins committed
212 213 214 215 216
tweaked config error handling


networks:
==========
217
fix to loggp latency calculation when using "receive queue"
218

Jonathan Jenkins's avatar
Jonathan Jenkins committed
219
made torus lps agnostic to groups and aware of annotations
220

Jonathan Jenkins's avatar
Jonathan Jenkins committed
221
miscellaneous fixes to dragonfly model
222

Jonathan Jenkins's avatar
Jonathan Jenkins committed
223 224
updates to simplep2p: support for having different latency/bw at sender &
  receiver end. See src/models/networks/model-net/doc/README.simplep2p.txt
225

Jonathan Jenkins's avatar
Jonathan Jenkins committed
226
minor fixes to usage of quickhash in replay tool
227

Jonathan Jenkins's avatar
Jonathan Jenkins committed
228
fixed RNG reverse computation bug in loggp
229

Jonathan Jenkins's avatar
Jonathan Jenkins committed
230 231 232 233
fixed swapped arguments in round-robin scheduler causing short circuit

workloads:
==========
234

Jonathan Jenkins's avatar
Jonathan Jenkins committed
235
minor changes to dumpi trace config files
236

Jonathan Jenkins's avatar
Jonathan Jenkins committed
237
resolving minor bug with reverse computation in dumpi traces
238

Jonathan Jenkins's avatar
Jonathan Jenkins committed
239
Updating network trace code to use the combined workload API
240

Jonathan Jenkins's avatar
Jonathan Jenkins committed
241
Adding synthetic traffic patterns (currently with dragonfly model)
242

Jonathan Jenkins's avatar
Jonathan Jenkins committed
243
Adding network workload test program for debugging
244

Jonathan Jenkins's avatar
Jonathan Jenkins committed
245
Updating MPI wait/wait_all code in replay tool
246

247 248 249
----------

0.3.0 (November 7, 2014)
250

251 252
codes-base

253 254
Initial "official" release. Against previous repository revisions, this release
includes more complete documentation.
255 256 257

codes-net

Jonathan Jenkins's avatar
Jonathan Jenkins committed
258 259 260 261
Initial "official" release. Against previous repository revisions, this release
includes more complete documentation and a rename of the "simplewan" model to
the "simplep2p" (simple point-to-point) model to more accurately represent
what it's modeling.