RELEASE_NOTES 6.51 KB
Newer Older
1 2 3 4 5 6 7 8
0.5.0

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
9
updated to ROSS revision d9cef53.
10 11 12 13 14 15 16 17

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
18
  Simulation", at SIGSIM-PADS'16. See README.slimfly.txt 
19
  (src/networks/model-net/doc).
20 21 22

modelnet now supports sampling at regular intervals. Dragonfly LPs can
  currently make use of this - others can be added based on demand. See
23
  (src/network-workloads/README_synthetic.txt).
24

25 26 27 28 29
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",
    M.Mubarak et al., at IEEE Trans. on Parallel and Distributed Systems (TPDS). 

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
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
59 60
  dumps" at Future Generation Computing Systems 2004. See README.codes-workload
  (src/workload).
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

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)


92 93
0.4.0

94 95
codes-base

96 97
general:
==========
98

99
significant source reorganization / refactoring
100

101
refactor some private headers out of the public eye
102

103 104 105 106
dead code removal

documentation:
==========
107

108
improved example_heterogeneous example program
109

Jonathan Jenkins's avatar
Jonathan Jenkins committed
110
added configuration to example_heterogeneous showing two torus networks
111

112 113
reorganized files to prevent name collisions on OSX. Top-level docs other than
  copyright now in doc directory
114

115 116 117 118
additions to best practice document

configurator:
==========
119

120
more stable file format for configurator output
121

122
ignore unrelated parameters passed into filter_configs
123

124 125 126 127
handle empty cfields in configurator

workloads:
==========
128

129
combined network and IO workload APIs into a single one
130

131
adding dumpi workload support in codes-workload-dump utility
132

133
workload dump utility option cleanup
134

135
renamed "bgp" workload generator to "iolang", significant cleanups
136

137
put network workload ops in workload dump util
138

139
removing one of the dumpi libraries from the build. It was generating some unwanted dumpi files.
140

141 142 143 144
network workload API more fleshed out

utilities:
==========
145

Jonathan Jenkins's avatar
Jonathan Jenkins committed
146
configuration bug fixes for larger LP type counts
147

148
resource LP annotation mapping hooks
149

150
local storage model API switch to use annotations
151

152
better configuration error handling
153

154
hedge against precision loss in codes_local_latency (see codes.h)
155

156 157 158
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
159

160 161
added simple GVT-aware stack with garbage collection (see rc-stack.h)

162 163
codes-net

Jonathan Jenkins's avatar
Jonathan Jenkins committed
164 165
general:
==========
166

Jonathan Jenkins's avatar
Jonathan Jenkins committed
167
cleanup of much of the code base
168

Jonathan Jenkins's avatar
Jonathan Jenkins committed
169
more informative error for failure to find modelnet lps
170

Jonathan Jenkins's avatar
Jonathan Jenkins committed
171
removed redundant include directory on install (was 'install/codes/codes/*.h')
Jonathan Jenkins's avatar
Jonathan Jenkins committed
172 173 174

documentation:
==========
175

Jonathan Jenkins's avatar
Jonathan Jenkins committed
176 177
reorganized files to prevent name collisions on OSX. Top-level docs other than
  copyright now in doc directory
178

Jonathan Jenkins's avatar
Jonathan Jenkins committed
179
updated code documentation
180

Jonathan Jenkins's avatar
Jonathan Jenkins committed
181
fix linker error in certain cases with codes-base
182

Jonathan Jenkins's avatar
Jonathan Jenkins committed
183 184 185 186 187
tweaked config error handling


networks:
==========
188
fix to loggp latency calculation when using "receive queue"
189

Jonathan Jenkins's avatar
Jonathan Jenkins committed
190
made torus lps agnostic to groups and aware of annotations
191

Jonathan Jenkins's avatar
Jonathan Jenkins committed
192
miscellaneous fixes to dragonfly model
193

Jonathan Jenkins's avatar
Jonathan Jenkins committed
194 195
updates to simplep2p: support for having different latency/bw at sender &
  receiver end. See src/models/networks/model-net/doc/README.simplep2p.txt
196

Jonathan Jenkins's avatar
Jonathan Jenkins committed
197
minor fixes to usage of quickhash in replay tool
198

Jonathan Jenkins's avatar
Jonathan Jenkins committed
199
fixed RNG reverse computation bug in loggp
200

Jonathan Jenkins's avatar
Jonathan Jenkins committed
201 202 203 204
fixed swapped arguments in round-robin scheduler causing short circuit

workloads:
==========
205

Jonathan Jenkins's avatar
Jonathan Jenkins committed
206
minor changes to dumpi trace config files
207

Jonathan Jenkins's avatar
Jonathan Jenkins committed
208
resolving minor bug with reverse computation in dumpi traces
209

Jonathan Jenkins's avatar
Jonathan Jenkins committed
210
Updating network trace code to use the combined workload API
211

Jonathan Jenkins's avatar
Jonathan Jenkins committed
212
Adding synthetic traffic patterns (currently with dragonfly model)
213

Jonathan Jenkins's avatar
Jonathan Jenkins committed
214
Adding network workload test program for debugging
215

Jonathan Jenkins's avatar
Jonathan Jenkins committed
216
Updating MPI wait/wait_all code in replay tool
217

218 219
0.3.0

220 221
codes-base

222 223
Initial "official" release. Against previous repository revisions, this release
includes more complete documentation.
224 225 226

codes-net

Jonathan Jenkins's avatar
Jonathan Jenkins committed
227 228 229 230
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.