CHANGELOG.md 13.4 KB
Newer Older
Gilbert Lee's avatar
Gilbert Lee committed
1
2
# Changelog for TransferBench

3
Documentation for TransferBench is available at
Lisa's avatar
Lisa committed
4
[https://rocm.docs.amd.com/projects/TransferBench](https://rocm.docs.amd.com/projects/TransferBench).
Lisa Delaney's avatar
Lisa Delaney committed
5

6
7
8
9
10
11
12
13
14
## v1.41

### Additions
* Adding schmoo preset config benchmarks local/remote reads/writes/copies
  * Usage: ./TransferBench schmoo <numBytes=64M> <localIdx=0> <remoteIdx=1> <maxNumCUs=32>

### Fixes
* Fixing some misreported timings when running with non-fixed number of iterations

15
16
17
18
19
## v1.40

### Fixes
* Fixing XCC defaulting to 0 instead of random for preset configs, ignoring XCC_PREF_TABLE

20
21
22
23
24
25
26
## v1.39

### Additions
* (Experimental) Adding support for Executor sub-index
### Fixes
- Remove deprecated gcnArch code.  ROCm version must include support for hipDeviceMallocUncached

27
28
29
30
31
## v1.38

### Fixes
* Adding missing threadfence which could cause non-fine-grained Transfers to report higher speeds

32
33
34
35
36
37
38
39
## v1.37

### Changes
* USE_SINGLE_STREAM is enabled by default now.  (Disable via USE_SINGLE_STREAM=0)

### Fixes
* Fix unrecognized token error when XCC_PREF_TABLE is unspecified

40
41
42
43
44
45
46
## v1.36

### Additions

* (Experimental) Adding XCC filtering - combined with XCC_PREF_TABLE, this tries to select
  specific XCCs to use for specific (SRC->DST) Transfers

47
48
49
50
51
## v1.35

### Additions

* USE_FINE_GRAIN also applies to a2a preset
52

53
## v1.34
54
55
56
57

### Additions

* Set `GPU_KERNEL=3` as default for gfx942
58

59
## v1.33
Lisa Delaney's avatar
Lisa Delaney committed
60
61
62
63
64

### Additions

* Added the `ALWAYS_VALIDATE` environment variable to allow for validation after every iteration, instead
  of only once at the end of all iterations
65

66
## v1.32
Lisa Delaney's avatar
Lisa Delaney committed
67
68
69
70

### Changes

* Increased the line limit from 2048 to 32768
71

gilbertlee-amd's avatar
gilbertlee-amd committed
72
## v1.31
Lisa Delaney's avatar
Lisa Delaney committed
73
74
75
76
77

### Changes

* `SHOW_ITERATIONS` now shows XCC:CU instead of just CU ID
* `SHOW_ITERATIONS` is printed when `USE_SINGLE_STREAM`=1
gilbertlee-amd's avatar
gilbertlee-amd committed
78

79
## v1.30
Lisa Delaney's avatar
Lisa Delaney committed
80
81
82
83
84
85
86
87
88

### Additions

* `BLOCK_SIZE` has been added to control the threadblock size (must be a multiple of 64, up to 512)
* `BLOCK_ORDER` has been added to control how work is ordered for GFX-executors running
  `USE_SINGLE_STREAM`=1
  * 0 - Threadblocks for transfers are ordered sequentially (default)
  * 1 - Threadblocks for transfers are interleaved
  * 2 - Threadblocks for transfers are ordered randomly
89

90
## v1.29
Lisa Delaney's avatar
Lisa Delaney committed
91
92
93
94
95
96
97
98
99
100
101
102
103
104

### Additions

* A2A preset config now responds to `USE_REMOTE_READ`

### Fixes

* Race-condition during wall-clock initialization caused "inf" during single-stream runs
* CU numbering output after CU masking

### Changes

* The default number of warmups has been reverted to 3
* The default unroll factor for gfx940/941 has been set to 6
105

106
## v1.28
Lisa Delaney's avatar
Lisa Delaney committed
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123

### Additions

* Added `A2A_DIRECT`, which only runs all-to-all on directly connected GPUs (now on by default)
* Added average statistics for P2P and A2A benchmarks
* Added `USE_FINE_GRAIN` for P2P benchmark
  * With older devices, P2P performance with default coarse-grain device memory stops timing as soon
    as a request is sent to data fabric, and not actually when it arrives remotely. This can artificially
    inflate bandwidth numbers, especially when sending small amounts of data.

### Changes

* Modified P2P output to help distinguish between CPU and GPU devices

### Fixes

* Fixed Makefile target to prevent unnecessary re-compilation
124

125
## v1.27
Lisa Delaney's avatar
Lisa Delaney committed
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144

### Additions

* Added cmdline preset to allow specification of  simple tests on command line (e.g.,
  `./TransferBench cmdline 64M "1 4 G0->G0->G1"`)
* Adding the `HIDE_ENV` environment variable, which stops environment variable values from printing
* Adding the `CU_MASK` environment variable, which allows you to select the CUs to run on
* `CU_MASK` is specified in CU indices (0-#CUs-1), where ' - ' can be used to denote ranges of values
  (e.g., `CU_MASK`=3-8,16 requests that transfer be run only on CUs 3,4,5,6,7,8,16)
  * Note that this is somewhat experimental and may not work on all hardware
* `SHOW_ITERATIONS` now shows CU usage for that iteration (experimental)

### Changes

* Added extra comments on commonly missing includes with details on how to install them

### Fixes

* CUDA compilation works again (the `wall_clock64` CUDA alias was not defined)
145

146
147
## v1.26

Lisa Delaney's avatar
Lisa Delaney committed
148
149
150
151
152
153
154
155
156
157
158
159
### Additions

* Setting SHOW_ITERATIONS=1 provides additional information about per-iteration timing for file and
  P2P configs
  * For file configs, iterations are sorted from min to max bandwidth and displayed with standard
    deviation
  * For P2P, min/max/standard deviation is shown for each direction

### Changes

* P2P benchmark formatting now reports bidirectional bandwidth in each direction (as well as sum) for
  clarity
160

161
## v1.25
Lisa Delaney's avatar
Lisa Delaney committed
162
163
164
165
166

### Fixes

* Fixed a bug in the P2P bidirectional benchmark that used the incorrect number of `subExecutors` for
  CPU<->GPU tests
167

gilbertlee-amd's avatar
gilbertlee-amd committed
168
## v1.24
Lisa Delaney's avatar
Lisa Delaney committed
169
170
171
172
173

### Additions

* New All-To-All GPU benchmark accessed by preset "A2A"
* Added gfx941 wall clock frequency
gilbertlee-amd's avatar
gilbertlee-amd committed
174

175
## v1.23
Lisa Delaney's avatar
Lisa Delaney committed
176
177
178
179
180

### Additions

* New GPU subexec scaling benchmark accessed by preset "scaling"
  * Tests GPU-GFX copy performance based on # of CUs used
181

182
## v1.22
Lisa Delaney's avatar
Lisa Delaney committed
183
184
185
186

### Changes

* Switched the kernel timing function to `wall_clock64`
187

gilbertlee-amd's avatar
gilbertlee-amd committed
188
## v1.21
Lisa Delaney's avatar
Lisa Delaney committed
189
190
191
192

### Fixes

* Fixed a bug with `SAMPLING_FACTOR`
gilbertlee-amd's avatar
gilbertlee-amd committed
193

194
## v1.20
Lisa Delaney's avatar
Lisa Delaney committed
195
196
197
198
199

### Fixes

* `VALIDATE_DIRECT` can now be used with `USE_PREP_KERNEL`
* Switched to local GPU for validating GPU memory
200

gilbertlee-amd's avatar
gilbertlee-amd committed
201
## v1.19
Lisa Delaney's avatar
Lisa Delaney committed
202
203
204
205
206

### Additions

* `VALIDATE_DIRECT` now also applies to source memory array checking
* Added null memory pointer check prior to deallocation
gilbertlee-amd's avatar
gilbertlee-amd committed
207

208
## v1.18
Lisa Delaney's avatar
Lisa Delaney committed
209
210
211
212
213
214
215
216
217
218
219
220
221

### Additions

* Adding the ability to validate GPU destination memory directly without going through the CPU
  staging buffer (`VALIDATE_DIRECT`)
  * Note that this only works on AMD devices with large-bar access enabled, and may slow things down
    considerably

### Changes

* Refactored how environment variables are displayed
* Mismatch stops after the first detected error within an array instead of listing all mismatched
  elements
222

223
## v1.17
Lisa Delaney's avatar
Lisa Delaney committed
224
225
226
227
228
229
230
231
232
233
234
235
236
237

### Additions

* Allowed switch to GFX kernel for source array initialization (`USE_PREP_KERNEL`)
  * Note that `USE_PREP_KERNEL` can't be used with `FILL_PATTERN`
* Added the ability to compile with nvcc only (`TransferBenchCuda`)

### Changes

* The default pattern was set to [Element i = ((i * 517) modulo 383 + 31) * (srcBufferIdx + 1)]

### Fixes

* Added the `example.cfg` file
238

239
## v1.16
Lisa Delaney's avatar
Lisa Delaney committed
240
241
242
243
244
245
246

### Additions

* Additional src array validation during preparation
* Added a new environment variable (`CONTINUE_ON_ERROR`) to resume tests after a mis-match
  detection
* Initialized GPU memory to 0 during allocation
247

248
## v1.15
Lisa Delaney's avatar
Lisa Delaney committed
249
250
251
252
253
254
255
256
257

### Fixes

* Fixed a bug that prevented single transfers greater than 8 GB

### Changes

* Removed "check for latest ROCm" warning when allocating too much memory
* Off-source memory value is now printed when a mis-match is detected
258

PedramAlizadeh's avatar
PedramAlizadeh committed
259
## v1.14
Lisa Delaney's avatar
Lisa Delaney committed
260
261
262
263
264
265

### Additions

* Added documentation
* Added pthread linking in src/Makefile and CMakeLists.txt
* Added printing off the hex value of the floats for output and reference
266

PedramAlizadeh's avatar
PedramAlizadeh committed
267
268
## v1.13

Lisa Delaney's avatar
Lisa Delaney committed
269
270
271
272
273
274
275
### Additions

* Added support for cmake

### Changes

* Converted to the Pitchfork layout standard
PedramAlizadeh's avatar
PedramAlizadeh committed
276

277
## v1.12
Lisa Delaney's avatar
Lisa Delaney committed
278
279
280
281
282

### Additions

* Added support for TransferBench on NVIDIA platforms (via `HIP_PLATFORM`=nvidia)
  * Note that CPU executors on NVIDIA platform cannot access GPU memory (no large-bar access)
283

gilbertlee-amd's avatar
gilbertlee-amd committed
284
## v1.11
Lisa Delaney's avatar
Lisa Delaney committed
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330

### Additions

* Added multi-input/multi-output (MIMO) support: transfers now can reduce (element-wise summation)
  multiple input memory arrays and write sums to multiple outputs
* Added GPU-DMA executor 'D', which uses `hipMemcpy` for SDMA copies
  * Previously, this was done using `USE_HIP_CALL`, but now GPU-GFX kernel can run in parallel with
    GPU-DMA, instead of applying to all GPU executors globally
  * GPU-DMA executor can only be used for single-input/single-output transfers
  * GPU-DMA executor can only be associated with one SubExecutor
* Added new "Null" memory type 'N', which represents empty memory. This allows for read-only or
  write-only transfers
* Added new `GPU_KERNEL` environment variable that allows switching between various GPU-GFX
  reduction kernels

### Optimizations

* Improved GPU-GFX kernel performance based on hardware architecture when running with
  fewer CUs

### Changes

* Updated the `example.cfg` file to cover new features
* Updated output to support MIMO
* Changed CU and CPU thread naming to SubExecutors for consistency
* Sweep Preset: default sweep preset executors now includes DMA
* P2P benchmarks:
  * Removed `p2p_rr`, `g2g` and `g2g_rr` (now only works via P2P)
    * Setting `NUM_CPU_DEVICES`=0 can only be used to benchmark GPU devices (like `g2g`)
    * The new `USE_REMOTE_READ` environment variable replaces `_rr` presets
  * New environment variable `USE_GPU_DMA`=1 replaces `USE_HIP_CALL`=1 for benchmarking with
    GPU-DMA Executor
  * Number of GPU SubExecutors for benchmark can be specified via `NUM_GPU_SE`
    * Defaults to all CUs for GPU-GFX, 1 for GPU-DMA
  * Number of CPU SubExecutors for benchmark can be specified via `NUM_CPU_SE`
* Psuedo-random input pattern has been slightly adjusted to have different patterns for each input
  array within same transfer

### Removals

* `USE_HIP_CALL`: use `GPU-DMA` executor 'D' or set `USE_GPU_DMA`=1 for P2P
  benchmark presets
  * Currently, a warning will be issued if `USE_HIP_CALL` is set to 1 and the program will stop
* `NUM_CPU_PER_TRANSFER`: the number of CPU SubExecutors will be whatever is specified for the
  transfer
* `USE_MEMSET`: this function can now be done via a transfer using the null memory type
gilbertlee-amd's avatar
gilbertlee-amd committed
331

332
## v1.10
Lisa Delaney's avatar
Lisa Delaney committed
333
334
335
336

### Fixes

* Fixed incorrect bandwidth calculation when using single stream mode and per-transfer data sizes
337

338
## v1.09
Lisa Delaney's avatar
Lisa Delaney committed
339
340
341
342
343
344
345
346

### Additions

* Printing off src/dst memory addresses during interactive mode

### Changes

* Switching to `numa_set_preferred` instead of `set_mempolicy`
347

348
## v1.08
Lisa Delaney's avatar
Lisa Delaney committed
349
350
351
352
353
354

### Changes

* Fixed handling of non-configured NUMA nodes
* Topology detection now shows actual NUMA node indices
* Fixed 'for' issue with `NUM_GPU_DEVICES`
355

356
## v1.07
Lisa Delaney's avatar
Lisa Delaney committed
357
358
359
360

### Fixes

* Fixed bug with allocations involving non-default CPU memory types
361

gilbertlee-amd's avatar
gilbertlee-amd committed
362
## v1.06
Lisa Delaney's avatar
Lisa Delaney committed
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378

### Additions

* Unpinned CPU memory type ('U'), which may require `HSA_XNACK`=1 in order to access via
  GPU executors
* Added sweep configuration logging to `lastSweep.cfg`
* Ability to specify the number of CUs to use for sweep-based presets

### Changes

* Modified advanced configuration file format to accept bytes-per-transfer

### Fixes

* Fixed random sweep repeatability
* Fixed bug with CPU NUMA node memory allocation
gilbertlee-amd's avatar
gilbertlee-amd committed
379

380
## v1.05
Lisa Delaney's avatar
Lisa Delaney committed
381
382
383
384
385
386
387
388
389

### Additions

* Topology output now includes NUMA node information
* Support for NUMA nodes with no CPU cores (e.g., CXL memory)

### Removals

* The `SWEEP_SRC_IS_EXE` environment variable was removed
390

391
## v1.04
Lisa Delaney's avatar
Lisa Delaney committed
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406

### Additions

* There are new environment variables for sweep based presets:
  * `SWEEP_XGMI_MIN`: The minumum number of XGMI hops for transfers
  * `SWEEP_XGMI_MAX`: The maximum number of XGMI hops for transfers
  * `SWEEP_SEED`: Uses a random seed
  * `SWEEP_RAND_BYTES`: Uses a random amount of bytes (up to pre-specified N) for each transfer

### Changes

* CSV output for sweep now includes an environment variables section followed by output
* CSV output no longer lists environment variable parameters in columns
* We changed the default number of warmup iterations from 3 to 1
* Split CSV output of link type to `ExeToSrcLinkType` and `ExeToDstLinkType`
407

Gilbert Lee's avatar
Gilbert Lee committed
408
## v1.03
Lisa Delaney's avatar
Lisa Delaney committed
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431

### Additions

* There are new preset modes stress-test benchmarks: `sweep` and `randomsweep`
  * `sweep` iterates over all possible sets of transfers to test
  * `randomsweep` iterates over random sets of transfers
  * New sweep-only environment variables can modify `sweep`
    * `SWEEP_SRC`: String containing only "B","C","F", or "G" that defines possible source memory types
    * `SWEEP_EXE`: String containing only "C" or "G" that defines possible executors
    * `SWEEP_DST`: String containing only "B","C","F", or "G" that defines possible destination memory types
    * `SWEEP_SRC_IS_EXE`: Restrict the executor to be the same as the source, if non-zero
    * `SWEEP_MIN`: Minimum number of parallel transfers to test
    * `SWEEP_MAX`: Maximum number of parallel transfers to test
    * `SWEEP_COUNT`: Maximum number of tests to run
    * `SWEEP_TIME_LIMIT`: Maximum number of seconds to run tests
* New environment variables to restrict number of available devices to test on (primarily for sweep
  runs)
  * `NUM_CPU_DEVICES`: Number of CPU devices
  * `NUM_GPU_DEVICES`: Number of GPU devices

### Fixes

* Fixed timing display for CPU executors when using single-stream mode
Gilbert Lee's avatar
Gilbert Lee committed
432

Gilbert Lee's avatar
Gilbert Lee committed
433
## v1.02
Lisa Delaney's avatar
Lisa Delaney committed
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450

### Additions

* Setting `NUM_ITERATIONS` to a negative number indicates a run of -`NUM_ITERATIONS` seconds per
  test

### Changes

* Copies are now referred to as 'transfers' instead of 'links'
* Reordered how environment variables are displayed (alphabetically now)

### Removals

* Combined timing is now always on for kernel-based GPU copies; the `COMBINED_TIMING`
  environment variable has been removed
* Single sync is no longer supported for facility variable iterations; the `USE_SINGLE_SYNC`
  environmental variable has been removed
Gilbert Lee's avatar
Gilbert Lee committed
451

Gilbert Lee's avatar
Gilbert Lee committed
452
## v1.01
Lisa Delaney's avatar
Lisa Delaney committed
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471

### Additions

* Added the `USE_SINGLE_STREAM` feature
  * All Links that run on the same GPU device are run with a single kernel launch on a single stream
  * This doesn't work with `USE_HIP_CALL`, and it forces `USE_SINGLE_SYNC` to collect timings
  * Added the ability to request coherent or fine-grained host memory ('B')

### Changes

* Separated the TransferBench repository from the RCCL repository
* Peer-to-peer benchmark mode now works with `OUTPUT_TO_CSV`
* Toplogy display now works with `OUTPUT_TO_CSV`
* Moved the documentation about the config file into `example.cfg`

### Removals

* Removed config file generation
* Removed the 'show pointer address' (`SHOW_ADDR`) environment variable