kernbench2/docs/diagrams/allreduce_latency_plots/summary.csv at master

Files

T

mukesh b610cb0d9a sccl: drive allreduce tests via torch.distributed; reorganize into tests/sccl/

Convert the multidevice allreduce correctness + latency/buffer-kind sweeps
to run through the real PyTorch-distributed path
(init_process_group(backend="ahbm") -> mp.spawn -> dist.all_reduce) instead
of direct ctx.launch, and reorganize the CCL/allreduce tests into a
tests/sccl/ package split one test per file.

Production change (required for the distributed path on non-square SIP grids):
- AhbmCCLBackend now reads explicit system.sips.w/h from the spec, with a
  square-only sqrt fallback that raises on ambiguity, instead of silently
  guessing round(sqrt(count)). This fixes the 2x3 / 3x2 torus + mesh cases,
  which previously resolved to a wrong 2x2 grid. Mirrors the test helper's
  _sip_topo_dims precedence (explicit w/h > square fallback > raise).

Test reorganization (tests/sccl/):
- _allreduce_helpers.py: shared plumbing (distributed driver, config writers,
  direct-launch run_allreduce parity reference, sweep/buffer-kind constants,
  plot aggregators, topology-diagram + FSIM-comparison emitters).
- test_allreduce_ring_torus_mesh.py: correctness across ring/torus/mesh.
- test_distributed_default_topology.py: full distributed path on topology.yaml.
- test_plot_latency_sweep.py / test_plot_buffer_kind_sweep.py: sweep rows.
- test_plot_topology_diagram.py / test_plot_comparison_fsim.py: plot emitters.
- test_intercube_root_center.py: moved in (ADR-0032 center-root latency guard).

Also:
- Move the FSIM comparison plot generator out of scripts/ into the sccl suite.
- Delete superseded test files (test_allreduce_multidevice,
  test_distributed_lrab_hierarchical_allreduce, test_allreduce_buffer_kind_sweep)
  and repoint conftest aggregators + the ipcq buffer-kind importers.
- Regenerate the allreduce_latency_plots derived artifacts from the full sweep.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-20 22:24:43 -07:00

2.7 KiB

Raw Permalink Blame History

1	algorithm	sip_topology	n_sips	n_elem	bytes_per_pe	bytes_per_sip	latency_ns
2	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	8	16	256	2666.552500000015
3	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	32	64	1024	2747.7400000000152
4	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	64	128	2048	2855.990000000018
5	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	128	256	4096	3072.490000000019
6	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	512	1024	16384	3337.1133333333582
7	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	1024	2048	32768	3708.0333333333692
8	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	2048	4096	65536	4449.873333333393
9	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	4096	8192	131072	5933.020000000124
10	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	8192	16384	262144	8900.379999999863
11	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	16384	32768	524288	14835.099999999224
12	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	32768	65536	1048576	26704.540000000765
13	lrab_hierarchical_allreduce	mesh_2d_no_wrap	6	49152	98304	1572864	38573.97999999701
14	lrab_hierarchical_allreduce	ring_1d	6	8	16	256	2365.255833333347
15	lrab_hierarchical_allreduce	ring_1d	6	32	64	1024	2436.9433333333473
16	lrab_hierarchical_allreduce	ring_1d	6	64	128	2048	2532.526666666683
17	lrab_hierarchical_allreduce	ring_1d	6	128	256	4096	2723.693333333349
18	lrab_hierarchical_allreduce	ring_1d	6	512	1024	16384	3048.635000000021
19	lrab_hierarchical_allreduce	ring_1d	6	1024	2048	32768	3393.4016666666957
20	lrab_hierarchical_allreduce	ring_1d	6	2048	4096	65536	4082.401666666714
21	lrab_hierarchical_allreduce	ring_1d	6	4096	8192	131072	5458.80166666677
22	lrab_hierarchical_allreduce	ring_1d	6	8192	16384	262144	8216.934999999943
23	lrab_hierarchical_allreduce	ring_1d	6	16384	32768	524288	13733.201666665835
24	lrab_hierarchical_allreduce	ring_1d	6	32768	65536	1048576	24765.73500000064
25	lrab_hierarchical_allreduce	ring_1d	6	49152	98304	1572864	35798.268333331536
26	lrab_hierarchical_allreduce	torus_2d	6	8	16	256	1700.6025000000095
27	lrab_hierarchical_allreduce	torus_2d	6	32	64	1024	1753.2900000000102
28	lrab_hierarchical_allreduce	torus_2d	6	64	128	2048	1823.540000000012
29	lrab_hierarchical_allreduce	torus_2d	6	128	256	4096	1964.040000000012
30	lrab_hierarchical_allreduce	torus_2d	6	512	1024	16384	2196.8183333333463
31	lrab_hierarchical_allreduce	torus_2d	6	1024	2048	32768	2477.2783333333473
32	lrab_hierarchical_allreduce	torus_2d	6	2048	4096	65536	3038.1983333333583
33	lrab_hierarchical_allreduce	torus_2d	6	4096	8192	131072	4159.5050000000665
34	lrab_hierarchical_allreduce	torus_2d	6	8192	16384	262144	6403.185000000109
35	lrab_hierarchical_allreduce	torus_2d	6	16384	32768	524288	10890.5449999995
36	lrab_hierarchical_allreduce	torus_2d	6	32768	65536	1048576	19865.265000000378
37	lrab_hierarchical_allreduce	torus_2d	6	49152	98304	1572864	28839.98500000059

2.7 KiB Raw Permalink Blame History

2.7 KiB

Raw Permalink Blame History