-
Notifications
You must be signed in to change notification settings - Fork 10
Hnimrama/ix atom multinode #247
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Open
hnimra-amd
wants to merge
142
commits into
dev/dtni
Choose a base branch
from
hnimrama/ix-atom-multinode
base: dev/dtni
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
Open
Changes from 121 commits
Commits
Show all changes
142 commits
Select commit
Hold shift + click to select a range
02a05ae
docs(dtni): add vllm_single PoC plan and DTNI suite developer guide
atnair-amd 438aafc
docs(dtni): expand dev guide with Background, before/after, config/th…
atnair-amd 9926ab4
docs(dtni): rewrite dev guide with lifecycle flowcharts, test skeleto…
atnair-amd 9466ce7
fix(cli): exclude conftest.py and _-prefixed files from cvs list [AIM…
atnair-amd 5878aa7
vLLM single node refactor (#223)
atnair-amd 8b3c756
feat(dtni): move vLLM bench client to stock vllm bench serve (Spec 0)…
atnair-amd cfd59f9
Merge pull request #225 from ROCm/hnimrama/inferencemax-uplift
hnimra-amd 0207833
Revert "Merge pull request #225 from ROCm/hnimrama/inferencemax-uplif…
hnimra-amd 831b19f
Restructure shared/inference libs + sweep selector + artifact-based v…
atnair-amd 15b1ec3
Updating sglanf for multinode
amd-droy 1f6a0f3
Updating smoke tests-1
amd-droy 89894bd
Updating smoke tests-2
amd-droy de777a7
Updating smoke tests-3
amd-droy 5120a41
Updating smoke tests-4
amd-droy e4c6cd3
Updating smoke tests-5
amd-droy 5e430cd
Updating smoke tests-6
amd-droy e1840dd
Updating smoke tests-7
amd-droy e991b4a
Updating smoke tests-8
amd-droy 2fd8228
Updating smoke tests-9
amd-droy ce7ce0b
Updating smoke tests-10
amd-droy 341c537
Updating smoke tests-11
amd-droy 917333d
Updating smoke tests-12
amd-droy dd0fe25
Updating smoke tests-13
amd-droy f168ae5
Updating smoke tests-14
amd-droy 0b4af44
Updating smoke tests-15
amd-droy a6e7f26
Updating smoke tests-16
amd-droy 0fa1356
Updating hellaswag-2
amd-droy d8c9e18
Updating hellaswag-3
amd-droy ae347f1
Updating hellaswag-4
amd-droy d1e7fed
Updating hellaswag-5
amd-droy 6134bcf
Updating hellaswag-6
amd-droy bcaef40
Updating hellaswag-7
amd-droy f006160
Updating hellaswag-8
amd-droy 97d035a
framework update -1
amd-droy 4d34a69
framework update -2
amd-droy 07a4954
Framework update -3
amd-droy e78cd77
Framework update -5
amd-droy 5d343c1
Framework update -6
amd-droy c0773d1
Framework update -7'
amd-droy c212ae3
Framework update -8
amd-droy 8d25045
Framework update -9
amd-droy 48ffde1
Framework update -10
amd-droy d546aed
Framework update -11
amd-droy 67d86a8
Framework update -13
amd-droy a6fff21
Framework update -14
amd-droy 5ede28b
Framework update -15
amd-droy ff2e9a1
Framework update -17
amd-droy 9453fec
Framework update -18
amd-droy c0cce0a
Framework update -18
amd-droy 11a8be0
Framework update -19
amd-droy e98cf90
Framework update -20
amd-droy f707dde
Framework update -20
amd-droy f48d7d7
Framework update -21
amd-droy 00333ab
Framework update -22
amd-droy 4fbebf3
Framework update -23
amd-droy 7ed71da
Framework update -24
amd-droy e2469a1
Framework update -25
amd-droy be762a9
Framework update -26
amd-droy be86f73
Framework update -27
amd-droy ab98d11
Framework update -27
amd-droy 44396a2
Framework update -28
amd-droy d21ed9f
Framework update -29
amd-droy c65b784
Framework update -30
amd-droy 369f315
Framework update -31
amd-droy 3e1e031
xdit
amd-droy 3d7bc88
Framework update -1
amd-droy 13347a1
Framework update -2
amd-droy 9c68d39
Framework update -3
amd-droy 093502c
Framework update -4
amd-droy 6db0dc9
Framework update -5
amd-droy 474efc8
Framework update -6
amd-droy ff23d34
Framework update -7
amd-droy 932ce55
Framework update -8
amd-droy 7d44703
Framework update -17
amd-droy 5a862ae
Framework update -18
amd-droy dce67d9
Framework update -19
amd-droy 48b36b6
Framework update -20
amd-droy cd4fc82
Framework update -21
amd-droy 1925377
Framework update -22
amd-droy 1c0a1c1
Framework update -22
amd-droy 8790cb2
Framework update -23
amd-droy 78a303b
Framework update -25
amd-droy fce5e24
Framework update -26
amd-droy 17077a1
Framework update -27
amd-droy c9cba51
Hnimrama/ix atom (#238) AIMVT-244/Add ATOM framework and inference …
hnimra-amd ea77382
docs(plans): add inference suite architecture changes presentation doc
atnair-amd 1d3f7ea
Revert "docs(plans): add inference suite architecture changes present…
atnair-amd 43ae95b
Hnimrama/Added ix atom reporting run deck module (#244)
hnimra-amd 4332727
refactor(inference): move shared suite helpers under inference/utils …
hnimra-amd 1fc1adb
[CVS] GPU metrics polling integration for inference validation suites…
atnair-amd 8e3fc46
Integrating nodesmoke tier 1 tests using Primus cli (#250)
urtiwari d50fa0a
feat(vllm): unify vllm_single + vllm_distributed into one topology-pa…
atnair-amd e7bce93
feat(vllm): wire unified vLLM suite into inference report engine (#261)
atnair-amd a03823d
feat(inferencex_atom): add multinode params and scaling efficiency me…
hnimra-amd 658efd3
feat(inferencex_atom): run distributed ATOM serve across cluster ranks
hnimra-amd bd5a225
feat(inferencex_atom): add MI300X 2-node cluster and W1 perf_multi va…
hnimra-amd cb33d56
test(inferencex_atom): cover multinode job routing and config loading
hnimra-amd ac05ac4
docs(inferencex_atom): document multinode cluster and variant params
hnimra-amd b69d698
feat(orchestrator): install openssh-server when image lacks sshd
hnimra-amd 8699fd8
fix(inference): restore test_launch_container in inference_suite_life…
hnimra-amd b6c58ad
fix(inferencex_atom): use ATOM-native multinode serve instead of vLLM…
hnimra-amd 4f55c31
fix(inferencex_atom): isolate ATOM multinode argv from vLLM distribut…
hnimra-amd 1edba1e
config(inferencex_atom): expand multinode sweep to 5 shapes x conc 16…
hnimra-amd 3ea052a
feat(inferencex_atom): enforce multinode scaling thresholds in gates …
hnimra-amd 624d794
fix(inference): repair syntax error in test_launch_container
hnimra-amd 126863a
fix(inference): tolerate CollectReport in html_metric_table_row
hnimra-amd ae2b647
refactor(inferencex_atom): rename suite from inferencex_atom_single
hnimra-amd 81fbe06
docs(inferencex_atom): run make install before activating venv
hnimra-amd 36cea47
style(inferencex_atom): fix ruff format and lint on multinode changes
hnimra-amd 5f68413
feat(inferencex_atom): add MI355X multinode cluster and perf_multi va…
hnimra-amd 979a41e
config(inferencex_atom): add MI300X baseline perf sweep variant
hnimra-amd b44c9b8
config(inferencex_atom): add MI355X baseline perf sweep variant
hnimra-amd 8657ad3
test(inferencex_atom): cover baseline perf sweep config loading
hnimra-amd 3a9f83a
docs(inferencex_atom): document baseline perf sweep runbook
hnimra-amd 5dbc306
config(inferencex_atom): point models_dir at /home/models
hnimra-amd a69ef0b
config(cluster): note /home/models expectation for IX-atom templates
hnimra-amd ac38280
docs(inferencex_atom): document /home/models cache layout
hnimra-amd 164c6ad
config(cluster): consolidate IX-atom cluster templates
hnimra-amd 3b041bd
feat(inferencex_atom): add multinode baseline sweep and drop inferenc…
hnimra-amd 2b37fd0
refactor(inferencex_atom): rename config stems from inferencex-atom-s…
hnimra-amd b71e9d8
chore(inferencex_atom): refresh multinode configs, docs, and loader t…
hnimra-amd d7726bf
fix(inferencex_atom): calibrate perf_multi C=16 thresholds from lab r…
hnimra-amd 0f412be
fix(container): fail fast when openssh-server is missing from image
hnimra-amd 9876b02
feat(inferencex_atom): add vllm_atom and sglang drivers for true PP m…
hnimra-amd 0cb8a31
test(inferencex_atom): cover vllm_atom PP2 and sglang distributed launch
hnimra-amd d660459
config(inferencex_atom): switch multinode variants to vllm_atom PP=2
hnimra-amd 44ba9fa
config(inferencex_atom): add SGLang PP=2 multinode W1 variant
hnimra-amd 0c45f6d
docs(inferencex_atom): document vllm_atom and sglang PP multinode paths
hnimra-amd 3f3dc1b
docs(inferencex_atom): align plan and runbooks with PP=2 driver model
hnimra-amd af127ef
docs(cluster): improve inferencex_atom cluster template for multinode…
hnimra-amd 155c1a0
Rename inferencex_atom configs to flat single/distributed layout.
hnimra-amd 021ade4
Fix multinode vLLM-ATOM readiness and align configs with flat layout.
hnimra-amd 881193b
Fix multinode NCCL/Gloo networking for inferencex_atom.
hnimra-amd e46c76a
fix(container): repair setup_sshd docstring and drop sshd preflight c…
hnimra-amd 200ff0b
Auto-discover socket netdev for multinode inferencex_atom.
hnimra-amd 0e73bd3
Coerce legacy mlx5 ib_netdev configs to auto discovery.
hnimra-amd 0227f00
Resolve multinode fabric lazily when topology test is skipped in smok…
hnimra-amd 6ea2829
Probe multinode fabric on host OS, not inside container.
hnimra-amd d15edfa
Fix host topology probes: valid bash ip syntax and ibv banner parsing.
hnimra-amd 0975c96
Recalibrate MI300X multinode thresholds from 20260723 lab run.
hnimra-amd c0d2cbc
Recalibrate CONC=16 TTFT gates from 20260724 multinode run.
hnimra-amd ac88bed
Raise W1 multinode max_model_length to 8192 for 2k sweep cells.
hnimra-amd File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -17,3 +17,6 @@ docs/sphinx/_toc.yml | |
|
|
||
| # Build distributions | ||
| dist/ | ||
|
|
||
| # Local sample output | ||
| sample_reports/ | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,34 @@ | ||
| { | ||
| "_comment": "InferenceX ATOM cluster template. Edit node_dict to match variant params.nnodes (1 host for single-node sweeps; 2+ for multinode). Set head_node_dict.mgmt_ip to the rank-0 VPC IP (params.master_addr for nnodes>1). Model weights under /home/models on each GPU node. Variant config overrides container image/name.", | ||
| "orchestrator": "container", | ||
| "username": "{user-id}", | ||
| "priv_key_file": "/home/{user-id}/.ssh/id_rsa", | ||
| "head_node_dict": { | ||
| "mgmt_ip": "{head-node-ip}" | ||
| }, | ||
| "env_vars": {}, | ||
| "node_dict": { | ||
| "{head-node-ip}": { | ||
| "bmc_ip": "NA", | ||
| "vpc_ip": "{head-node-ip}" | ||
| }, | ||
| "{worker-node-ip}": { | ||
| "bmc_ip": "NA", | ||
| "vpc_ip": "{worker-node-ip}" | ||
| } | ||
| }, | ||
| "container": { | ||
| "lifetime": "per_run", | ||
| "image": "rocm/atom-dev:latest", | ||
| "name": "inferencex_atom", | ||
| "runtime": { | ||
| "name": "docker", | ||
| "args": { | ||
| "network": "host", | ||
| "ipc": "host", | ||
| "privileged": true, | ||
| "shm_size": "128G" | ||
| } | ||
| } | ||
| } | ||
| } |
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Uh oh!
There was an error while loading. Please reload this page.