feat(inferencex_atom): add M4 vLLM parity suite and W1 variant#249

Draft

hnimra-amd wants to merge 20 commits into

hnimrama/ix-atom-multinodefrom

hnimrama/AIMVT-250-add-vllm-parity

Commits on Jul 13, 2026

feat(inferencex_atom): add multinode params and scaling efficiency metric
hnimra-amd
committed
feat(inferencex_atom): run distributed ATOM serve across cluster ranks
hnimra-amd
committed
feat(inferencex_atom): add MI300X 2-node cluster and W1 perf_multi variant
hnimra-amd
committed
test(inferencex_atom): cover multinode job routing and config loading
hnimra-amd
committed
docs(inferencex_atom): document multinode cluster and variant params
hnimra-amd
committed
feat(orchestrator): install openssh-server when image lacks sshd
hnimra-amd
committed
fix(inference): restore test_launch_container in inference_suite_lifecycle
hnimra-amd
committed
fix(inferencex_atom): use ATOM-native multinode serve instead of vLLM PP flags
hnimra-amd
committed
fix(inferencex_atom): isolate ATOM multinode argv from vLLM distributed flags
hnimra-amd
committed
config(inferencex_atom): expand multinode sweep to 5 shapes x conc 16/64/128
hnimra-amd
committed
feat(inferencex_atom): enforce multinode scaling thresholds in gates and report
hnimra-amd
committed
fix(inference): repair syntax error in test_launch_container
hnimra-amd
committed
fix(inference): tolerate CollectReport in html_metric_table_row
hnimra-amd
committed
refactor(inferencex_atom): rename suite from inferencex_atom_single
hnimra-amd
committed
docs(inferencex_atom): run make install before activating venv
hnimra-amd
committed
style(inferencex_atom): fix ruff format and lint on multinode changes
hnimra-amd
committed
feat(inferencex_atom): add MI355X multinode cluster and perf_multi variant
hnimra-amd
committed
feat(inferencex_atom): add M4 vLLM parity suite and W1 variant
hnimra-amd
committed
refactor(inferencex_atom): rename vLLM parity suite to inferencex_atom_vllm
hnimra-amd
committed
feat(inferencex_atom): add vLLM parity multinode W1 variants
hnimra-amd
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

feat(inferencex_atom): add M4 vLLM parity suite and W1 variant#249

feat(inferencex_atom): add M4 vLLM parity suite and W1 variant#249
hnimra-amd wants to merge 20 commits into
hnimrama/ix-atom-multinodefrom
hnimrama/AIMVT-250-add-vllm-parity

Commits on Jul 13, 2026

feat(inferencex_atom): add multinode params and scaling efficiency metric

feat(inferencex_atom): run distributed ATOM serve across cluster ranks

feat(inferencex_atom): add MI300X 2-node cluster and W1 perf_multi variant

test(inferencex_atom): cover multinode job routing and config loading

docs(inferencex_atom): document multinode cluster and variant params

feat(orchestrator): install openssh-server when image lacks sshd

fix(inference): restore test_launch_container in inference_suite_lifecycle

fix(inferencex_atom): use ATOM-native multinode serve instead of vLLM PP flags

fix(inferencex_atom): isolate ATOM multinode argv from vLLM distributed flags

config(inferencex_atom): expand multinode sweep to 5 shapes x conc 16/64/128

feat(inferencex_atom): enforce multinode scaling thresholds in gates and report

fix(inference): repair syntax error in test_launch_container

fix(inference): tolerate CollectReport in html_metric_table_row

refactor(inferencex_atom): rename suite from inferencex_atom_single

docs(inferencex_atom): run make install before activating venv

style(inferencex_atom): fix ruff format and lint on multinode changes

feat(inferencex_atom): add MI355X multinode cluster and perf_multi variant

feat(inferencex_atom): add M4 vLLM parity suite and W1 variant

refactor(inferencex_atom): rename vLLM parity suite to inferencex_atom_vllm

feat(inferencex_atom): add vLLM parity multinode W1 variants