[AMD] CI - Add MI35x nightly/PR tests for kv-cache-fp8 and allreduce-fusion (DeepSeek) (#19834)

Co-authored-by: bingxche <Bingxu.Chen@amd.com>
This commit is contained in:
YC Tseng
2026-03-05 07:09:57 -08:00
committed by GitHub
co-authored by bingxche
parent 0de0d74195
commit b5edab57f2
13 changed files with 1614 additions and 177 deletions
+31 -15
View File
@@ -25,7 +25,7 @@ on:
workflow_dispatch:
inputs:
target_stage:
description: "Specific stage to run (optional, for quick testing)"
description: "Specific stage(s) to run, comma-separated (e.g. 'stage-a-test-1-amd,stage-b-test-small-1-gpu-amd')"
required: false
type: string
default: ""
@@ -141,7 +141,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'sgl-kernel-unit-test-amd') ||
(contains(format(',{0},', inputs.target_stage), ',sgl-kernel-unit-test-amd,')) ||
(
!inputs.target_stage &&
needs.check-changes.outputs.sgl_kernel == 'true'
@@ -188,7 +188,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'sgl-kernel-unit-test-2-gpu-amd') ||
(contains(format(',{0},', inputs.target_stage), ',sgl-kernel-unit-test-2-gpu-amd,')) ||
(
!inputs.target_stage &&
needs.check-changes.outputs.sgl_kernel == 'true'
@@ -230,7 +230,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'stage-a-test-1-amd') ||
(contains(format(',{0},', inputs.target_stage), ',stage-a-test-1-amd,')) ||
(
!inputs.target_stage &&
(!failure() && !cancelled()) &&
@@ -270,7 +270,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'jit-kernel-unit-test-amd') ||
(contains(format(',{0},', inputs.target_stage), ',jit-kernel-unit-test-amd,')) ||
(
!inputs.target_stage &&
needs.check-changes.outputs.jit_kernel == 'true'
@@ -309,7 +309,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'stage-b-test-small-1-gpu-amd') ||
(contains(format(',{0},', inputs.target_stage), ',stage-b-test-small-1-gpu-amd,')) ||
(
!inputs.target_stage &&
(!failure() && !cancelled()) &&
@@ -349,7 +349,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'stage-b-test-small-1-gpu-amd-nondeterministic') ||
(contains(format(',{0},', inputs.target_stage), ',stage-b-test-small-1-gpu-amd-nondeterministic,')) ||
(
!inputs.target_stage &&
(!failure() && !cancelled()) &&
@@ -388,7 +388,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'stage-b-test-small-1-gpu-amd-mi35x') ||
(contains(format(',{0},', inputs.target_stage), ',stage-b-test-small-1-gpu-amd-mi35x,')) ||
(
!inputs.target_stage &&
(!failure() && !cancelled()) &&
@@ -427,7 +427,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'stage-b-test-large-1-gpu-amd') ||
(contains(format(',{0},', inputs.target_stage), ',stage-b-test-large-1-gpu-amd,')) ||
(
!inputs.target_stage &&
(!failure() && !cancelled()) &&
@@ -467,7 +467,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'stage-b-test-large-2-gpu-amd') ||
(contains(format(',{0},', inputs.target_stage), ',stage-b-test-large-2-gpu-amd,')) ||
(
!inputs.target_stage &&
(!failure() && !cancelled()) &&
@@ -504,7 +504,15 @@ jobs:
multimodal-gen-test-1-gpu-amd:
needs: [check-changes]
if: needs.check-changes.outputs.multimodal_gen == 'true'
if: |
always() &&
(
(contains(format(',{0},', inputs.target_stage), ',multimodal-gen-test-1-gpu-amd,')) ||
(
!inputs.target_stage &&
needs.check-changes.outputs.multimodal_gen == 'true'
)
)
strategy:
fail-fast: false
max-parallel: 1 # Run one at a time to avoid eviction from resource exhaustion during AITER kernel JIT
@@ -624,7 +632,15 @@ jobs:
multimodal-gen-test-2-gpu-amd:
needs: [check-changes]
if: needs.check-changes.outputs.multimodal_gen == 'true'
if: |
always() &&
(
(contains(format(',{0},', inputs.target_stage), ',multimodal-gen-test-2-gpu-amd,')) ||
(
!inputs.target_stage &&
needs.check-changes.outputs.multimodal_gen == 'true'
)
)
strategy:
fail-fast: false
max-parallel: 1 # Run one at a time to avoid eviction from resource exhaustion during AITER kernel JIT
@@ -746,7 +762,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'stage-c-test-large-8-gpu-amd') ||
(contains(format(',{0},', inputs.target_stage), ',stage-c-test-large-8-gpu-amd,')) ||
(
!inputs.target_stage &&
(!failure() && !cancelled()) &&
@@ -794,7 +810,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'stage-c-test-large-8-gpu-amd-mi35x') ||
(contains(format(',{0},', inputs.target_stage), ',stage-c-test-large-8-gpu-amd-mi35x,')) ||
(
!inputs.target_stage &&
(!failure() && !cancelled()) &&
@@ -835,7 +851,7 @@ jobs:
if: |
always() &&
(
(inputs.target_stage == 'stage-b-test-large-8-gpu-disaggregation-amd') ||
(contains(format(',{0},', inputs.target_stage), ',stage-b-test-large-8-gpu-disaggregation-amd,')) ||
(
!inputs.target_stage &&
(!failure() && !cancelled()) &&