Merge branch 'main' into feat/shard_id_with_node

Author: git-flyer

.github/new-prs-labeler.yml

@@ -30,7 +30,6 @@ CORE:
           - 'cmake/**'
           - 'include/**'
           - 'lib/**'
-          - 'python/**'
           - 'scripts/**'
           - 'test/**'
           - 'unittest/**'
@@ -45,6 +44,19 @@ CORE:
           - 'setup.py'
           - '.github/actions/**'
           - '.github/workflows/code-format-check.yml'
+      - all-globs-to-any-file:
+          - 'python/**'
+          - '!python/setup_tools/utils/cambricon.py'
+          - '!python/setup_tools/utils/enflame.py'
+          - '!python/setup_tools/utils/hcu.py'
+          - '!python/setup_tools/utils/iluvatar.py'
+          - '!python/setup_tools/utils/metax.py'
+          - '!python/setup_tools/utils/mthreads.py'
+          - '!python/setup_tools/utils/ppu.py'
+          - '!python/setup_tools/utils/rpu.py'
+          - '!python/setup_tools/utils/sunrise.py'
+          - '!python/setup_tools/utils/thrive.py'
+          - '!python/setup_tools/utils/xpu.py'
 
 # ==========================================
 # 2. CI/CD Labels
@@ -128,6 +140,7 @@ cambricon:
           - 'third_party/cambricon/**'
           - '.github/workflows/cambricon*'
           - '.github/workflows/benchmark/cambricon/**'
+          - 'python/setup_tools/utils/cambricon.py'
 
 cpu:
   - changed-files:
@@ -142,40 +155,45 @@ enflame:
           - 'third_party/enflame/**'
           - '.github/workflows/enflame*'
           - '.github/workflows/benchmark/enflame/**'
+          - 'python/setup_tools/utils/enflame.py'
 
 hcu:
   - changed-files:
       - any-glob-to-any-file:
           - 'third_party/hcu/**'
           - '.github/workflows/hcu*'
           - '.github/workflows/benchmark/hcu/**'
+          - 'python/setup_tools/utils/hcu.py'
 
 iluvatar:
   - changed-files:
       - any-glob-to-any-file:
           - 'third_party/iluvatar/**'
           - '.github/workflows/iluvatar*'
           - '.github/workflows/benchmark/iluvatar/**'
+          - 'python/setup_tools/utils/iluvatar.py'
 
 metax:
   - changed-files:
       - any-glob-to-any-file:
           - 'third_party/metax/**'
           - '.github/workflows/metax*'
           - '.github/workflows/benchmark/metax/**'
+          - 'python/setup_tools/utils/metax.py'
 
 mthreads:
   - changed-files:
       - any-glob-to-any-file:
           - 'third_party/mthreads/**'
           - '.github/workflows/mthreads*'
           - '.github/workflows/benchmark/mthreads/**'
+          - 'python/setup_tools/utils/mthreads.py'
 
 nvidia:
   - changed-files:
       - any-glob-to-any-file:
           - 'third_party/nvidia/**'
-          - '.github/workflows/nv*'
+          - '.github/workflows/nvidia*'
           - '.github/workflows/benchmark/nvidia/**'
 
 ppu:
@@ -184,27 +202,31 @@ ppu:
           - 'third_party/ppu/**'
           - '.github/workflows/ppu*'
           - '.github/workflows/benchmark/ppu/**'
+          - 'python/setup_tools/utils/ppu.py'
 
 rpu:
   - changed-files:
       - any-glob-to-any-file:
           - 'third_party/rpu/**'
           - '.github/workflows/rpu*'
           - '.github/workflows/benchmark/rpu/**'
+          - 'python/setup_tools/utils/rpu.py'
 
 sunrise:
   - changed-files:
       - any-glob-to-any-file:
           - 'third_party/sunrise/**'
           - '.github/workflows/sunrise*'
           - '.github/workflows/benchmark/sunrise/**'
+          - 'python/setup_tools/utils/sunrise.py'
 
 thrive:
   - changed-files:
       - any-glob-to-any-file:
           - 'third_party/thrive/**'
           - '.github/workflows/thrive*'
           - '.github/workflows/benchmark/thrive/**'
+          - 'python/setup_tools/utils/thrive.py'
 
 tileir:
   - changed-files:
@@ -233,6 +255,7 @@ xpu:
           - 'third_party/xpu/**'
           - '.github/workflows/xpu*'
           - '.github/workflows/benchmark/xpu/**'
+          - 'python/setup_tools/utils/xpu.py'
 
 # ====================================================================================
 # 6. TLE Labels

.github/workflows/nvidia3.6-build-and-test.yml

@@ -81,8 +81,6 @@ jobs:
         run: |
           set -x
           source ~/env-3.6.sh
-          export FLAGCX_RECOMPILE=ON
-          export USE_FLAGCX=ON
           MAX_JOBS=32 python3 -m pip install . --no-build-isolation -v
 
       - name: Clear cache if ClearCache label present

.github/workflows/rpu3.6-build-and-test.yml

@@ -39,7 +39,7 @@ concurrency:
   # (cancel-in-progress: false) so an in-flight board test is never interrupted
   # mid-run (a cancelled smoke can orphan a launch_kernel_runner that wedges
   # /dev/rpu for the next job).
-  group: rpu3.6-board
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
   cancel-in-progress: false
 
 # The RPU CI runner is a small x86 host that orchestrates the build/test on a