feat: CI Stream for NVIDIA platform

[zhangyue207]

No description provided.

[zhangyue207]

function display

(python3.10) zhangyue@server:~/InfiniOps$ python .ci/agent.py run --branch feat/dev-infra
==> dispatching nvidia gpu job to http://localhost:8080
    job_id: f088f369
==> dispatching iluvatar gpu job to http://172.22.162.13:8080
    job_id: 9f280eae
==> dispatching metax gpu job to http://172.22.162.49:8080
    job_id: e7a94c55
==> dispatching moore gpu job to http://172.22.162.95:8080
    job_id: a8de4bc5
==> dispatching cambricon gpu job to http://172.22.162.16:8080
error: failed to dispatch to http://172.22.162.16:8080: HTTP Error 502: Bad Gateway
    failed to dispatch cambricon_gpu
<== FAIL  moore_gpu     (24s)
--- error output (last 50 lines) ---
          -- Found OpenMP_CXX: -fopenmp (found version "4.5")
          -- Found OpenMP: TRUE (found version "4.5")
          -- Found Python: /usr/bin/python (found version "3.10.12") found components: Interpreter
          -- Generating wrappers - done
          -- Found Python: /usr/bin/python (found version "3.10.12") found components: Interpreter Development Development.Module Development.Embed
          -- Performing Test HAS_FLTO_AUTO
          -- Performing Test HAS_FLTO_AUTO - Success
          -- Found pybind11: /usr/local/lib/python3.10/dist-packages/pybind11/include (found version "3.0.1")
          -- Configuring done (10.9s)
          -- Generating done (0.0s)
          -- Build files have been written to: /tmp/tmpmqa_49uf/build
          *** Building project with Ninja...
          [0/2] Re-checking globbed directories...
          [1/10] Building CXX object src/CMakeFiles/ops.dir/__/generated/bindings/ops.cc.o
          FAILED: src/CMakeFiles/ops.dir/__/generated/bindings/ops.cc.o
          /workspace/repo/scripts/mcc_wrapper.sh -DWITH_CPU=1 -DWITH_MOORE=1 -Dops_EXPORTS -I/workspace/repo/src -I/usr/local/musa/include -I/workspace/repo -isystem /usr/include/python3.10 -isystem /usr/local/lib/python3.10/dist-packages/pybind11/include -O3 -DNDEBUG -std=gnu++17 -fPIC -fvisibility=hidden -x musa -MD -MT src/CMakeFiles/ops.dir/__/generated/bindings/ops.cc.o -MF src/CMakeFiles/ops.dir/__/generated/bindings/ops.cc.o.d -o src/CMakeFiles/ops.dir/__/generated/bindings/ops.cc.o -c /workspace/repo/generated/bindings/ops.cc
          In file included from /workspace/repo/generated/bindings/ops.cc:4:
          In file included from /workspace/repo/src/moore/add/kernel.h:14:
          In file included from /workspace/repo/src/cuda/add/kernel.h:6:
          In file included from /workspace/repo/src/base/add.h:6:
          In file included from /workspace/repo/src/operator.h:10:
          In file included from /workspace/repo/src/dispatcher.h:10:
          In file included from /workspace/repo/src/data_type.h:19:
          In file included from /usr/local/musa/include/musa_fp16.h:3540:
          /usr/local/musa/include/musa_fp16_mtgpu.h:1609:26: error: out-of-line declaration of 'hrcp' does not match any declaration in namespace 'infini::ops'
          inline __device__ __half hrcp(__half x);
                                   ^~~~
          /workspace/repo/src/moore/polyfills.cuh:39:27: note: expanded from macro 'hrcp'
          #define hrcp infini::ops::hrcp
                                    ^~~~
          1 error generated when compiling for mp_21.
          [2/10] Building CXX object examples/CMakeFiles/data_type.dir/data_type.cc.o
          [3/10] Building CXX object examples/CMakeFiles/tensor.dir/tensor.cc.o
          [4/10] Building CXX object src/CMakeFiles/infiniops.dir/tensor.cc.o
          [5/10] Building CXX object examples/CMakeFiles/gemm.dir/gemm/gemm.cc.o
          ninja: build stopped: subcommand failed.
          
          *** CMake build failed
          [end of output]
      
      note: This error originates from a subprocess, and is likely not a problem with pip.
      ERROR: Failed building wheel for InfiniOps
    Failed to build InfiniOps
    
    [notice] A new release of pip is available: 25.2 -> 26.0.1
    [notice] To update, run: python -m pip install --upgrade pip
    error: failed-wheel-build-for-install
    
    × Failed to build installable wheels for some pyproject.toml based projects
    ╰─> InfiniOps
---
<== PASS  iluvatar_gpu  (93s)
<== PASS  nvidia_gpu    (114s)
<== PASS  metax_gpu     (207s)

========== Failed ==========
  FAIL  cambricon_gpu  error (0s)
  FAIL  moore_gpu      failure (24s)