-
Notifications
You must be signed in to change notification settings - Fork 230
Description
I have seen #975 and I set:
export CUDA_SOURCE_DIR=/usr/local/cuda-12.9
export PATH="/usr/local/cuda-12.9/bin:$PATH"
after installing 12.9 toolkit
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/sbsa/cuda-ubuntu2404.pin
sudo mv cuda-ubuntu2404.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/12.9.0/local_installers/cuda-repo-ubuntu2404-12-9-local_12.9.0-575.51.03-1_arm64.deb
sudo dpkg -i cuda-repo-ubuntu2404-12-9-local_12.9.0-575.51.03-1_arm64.deb
sudo cp /var/cuda-repo-ubuntu2404-12-9-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-9
nvidia-smi
Fri Dec 19 17:54:01 2025
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 575.57.08 Driver Version: 575.57.08 CUDA Version: 12.9 |
|-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GH200 480GB On | 00000009:01:00.0 Off | 0 |
| N/A 25C P0 79W / 900W | 0MiB / 97871MiB | 0% Default |
| | | Disabled |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| No running processes found |
+-----------------------------------------------------------------------------------------+
uv venv
source .venv/bin/activate
uv pip install torch --reinstall --index-url=https://download.pytorch.org/whl/cu129
uv pip install --reinstall cupti-python setuptools
Resolved 6 packages in 68ms
Prepared 6 packages in 0.30ms
Uninstalled 6 packages in 15ms
Installed 6 packages in 13ms
~ cuda-bindings==13.0.0
~ cuda-pathfinder==1.3.3
~ cupti-python==13.0.0
~ numpy==2.3.5
~ nvidia-cuda-cupti==13.0.85
~ setuptools==80.9.0
$ cmake .
-- ROCM_SOURCE_DIR =
-- Found CUPTI
-- Found NVPERF: Using built-in FindCUDAToolkit target
-- {fmt} version: 11.2.0
-- Build type: Release
-- Kineto: FMT_SOURCE_DIR = /home/user/kineto/libkineto/third_party/fmt
-- ROCTRACER_INCLUDE_DIR = /include/roctracer
-- DYNOLOG_INCLUDE_DIR = /home/user/kineto/libkineto/third_party/dynolog/
-- IPCFABRIC_INCLUDE_DIR = /home/user/kineto/libkineto/third_party/dynolog//dynolog/src/ipcfabric/
-- Configuring done (0.4s)
-- Generating done (0.0s)
-- Build files have been written to: /home/user/kineto/libkineto
$ make
[ 0%] Built target libkineto_defs.bzl
[ 54%] Built target kineto_base
[ 57%] Built target kineto_api
[ 59%] Built target kineto
[ 63%] Built target fmt
[ 67%] Built target gtest
[ 70%] Built target gmock
[ 73%] Built target gmock_main
[ 77%] Built target gtest_main
[ 78%] Linking CXX executable ConfigTest
/home/user/kineto/libkineto/test/ConfigTest: symbol lookup error: /home/user/kineto/libkineto/test/ConfigTest: undefined symbol: NVPW_MetricsContext_GetMetricProperties_Begin
CMake Error at /usr/share/cmake-3.28/Modules/GoogleTestAddTests.cmake:112 (message):
Error running test executable.
Path: '/home/user/kineto/libkineto/test/ConfigTest'
Result: 127
Output:
Call Stack (most recent call first):
/usr/share/cmake-3.28/Modules/GoogleTestAddTests.cmake:225 (gtest_discover_tests_impl)
make[2]: *** [test/CMakeFiles/ConfigTest.dir/build.make:173: test/ConfigTest] Error 1
make[2]: *** Deleting file 'test/ConfigTest'
make[1]: *** [CMakeFiles/Makefile2:446: test/CMakeFiles/ConfigTest.dir/all] Error 2
make: *** [Makefile:146: all] Error 2
Where is this `NVPW_MetricsContext_GetMetricProperties_Begin` symbol?