You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I'm so sorry to take your time, but I would really like to make this work.
I'm on Ubuntu with AMD RX 7800XT. It's not officially supported by ROCm, and I previously struggled to install ROCm, however I installed Ollama with ease and I noticed it installed ROCm as well and it seems to work, however I would like to move to llama.cpp as it seems to be more efficient. rocminfo | grep gfx | head -1 | awk '{print $2}' returns gfx1101, so it feels like it should work.
I can also build llama.cpp for CPU without any problems.
However, when I attempt to build with HIP: HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -R)" \ cmake -S . -B build -DGGML_HIP=ON -DAMDGPU_TARGETS=gfx1100 -DCMAKE_BUILD_TYPE=Release \ && cmake --build build --config Release -- -j 16
I get: CMake Error at /usr/share/cmake-3.30/Modules/CMakeDetermineHIPCompiler.cmake:49 (message): Could not find compiler set in environment variable HIPCXX: /usr/bin/clang. Call Stack (most recent call first): ggml/src/ggml-hip/CMakeLists.txt:36 (enable_language)
And indeed /usr/bin/clang does not exist, I have a shit tone of other files starting with clang there. My first intuition would be to do sudo apt install clang but I already read somewhere that it's not recommendable, as the build might require a specific version.
Additionally, I might immediately ask about something I will need if this is successful - the build documentation says:
The environment variable [HIP_VISIBLE_DEVICES] can be used to specify which GPU(s) will be used. If your GPU is not officially supported you can use the environment variable [HSA_OVERRIDE_GFX_VERSION] set to a similar GPU, for example 10.3.0 on RDNA2 (e.g. gfx1030, gfx1031, or gfx1035) or 11.0.0 on RDNA3.
This tells me I should in the future set [HSA_OVERRIDE_GFX_VERSION]=11.0.0 however I'm not sure where this environmental variable is going to be located, I'm only familiar with environmental variables stored in an .env file...
I would love to get some help with this, but I'm afraid the answer is either super simple or that it's impossible without ROCm support.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hi, I'm so sorry to take your time, but I would really like to make this work.
I'm on Ubuntu with AMD RX 7800XT. It's not officially supported by ROCm, and I previously struggled to install ROCm, however I installed Ollama with ease and I noticed it installed ROCm as well and it seems to work, however I would like to move to llama.cpp as it seems to be more efficient.
rocminfo | grep gfx | head -1 | awk '{print $2}'
returnsgfx1101
, so it feels like it should work.I can also build llama.cpp for CPU without any problems.
However, when I attempt to build with HIP:
HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -R)" \ cmake -S . -B build -DGGML_HIP=ON -DAMDGPU_TARGETS=gfx1100 -DCMAKE_BUILD_TYPE=Release \ && cmake --build build --config Release -- -j 16
I get:
CMake Error at /usr/share/cmake-3.30/Modules/CMakeDetermineHIPCompiler.cmake:49 (message): Could not find compiler set in environment variable HIPCXX: /usr/bin/clang. Call Stack (most recent call first): ggml/src/ggml-hip/CMakeLists.txt:36 (enable_language)
And indeed /usr/bin/clang does not exist, I have a shit tone of other files starting with clang there. My first intuition would be to do
sudo apt install clang
but I already read somewhere that it's not recommendable, as the build might require a specific version.Additionally, I might immediately ask about something I will need if this is successful - the build documentation says:
The environment variable [HIP_VISIBLE_DEVICES] can be used to specify which GPU(s) will be used. If your GPU is not officially supported you can use the environment variable [HSA_OVERRIDE_GFX_VERSION] set to a similar GPU, for example 10.3.0 on RDNA2 (e.g. gfx1030, gfx1031, or gfx1035) or 11.0.0 on RDNA3.
This tells me I should in the future set
[HSA_OVERRIDE_GFX_VERSION]=11.0.0
however I'm not sure where this environmental variable is going to be located, I'm only familiar with environmental variables stored in an .env file...I would love to get some help with this, but I'm afraid the answer is either super simple or that it's impossible without ROCm support.
Beta Was this translation helpful? Give feedback.
All reactions