{"id":25616685,"url":"https://github.com/eliranwong/amd_igpu_ai_setup","last_synced_at":"2026-06-17T06:32:09.350Z","repository":{"id":278424396,"uuid":"934452011","full_name":"eliranwong/AMD_iGPU_AI_Setup","owner":"eliranwong","description":"AMD iGPU AI Setup and Speed Test - GPD Pocket 4 - Linux + ROCm + AgentMake AI","archived":false,"fork":false,"pushed_at":"2025-02-19T18:24:34.000Z","size":27,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-02-19T18:25:58.602Z","etag":null,"topics":["agentmake","ai","amd","gpd","gpdpocket","llamacpp","ollama","perplexica","rocm","searxng","ubuntu"],"latest_commit_sha":null,"homepage":"","language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/eliranwong.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2025-02-17T21:25:00.000Z","updated_at":"2025-02-19T18:24:38.000Z","dependencies_parsed_at":"2025-02-19T18:36:06.653Z","dependency_job_id":null,"html_url":"https://github.com/eliranwong/AMD_iGPU_AI_Setup","commit_stats":null,"previous_names":["eliranwong/amd_igpu_ai_setup"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/eliranwong%2FAMD_iGPU_AI_Setup","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/eliranwong%2FAMD_iGPU_AI_Setup/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/eliranwong%2FAMD_iGPU_AI_Setup/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/eliranwong%2FAMD_iGPU_AI_Setup/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/eliranwong","download_url":"https://codeload.github.com/eliranwong/AMD_iGPU_AI_Setup/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":240122812,"owners_count":19751178,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agentmake","ai","amd","gpd","gpdpocket","llamacpp","ollama","perplexica","rocm","searxng","ubuntu"],"created_at":"2025-02-22T04:17:59.544Z","updated_at":"2026-06-17T06:32:09.341Z","avatar_url":"https://github.com/eliranwong.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"# AMD_iGPU_AI_Setup\n\nAMD-iGPU-device setup for AI development.  We also record some speed test results.\n\nFor multiple AMD-GPU setup, please visit https://github.com/eliranwong/MultiAMDGPU_AIDev_Ubuntu\n\n# Tested Device\n\nDevice: [GPD Pocket 4](https://gpd.hk/gpdpocket4)\n\nHardware: AMD Ryzen™ AI 9 HX 370; AMD Radeon™ 890M (RDNA 3.5)\n\n## Memory Setting\n\nBIOS Memory Setting (reboot+DEL key):\n\nUEFI/BIOS -\u003e Advanced -\u003e AMD CBS -\u003e NBIO -\u003e GFX Configuration \u003e \n\nDefault settings:\n\n```\niGPU Advanced Control \u003e Disabled\nDedicated Graphics Memory \u003e Medium (16GB)\nRemaining System Memory \u003e 48GB\n```\n\nSettings for the best performance:\n\nFor comparison in performance, please refer to our [speed test results](https://github.com/eliranwong/AMD_iGPU_AI_Setup/blob/main/README.md#speed-tests)\n\n```\niGPU Advanced Control \u003e Enabled\nDedicated Graphics Memory \u003e 32GB\nRemaining System Memory \u003e 32GB\n```\n\n## Operating System\n\nOperating System: [Ubuntu 24.04.1 LTS](https://ubuntu.com/tutorials/install-ubuntu-desktop#1-overview) (reboot+F7 to install)\n\n\u003e uname -a\n\n```\nLinux ai 6.14.0-37-generic #37~24.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Nov 20 10:25:38 UTC 2 x86_64 x86_64 x86_64 GNU/Linux\n```\n\n# Install Basic Tools\n\n\u003e sudo apt update \u0026\u0026 sudo apt full-upgrade\n\n\u003e sudo apt install \"linux-headers-$(uname -r)\" \"linux-modules-extra-$(uname -r)\"\n\n\u003e sudo apt -y install software-properties-common dirmngr apt-transport-https lsb-release ca-certificates apt-utils build-essential make cmake tree wget curl git zip unzip xz-utils nano micro w3m lynx sqlite3 libsqlite3-dev sqlitebrowser libnss3 libnss3-dev libgl1-mesa-dev mesa-utils libglu1-mesa lsb-release binutils ffmpeg gawk opencc plocate gnome-keyring libssl-dev libffi-dev libpci3 libpci-dev python3 python3-setuptools python3-pip python3-dev python3-venv zlib1g-dev libgdbm-dev libreadline-dev libbz2-dev gcc xorg-dev exo-utils dex xdg-utils libavcodec-extra libportaudio2 moreutils llvm tk-dev liblzma-dev python3-openssl libxml2-dev libxmlsec1-dev protobuf-compiler libc6-dev libstdc++-12-dev libxcb-cursor-dev libxcb-cursor0 libncurses-dev libncurses6 ubuntu-restricted-addons ubuntu-restricted-extras gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-plugins-ugly xsel portaudio19-dev vlc python3-wheel python3-wheel-whl twine libomp-dev gnome-shell-extension-manager\n\n## Touchscreen Tools\n\nLaunch `Extension Manager` and install:\n\n* Screen Rotate\n\n* GJS OSK\n\n# Install ROCM\n\nFind details at: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installryz/native_linux/install-ryzen.html\n\nTested ROCm version: 7.1.1\n\n```\nsudo usermod -a -G render,video $LOGNAME\nsudo apt update\nsudo apt install -y libstdc++-12-dev\nwget https://repo.radeon.com/amdgpu-install/7.1.1/ubuntu/noble/amdgpu-install_7.1.1.70101-1_all.deb\nsudo apt install ./amdgpu-install_7.1.1.70101-1_all.deb\nsudo amdgpu-install --usecase=graphics,multimedia,rocm,rocmdev,rocmdevtools,lrt,opencl,openclsdk,hip,hiplibsdk,openmpsdk,mllib,mlsdk --no-dkms -y\nsudo reboot\n```\n\nRead more at: https://github.com/eliranwong/MultiAMDGPU_AIDev_Ubuntu/blob/main/README.md\n\n# Check gfx driver version\n\n\u003e rocminfo\n\nExpected output:\n\n```\n...\nName: gfx1150\n...\n```\n\n# Environment variables\n\nAdd the following lines to `~/.bashrc`:\n\n```\nexport ROCM_HOME=/opt/rocm-7.1.1\nexport LD_LIBRARY_PATH=/opt/rocm-7.1.1/include:/opt/rocm-7.1.1/lib:$LD_LIBRARY_PATH\nexport PATH=$HOME/.local/bin:/opt/rocm-7.1.1/bin:/opt/rocm-7.1.1/llvm/bin:$PATH\n```\n\n```\n# rocm\nexport GFX_ARCH=gfx1150\nexport HCC_AMDGPU_TARGET=gfx1150\nexport CUPY_INSTALL_USE_HIP=1\nexport ROCM_VERSION=7.1\nexport ROCM_HOME=/opt/rocm\nexport LD_LIBRARY_PATH=/usr/include/vulkan:/opt/rocm/include:/opt/rocm/lib:/opt/rocm/lib/llvm/lib:/opt/rocm/lib/migraphx/lib:$LD_LIBRARY_PATH\nexport PATH=/home/eliran/.local/bin:/opt/rocm/bin:/opt/rocm/llvm/bin:$PATH\nexport HSA_OVERRIDE_GFX_VERSION=11.5.0\n#export ROCR_VISIBLE_DEVICES=GPU-XX\nexport GPU_DEVICE_ORDINAL=0\nexport HIP_VISIBLE_DEVICES=0\nexport CUDA_VISIBLE_DEVICES=0\nexport LLAMA_HIPLAS=0\nexport DRI_PRIME=0\nexport OMP_DEFAULT_DEVICE=0\n# vulkan\nexport GGML_VULKAN_DEVICE=0\nexport GGML_VK_VISIBLE_DEVICES=0\nexport VULKAN_SDK=/usr/share/vulkan\nexport VK_LAYER_PATH=$VULKAN_SDK/explicit_layer.d\n```\n\n## Remarks about ROCR_VISIBLE_DEVICES\n\nDo NOT set ROCR_VISIBLE_DEVICES for iGPU.\n\nSet it ONLY for discrete GPUs, e.g. https://github.com/eliranwong/MultiAMDGPU_AIDev_Ubuntu#overview\n\n## Remarks about HSA_OVERRIDE_GFX_VERSION:\n\n1. Check `rocminfo` output first\n\n2. General workaround if gfx version is not available in `rocminfo` output:\n\nfor GCN 5th gen based GPUs and APUs HSA_OVERRIDE_GFX_VERSION=9.0.0\n\nfor RDNA 1 based GPUs and APUs HSA_OVERRIDE_GFX_VERSION=10.1.0\n\nfor RDNA 2 based GPUs and APUs HSA_OVERRIDE_GFX_VERSION=10.3.0\n\nfor RDNA 3 based GPUs and APUs HSA_OVERRIDE_GFX_VERSION=11.0.0\n\nfor RDNA 3.5 based GPUs and APUs HSA_OVERRIDE_GFX_VERSION=11.5.0\n\n3. Read more at: https://llvm.org/docs/AMDGPUUsage.html#processors\n\nIn my case, I am running ROCm version 7.1.1.\n\nWhen I run:\n\n\u003e ls /opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx*.dat\n\nI got the following output:\n\n```\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1030.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1100.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1101.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1102.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1150.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1151.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1200.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx1201.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx908.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx90a.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx942.dat\n/opt/rocm/lib/rocblas/library/TensileLibrary_lazy_gfx950.dat\n```\n\nThis confirms that ROCm version 7.1.1 does support `gfx1150`.\n\n# Install Docker Engine\n\n```\n# uninstall any old versions\nfor pkg in docker.io docker-doc docker-compose docker-compose-v2 podman-docker containerd runc; do sudo apt-get remove $pkg; done\n\n# Add Docker's official GPG key:\nsudo apt update\nsudo apt install ca-certificates curl\nsudo install -m 0755 -d /etc/apt/keyrings\nsudo curl -fsSL https://download.docker.com/linux/ubuntu/gpg -o /etc/apt/keyrings/docker.asc\nsudo chmod a+r /etc/apt/keyrings/docker.asc\n\n# Add the repository to Apt sources:\nsudo tee /etc/apt/sources.list.d/docker.sources \u003c\u003cEOF\nTypes: deb\nURIs: https://download.docker.com/linux/ubuntu\nSuites: $(. /etc/os-release \u0026\u0026 echo \"${UBUNTU_CODENAME:-$VERSION_CODENAME}\")\nComponents: stable\nSigned-By: /etc/apt/keyrings/docker.asc\nEOF\n\nsudo apt update\n\n# Install the docker packages\nsudo apt-get install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin\n\n# add user to docker group\nsudo usermod -aG docker $LOGNAME\nnewgrp docker\n```\n\nRead more at https://docs.docker.com/engine/install/ubuntu/\n\n# Install Perplexica\n\nRun in terminal:\n\n```\nsudo apt install -y git\ngit clone https://github.com/ItzCrazyKns/Perplexica.git\ncd Perplexica\ndocker build -t perplexica .\ndocker run -d --restart unless-stopped -p 3000:3000 -p 4000:8080 -v perplexica-data:/home/perplexica/data -v perplexica-uploads:/home/perplexica/uploads --name perplexica perplexica\n```\n\nTo open Perplexica and set up providers, run:\n\n\u003e open http://localhost:3000\n\nTo open SearXNG, run:\n\n\u003e open http://localhost:4000\n\n# Install Ollama\n\nStandard installation: https://ollama.com/download\n\nConfigure Ollama, run:\n\n\u003e sudo nano /etc/systemd/system/ollama.service\n\nAdd the following three lines at the end of the [Service] session:\n\n```\nEnvironment=\"OLLAMA_NUM_PARALLEL=2\"\nEnvironment=\"OLLAMA_MAX_LOADED_MODELS=2\"\nEnvironment=\"OLLAMA_HOST=0.0.0.0\"\n```\n\nReload Ollama, run:\n\n\u003e sudo systemctl daemon-reload\n\n\u003e sudo systemctl restart ollama\n\nAdd user to group `ollama` for access of Ollama directory:\n\n\u003e sudo usermod -a -G ollama $LOGNAME\n\n\u003e sudo reboot\n\n# Build llama.cpp that runs ROCm backend\n\nRun in terminal:\n\n```\ngit clone https://github.com/ggml-org/llama.cpp\ncd llama.cpp\nHIPCXX=\"$(hipconfig -l)/clang\" HIP_PATH=\"$(hipconfig -R)\" cmake -S . -B build -DGGML_HIP=ON -DGGML_HIP_UMA=ON -DAMDGPU_TARGETS=gfx1150 -DCMAKE_BUILD_TYPE=Release \u0026\u0026 cmake --build build --config Release -- -j $(lscpu | grep -m 1 '^Core(s)' | awk '{print $NF}')\n```\n\nExpected lines in the terminal output:\n\n```\n...\n-- Adding CPU backend variant ggml-cpu: -march=native \n-- The HIP compiler identification is Clang 18.0.0\n-- Detecting HIP compiler ABI info\n-- Detecting HIP compiler ABI info - done\n-- Check for working HIP compiler: /opt/rocm-7.1.1/lib/llvm/bin/clang - skipped\n-- Detecting HIP compile features\n-- Detecting HIP compile features - done\n-- HIP and hipBLAS found\n-- Including HIP backend\n...\n```\n\n# Build llama.cpp that runs Vulkan backend\n\nAs an alternative to ROCm backend, you may build a copy of llama.cpp that runs Vulkan backend. In our tested device with iGPU, Vulkan backends performs better than ROCm backend. For details, please refer to [speed test results](https://github.com/eliranwong/AMD_iGPU_AI_Setup/blob/main/README.md#speed-tests).\n\nTo set up Vulkan driver:\n\n```\nsudo apt install glslc glslang-tools glslang-dev mesa-vulkan-drivers vulkan-tools libvulkan-dev libvulkan-memory-allocator-dev libvulkan-volk-dev vulkan-validationlayers vulkan-utility-libraries-dev\n```\n\nInstall libcurl development headers (if on Ubuntu/Debian):\n\n```\nsudo apt-get install libcurl4-openssl-dev\n```\n\nTo build run:\n\n```\ngit clone https://github.com/ggml-org/llama.cpp\ncd llama.cpp\ncmake -S . -B build -DLLAMA_CURL=ON -DGGML_VULKAN=ON -DCMAKE_BUILD_TYPE=Release \u0026\u0026 cmake --build build --config Release -- -j $(lscpu | grep -m 1 '^Core(s)' | awk '{print $NF}')\n```\n\nExpected lines in the terminal output:\n\n```\n...\n-- Adding CPU backend variant ggml-cpu: -march=native \n-- Found Vulkan: /usr/lib/x86_64-linux-gnu/libvulkan.so (found version \"1.3.275\") found components: glslc glslangValidator \n-- Vulkan found\n-- GL_KHR_cooperative_matrix supported by glslc\n-- GL_NV_cooperative_matrix2 not supported by glslc\n-- Including Vulkan backend\n...\n```\n\nMake sure you set the vulkan-related variables, e.g. https://github.com/eliranwong/AMD_iGPU_AI_Setup#environment-variables\n\n## Alias for launching llama-server\n\nRun in terminal:\n\n```\ncd llama.cpp\necho \"alias llamacpp=\\\"cd /home/$USER/agentmake/models/gguf/ \u0026\u0026 $(pwd)/build/bin/llama-server --threads $(lscpu | grep -m 1 '^Core(s)' | awk '{print $NF}') -ngl 99 --model\\\"\" \u003e\u003e $HOME/.bashrc\n```\n\nRemarks: We add `-ngl 99` to offload as many layers as available to GPU. Using `vulkan` backend, we managed to run `70b` models on the tested device with `-ngl 99` specified. Depending on your hardware, you may need to reduce the value of ngl to load large-sized models.\n\n# Install Open-notebook\n\n```\n# Create project directory\nmkdir -p ~/dev/open-notebook \u0026\u0026 cd ~/dev/open-notebook\n\n# Download configuration files\ncurl -O https://raw.githubusercontent.com/lfnovo/open-notebook/main/docker-compose.full.yml\ncurl -O https://raw.githubusercontent.com/lfnovo/open-notebook/main/.env.example\n\n# Rename and configure environment\nmv .env.example docker.env\n```\n\n```\n# Edit docker.env with your API keys\nnano docker.env\n```\n# Rename and edit the docker-compose file\nmv docker-compose.full.yml docker-compose.yml\n# In my case the port `8000` is already used by another service, so\nsed -i 's/8000:8000/9000:8000/g' docker-compose.yml\n\n# Start Open Notebook\ndocker compose up -d\n\n# Make save data accessible\nsudo chmod a+w notebook_data/ surreal_data/\n\n# Access open-notebook\nopen http://localhost:8502\n```\n\n# Install Fabric\n\nRun in terminal:\n\n```\nmkdir -p ~/.local/bin\ncd ~/.local/bin\ncurl -L https://github.com/danielmiessler/fabric/releases/latest/download/fabric-linux-amd64 \u003e fabric \u0026\u0026 chmod +x fabric \u0026\u0026 ./fabric --version\nfabric --setup\n```\n\n# Install Agentmake AI\n\nRun in terminal:\n\n```\n# optional: navigate to home directory\ncd\n# install in a virtual environment\npython3 -m venv ai\nsource ai/bin/activate\npip install --upgrade agentmake[genai]\necho \"export PATH=$PATH:$HOME/ai/bin\" \u003e\u003e ~/.bashrc\n# To test\nai Hi!\n```\n\n## Edit Configurations\n\nTo edit configurations, run in terminal:\n\n\u003e ai -ec\n\n## Test with Ollama\n\n\u003e ai Hi!\n\nRemarks: Ollama is set as the default backend, so you can use the `ai` or `aic` commands without specifying the backend option. Run `ai -ec` to edit configurations.\n\n## Test with Llama.cpp\n\nTo access model files, downloaded via ollama, add user to group `ollama`:\n\n\u003e sudo usermod -a -G ollama $LOGNAME\n\n\u003e sudo reboot\n\nTo download a model via Ollama and save a copy of it in `~/agentmake/models/gguf/` by default, e.g.:\n\n\u003e ai --get_model deepseek-r1 -gm llama3.3:70b -gm aya-expanse\n\nTo run an instance of llama-server, assuming that you have set up an alias as mentioned [here](https://github.com/eliranwong/AMD_iGPU_AI_Setup#alias-for-launching-llama-server), e.g.:\n\n\u003e llamacpp deepseek-r1.gguf\n\nTo run agentmake with llama.cpp, e.g.:\n\n\u003e ai -b llamacpp Hi!\n\n## Test with Perplexica\n\nTo list available tools that work with perplexica, run:\n\n\u003e ai -lt | grep perplexica\n\nExpected output:\n\n```\nperplexica/openai\nperplexica/groq\nperplexica/xai\nperplexica/googleai\nperplexica/anthropic\nperplexica/github\n```\n\nTo use one of them, e.g.:\n\n\u003e ai -t perplexica/github What is AgentMake AI?\n\n## Test with SearXNG\n\nSearXNG is automatically installed with Perplexica, to get real-time information, e.g.:\n\n\u003e ai -t search/searxng Give me news updates in London today.\n\n## Test with Fabric Integration\n\nAssuming fabric patterns are downloaded, e.g.:\n\n\u003e ai What are AI agents? -sys fabric.write_micro_essay -b genai\n\n## Test with Selected Text in Any Applicaitons\n\nFirst, make sure `xsel` is installed:\n\n\u003e sudo apt install xsel\n\nLaunch `Settings` \u003e Keyboard \u003e View and Customise Shortcuts \u003e Custom Shortcuts \u003e +\n\nFill in content, like below (replace `username` with your `username`: \n\n```\nName: AgentMake AI\nCommand: gnome-terminal -- bash -c \"/home/username/ai/bin/ai -i -eo -py\"\nShift+Ctrl+A\n```\n\n![Image](https://github.com/user-attachments/assets/d21fea9a-2288-4e85-96ad-dfbee7ce160d)\n\nSelect some text in an application, then press `Shift+Ctrl+A`.\n\nChoose a predefined instruction:\n\n![Image](https://github.com/user-attachments/assets/e4872498-0cef-48e7-a550-55c0c4234929)\n\nAssistant response is automatically copied to clipboard.\n\nRemarks: You can define up to 10 custom instructions for being selected in the dialog, by specifying the values of `CUSTOM_INSTRUCTION_1`, `CUSTOM_INSTRUCTION_2`, `CUSTOM_INSTRUCTION_3`, ... `CUSTOM_INSTRUCTION_10` in AgentMake configurations (run `ai -ec` to edit).\n\n## Note about Azure AI Setup\n\nAn easy way to deploy AI models via Azure service:\n\n1. Sign in https://ai.azure.com/github\n2. All resources \u003e Create New\n3. Overview \u003e copy an API key, Azure OpenAI Service and Azure AI inference endpoints\n\n* Use Azure OpenAI Service endpoint for running OpenAI models; the endpoint should look like https://resource_name.openai.azure.com/\n\n* Use Azure AI inference endpoint for running DeepSeek-R1 and Phi-4; the endpoint should look like https://resource_name.services.ai.azure.com/models\n\nTo configure AgentMake AI, run:\n\n\u003e ai -ec\n\n## Note about Vertex AI\n\nMake sure the extra package `genai` is installed with the command mentioned above:\n\n\u003e pip install --upgrade \"agentmake[genai]\"\n\nTo configure, run:\n\n\u003e ai -ec\n\nEnter the path of your Google application credentials JSON file as the value of `VERTEXAI_API_KEY`. You need to specify your project ID and service location, in the configurations, as well. e.g.:\n\n```\nVERTEXAI_API_KEY=~/agentmake/google_application_credentials.json\nVERTEXAI_API_PROJECT_ID=my_project_id\nVERTEXAI_API_SERVICE_LOCATION=us-central1\n```\n\nTo test Gemini 2.0 with Vertex AI, e.g.:\n\n\u003e ai -b vertexai -m gemini-2.0-flash Hi!\n\n## Using other backends and tools\n\nAgentMake AI supports 14 AI backends and 7 agentic components.\n\nRead more at https://github.com/eliranwong/agentmake\n\n# Install OpenClaw with Discord Integration\n\n[OpenClaw Setup](openclaw.md)\n\n# Speed Tests\n\nTested with the same prompt `\"What is machine learning?\"`:\n\n## Dedicated Graphics Memory: 16G\n\nLlama.cpp+ROCM [ Prompt: 448.5 t/s | Generation: 28.7 t/s ]\nLlama.cpp+Vulkan [ Prompt: 341.1 t/s | Generation: 29.4 t/s ]\nOllama [ prompt eval rate: 64.58 tokens/s | eval rate: 16.93 tokens/s ]\n\n## Dedicated Graphics Memory: 32G\n\nLlama.cpp+ROCM [ Prompt: 562.1 t/s | Generation: 29.7 t/s ]\nLlama.cpp+Vulkan [ Prompt: 410.7 t/s | Generation: 30.4 t/s ]\nOllama [ prompt eval rate: 63.77 tokens/s | eval rate: 16.83 tokens/s ]\n\n## Dedicated Graphics Memory: 48G\n\nLlama.cpp+ROCM [ Prompt: 551.6 t/s | Generation: 29.8 t/s ]\nLlama.cpp+Vulkan [ Prompt: 399.0 t/s | Generation: 30.5 t/s ]\nOllama [ prompt eval rate: 21.56 tokens/s | eval rate: 14.92 tokens/s ]\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Feliranwong%2Famd_igpu_ai_setup","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Feliranwong%2Famd_igpu_ai_setup","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Feliranwong%2Famd_igpu_ai_setup/lists"}