[PR #12008] [CLOSED] [WIP]: remove volkan header dependencies #44932

Closed
opened 2026-04-25 00:36:35 -05:00 by GiteaMirror · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/12008
Author: @fishingfly
Created: 8/21/2025
Status: Closed

Base: mainHead: enhancement/bulkan


📝 Commits (10+)

  • 7b90154 musa: support Moore Threads GPU
  • 1092984 hack: download go from aliyun/use goproxy.cn/only build musa image (#3)
  • 9a19bd5 hack: add archive build script and update install.sh to support archive install (#4)
  • 6816ecd hack: install.sh detects aibook (#5)
  • 6bf5f2a musa: remove hardcoded platform constraints from Docker target musa-4 (#6)
  • 6a1d9bd musa: support flash attention (#9)
  • 6b06001 hack: update version for building docker image and tgz (#11)
  • 163587b musa: use mudnn::Unary::IDENTITY op to accelerate D2D memory copy (#12)
  • d5834bd musa: update doc for musa sdk to rc4.2.0 (#14)
  • ab04906 musa: upgrade musa sdk to rc4.2.0 (#13)

📊 Changes

167 files changed (+29332 additions, -37 deletions)

View changed files

📝 .gitignore (+1 -0)
📝 CMakeLists.txt (+43 -0)
📝 CMakePresets.json (+41 -0)
📝 Dockerfile (+63 -3)
📝 LICENSE (+11 -0)
📝 README.md (+37 -1)
discover/Makefile.vulkan (+41 -0)
📝 discover/gpu.go (+490 -8)
📝 discover/gpu_info.h (+5 -1)
discover/gpu_info_mtml.c (+141 -0)
discover/gpu_info_mtml.h (+48 -0)
discover/gpu_info_mtmusa.c (+248 -0)
discover/gpu_info_mtmusa.h (+69 -0)
discover/gpu_info_musart.c (+183 -0)
discover/gpu_info_musart.h (+137 -0)
📝 discover/gpu_info_nvml.c (+1 -1)
discover/gpu_info_vulkan.c (+183 -0)
discover/gpu_info_vulkan.h (+1380 -0)
📝 discover/gpu_linux.go (+30 -0)
📝 discover/gpu_test.go (+1 -1)

...and 80 more files

📄 Description

No description provided


🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/12008 **Author:** [@fishingfly](https://github.com/fishingfly) **Created:** 8/21/2025 **Status:** ❌ Closed **Base:** `main` ← **Head:** `enhancement/bulkan` --- ### 📝 Commits (10+) - [`7b90154`](https://github.com/ollama/ollama/commit/7b90154591551ff0679e75fe2d24e2b38dec4682) musa: support Moore Threads GPU - [`1092984`](https://github.com/ollama/ollama/commit/10929848f3ca5185ae88f25e022fac7aa47d65cf) hack: download go from aliyun/use goproxy.cn/only build musa image (#3) - [`9a19bd5`](https://github.com/ollama/ollama/commit/9a19bd51ecc38bf5f32f2ced426b82071de4ff5a) hack: add archive build script and update install.sh to support archive install (#4) - [`6816ecd`](https://github.com/ollama/ollama/commit/6816ecd8aea9c545e277df3801a3afa51489a03d) hack: install.sh detects aibook (#5) - [`6bf5f2a`](https://github.com/ollama/ollama/commit/6bf5f2a894056e50b77b84dae1f010d0622175a8) musa: remove hardcoded platform constraints from Docker target musa-4 (#6) - [`6a1d9bd`](https://github.com/ollama/ollama/commit/6a1d9bd51ecfaf895d865c5255e872bf9f1192ef) musa: support flash attention (#9) - [`6b06001`](https://github.com/ollama/ollama/commit/6b0600118a01da2e09e07413885aecb795cbc0ce) hack: update version for building docker image and tgz (#11) - [`163587b`](https://github.com/ollama/ollama/commit/163587b6c1690def3ddbc8934918d87f01d8c122) musa: use mudnn::Unary::IDENTITY op to accelerate D2D memory copy (#12) - [`d5834bd`](https://github.com/ollama/ollama/commit/d5834bd0c3e969c99047b63f4c9b615d1bf5583a) musa: update doc for musa sdk to rc4.2.0 (#14) - [`ab04906`](https://github.com/ollama/ollama/commit/ab04906bcf566c56a434b75e3f18c3c661aa6c75) musa: upgrade musa sdk to rc4.2.0 (#13) ### 📊 Changes **167 files changed** (+29332 additions, -37 deletions) <details> <summary>View changed files</summary> 📝 `.gitignore` (+1 -0) 📝 `CMakeLists.txt` (+43 -0) 📝 `CMakePresets.json` (+41 -0) 📝 `Dockerfile` (+63 -3) 📝 `LICENSE` (+11 -0) 📝 `README.md` (+37 -1) ➕ `discover/Makefile.vulkan` (+41 -0) 📝 `discover/gpu.go` (+490 -8) 📝 `discover/gpu_info.h` (+5 -1) ➕ `discover/gpu_info_mtml.c` (+141 -0) ➕ `discover/gpu_info_mtml.h` (+48 -0) ➕ `discover/gpu_info_mtmusa.c` (+248 -0) ➕ `discover/gpu_info_mtmusa.h` (+69 -0) ➕ `discover/gpu_info_musart.c` (+183 -0) ➕ `discover/gpu_info_musart.h` (+137 -0) 📝 `discover/gpu_info_nvml.c` (+1 -1) ➕ `discover/gpu_info_vulkan.c` (+183 -0) ➕ `discover/gpu_info_vulkan.h` (+1380 -0) 📝 `discover/gpu_linux.go` (+30 -0) 📝 `discover/gpu_test.go` (+1 -1) _...and 80 more files_ </details> ### 📄 Description _No description provided_ --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-04-25 00:36:35 -05:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#44932