[PR #12070] [CLOSED] PowerPC:Fix for issue #12076 #13701

Closed
opened 2026-04-13 00:33:13 -05:00 by GiteaMirror · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/12070
Author: @shalinib-ibm
Created: 8/25/2025
Status: Closed

Base: mainHead: ppc_bug_fix


📝 Commits (1)

  • f08cecd PowerPC:Fix build failure

📊 Changes

7 files changed (+2397 additions, -4 deletions)

View changed files

📝 ml/backend/ggml/ggml/src/CMakeLists.txt (+1 -1)
📝 ml/backend/ggml/ggml/src/ggml-cpu/arch-fallback.h (+0 -1)
ml/backend/ggml/ggml/src/ggml-cpu/arch/powerpc/cpu-feats.cpp (+82 -0)
ml/backend/ggml/ggml/src/ggml-cpu/arch/powerpc/powerpc.go (+5 -0)
ml/backend/ggml/ggml/src/ggml-cpu/arch/powerpc/quants.c (+2305 -0)
📝 ml/backend/ggml/ggml/src/ggml-cpu/llamafile/sgemm.cpp (+3 -0)
📝 ml/backend/ggml/ggml/src/ggml-cpu/simd-mappings.h (+1 -2)

📄 Description

This patch adds cpu backend support and fix build failures on POWER systems

  1. Added ggml-cpu/arch/powerpc directory to host Power-specific code.
  2. Fixed compilation errors in sgemm.cpp and simd-mappings.h where vector was not defined by explicitly mapping it to __vector.
  3. Resolved build errors from -mcpu=power11 by replacing it with the supported -mcpu=power10.

Build steps:
cmake -B build
cmake --build build
export CGO_LDFLAGS="-L$(pwd)/build/lib/ollama/ -lggml-cpu-power10"
go build .


🔄 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/12070 **Author:** [@shalinib-ibm](https://github.com/shalinib-ibm) **Created:** 8/25/2025 **Status:** ❌ Closed **Base:** `main` ← **Head:** `ppc_bug_fix` --- ### 📝 Commits (1) - [`f08cecd`](https://github.com/ollama/ollama/commit/f08cecd7418ef56d53acfa6469581c4c57dac5d1) PowerPC:Fix build failure ### 📊 Changes **7 files changed** (+2397 additions, -4 deletions) <details> <summary>View changed files</summary> 📝 `ml/backend/ggml/ggml/src/CMakeLists.txt` (+1 -1) 📝 `ml/backend/ggml/ggml/src/ggml-cpu/arch-fallback.h` (+0 -1) ➕ `ml/backend/ggml/ggml/src/ggml-cpu/arch/powerpc/cpu-feats.cpp` (+82 -0) ➕ `ml/backend/ggml/ggml/src/ggml-cpu/arch/powerpc/powerpc.go` (+5 -0) ➕ `ml/backend/ggml/ggml/src/ggml-cpu/arch/powerpc/quants.c` (+2305 -0) 📝 `ml/backend/ggml/ggml/src/ggml-cpu/llamafile/sgemm.cpp` (+3 -0) 📝 `ml/backend/ggml/ggml/src/ggml-cpu/simd-mappings.h` (+1 -2) </details> ### 📄 Description This patch adds cpu backend support and fix build failures on POWER systems 1. Added `ggml-cpu/arch/powerpc` directory to host Power-specific code. 2. Fixed compilation errors in `sgemm.cpp` and simd-mappings.h where `vector` was not defined by explicitly mapping it to `__vector`. 3. Resolved build errors from `-mcpu=power11` by replacing it with the supported `-mcpu=power10`. Build steps: cmake -B build cmake --build build export CGO_LDFLAGS="-L$(pwd)/build/lib/ollama/ -lggml-cpu-power10" go build . --- <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 2026-04-13 00:33:13 -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#13701