feat: NUMA-aware model sharding package for POWER8 llama.cpp (issue #2277)#1799
Merged
Scottcjn merged 1 commit intoMar 25, 2026
Merged
Conversation
github-actions
Bot
added
documentation
Contributor
Author
|
RTC1d48d848a5aa5ecf2c5f01aa5fb64837daaf2f35 |
kuanglaodi2-sudo
added a commit
to kuanglaodi2-sudo/Rustchain
that referenced
this pull request
Mar 26, 2026
…m benchmarks, GGUF analyzer Enhanced implementation for Scottcjn/rustchain-bounties Scottcjn#2277 New additions: - ggml_numa_bindings.py: Python ctypes bindings for NUMA sharding API - GGMLNUMABindings class with full C API wrapper - Pure Python fallbacks when native library unavailable - recommend_shard_map() auto-generator for any model/layer count - CLI: topology, recommend, analyze commands - benchmark_numa.ps1: Cross-platform PowerShell benchmark harness - Works on Linux, macOS, Windows (PowerShell 7+) - Auto-detects NUMA topology and POWER8 architecture - Supports compare/baseline/numa modes - gguf_analyze.py: GGUF model tensor analyzer - Parses GGUF magic/version, extracts tensor metadata - Per-layer memory footprint analysis - Auto-generates NUMA shard recommendations - JSON and text output modes - power8_llama2_70b.json: Preset for LLaMA 2 70B (80 layers, 4-node) - power8_mixtral_8x7b.json: Preset for Mixtral-8x7B MoE (44 layers) - Updated README.md and FINAL_SUMMARY.md with new additions Based on merged PR Scottcjn#1799 (createkr) implementation. Bounty: Scottcjn/rustchain-bounties Scottcjn#2277 Wallet: C4c7r9WPsnEe6CUfegMU9M7ReHD1pWg8qeSfTBoRcLbg
Owner
|
Transfer confirmed — this was included in the batch settlement of 1,091 RTC to @createkr's wallet. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Implements a NUMA-aware sharding package for llama.cpp with layer routing, benchmark harness, reproducible presets, validation reports, and integration docs.\n\nThis submission is the working sharding package intended for the partial payout track while waiting for direct POWER8 hardware benchmark execution.\n\nValidation:\n- artifact package + benchmark harness included\n- tuning presets included\n- integration docs included\n\nCloses #2277