HIP/ROCm fork optimized for AMD RDNA2 (gfx1030) with PrismML Q1_0_G128 1-bit quant support, RotorQuant, TurboQuant, EAGLE3 and P-EAGLE speculative decoding, and full Wave32 kernel optimizations.
-
Updated
Apr 16, 2026 - C++
#
gfx1030
Here are 3 public repositories matching this topic...
AMD ROCm (gfx1030) inference fork with RotorQuant/TurboQuant KV compression, PHANTOM-X zero-copy draft speculation, EAGLE3 speculative decoding, 12 RDNA2 crash fixes, and PrismML Bonsai Q1_0_G128 1-bit GGUF support.
triton hip bonsai rocm amd-gpu gguf speculative-decoding sglang rdna2 eagle3 turboquant prismml gfx1030 p-eagle radix-cache
-
Updated
Apr 16, 2026 - Python
Custom SPIR-V kernel factory for PHANTOM speculative decoding — LLVM IR to GPU (SPIR-V/HIP) and CPU (native x86) cross-target compilation, RDNA2/gfx1030 optimized pre-compiled kernels, dynamic kernel swapping, zero-JIT inference pipeline
-
Updated
Apr 13, 2026 - LLVM
Improve this page
Add a description, image, and links to the gfx1030 topic page so that developers can more easily learn about it.
Add this topic to your repo
To associate your repository with the gfx1030 topic, visit your repo's landing page and select "manage topics."