End-to-End ML Graph Compiler Fused with Triton Kernel Compiler for RISC-V
2026-06-09 , Poster Island C

RISC-V AI acceleration faces a combinatorial explosion: hundreds of kernel variations across shapes, data types, and vendor platforms create unsustainable complexity. We present an end-to-end compilation solution fusing ML graph compilation with Triton DSL kernel compilation in a unified MLIR-based framework targeting RISC-V scalar (RV64IM) and vector (RVV) instruction sets.

Co-founder & CTO of Terapines Technology. More than 15 years compiler design and development experience in Andes, S3 Graphics, Imagination and Terapines. Specialized in CPU, GPU, GPGPU, AI compilers based on MLIR, LLVM and GCC.