Accelerating GenAI Workloads by Enabling RISC-V Microkernel Support in IREE

Adeel Ahmad, Ahmad Tameem Kamal, Nouman Amir, Bilal Zafar, Saad Bin Nasir · 2025

Read Accelerating GenAI Workloads by Enabling RISC-V Microkernel Support in IREE by Adeel Ahmad, Ahmad Tameem Kamal, Nouman Amir, Bilal Zafar, Saad Bin Nasir (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-14899.txt.

Description

Adeel Ahmad, Ahmad Tameem Kamal, Nouman Amir, Bilal Zafar, and Saad Bin Nasir speed up AI workloads on RISC-V chips, in a 2025 report. Their target is IREE, a compiler and runtime for machine learning models. RISC-V is an open chip design gaining ground. The team first enabled a key step in the compiler. Matrix work is lowered into a packed form suited to RISC-V targets.

They then wrote tuned small kernels, the inner loops that do most of the math. These kernels exploit the chip's vector features. The team compared speed against the standard IREE build. They also compared against a popular local model runner. The test model was a one-billion-parameter chat model from the Llama family. The work shows open hardware can run modern AI models well. It gives other builders a path to the same gains.

Catalog record

CREATORAdeel Ahmad, Ahmad Tameem Kamal, Nouman Amir, Bilal Zafar, Saad Bin Nasir
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE6 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-14899

Browse the full library · About the Tower of Babel