arXiv subjects

Manman Ren

Publications and source records attributed to Manman Ren.

3 recordsLinked to original sources

arXiv2026

TLX: Hardware-Native, Evolvable MIMW GPU Compiler for Large-scale Production Environments

Modern GPUs increasingly rely on specialized hardware units and asynchronous coordination mechanisms, so performance depends on orchestrating data movement, tensor-core computation, and synchronization rather than exposing more thread-level parallelism. This creates a programming-model tension: if too much execution structure is hidden, the compiler must catch up to new hardware mechanisms; if too much is exposed, the burden of orchestration falls back onto the programmer. We present TLX (Triton Low-level Language Extensions), built around MIMW (Multi-Instruction, Multi-Warp), which expresses orchestration at warp-group granularity while preserving Triton's productive blocked programming model for regular computation. TLX realizes this idea as an embedded extension to Triton, exposing explicit interfaces for multi-warp execution, local-memory orchestration, asynchronous operations, and cluster-aware control. Our evaluation shows that TLX supports substantial customization with limited development effort while remaining competitive with state-of-the-art implementations. TLX-authored kernels have been deployed in large-scale training and inference production systems. Our code is open sourced at https://github.com/facebookexperimental/triton.

cs.AR↗

arXiv2025

KPerfIR: Towards an Open and Compiler-centric Ecosystem for GPU Kernel Performance Tooling on Modern AI Workloads

In this work, we propose KPerfIR, a novel multilevel compiler-centric infrastructure to enable the development of customizable, extendable, and portable profiling tools tailored for modern artificial intelligence (AI) workloads on modern GPUs. Our approach integrates profiling capabilities directly into the compiler workflow, allowing profiling functionalities to be implemented as compiler passes, offering a programmable and reusable framework for performance analysis. This design bridges the gap between compilers and profilers, enabling fine-grained insights into complex optimization challenges such as overlapping the execution of fine-grained function units on GPUs. KPerfIR is integrated into the Triton infrastructure to highlight the power of a compiler-centric approach to advance performance analysis and optimization in the ever-evolving landscape of AI compilers. Our evaluation shows that our tool incurs low overhead (8.2%), provides accurate measurements (2% relative error), and delivers actionable insights into complicated GPU intra-kernel optimizations.

cs.DC↗

arXiv2023

Optimistic Global Function Merger

Function merging is a pivotal technique for reducing code size by combining identical or similar functions into a single function. While prior research has extensively explored this technique, it has not been assessed in conjunction with function outlining and linker's identical code folding, despite substantial common ground. The traditional approaches necessitate the complete intermediate representation to compare functions. Consequently, none of these approaches offer a scalable solution compatible with separate compilations while achieving global function merging, which is critical for large app development. In this paper, we introduce our global function merger, leveraging global merge information from previous code generation runs to optimistically create merging instances within each module context independently. Notably, our approach remains sound even when intermediate representations change, making it well-suited for distributed build environments. We present a comprehensive code generation framework that can run both the state-of-the-art global function outliner and our global function merger. These components complement each other, resulting in a positive impact on code size reduction. Our evaluation demonstrates that when integrating the global function merger with a state-of-the-art global function outliner that is fully optimized with ThinLTO, a further reduction of up to 3.5% in code size can be attained. This is in addition to the initial average reduction of 17.3% achieved through global function outlining for real-world iOS apps, all with minimal extra build time.

cs.PL↗