Metraipolis › archive › dc_527116f4e007
A Thread-Register Decoupled GPU Execution Model for Efficient Tensor Computation — arXiv:2608.19628
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.19628
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T11:44:00.215Z
- SHA-256 of the archived text
0ac2659a1244b44aa7e2f411ea505b7deacb3e40ed601b4a165f16847b615033- Stable URL of this page
https://metraipolis.com/documento/dc_527116f4e007
How to cite
A Thread-Register Decoupled GPU Execution Model for Efficient Tensor Computation — arXiv:2608.19628 — arXiv (en), https://arxiv.org/abs/2608.19628 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_527116f4e007 (Metraipolis archive dc_527116f4e007, sha256 0ac2659a1244b44aa7e2f411ea505b7deacb3e40ed601b4a165f16847b615033)
BibTeX
@misc{dc527116f4e007,
title = {A Thread-Register Decoupled GPU Execution Model for Efficient Tensor Computation — arXiv:2608.19628},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.19628},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_527116f4e007; sha256 0ac2659a1244b44aa7e2f411ea505b7deacb3e40ed601b4a165f16847b615033},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.19628. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Modern GPUs increasingly integrate Tensor Cores into the execution pipeline. Although aggregate tensor throughput continues to grow, aided by an operand supply that has evolved from register-based in Ampere to redundancy-free, memory-based in Hopper and Blackwell, efficiently orchestrating the complete tensor compute pipeline for the modern AI workloads remains challenging. We identify the fundamental bottlenecks as fixed parallelism and coarse-grained scheduling, both of which are exposed by modern AI workloads that interleave diverse non-GEMM operations with GEMM. To orchestrate tensor computation efficiently, we propose FIBER, a new architecture that extends the GPU SIMT (single instruction, multiple thread) model. Its basic execution instance, the \emph{fiber}, is decoupled from private register ownership, carrying only minimal control state while accessing an SM's registers through a shared view. This enables dynamic parallelism scaling, fine-grained register-level dataflow scheduling, and offers a redundancy-free alternative for matrix operand supply. We extend the ISA, microarchitecture, and compiler to realize shared-register addressing, conflict-free operand delivery, and fiber-based program mapping. Under a typical mixed-precision LLM serving scenario, FIBER achieves a 2.25x end-to-end speedup on Ampere (1.15x for the original FP16 computation), with 1.8x and 2.09x on Hopper and Blackwell respectively, and kernel-level gains up to 2.49x.
Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-08-21 in cs.AR).
Categorie arXiv: Hardware Architecture (cs.AR).
Pagina del lavoro: https://arxiv.org/abs/2608.19628 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-18T11:39:29.374Z dalla pagina https://arxiv.org/catchup/cs.AR/2026-08-21?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).