跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.24458 · v1

A Runtime System for Mitigating Crossbar Port Contention on SM-Partitioned GPUs

首次提交 2026-11-18 22:02
版本更新 2026-11-18 22:02
雷达观测 2026-11-19 15:04
arXiv ↗PDF ↗

摘要

To avoid scheduling delays when running concurrent workloads on a GPU, Streaming Multiprocessor (SM) partitioning mechanisms such as CUDA Green Contexts restrict each workload to separate SMs. Despite this partitioning, concurrent workloads still significantly slow down latency-critical workloads, which recent studies attribute to shared L2 cache and memory contention. However, contention also occurs at the crossbar ports connecting each Graphics Processing Cluster (GPC) of SMs to the L2 cache, as all SMs in a GPC share one port. Green Contexts spread workloads across GPCs, placing concurrent workloads in the same GPCs and forcing latency-critical memory requests to queue behind background requests. Packing workloads into separate GPCs avoids this contention, but forces latency-critical memory requests through fewer ports, limiting memory bandwidth. Consequently, neither spreading nor packing is always faster across workloads and GPUs. Based on these observations, we present GPCPlace, a runtime system that dynamically selects between spreading and packing by profiling workloads on the target GPU to estimate both slowdowns, and hooks kernel launches in the CUDA driver to set SM masks without application modifications. Across 14 GPUs and 112 workload cases, GPCPlace speeds up latency-critical workloads by up to 1.15x on average and 2.38x maximum over Green Contexts without degrading throughput.

作者备注:7 pages, 7 figures, 1 table

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-18 22:02 · 最新观测版本

A Runtime System for Mitigating Crossbar Port Contention on SM-Partitioned GPUs

To avoid scheduling delays when running concurrent workloads on a GPU, Streaming Multiprocessor (SM) partitioning mechanisms such as CUDA Green Contexts restrict each workload to separate SMs. Despite this partitioning, concurrent workloads still significantly slow down latency-critical workloads, which recent studies attribute to shared L2 cache and memory contention. However, contention also occurs at the crossbar ports connecting each Graphics Processing Cluster (GPC) of SMs to the L2 cache, as all SMs in a GPC share one port. Green Contexts spread workloads across GPCs, placing concurrent workloads in the same GPCs and forcing latency-critical memory requests to queue behind background requests. Packing workloads into separate GPCs avoids this contention, but forces latency-critical memory requests through fewer ports, limiting memory bandwidth. Consequently, neither spreading nor packing is always faster across workloads and GPUs. Based on these observations, we present GPCPlace, a runtime system that dynamically selects between spreading and packing by profiling workloads on the target GPU to estimate both slowdowns, and hooks kernel launches in the CUDA driver to set SM masks without application modifications. Across 14 GPUs and 112 workload cases, GPCPlace speeds up latency-critical workloads by up to 1.15x on average and 2.38x maximum over Green Contexts without degrading throughput.

查看此版本解读 →查看此版本原文 ↗