Global Optimizations & Lightweight Dynamic Logic for Concurrency

TL;DR

GOLDYLOC enhances GPU performance for concurrent GEMMs by globally optimizing kernel selection and employing lightweight dynamic logic to adapt to runtime conditions, achieving significant speedups.

Contribution

It introduces GOLDYLOC, a novel approach that considers global resources and dynamic execution environments for optimizing concurrent GEMMs on GPUs.

Findings

Up to 2× performance improvement for concurrent GEMMs.

Up to 2.5× speedup over sequential execution.

Significant reduction in resource contention and slowdown.

Abstract

Modern accelerators like GPUs are increasingly executing independent operations concurrently to improve the device's compute utilization. However, effectively harnessing it on GPUs for important primitives such as general matrix multiplications (GEMMs) remains challenging. Although modern GPUs have significant hardware and software support for GEMMs, their kernel implementations and optimizations typically assume each kernel executes in isolation and can utilize all GPU resources. This approach is highly efficient when kernels execute in isolation, but causes significant resource contention and slowdowns when kernels execute concurrently. Moreover, current approaches often only statically expose and control parallelism within an application, without considering runtime information such as varying input size and concurrent applications -- often exacerbating contention. These issues limit performance benefits from concurrently executing independent operations. Accordingly, we propose GOLDYLOC, which considers the global resources across all concurrent operations to identify performant GEMM kernels, which we call globally optimized (GO)-Kernels. Moreover, GOLDYLOC introduces a lightweight dynamic logic which considers the dynamic execution environment for available parallelism and input sizes to execute performant combinations of concurrent GEMMs on the GPU. Overall, GOLDYLOC improves performance of concurrent GEMMs on a real GPU by up to 2$\times$ (18% geomean per workload) and provides up to 2.5$\times$ (43% geomean per workload) speedups over sequential execution.