Affiliation:
1. University of Illinois at Urbana-Champaign, Urbana, IL, USA
Abstract
With the emergence of highly multithreaded architectures, performance monitoring techniques face new challenges in efficiently locating sources of performance discrepancies in the program source code. For example, the state-of-the-art performance counters in highly multithreaded graphics processing units (GPUs) report only the overall occurrences of microarchitecture events at the end of program execution. Furthermore, even if supported, any fine-grained sampling of performance counters will distort the actual program behavior and will make the sampled values inaccurate. On the other hand, it is difficult to achieve high resolution performance information at low sampling rates in the presence of thousands of concurrently running threads. In this paper, we present a novel software-based approach for monitoring the memory hierarchy performance in highly multithreaded general-purpose graphics processors. The proposed analysis is based on memory traces collected for snapshots of an application execution. A trace-based memory hierarchy model with a Monte Carlo experimental methodology generates statistical bounds of performance measures without being concerned about the exact inter-thread ordering of individual events but rather studying the behavior of the overall system. The statistical approach overcomes the classical problem of disturbed execution timing due to fine-grained instrumentation. The approach scales well as we deploy an efficient parallel trace collection technique to reduce the trace generation overhead and a simple memory hierarchy model to reduce the simulation time. The proposed scheme also keeps track of individual memory operations in the source code and can quantify their efficiency with respect to the memory system. A cross-validation of our results shows close agreement with the values read from the hardware performance counters on an NVIDIA Tesla C2050 GPU. Based on the high resolution profile data produced by our model we optimized memory accesses in the sparse matrix vector multiply kernel and achieved speedups ranging from 2.4 to 14.8 depending on the characteristics of the input matrices.
Publisher
Association for Computing Machinery (ACM)
Subject
Computer Graphics and Computer-Aided Design,Software
Reference22 articles.
1. http://clang.llvm.org/. http://clang.llvm.org/.
2. The OpenCL Specification 2009. The OpenCL Specification 2009.
3. An adaptive performance modeling tool for GPU architectures
4. Analyzing CUDA workloads using a detailed GPU simulator
5. G. Diamos A. Kerr and M. Kesavan. A dynamic compilation framework for ptx. http://code.google.com/p/gpuocelot. G. Diamos A. Kerr and M. Kesavan. A dynamic compilation framework for ptx. http://code.google.com/p/gpuocelot.
Cited by
23 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. DELTA: Validate GPU Memory Profiling with Microbenchmarks;The International Symposium on Memory Systems;2020-09-28
2. A Study of Single and Multi-device Synchronization Methods in Nvidia GPUs;2020 IEEE International Parallel and Distributed Processing Symposium (IPDPS);2020-05
3. Understanding the Performance of GPGPU Applications from a Data-Centric View;2019 IEEE/ACM International Workshop on Programming and Performance Visualization Tools (ProTools);2019-11
4. GPUs Cache Performance Estimation using Reuse Distance Analysis;2019 IEEE 38th International Performance Computing and Communications Conference (IPCCC);2019-10
5. A Performance Model for GPU Architectures that Considers On-Chip Resources: Application to Medical Image Registration;IEEE Transactions on Parallel and Distributed Systems;2019-09-01