D5.5.3 – Design and implementation of the SIMD-MIMD GPU architecture

J. Lucas, S. Lal, M. Alvarez-Mesa, A. Alhossini, B. Juurlink


   title={D5. 5.3–Design and implementation of the SIMD-MIMD GPU architecture},

   author={TUB, Jan Lucas and TUB, Sohan Lal and TUB, Mauricio Alvarez-Mesa and TUB, Ahmed Alhossini and TUB, Ben Juurlink and Keir, Paul and Stamoulis, Iakovos and Keramidas, Reviewers Georgios}


Download Download (PDF)   View View   Source Source   



To develop a new SIMD-MIMD architecture we first characterized GPGPU workloads using simple and well known workload metrics to identify the performance bottlenecks. We found that the benchmarks with branch divergence do not utilize the SIMD width optimally on conventional GPUs. We also studied the performance bottlenecks of motion compensation kernel developed in Task 3.2 and showed that increasing the maximum limit on CTA and shared memory can signi cantly increase in performance and save energy. We also studied the correlation between workload characteristics and GPU component power consumption. In addition we categorize the workload into high, medium, and low IPC category to study the power consumption behavior of each category. The results show a significant change in components power consumption across the three categories of kernels. We believe this is a vital information for computer architects and application programmers to prioritize the components for power and performance optimizations. Guided by this information we proposed a new architecture which can handle branch divergence eciently.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2021 hgpu.org

All rights belong to the respective authors

Contact us: