Vladimir Pentkovski

Implementing streaming SIMD extensions on the Pentium III processor

This paper describes the streaming SIMD extensions (SSE) provides a rich set of instructions to meet the requirements of demanding multimedia and Internet applications. In implementing the SSE, the Pentium III developers made a number of design trade-offs to satisfy tight die size constraints and attain frequency goals.

ZR: A 3Dd API transparent technology for chunk rendering

This paper presents ZR (Zone Rendering), a 3D graphics technology that addresses ever-increasing bandwidth requirements using chunk rendering technique, and at the same time solves 3D API compatibility issues commonly associated with chunk rendering graphics devices. We apply a pipeline serialization technique to handle the cases causing compatibility issues. However, excessive frequency of serializations may offset the performance advantage of ZR. In order to manage potential performance problems we developed software and hardware techniques to optimize ZR performance for most events that might cause the serialization. Comprehensive validation experiments were conducted for popular 3D applications to show that after the optimizations the residual impact of the serialization is very small. Finally, using the results of Intel® 830 graphics, which implements ZR, we demonstrate that ZR provides significant 3D graphics performance improvement. This result is achieved within a limited bandwidth budget, and at the cost of modest micro-architectural changes to traditional graphics pipeline.

Simulation and Performance Study of Large Scale Computer Cluster Configuration: Combined Multi-level Approach

Abstract In this paper we describe an approach used to study the functional aspects and estimate the performance of large (read “ten times more cores than currently available”) cluster computing system configurations running MPI applications. A combination of a functional simulation, performance models, traces analysis and real-world measurements is used to achieve these goals. We also present the first results of applying this methodology to actual applications of interest.