S
Scott Mahlke
Researcher at University of Michigan
Publications - 294
Citations - 15433
Scott Mahlke is an academic researcher from University of Michigan. The author has contributed to research in topics: Compiler & Scheduling (computing). The author has an hindex of 69, co-authored 289 publications receiving 14764 citations. Previous affiliations of Scott Mahlke include Samsung & Hewlett-Packard.
Papers
More filters
Journal ArticleDOI
The superblock: an effective technique for VLIW and superscalar compilation
Wen-mei W. Hwu,Scott Mahlke,William Y. Chen,Pohua P. Chang,Nancy J. Warter,Roger A. Bringmann,Roland G. Ouellette,Richard E. Hank,Tokuzo Kiyohara,Grant Haab,John G. Holm,Daniel M. Lavery +11 more
TL;DR: Superblocks as discussed by the authors enable the optimizer and scheduler to extract more ILP along the important execution paths by systematically removing constraints due to the unimportant paths, which is useful for control-intensive programs.
Journal ArticleDOI
Effective compiler support for predicated execution using the hyperblock
TL;DR: In this paper, a new structure, referred to as the hyperblock, is proposed to combine speculative execution with predicated execution for both compile-time optimization and scheduling of conditional branches.
Proceedings ArticleDOI
COMET: code offload by migrating execution transparently
TL;DR: The prototype of COMET (Code Offload by Migrating Execution Transparently), a realization of this design built on top of the Dalvik Virtual Machine, leverages the underlying memory model of the runtime to implement distributed shared memory (DSM) with as few interactions between machines as possible.
Proceedings ArticleDOI
Scalpel: Customizing DNN Pruning to the Underlying Hardware Parallelism
TL;DR: This work implemented weight pruning for several popular networks on a variety of hardware platforms and observed surprising results, including mean speedups of 3.54x, 2.61x, and 1.25x while reducing the model sizes by 88, 82%, and 53%.
Proceedings ArticleDOI
IMPACT: an architectural framework for multiple-instruction-issue processors
TL;DR: The optimization capabilities of the IMPACT-I C compiler are summarized in this paper and experiments to analyze the performance of multiple-instruction-issue processors executing some important non-numerical programs are ran.