Filters








102 Hits in 2.9 sec

Scaling up Superoptimization

Phitchaya Mangpo Phothilimthana, Aditya Thakur, Rastislav Bodik, Dinakar Dhurjati
2016 SIGARCH Computer Architecture News  
We develop LENS, a search algorithm that increases the size of code a superoptimizer can synthesize by rapidly pruning away invalid candidate programs.  ...  If we can increase the size of code fragments a superoptimizer can optimize, we will be able to discover more optimizations.  ...  In summary, we introduced strategies to scale up superoptimization to optimize real-world programs.  ... 
doi:10.1145/2980024.2872387 fatcat:jjkzuerpojf47dxpwdtp5jqjzi

Scaling up Superoptimization

Phitchaya Mangpo Phothilimthana, Aditya Thakur, Rastislav Bodik, Dinakar Dhurjati
2016 Proceedings of the Twenty-First International Conference on Architectural Support for Programming Languages and Operating Systems - ASPLOS '16  
We develop LENS, a search algorithm that increases the size of code a superoptimizer can synthesize by rapidly pruning away invalid candidate programs.  ...  If we can increase the size of code fragments a superoptimizer can optimize, we will be able to discover more optimizations.  ...  In summary, we introduced strategies to scale up superoptimization to optimize real-world programs.  ... 
doi:10.1145/2872362.2872387 dblp:conf/asplos/PhothilimthanaT16 fatcat:z67r6qsje5cijggqxonmwtpkqu

Scaling up Superoptimization

Phitchaya Mangpo Phothilimthana, Aditya Thakur, Rastislav Bodik, Dinakar Dhurjati
2016 ACM SIGOPS Operating Systems Review  
We develop LENS, a search algorithm that increases the size of code a superoptimizer can synthesize by rapidly pruning away invalid candidate programs.  ...  If we can increase the size of code fragments a superoptimizer can optimize, we will be able to discover more optimizations.  ...  In summary, we introduced strategies to scale up superoptimization to optimize real-world programs.  ... 
doi:10.1145/2954680.2872387 fatcat:dyudxbtnm5hghfeg4pwngld2va

Scaling up Superoptimization

Phitchaya Mangpo Phothilimthana, Aditya Thakur, Rastislav Bodik, Dinakar Dhurjati
2016 SIGPLAN notices  
We develop LENS, a search algorithm that increases the size of code a superoptimizer can synthesize by rapidly pruning away invalid candidate programs.  ...  If we can increase the size of code fragments a superoptimizer can optimize, we will be able to discover more optimizations.  ...  In summary, we introduced strategies to scale up superoptimization to optimize real-world programs.  ... 
doi:10.1145/2954679.2872387 fatcat:lhu2j7hcjjdqxezwrnz2gfog74

Future Directions for Optimizing Compilers [article]

Nuno P. Lopes, John Regehr
2018 arXiv   pre-print
As software becomes larger, programming languages become higher-level, and processors continue to fail to be clocked faster, we'll increasingly require compilers to reduce code bloat, eliminate abstraction  ...  Acknowledgments The authors would like to thank Dan Gohman and members of the Utah PL reading group for their valuable feedback on drafts of this paper.  ...  The process of formalization is itself (ironically) error-prone and, worse, it often exposes latent ambiguities in real-world systems that can be hard to resolve definitively.  ... 
arXiv:1809.02161v1 fatcat:hwn23mp7rza4dfgnygdcmu64ri

Stochastic superoptimization

Eric Schkufza, Rahul Sharma, Alex Aiken
2013 SIGPLAN notices  
Although our method sacrifices completeness, the scope of programs we are able to consider, and the resulting quality of the programs that we produce, far exceed those of existing superoptimizers.  ...  of all possible programs to find one that is an optimization of a given target program.  ...  Acknowledgments The authors would like to thank Peter Johnston, Juan Manuel Tamayo, and Kushal Tayal for their assistance in the implementation of STOKE, Ankur Taly and David Ramos for their advice regarding  ... 
doi:10.1145/2499368.2451150 fatcat:l33axgntmrb5plmfkeiyjpjgse

Stochastic superoptimization

Eric Schkufza, Rahul Sharma, Alex Aiken
2013 Proceedings of the eighteenth international conference on Architectural support for programming languages and operating systems - ASPLOS '13  
Although our method sacrifices completeness, the scope of programs we are able to consider, and the resulting quality of the programs that we produce, far exceed those of existing superoptimizers.  ...  of all possible programs to find one that is an optimization of a given target program.  ...  Acknowledgments The authors would like to thank Peter Johnston, Juan Manuel Tamayo, and Kushal Tayal for their assistance in the implementation of STOKE, Ankur Taly and David Ramos for their advice regarding  ... 
doi:10.1145/2451116.2451150 dblp:conf/asplos/Schkufza0A13 fatcat:ua35dzg7vvbgbnuzum3yq7k7sq

Stochastic Superoptimization [article]

Eric Schkufza, Rahul Sharma, Alex Aiken
2012 arXiv   pre-print
Although our method sacrifices com- pleteness, the scope of programs we are able to reason about, and the quality of the programs we produce, far exceed those of existing superoptimizers.  ...  of all possible programs to find one that is an optimization of a given target program.  ...  The authors would like to thank Peter Johnston, Juan Manuel Tamayo, and Kushal Tayal for their assistance in the implementation of STOKE, Ankur Taly for his help with semantics of X86 opcodes, David Ramos  ... 
arXiv:1211.0557v1 fatcat:7t46jz7aqrflfhjvb6kuavis3e

Conditionally correct superoptimization

Rahul Sharma, Eric Schkufza, Berkeley Churchill, Alex Aiken
2015 Proceedings of the 2015 ACM SIGPLAN International Conference on Object-Oriented Programming, Systems, Languages, and Applications - OOPSLA 2015  
However, both general purpose compilers and highly specialized tools such as superoptimizers often do not have sufficient static knowledge of restrictions on program inputs that could be exploited to produce  ...  This approach yields a superoptimizer for x86 that in our experiments produces binaries that are often multiple times faster than those produced by production compilers.  ...  Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon.  ... 
doi:10.1145/2814270.2814278 dblp:conf/oopsla/0001SCA15 fatcat:neztmd3kmre7jdlzqoxf6hc3fy

Conditionally correct superoptimization

Rahul Sharma, Eric Schkufza, Berkeley Churchill, Alex Aiken
2015 SIGPLAN notices  
However, both general purpose compilers and highly specialized tools such as superoptimizers often do not have sufficient static knowledge of restrictions on program inputs that could be exploited to produce  ...  This approach yields a superoptimizer for x86 that in our experiments produces binaries that are often multiple times faster than those produced by production compilers.  ...  Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon.  ... 
doi:10.1145/2858965.2814278 fatcat:sj66k7gds5cghemf767zxkbram

Adversarial Binaries for Authorship Identification [article]

Xiaozhu Meng, Barton P. Miller, Somesh Jha
2018 arXiv   pre-print
Binary code authorship identification determines authors of a binary program. Existing techniques have used supervised machine learning for this task.  ...  We develop two key attack capabilities: feature vector modification, generating an adversarial feature vector that both corresponds to a real binary and causes the required misprediction, and input binary  ...  Our attack experiences show that it is not secure to rely on features derived from program properties that are easy to modify, such as function calls, symbols, data, and instructions.  ... 
arXiv:1809.08316v2 fatcat:lc67mss26fdtnhws5zk2xyavga

GEVO: GPU Code Optimization using Evolutionary Computation [article]

Jhe-Yu Liou, Xiaodong Wang, Stephanie Forrest, Carole-Jean Wu
2020 arXiv   pre-print
We demonstrate that GEVO improves the execution time of the GPU programs in the Rodinia benchmark suite and the machine learning models, SVM and ResNet18, on NVIDIA Tesla P100.  ...  As the programming stack and tool support have matured, GPUs have also become accessible to programmers, who may lack detailed knowledge of the underlying architecture and fail to fully leverage the GPU's  ...  In some cases, the goal is to produce programs that run faster, which is known as superoptimization.  ... 
arXiv:2004.08140v2 fatcat:rtn27juoondpjkg6b3ddtgzr34

Data-Driven Synthesis of Full Probabilistic Programs [chapter]

Sarah Chasins, Phitchaya Mangpo Phothilimthana
2017 Lecture Notes in Computer Science  
We introduce a data-guided approach to the program mutation stage of simulated annealing; this innovation allows our tool to scale to synthesizing complete probabilistic programs from scratch.  ...  Our synthesizer leverages the input data to generate a program sketch, then applies simulated annealing to complete the sketch.  ...  Department of Energy, Office of Science, Office of Basic Energy Sciences Energy Frontier Research Centers program under Award Number FOA-0000619, and grants from DARPA FA8750-14-C-0011 and DARPA FA8750  ... 
doi:10.1007/978-3-319-63387-9_14 fatcat:idnaf7svwrc2thcmndstn6gjxm

Ithemal: Accurate, Portable and Fast Basic Block Throughput Estimation using Deep Neural Networks [article]

Charith Mendis, Alex Renda, Saman Amarasinghe, Michael Carbin
2019 arXiv   pre-print
In this paper we present Ithemal, the first tool which learns to predict the throughput of a set of instructions.  ...  Ithemal uses a hierarchical LSTM--based approach to predict throughput based on the opcodes and operands of instructions in a basic block.  ...  We designed the dataset to include real world benchmarking programs as well as programs that exercise a wide variety of x86-64 instructions, such as different vector instructions sets.  ... 
arXiv:1808.07412v2 fatcat:rpebyqwf5jdf5nqvzfypmkotzi

The perception of a face can be greater than the sum of its parts

Jianhong Shen, Thomas J. Palmeri
2014 Psychonomic Bulletin & Review  
We tested whether the facial-feature integration index would also equal 1 when faces have a range of configurations mirroring the range of variability in real-world faces, using the same experimental procedure  ...  Holistic processing is often used as a construct to characterize face recognition.  ...  We thank Jason Gold and his colleagues for providing the programs and supporting materials used in their original study as well as for several discussions regarding our findings.  ... 
doi:10.3758/s13423-014-0726-y pmid:25228169 fatcat:6hgmcufk4rfkxpfbjrajg27pdy
« Previous Showing results 1 — 15 out of 102 results