MISIM: A Neural Code Semantics Similarity System Using the Context-Aware Semantics Structure

Autor: Ye, Fangke, Zhou, Shengtian, Venkat, Anand, Marcus, Ryan, Tatbul, Nesime, Tithi, Jesmin Jahan, Hasabnis, Niranjan, Petersen, Paul, Mattson, Timothy, Kraska, Tim, Dubey, Pradeep, Sarkar, Vivek, Gottschlich, Justin

Code semantics similarity can be used for many tasks such as code recommendation, automated software defect correction, and clone detection. Yet, the accuracy of such systems has not yet reached a level of general purpose reliability. To help address

Externí odkaz: http://arxiv.org/abs/2006.05265

High-Performance Deep Learning via a Single Building Block

Autor: Georganas, Evangelos, Banerjee, Kunal, Kalamkar, Dhiraj, Avancha, Sasikanth, Venkat, Anand, Anderson, Michael, Henry, Greg, Pabst, Hans, Heinecke, Alexander

Deep learning (DL) is one of the most prominent branches of machine learning. Due to the immense computational cost of DL workloads, industry and academia have developed DL libraries with highly-specialized kernels for each workload/architecture, lea

Externí odkaz: http://arxiv.org/abs/1906.06440

Sparse Matrix Code Dependence Analysis Simplification at Compile Time

Autor: Mohammadi, Mahdi Soltan, Cheshmi, Kazem, Gopalakrishnan, Ganesh, Hall, Mary, Dehnavi, Maryam Mehri, Venkat, Anand, Yuki, Tomofumi, Strout, Michelle Mills

Analyzing array-based computations to determine data dependences is useful for many applications including automatic parallelization, race detection, computation and communication overlap, verification, and shape analysis. For sparse matrix codes, ar

Externí odkaz: http://arxiv.org/abs/1807.10852

An integrated compiler and runtime framework for sparse matrix codes

Autor: Venkat, Anand

Sparse matrix codes are found in numerous applications ranging from iterative numerical solvers to graph analytics. Achieving high performance on these codes has however been a significant challenge, mainly due to array access indirection, for exampl

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_________::7d66d8968e529f6a430ecb1f85975d1d

Optimizing LOBPCG: Sparse Matrix Loop and Data Transformations in Action.

Autor: Ahmad, Khalid, Venkat, Anand, Hall, Mary

Publikováno v: Languages & Compilers for Parallel Computing: 29th International Workshop, LCPC 2016, Rochester, NY, USA, September 28-30, 2016, Revised Papers; 2017, p218-232, 15p