PPT-FAT-GPU: Formal Analysis Techniques for GPU Kernels

Alastair Donaldson Imperial College London wwwdocicacukafd afdimperialacuk Tutorial at HiPEAC 2013 Berlin Supported by the FP7 project CARP Correct and Efficient Accelerator Programming. There are two types of trans fat Naturally occurring trans fats are caused by the way that some animals such a s cattle Artificial synthetic industrial or manufactured trans fats are caused by the way some fats and oils are processed Why is trans fa g BinetCauchy kernels However such approaches are only applicable to time series data living in a Euclidean space eg joint trajectories extracted from motion capture data or feature point trajectories extracted from video Much of the success of rec Ankit Sethia* Scott . Mahlke. University of Michigan. Graphics. Simulation . Linear Algebra. Data Analytics. Machine Learning. Computer Vision. Resource Requirements of GPU applications are diverging. Popcorn project. Problem. What brand of microwave popcorn do you get your moneys worth?. Variables. Same size bags of popcorn were popped, Each bag was popped 2 minutes 10 seconds. All bags were butter lovers. Performance Characterization of. Irregular . GPU Kernels. Molly A. O’Neil and Martin Burtscher. Department of Computer Science. Introduction. GPUs as general-purpose accelerators. Ubiquitous. in high performance computing. with Multiple Labels. Lei Tang. , . Jianhui. Chen and . Jieping. Ye. Kernel-based Methods. Kernel-based methods . Support Vector Machine (SVM). Kernel Linear Discriminate Analysis (KLDA). Demonstrate success in various domains. OpenACC. Directives. subroutine . saxpy. (n, a, x, y) . real :: x(:), y(:), a. integer :: n, . i. $!. acc. kernels. do . i. =1,n.   y(. i. ) = a*x(. i. )+y(. i. ). . enddo. $!. acc. end kernels. April 2016. Agenda. Overview. Kernel architecture. Producing kernels. Using kernels. Introduction to Kernels. 2. Introduction to Kernels. 3. What is a SPICE “Kernel”. “Kernel” means file. “Kernel” means a file . Ross McIlroy, Chris Hawblitzel, Galen Hunt. Microsoft Research. Helios: Heterogeneous Multiprocessing with Satellite Kernels. 1. Problem: HW now heterogeneous. Heterogeneity . ignored. by operating systems. Perceptrons. The . perceptron. A. B. instance. . x. i. Compute: . y. i. = . sign(. v. k. . . . x. i. . ). ^. y. i. ^. y. i. If mistake: . v. k+1. = . v. k. + . y. i. . x. i. . x . is a vector. Machine Learning. March 25, 2010. Last Time. Basics of the Support Vector Machines. Review: Max . Margin. How can we pick which is best?. Maximize the size of the margin.. 3. Are these really . “equally valid”?. Janghaeng Lee. , . Mehrzad. . Samadi. , and Scott . Mahlke. October, 2015. University . of Michigan - Ann . Arbor. University of Michigan. Electrical Engineering . and . Computer Science. Financial Modeling. CS 179: GPU Programming Lecture 7 Week 3 Goals: Advanced GPU- accelerable algorithms CUDA libraries and tools This Lecture GPU- accelerable algorithms: Reduction Prefix sum Stream compaction Sorting (quicksort) 1 Sodium 12%12% 5% Calories 150 Calories from Fat 70 % Daily Value*Total Fat Saturated Fat 0gTrans Fat 0gCholesterol Sodium 0%2% 5% Calories 90 Calories from Fat 0 Nutrition Fac