EFFICIENT BREADTH FIRST SEARCH ON MULTI-GPU SYSTEMS USING GPU-CENTRIC OPENSHMEM
|
|
- Walter McDonald
- 5 years ago
- Views:
Transcription
1 EFFICIENT BREADTH FIRST SEARCH ON MULTI-GPU SYSTEMS USING GPU-CENTRIC OPENSHMEM Sreeram Potluri, Anshuman Goswami NVIDIA Manjunath Gorentla Venkata, Neena Imam - ORNL
2 SCOPE OF THE WORK Reliance on CPU for communication limits scaling on GPU clusters NVSHMEM provides GPU kernel-side OpenSHMEM API NVLink is a high-bandwidth memory-like fabric between GPUs Effectiveness of NVSHMEM over NVLink with Breadth First Search CPU-initiated MPI vs GPU-initiated SHMEM in a particular implementation of BFS 2
3 GPU Programming Models: MPI and SHMEM NVSHMEM Overview AGENDA DGX and NVLink MPI+CUDA implementation of BFS SHMEM implementation of BFS Performance Evaluation Conclusion and Future Work 3
4 GPU CLUSTER PROGRAMMING Offload model Compute on GPU Communication from CPU Synchronization at boundaries Overheads on GPU Clusters Offload latencies Synchronization overheads Limits strong scaling More CPU means more power void 2dstencil (u, v, ) { for (timestep = 0; ) { interior_compute_kernel <<< >>> ( ) pack_kernel <<< >>> ( ) cudastreamsynchronize( ) MPI_Irecv( ) MPI_Isend( ) MPI_Waitall( ) unpack_kernel <<< >>> ( ) boundary_compute_kernel <<< >>> ( ) } } 4
5 GPU-CENTRIC COMMUNICATION GPU capabilities Warp scheduling to hide latencies to global memory Implicit coalescing of loads/stores to achieve efficiency CUDA helps to program to these Should also benefit when accessing data over the network Direct accesses to remote memory simplifies programming Achieving efficiency while making it easier to program Continuous fine-grained accesses smooths traffic over the network 5
6 Symmetric Heap GPU GLOBAL ADDRESS SPACE 6
7 COMMUNICATION FROM CUDA KERNELS Long running CUDA kernels Communication within parallelism void 2dstencil (u, v, ) { stencil_kernel <<< >>> ( ) } global void 2dstencil (u, v, sync, ) { for(timestep = 0; ) { if (i+1 > nx) { v[i+1] = shmem_float_g (v[1], rightpe); } if (i-1 < 1) { v[i-1] = shmem_float_g (v[nx], leftpe); } } } u[i] = (u[i] + (v[i+1] + v[i-1]... if (i < 2) { shmem_int_p (sync + i, 1, peers[i]); shmem_quiet(); shmem_wait_until (sync + i, EQ, 1); } //intra-kernel sync 7
8 NVSHMEM Implementation of OpenSHMEM Specification (1.3) Symmetric heap on GPU memory Implements multi-threading support as being discussed in PR #43 HOST ONLY Library setup, exit and query Memory management HOST/GPU shmem_ptr Remote memory access Atomic memory operations Collectives Point-to-point synchronization Memory ordering Locking routines Extensions for GPUs CUDA stream-based SHMEM communication Thread cooperative SHMEM communication Expected availability: Q (single OS Multi-GPU nodes: NVLink, PCIe, QPI) 8
9 DGX-1 ARCHITECTURE CPU CPU IB 100 GB/sec PLX switch PLX switch PLX switch PLX switch IB 100 GB/sec IB 100 GB/sec IB 100 GB/sec P100 P100 P100 P100 P100 P100 P100 P100 9
10 CPU Inter-GPU Copies Single Process Unified Virtual Addressing available since CUDA 4.0 (2011) cudasetdevice(0); cudamalloc(&buf0, size) cudasetdevice(1); cudamalloc(&buf1, size) cudamemcpy (buf0, buf1, size, cudamemcpydefault ) GPU 0 GPU 1 10
11 CPU CPU GPUDirect P2P: by-passing CPU memory cudadevicecanaccesspeer(int *access, int device, int peerdevice); cudaenablepeeraccess(int peerdevice, unsigned int flags) cudasetdevice(0); cudamalloc(&buf0, size); cudacanaccesspeer (&access, 0, 1); assert(access == 1); cudaenablepeeraccess (1, 0); GPU 0 GPU 1 PCIe P2P (copy/ld/st) cudasetdevice(1); cudamalloc(&buf1, size); cudasetdevice (0); cudamemcpy (buf0, buf1, size, cudamemcpydefault); GPU 0 GPU 1 NVLink (copy/ld/st /atomics) Can also access (LD/ST) the peer GPU memory from inside a CUDA kernel 11
12 cudaipcgetmemhandle(), cudaipcopenmemhandle() & cudadeviceenablepeeraccess() Using GPU memory from another process GPU-0 VAs for Process 1 Physical GPU Memory Existing allocation Get handle and send to other process GPU-0 H Opening handle creates mapping in receiving process GPU-1 GPU-1 VAs for Process 2 Available for access by another process AND from a second GPU use peering 12
13 GB/sec GB/sec GB/sec Copy using DMA NVLINK1 W/ PASCAL PCIe 1 NVLink1 4x NVLink x 5.89x 0 PCIe 1 NVLink1 4x NVLink1 PCIe 1 NVLink1 4x NVLink LD/ST - Linear PCIe 1 NVLink1 4x NVLink1 5.43x 1.35x LD/ST - Random Access PCIe 1 NVLink1 4x NVLink1 2.64x PCIe 1 NVLink1 4x NVLink x GPU memory model is relaxed and scoped Same memory model works within/across GPUs 13
14 BREADTH FIRST SEARCH (BFS) Baseline: an MPI-based multi-gpu implementation of BFS by M.Bisson et. al. Has been shown to scale to thousands of GPUs on the Tsubame supercomputer Loosely follows the Graph 500 specification R-MAT generator, mean performance over 64 BFS operations Uses 32-bit vertices instead of required 48-bit representation Uses a parallel level-synchronous algorithm Frontier: list of vertices from which traversal starts in a give level Expand: traverse graph to find un-visited neighbors of frontier vertices Exchange: exchange list of newly found vertices Append: remove redundancy and build the frontier for next level 14
15 DATA PARTITIONING 0 2 N/C Process grid: R rows x C columns 0 1 C-1 P0,0 P0,1 P0,C-1 P1,0 P1,1 P1,C-1 R-1 PR-1,0 PR-1,1 PR-1,C-1 0R + i 1R + i R R+1 P0,0 P1,0 P0,1 P1,1 P0,C-1 P1,C-1 jr + i (C-1)R + i Partition at Process Pi,j RC 2R-1 PR-1,0 PR-1,1 PR-1,C-1 Destination vertex of each edge is in the same process row Adjacency list of a vertex is distributed along process column R(C-1) P0,0 P0,1 P0,C-1 R(C-1)+1 P1,0 P1,1 P1,C-1 RC-1 PR-1,0 PR-1,1 PR-1,C-1 15
16 BFS USING MPI frontier list at one processor column is populated with the seed vertex while (frontier not empty) { Expand: local expansion Horizontal Exchange: Alltoall exchange newly discovered vertices (happens only along grid rows) Append: build new frontier vertex list Vertical exchange: Allgather frontier list along grid columns (happens only along columns) Convergence check: Allreduce frontier list size among all processors } 16
17 EXCHANGE USING MPI Vertex lists are exchanged in two forms: Using a vertex list (one integer per vertex) Using a bitmap (one bit per vertex) Format is statically picked based on the level of traversal bitmap when number of discovered vertices is expected to be large vertex list when number of discovered vertices is expected to be large Communication implemented using non-blocking MPI sends and receives Enhanced the base version to using CUDA-aware MPI internally uses CUDA IPC/P2P to take advantage of direct PCIe or NVLink connections 17
18 GTEPS GTEPS PERFORMANCE WITH CUDA-AWARE MPI baseline cumpi baseline cumpi % % Graph Size (scale) Graph Size (scale) 4 K40 GPUs + PCIe 4 P100 GPUs + NVLink Supermicro server w/ IvyBridge CPU and Broadcom PCIe switch, CUDA 8.0, MVAPICH2 2.2 DGX-1 w/ Broadwell CPU, CUDA 8.0, MVAPICH
19 BFS USING NVSHMEM Fuse communication into CUDA kernels expand <- exchange of newly discovered vertices append <- exchange frontier vertices while (frontier not empty) { Expand on GPU: local expansion Append on GPU: build new frontier vertex list Convergence check on CPU: Allreduce frontier list size among all processors } atomicor(sbuf + r[k]/bits(msk), m[k]); int peer = r[k] / drow_bl; uint32_t off = disp + r[k]%drow_bl; q[k] = shmem_int_or ((int *)(msk + off/bits(msk)), m[k], peer); gets rid of the MPI exchange code on the host 19
20 WORKAROUND FOR PCIE GPU-GPU atomics are not supported over PCIE We take advantage of fact that 32-bit writes are atomic Vertex list is represented as integers, marked with shmem_int_p Higher processing overhead as the vertex map is 32x larger Provides a common version to compare behavior over PCIe and NVLink 20
21 GTEPS GTEPS PERFORMANCE WITH PUTS cumpi SHMEM-Put cumpi SHMEM-Put Graph Size (Scale) Graph Size (Scale) 4 K40 GPUs + PCIe 4 P100 GPUs + NVLink 21
22 GTEPS GTEPS PERFORMANCE WITH ATOMICS cumpi SHMEM-Atomics cumpi SHMEM-Atomics 50 Overall: 12% % 52% 22% Graph Size (scale) Graph Size (scale) 8 P100 GPUs + NVLink (2x4) 8 P100 GPUs + NVLink (4x2) 22
23 FUTURE WORK Fairer comparison by removing tracking of predecessor information Hybrid vertex list/bitmap design like in the case of MPI Can offload multiple levels of BFS using stream-based collectives in NVSHMEM Possibility of kernel fusion, using kernel-side collectives in NVSHMEM Evaluate the use of IB to extend NVSHEM across multiple nodes 23
24 CONCLUSION NVLink provides a high-bandwidth memory-like fabric between GPUS NVSHMEM provides GPU-side communication API based on OpenSHMEM We presented a study of using GPU-initiated communication in BFS Benefits from reduced overheads in strong scaling cases, compared to using MPI Tradeoff with efficient network utilization for weak scaling 24
25 THANKS We thank authors of the baseline BFS code Mauro Bisson, Massimo Bernaschi, Enrico Mastrostefano: Parallel Distributed Breadth First Search on the Kepler Architecture This has been supported in part by Oak Ridge National Lab under subcontract #
26
NVSHMEM: A PARTITIONED GLOBAL ADDRESS SPACE LIBRARY FOR NVIDIA GPU CLUSTERS. Sreeram Potluri, Anshuman Goswami - NVIDIA 3/28/2018
NVSHMEM: A PARTITIONED GLOBAL ADDRESS SPACE LIBRARY FOR NVIDIA GPU CLUSTERS Sreeram Potluri, Anshuman Goswami - NVIDIA 3/28/2018 GPU Programming Models AGENDA Overview of NVSHMEM Porting to NVSHMEM Performance
More informationTECHNOLOGIES FOR IMPROVED SCALING ON GPU CLUSTERS. Jiri Kraus, Davide Rossetti, Sreeram Potluri, June 23 rd 2016
TECHNOLOGIES FOR IMPROVED SCALING ON GPU CLUSTERS Jiri Kraus, Davide Rossetti, Sreeram Potluri, June 23 rd 2016 MULTI GPU PROGRAMMING Node 0 Node 1 Node N-1 MEM MEM MEM MEM MEM MEM MEM MEM MEM MEM MEM
More informationEnabling Efficient Use of UPC and OpenSHMEM PGAS models on GPU Clusters
Enabling Efficient Use of UPC and OpenSHMEM PGAS models on GPU Clusters Presentation at GTC 2014 by Dhabaleswar K. (DK) Panda The Ohio State University E-mail: panda@cse.ohio-state.edu http://www.cse.ohio-state.edu/~panda
More informationPERFORMANCE ANALYSIS AND DEBUGGING FOR VOLTA. Felix Schmitt 11 th Parallel Tools Workshop September 11-12, 2017
PERFORMANCE ANALYSIS AND DEBUGGING FOR VOLTA Felix Schmitt 11 th Parallel Tools Workshop September 11-12, 2017 INTRODUCING TESLA V100 Volta Architecture Improved NVLink & HBM2 Volta MPS Improved SIMT Model
More informationS8688 : INSIDE DGX-2. Glenn Dearth, Vyas Venkataraman Mar 28, 2018
S8688 : INSIDE DGX-2 Glenn Dearth, Vyas Venkataraman Mar 28, 2018 Why was DGX-2 created Agenda DGX-2 internal architecture Software programming model Simple application Results 2 DEEP LEARNING TRENDS Application
More informationExploiting Full Potential of GPU Clusters with InfiniBand using MVAPICH2-GDR
Exploiting Full Potential of GPU Clusters with InfiniBand using MVAPICH2-GDR Presentation at Mellanox Theater () Dhabaleswar K. (DK) Panda - The Ohio State University panda@cse.ohio-state.edu Outline Communication
More informationNVIDIA COLLECTIVE COMMUNICATION LIBRARY (NCCL)
NVIDIA COLLECTIVE COMMUNICATION LIBRARY (NCCL) RN-08645-000_v01 March 2018 Release Notes TABLE OF CONTENTS Chapter Chapter Chapter Chapter Chapter Chapter Chapter 1. 2. 3. 4. 5. 6. 7. NCCL NCCL NCCL NCCL
More informationOptimized Non-contiguous MPI Datatype Communication for GPU Clusters: Design, Implementation and Evaluation with MVAPICH2
Optimized Non-contiguous MPI Datatype Communication for GPU Clusters: Design, Implementation and Evaluation with MVAPICH2 H. Wang, S. Potluri, M. Luo, A. K. Singh, X. Ouyang, S. Sur, D. K. Panda Network-Based
More informationGPU-centric communication for improved efficiency
GPU-centric communication for improved efficiency Benjamin Klenk *, Lena Oden, Holger Fröning * * Heidelberg University, Germany Fraunhofer Institute for Industrial Mathematics, Germany GPCDP Workshop
More informationNVIDIA COLLECTIVE COMMUNICATION LIBRARY (NCCL)
NVIDIA COLLECTIVE COMMUNICATION LIBRARY (NCCL) RN-08645-000_v01 September 2018 Release Notes TABLE OF CONTENTS Chapter Chapter Chapter Chapter Chapter Chapter Chapter Chapter Chapter Chapter 1. NCCL Overview...1
More informationLatest Advances in MVAPICH2 MPI Library for NVIDIA GPU Clusters with InfiniBand
Latest Advances in MVAPICH2 MPI Library for NVIDIA GPU Clusters with InfiniBand Presentation at GTC 2014 by Dhabaleswar K. (DK) Panda The Ohio State University E-mail: panda@cse.ohio-state.edu http://www.cse.ohio-state.edu/~panda
More informationNVGRAPH,FIREHOSE,PAGERANK GPU ACCELERATED ANALYTICS NOV Joe Eaton Ph.D.
NVGRAPH,FIREHOSE,PAGERANK GPU ACCELERATED ANALYTICS NOV 2016 Joe Eaton Ph.D. Agenda Accelerated Computing nvgraph New Features Coming Soon Dynamic Graphs GraphBLAS 2 ACCELERATED COMPUTING 10x Performance
More informationMulti-GPU Programming
Multi-GPU Programming Paulius Micikevicius Developer Technology, NVIDIA 1 Outline Usecases and a taxonomy of scenarios Inter-GPU communication: Single host, multiple GPUs Multiple hosts Case study Multiple
More informationEfficient CPU GPU data transfers CUDA 6.0 Unified Virtual Memory
Institute of Computational Science Efficient CPU GPU data transfers CUDA 6.0 Unified Virtual Memory Juraj Kardoš (University of Lugano) July 9, 2014 Juraj Kardoš Efficient GPU data transfers July 9, 2014
More informationOak Ridge National Laboratory Computing and Computational Sciences
Oak Ridge National Laboratory Computing and Computational Sciences Preparing OpenSHMEM for Exascale Presented by: Pavel Shamis (Pasha) HPC Advisory Council Stanford Conference California Feb 2, 2015 Outline
More informationPortable SHMEMCache: A High-Performance Key-Value Store on OpenSHMEM and MPI
Portable SHMEMCache: A High-Performance Key-Value Store on OpenSHMEM and MPI Huansong Fu*, Manjunath Gorentla Venkata, Neena Imam, Weikuan Yu* *Florida State University Oak Ridge National Laboratory Outline
More informationPortland State University ECE 588/688. Graphics Processors
Portland State University ECE 588/688 Graphics Processors Copyright by Alaa Alameldeen 2018 Why Graphics Processors? Graphics programs have different characteristics from general purpose programs Highly
More informationInterconnection Network for Tightly Coupled Accelerators Architecture
Interconnection Network for Tightly Coupled Accelerators Architecture Toshihiro Hanawa, Yuetsu Kodama, Taisuke Boku, Mitsuhisa Sato Center for Computational Sciences University of Tsukuba, Japan 1 What
More informationGPU Fundamentals Jeff Larkin November 14, 2016
GPU Fundamentals Jeff Larkin , November 4, 206 Who Am I? 2002 B.S. Computer Science Furman University 2005 M.S. Computer Science UT Knoxville 2002 Graduate Teaching Assistant 2005 Graduate
More informationLDetector: A low overhead data race detector for GPU programs
LDetector: A low overhead data race detector for GPU programs 1 PENGCHENG LI CHEN DING XIAOYU HU TOLGA SOYATA UNIVERSITY OF ROCHESTER 1 Data races in GPU Introduction & Contribution Impact correctness
More informationCUDA Advanced Techniques 2 Mohamed Zahran (aka Z)
CSCI-GA.3033-004 Graphics Processing Units (GPUs): Architecture and Programming CUDA Advanced Techniques 2 Mohamed Zahran (aka Z) mzahran@cs.nyu.edu http://www.mzahran.com Alignment Memory Alignment Memory
More informationNVIDIA GPUDirect Technology. NVIDIA Corporation 2011
NVIDIA GPUDirect Technology NVIDIA GPUDirect : Eliminating CPU Overhead Accelerated Communication with Network and Storage Devices Peer-to-Peer Communication Between GPUs Direct access to CUDA memory for
More informationCUDA. Matthew Joyner, Jeremy Williams
CUDA Matthew Joyner, Jeremy Williams Agenda What is CUDA? CUDA GPU Architecture CPU/GPU Communication Coding in CUDA Use cases of CUDA Comparison to OpenCL What is CUDA? What is CUDA? CUDA is a parallel
More informationGPGPU LAB. Case study: Finite-Difference Time- Domain Method on CUDA
GPGPU LAB Case study: Finite-Difference Time- Domain Method on CUDA Ana Balevic IPVS 1 Finite-Difference Time-Domain Method Numerical computation of solutions to partial differential equations Explicit
More informationBreadth First Search on Cost efficient Multi GPU Systems
Breadth First Search on Cost efficient Multi Systems Takuji Mitsuishi Keio University 3 14 1 Hiyoshi, Yokohama, 223 8522, Japan mits@am.ics.keio.ac.jp Masaki Kan NEC Corporation 1753, Shimonumabe, Nakahara
More informationEnabling Efficient Use of UPC and OpenSHMEM PGAS Models on GPU Clusters. Presented at GTC 15
Enabling Efficient Use of UPC and OpenSHMEM PGAS Models on GPU Clusters Presented at Presented by Dhabaleswar K. (DK) Panda The Ohio State University E- mail: panda@cse.ohio- state.edu hcp://www.cse.ohio-
More informationEnterprise. Breadth-First Graph Traversal on GPUs. November 19th, 2015
Enterprise Breadth-First Graph Traversal on GPUs Hang Liu H. Howie Huang November 9th, 5 Graph is Ubiquitous Breadth-First Search (BFS) is Important Wide Range of Applications Single Source Shortest Path
More informationHow to Boost the Performance of Your MPI and PGAS Applications with MVAPICH2 Libraries
How to Boost the Performance of Your MPI and PGAS s with MVAPICH2 Libraries A Tutorial at the MVAPICH User Group (MUG) Meeting 18 by The MVAPICH Team The Ohio State University E-mail: panda@cse.ohio-state.edu
More informationScalable GPU Graph Traversal!
Scalable GPU Graph Traversal Duane Merrill, Michael Garland, and Andrew Grimshaw PPoPP '12 Proceedings of the 17th ACM SIGPLAN symposium on Principles and Practice of Parallel Programming Benwen Zhang
More informationS WHAT THE PROFILER IS TELLING YOU: OPTIMIZING GPU KERNELS. Jakob Progsch, Mathias Wagner GTC 2018
S8630 - WHAT THE PROFILER IS TELLING YOU: OPTIMIZING GPU KERNELS Jakob Progsch, Mathias Wagner GTC 2018 1. Know your hardware BEFORE YOU START What are the target machines, how many nodes? Machine-specific
More informationOptimizing MPI Communication on Multi-GPU Systems using CUDA Inter-Process Communication
Optimizing MPI Communication on Multi-GPU Systems using CUDA Inter-Process Communication Sreeram Potluri* Hao Wang* Devendar Bureddy* Ashish Kumar Singh* Carlos Rosales + Dhabaleswar K. Panda* *Network-Based
More informationCUDA Optimizations WS Intelligent Robotics Seminar. Universität Hamburg WS Intelligent Robotics Seminar Praveen Kulkarni
CUDA Optimizations WS 2014-15 Intelligent Robotics Seminar 1 Table of content 1 Background information 2 Optimizations 3 Summary 2 Table of content 1 Background information 2 Optimizations 3 Summary 3
More informationParalization on GPU using CUDA An Introduction
Paralization on GPU using CUDA An Introduction Ehsan Nedaaee Oskoee 1 1 Department of Physics IASBS IPM Grid and HPC workshop IV, 2011 Outline 1 Introduction to GPU 2 Introduction to CUDA Graphics Processing
More informationTightly Coupled Accelerators Architecture
Tightly Coupled Accelerators Architecture Yuetsu Kodama Division of High Performance Computing Systems Center for Computational Sciences University of Tsukuba, Japan 1 What is Tightly Coupled Accelerators
More informationIntroduction to parallel Computing
Introduction to parallel Computing VI-SEEM Training Paschalis Paschalis Korosoglou Korosoglou (pkoro@.gr) (pkoro@.gr) Outline Serial vs Parallel programming Hardware trends Why HPC matters HPC Concepts
More informationDirected Optimization On Stencil-based Computational Fluid Dynamics Application(s)
Directed Optimization On Stencil-based Computational Fluid Dynamics Application(s) Islam Harb 08/21/2015 Agenda Motivation Research Challenges Contributions & Approach Results Conclusion Future Work 2
More informationMPI and CUDA. Filippo Spiga, HPCS, University of Cambridge.
MPI and CUDA Filippo Spiga, HPCS, University of Cambridge Outline Basic principle of MPI Mixing MPI and CUDA 1 st example : parallel GPU detect 2 nd example: heat2d CUDA- aware MPI, how
More informationMULTI-GPU PROGRAMMING MODELS. Jiri Kraus, Senior Devtech Compute Sreeram Potluri, Senior CUDA Software Engineer
MULTI-GPU PROGRAMMING MODELS Jiri Kraus, Senior Devtech Compute Sreeram Potluri, Senior CUDA Software Engineer MOTIVATION Why use multiple GPUs? Need to compute larger, e.g. bigger networks, car models,
More informationNUMA-Aware Data-Transfer Measurements for Power/NVLink Multi-GPU Systems
NUMA-Aware Data-Transfer Measurements for Power/NVLink Multi-GPU Systems Carl Pearson 1, I-Hsin Chung 2, Zehra Sura 2, Wen-Mei Hwu 1, and Jinjun Xiong 2 1 University of Illinois Urbana-Champaign, Urbana
More informationOpenACC Course. Office Hour #2 Q&A
OpenACC Course Office Hour #2 Q&A Q1: How many threads does each GPU core have? A: GPU cores execute arithmetic instructions. Each core can execute one single precision floating point instruction per cycle
More informationAdvanced Software for the Supercomputer PRIMEHPC FX10. Copyright 2011 FUJITSU LIMITED
Advanced Software for the Supercomputer PRIMEHPC FX10 System Configuration of PRIMEHPC FX10 nodes Login Compilation Job submission 6D mesh/torus Interconnect Local file system (Temporary area occupied
More informationCUDA Lecture 2. Manfred Liebmann. Technische Universität München Chair of Optimal Control Center for Mathematical Sciences, M17
CUDA Lecture 2 Manfred Liebmann Technische Universität München Chair of Optimal Control Center for Mathematical Sciences, M17 manfred.liebmann@tum.de December 15, 2015 CUDA Programming Fundamentals CUDA
More informationWhen MPPDB Meets GPU:
When MPPDB Meets GPU: An Extendible Framework for Acceleration Laura Chen, Le Cai, Yongyan Wang Background: Heterogeneous Computing Hardware Trend stops growing with Moore s Law Fast development of GPU
More informationInterconnect Your Future
Interconnect Your Future Smart Interconnect for Next Generation HPC Platforms Gilad Shainer, August 2016, 4th Annual MVAPICH User Group (MUG) Meeting Mellanox Connects the World s Fastest Supercomputer
More informationdcuda: Distributed GPU Computing with Hardware Overlap
dcuda: Distributed GPU Computing with Hardware Overlap Tobias Gysi, Jeremia Bär, Lukas Kuster, and Torsten Hoefler GPU computing gained a lot of popularity in various application domains weather & climate
More informationBuilding the Most Efficient Machine Learning System
Building the Most Efficient Machine Learning System Mellanox The Artificial Intelligence Interconnect Company June 2017 Mellanox Overview Company Headquarters Yokneam, Israel Sunnyvale, California Worldwide
More informationTowards Automatic Heterogeneous Computing Performance Analysis. Carl Pearson Adviser: Wen-Mei Hwu
Towards Automatic Heterogeneous Computing Performance Analysis Carl Pearson pearson@illinois.edu Adviser: Wen-Mei Hwu 2018 03 30 1 Outline High Performance Computing Challenges Vision CUDA Allocation and
More informationCSE 599 I Accelerated Computing - Programming GPUS. Parallel Patterns: Graph Search
CSE 599 I Accelerated Computing - Programming GPUS Parallel Patterns: Graph Search Objective Study graph search as a prototypical graph-based algorithm Learn techniques to mitigate the memory-bandwidth-centric
More informationGPU Programming. Lecture 2: CUDA C Basics. Miaoqing Huang University of Arkansas 1 / 34
1 / 34 GPU Programming Lecture 2: CUDA C Basics Miaoqing Huang University of Arkansas 2 / 34 Outline Evolvements of NVIDIA GPU CUDA Basic Detailed Steps Device Memories and Data Transfer Kernel Functions
More informationGPU Sparse Graph Traversal. Duane Merrill
GPU Sparse Graph Traversal Duane Merrill Breadth-first search of graphs (BFS) 1. Pick a source node 2. Rank every vertex by the length of shortest path from source Or label every vertex by its predecessor
More informationEvaluating On-Node GPU Interconnects for Deep Learning Workloads
Evaluating On-Node GPU Interconnects for Deep Learning Workloads NATHAN TALLENT, NITIN GAWANDE, CHARLES SIEGEL ABHINAV VISHNU, ADOLFY HOISIE Pacific Northwest National Lab PMBS 217 (@ SC) November 13,
More informationGPGPUGPGPU: Multi-GPU Programming
GPGPUGPGPU: Multi-GPU Programming Fall 2012 HW4 global void cuda_transpose(const float *ad, const int n, float *atd) { } int i = threadidx.y + blockidx.y*blockdim.y; int j = threadidx.x + blockidx.x*blockdim.x;
More informationDesigning High-Performance MPI Collectives in MVAPICH2 for HPC and Deep Learning
5th ANNUAL WORKSHOP 209 Designing High-Performance MPI Collectives in MVAPICH2 for HPC and Deep Learning Hari Subramoni Dhabaleswar K. (DK) Panda The Ohio State University The Ohio State University E-mail:
More informationNEW DEVELOPER TOOLS FEATURES IN CUDA 8.0. Sanjiv Satoor
NEW DEVELOPER TOOLS FEATURES IN CUDA 8.0 Sanjiv Satoor CUDA TOOLS 2 NVIDIA NSIGHT Homogeneous application development for CPU+GPU compute platforms CUDA-Aware Editor CUDA Debugger CPU+GPU CUDA Profiler
More informationHigh-Performance Key-Value Store on OpenSHMEM
High-Performance Key-Value Store on OpenSHMEM Huansong Fu*, Manjunath Gorentla Venkata, Ahana Roy Choudhury*, Neena Imam, Weikuan Yu* *Florida State University Oak Ridge National Laboratory Outline Background
More informationGPU Sparse Graph Traversal
GPU Sparse Graph Traversal Duane Merrill (NVIDIA) Michael Garland (NVIDIA) Andrew Grimshaw (Univ. of Virginia) UNIVERSITY of VIRGINIA Breadth-first search (BFS) 1. Pick a source node 2. Rank every vertex
More informationOPEN MPI WITH RDMA SUPPORT AND CUDA. Rolf vandevaart, NVIDIA
OPEN MPI WITH RDMA SUPPORT AND CUDA Rolf vandevaart, NVIDIA OVERVIEW What is CUDA-aware History of CUDA-aware support in Open MPI GPU Direct RDMA support Tuning parameters Application example Future work
More informationSupport for GPUs with GPUDirect RDMA in MVAPICH2 SC 13 NVIDIA Booth
Support for GPUs with GPUDirect RDMA in MVAPICH2 SC 13 NVIDIA Booth by D.K. Panda The Ohio State University E-mail: panda@cse.ohio-state.edu http://www.cse.ohio-state.edu/~panda Outline Overview of MVAPICH2-GPU
More informationHigh-Performance Graph Primitives on the GPU: Design and Implementation of Gunrock
High-Performance Graph Primitives on the GPU: Design and Implementation of Gunrock Yangzihao Wang University of California, Davis yzhwang@ucdavis.edu March 24, 2014 Yangzihao Wang (yzhwang@ucdavis.edu)
More informationScalable Cluster Computing with NVIDIA GPUs Axel Koehler NVIDIA. NVIDIA Corporation 2012
Scalable Cluster Computing with NVIDIA GPUs Axel Koehler NVIDIA Outline Introduction to Multi-GPU Programming Communication for Single Host, Multiple GPUs Communication for Multiple Hosts, Multiple GPUs
More informationParallel Programming Libraries and implementations
Parallel Programming Libraries and implementations Partners Funding Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License.
More informationTuring Architecture and CUDA 10 New Features. Minseok Lee, Developer Technology Engineer, NVIDIA
Turing Architecture and CUDA 10 New Features Minseok Lee, Developer Technology Engineer, NVIDIA Turing Architecture New SM Architecture Multi-Precision Tensor Core RT Core Turing MPS Inference Accelerated,
More informationVOLTA: PROGRAMMABILITY AND PERFORMANCE. Jack Choquette NVIDIA Hot Chips 2017
VOLTA: PROGRAMMABILITY AND PERFORMANCE Jack Choquette NVIDIA Hot Chips 2017 1 TESLA V100 21B transistors 815 mm 2 80 SM 5120 CUDA Cores 640 Tensor Cores 16 GB HBM2 900 GB/s HBM2 300 GB/s NVLink *full GV100
More informationScaling with PGAS Languages
Scaling with PGAS Languages Panel Presentation at OFA Developers Workshop (2013) by Dhabaleswar K. (DK) Panda The Ohio State University E-mail: panda@cse.ohio-state.edu http://www.cse.ohio-state.edu/~panda
More informationImplementation of BFS on shared memory (CPU / GPU)
Kolganov A.S., MSU The BFS algorithm Graph500 && GGraph500 Implementation of BFS on shared memory (CPU / GPU) Predicted scalability 2 The BFS algorithm Graph500 && GGraph500 Implementation of BFS on shared
More informationGPUfs: Integrating a file system with GPUs
GPUfs: Integrating a file system with GPUs Mark Silberstein (UT Austin/Technion) Bryan Ford (Yale), Idit Keidar (Technion) Emmett Witchel (UT Austin) 1 Traditional System Architecture Applications OS CPU
More informationMPI Alltoall Personalized Exchange on GPGPU Clusters: Design Alternatives and Benefits
MPI Alltoall Personalized Exchange on GPGPU Clusters: Design Alternatives and Benefits Ashish Kumar Singh, Sreeram Potluri, Hao Wang, Krishna Kandalla, Sayantan Sur, and Dhabaleswar K. Panda Network-Based
More informationDEEP DIVE INTO DYNAMIC PARALLELISM
April 4-7, 2016 Silicon Valley DEEP DIVE INTO DYNAMIC PARALLELISM SHANKARA RAO THEJASWI NANDITALE, NVIDIA CHRISTOPH ANGERER, NVIDIA 1 OVERVIEW AND INTRODUCTION 2 WHAT IS DYNAMIC PARALLELISM? The ability
More informationCS 179: GPU Computing. Recitation 2: Synchronization, Shared memory, Matrix Transpose
CS 179: GPU Computing Recitation 2: Synchronization, Shared memory, Matrix Transpose Synchronization Ideal case for parallelism: no resources shared between threads no communication between threads Many
More informationLecture 6: odds and ends
Lecture 6: odds and ends Prof. Mike Giles mike.giles@maths.ox.ac.uk Oxford University Mathematical Institute Oxford e-research Centre Lecture 6 p. 1 Overview synchronicity multiple streams and devices
More informationParallel Programming. Libraries and Implementations
Parallel Programming Libraries and Implementations Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationGPU Programming Using CUDA. Samuli Laine NVIDIA Research
GPU Programming Using CUDA Samuli Laine NVIDIA Research Today GPU vs CPU Different architecture, different workloads Basics of CUDA Executing code on GPU Managing memory between CPU and GPU CUDA API Quick
More informationPorting The Spectral Element Community Atmosphere Model (CAM-SE) To Hybrid GPU Platforms
Porting The Spectral Element Community Atmosphere Model (CAM-SE) To Hybrid GPU Platforms http://www.scidacreview.org/0902/images/esg13.jpg Matthew Norman Jeffrey Larkin Richard Archibald Valentine Anantharaj
More informationMVAPICH2-GDR: Pushing the Frontier of HPC and Deep Learning
MVAPICH2-GDR: Pushing the Frontier of HPC and Deep Learning Talk at Mellanox Theater (SC 16) by Dhabaleswar K. (DK) Panda The Ohio State University E-mail: panda@cse.ohio-state.edu http://www.cse.ohio-state.edu/~panda
More informationEfficient Multi-GPU CUDA Linear Solvers for OpenFOAM
Efficient Multi-GPU CUDA Linear Solvers for OpenFOAM Alexander Monakov, amonakov@ispras.ru Institute for System Programming of Russian Academy of Sciences March 20, 2013 1 / 17 Problem Statement In OpenFOAM,
More informationBigger GPUs and Bigger Nodes. Carl Pearson PhD Candidate, advised by Professor Wen-Mei Hwu
Bigger GPUs and Bigger Nodes Carl Pearson (pearson@illinois.edu) PhD Candidate, advised by Professor Wen-Mei Hwu 1 Outline Experiences from working with domain experts to develop GPU codes on Blue Waters
More informationScalable Multi Agent Simulation on the GPU. Avi Bleiweiss NVIDIA Corporation San Jose, 2009
Scalable Multi Agent Simulation on the GPU Avi Bleiweiss NVIDIA Corporation San Jose, 2009 Reasoning Explicit State machine, serial Implicit Compute intensive Fits SIMT well Collision avoidance Motivation
More informationExploiting GPUDirect RDMA in Designing High Performance OpenSHMEM for NVIDIA GPU Clusters
2015 IEEE International Conference on Cluster Computing Exploiting GPUDirect RDMA in Designing High Performance OpenSHMEM for NVIDIA GPU Clusters Khaled Hamidouche, Akshay Venkatesh, Ammar Ahmad Awan,
More informationFCUDA: Enabling Efficient Compilation of CUDA Kernels onto
FCUDA: Enabling Efficient Compilation of CUDA Kernels onto FPGAs October 13, 2009 Overview Presenting: Alex Papakonstantinou, Karthik Gururaj, John Stratton, Jason Cong, Deming Chen, Wen-mei Hwu. FCUDA:
More informationAmgX 2.0: Scaling toward CORAL Joe Eaton, November 19, 2015
AmgX 2.0: Scaling toward CORAL Joe Eaton, November 19, 2015 Agenda Introduction to AmgX Current Capabilities Scaling V2.0 Roadmap for the future 2 AmgX Fast, scalable linear solvers, emphasis on iterative
More informationMulti-GPU Implementations of Parallel 3D Sweeping Algorithms with Application to Geological Folding
Procedia Computer Science Volume 51, 2015, Pages 1494 1503 ICCS 2015 International Conference On Computational Science Multi-GPU Implementations of Parallel 3D Sweeping Algorithms with Application to Geological
More informationGPU/CPU Heterogeneous Work Partitioning Riya Savla (rds), Ridhi Surana (rsurana), Jordan Tick (jrtick) Computer Architecture, Spring 18
ABSTRACT GPU/CPU Heterogeneous Work Partitioning Riya Savla (rds), Ridhi Surana (rsurana), Jordan Tick (jrtick) 15-740 Computer Architecture, Spring 18 With the death of Moore's Law, programmers can no
More informationTechnical Computing Suite supporting the hybrid system
Technical Computing Suite supporting the hybrid system Supercomputer PRIMEHPC FX10 PRIMERGY x86 cluster Hybrid System Configuration Supercomputer PRIMEHPC FX10 PRIMERGY x86 cluster 6D mesh/torus Interconnect
More informationCUDA Performance Optimization. Patrick Legresley
CUDA Performance Optimization Patrick Legresley Optimizations Kernel optimizations Maximizing global memory throughput Efficient use of shared memory Minimizing divergent warps Intrinsic instructions Optimizations
More informationNVIDIA COLLECTIVE COMMUNICATION LIBRARY (NCCL)
NVIDIA COLLECTIVE COMMUNICATION LIBRARY (NCCL) DU-08730-210_v01 March 2018 Installation Guide TABLE OF CONTENTS Chapter 1. Overview... 1 Chapter 2. Prerequisites...3 2.1. Software Requirements... 3 2.2.
More informationS WHAT THE PROFILER IS TELLING YOU OPTIMIZING WHOLE APPLICATION PERFORMANCE. Mathias Wagner, Jakob Progsch GTC 2017
S7445 - WHAT THE PROFILER IS TELLING YOU OPTIMIZING WHOLE APPLICATION PERFORMANCE Mathias Wagner, Jakob Progsch GTC 2017 BEFORE YOU START The five steps to enlightenment 1. Know your application What does
More informationPersistent RNNs. (stashing recurrent weights on-chip) Gregory Diamos. April 7, Baidu SVAIL
(stashing recurrent weights on-chip) Baidu SVAIL April 7, 2016 SVAIL Think hard AI. Goal Develop hard AI technologies that impact 100 million users. Deep Learning at SVAIL 100 GFLOP/s 1 laptop 6 TFLOP/s
More informationLecture 10. Efficient Host-Device Data Transfer
1 Lecture 10 Efficient Host-Device Data fer 2 Objective To learn the important concepts involved in copying (transferring) data between host and device System Interconnect Direct Memory Access Pinned memory
More informationApril 4-7, 2016 Silicon Valley INSIDE PASCAL. Mark Harris, October 27,
April 4-7, 2016 Silicon Valley INSIDE PASCAL Mark Harris, October 27, 2016 @harrism INTRODUCING TESLA P100 New GPU Architecture CPU to CPUEnable the World s Fastest Compute Node PCIe Switch PCIe Switch
More informationPerformance Optimization Part II: Locality, Communication, and Contention
Lecture 7: Performance Optimization Part II: Locality, Communication, and Contention Parallel Computer Architecture and Programming CMU 15-418/15-618, Spring 2015 Tunes Beth Rowley Nobody s Fault but Mine
More informationWHAT S NEW IN CUDA 8. Siddharth Sharma, Oct 2016
WHAT S NEW IN CUDA 8 Siddharth Sharma, Oct 2016 WHAT S NEW IN CUDA 8 Why Should You Care >2X Run Computations Faster* Solve Larger Problems** Critical Path Analysis * HOOMD Blue v1.3.3 Lennard-Jones liquid
More informationECE 408 / CS 483 Final Exam, Fall 2014
ECE 408 / CS 483 Final Exam, Fall 2014 Thursday 18 December 2014 8:00 to 11:00 Central Standard Time You may use any notes, books, papers, or other reference materials. In the interest of fair access across
More informationBuilding the Most Efficient Machine Learning System
Building the Most Efficient Machine Learning System Mellanox The Artificial Intelligence Interconnect Company June 2017 Mellanox Overview Company Headquarters Yokneam, Israel Sunnyvale, California Worldwide
More informationNVIDIA GTX200: TeraFLOPS Visual Computing. August 26, 2008 John Tynefield
NVIDIA GTX200: TeraFLOPS Visual Computing August 26, 2008 John Tynefield 2 Outline Execution Model Architecture Demo 3 Execution Model 4 Software Architecture Applications DX10 OpenGL OpenCL CUDA C Host
More informationIntroduction to CUDA Algoritmi e Calcolo Parallelo. Daniele Loiacono
Introduction to CUDA Algoritmi e Calcolo Parallelo References q This set of slides is mainly based on: " CUDA Technical Training, Dr. Antonino Tumeo, Pacific Northwest National Laboratory " Slide of Applied
More informationCuSha: Vertex-Centric Graph Processing on GPUs
CuSha: Vertex-Centric Graph Processing on GPUs Farzad Khorasani, Keval Vora, Rajiv Gupta, Laxmi N. Bhuyan HPDC Vancouver, Canada June, Motivation Graph processing Real world graphs are large & sparse Power
More informationCUDA Kernel based Collective Reduction Operations on Large-scale GPU Clusters
CUDA Kernel based Collective Reduction Operations on Large-scale GPU Clusters Ching-Hsiang Chu, Khaled Hamidouche, Akshay Venkatesh, Ammar Ahmad Awan and Dhabaleswar K. (DK) Panda Speaker: Sourav Chakraborty
More informationGRAPHICS PROCESSING UNITS
GRAPHICS PROCESSING UNITS Slides by: Pedro Tomás Additional reading: Computer Architecture: A Quantitative Approach, 5th edition, Chapter 4, John L. Hennessy and David A. Patterson, Morgan Kaufmann, 2011
More informationCartoon parallel architectures; CPUs and GPUs
Cartoon parallel architectures; CPUs and GPUs CSE 6230, Fall 2014 Th Sep 11! Thanks to Jee Choi (a senior PhD student) for a big assist 1 2 3 4 5 6 7 8 9 10 11 12 13 14 ~ socket 14 ~ core 14 ~ HWMT+SIMD
More informationHigh Performance Computing
High Performance Computing Dror Goldenberg, HPCAC Switzerland Conference March 2015 End-to-End Interconnect Solutions for All Platforms Highest Performance and Scalability for X86, Power, GPU, ARM and
More information