untitled

Size: px

Start display at page:

Download "untitled"

おきまさきせんばる
7 years ago
Views:

1 OpenCL GPU AMD JEL

2 -OpenCL - GPU - AMD/ATI GPU - 2

3 OpenCL 3

com OpenCL OpenCL Zone: http://developer.amd.

4 OpenCL : GPGPU API: CPU GPU : AMD, Apple, IBM, Intel, Nvidia, Sony : : GPGPU AMD developer.amd.com OpenCL OpenCL Zone: 4 OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos.

5 OpenCL Language Specification C-based cross-platform programming interface Subset of ISO C99 with language extensions - familiar to developers Well-defined numerical accuracy - IEEE 754 rounding behavior with defined maximum error Online or offline compilation and build of compute kernel executables Includes a rich set of built-in functions Platform Layer API A hardware abstraction layer over diverse computational resources Query, select and initialize compute devices Create compute contexts and work-queues Runtime API Execute compute kernels Manage scheduling, compute, and memory resources 5

6 OpenCL int main() { // ICD based init ICDinit(&platform); // Setup setupcl(cl_device_type_gpu, ); // Compile, Load Kernel compileloadcl(, vec_mult.cl ); // Create and Initialize buffers input0 = clcreatebuffer(context, ); input1 = clcreatebuffer(context, ); output = clcreatebuffer(context, ); CPU inputbuffer0 inputbuffer1 outputbuffer GPU GPU input0 input1 output // Execute kernel runclkernel(, &input0, &input1, &output, ); // Wait for kernel completion status = clfinish(cmd_queue); // Read data from CL device to CPU memory clenqueuereadbuffer(cmd_queue, ); // Wait for read completion status = clfinish(cmd_queue); 6 } main.cpp ( ) kernel void vec_mult ( global const float *a, global const float *b, global float *c) { int gid = get_global_id(0); c[gid] = a[gid] * b[gid]; } vec_mult.cl ( )

7 OpenCL C kernel void vec_mult ( global const float *a, global const float *b, global float *c) { int gid = get_global_id(0); c[gid] = a[gid] * b[gid]; } 1. kernel kernel 2. global, constant, local 3. ID 2 The OpenCL Specification, Version: 1.1 7

8 OpenCL in-order out-of-order (GPU) (CPU) N work-item work-group : private, local, constant, global The OpenCL Specification, Version: 1.1 8

9 PC CPU OS - Microsoft Windows - Linux - Microsoft Visual Studio -gcc CPU inputbuffer0 inputbuffer1 outputbuffer 9

10 GPGPU CPU OS - Microsoft Windows - Linux - Microsoft Visual Studio -gcc CPU GPU -GPU -GPGPU API (CPU ) GPU OS GPU -GPU CPU GPU OpenCL SDK AMD ATI Stream SDK inputbuffer0 input0 inputbuffer1 input1 outputbuffer output GPU 10

11 OpenCL : Khronos OpenCL C++ Bindings libstdcl Simplified OpenCL (SOCL) JavaCL OpenCL.NET PyOpenCL ScalaCL Wrapper Libraries (7+) Optional Extensions (15) cl_khr_gl_event 1.1 cl_khr_d3d10 2/26/2010 cl_khr_icd 1/29/2010 cl_khr_gl_sharing 8/28/2009 cl_khr_fp64 cl_khr_select_fprounding_mode cl_khr_fp16 cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_stores Multi-Vendor Extensions (2) cl_ext_device_fission 6/9/2010 cl_ext_migrate_memobject 6/8/2010 Core features in 1.1 Vendor-Specific Extensions (8) cl_amd_printf 7/2010 cl_amd_event_callback 7/2010 cl_amd_device_attribute_query 3/26/2010 cl_amd_fp64 3/26/2010 cl_amd_media_ops 3/26/ extensions from other vendors OpenCL Core Specification /9/ /14/

12 GPU -AMD/ATI GPU - 12

13 AMD/ATI GPU Stream Computing R520 ATI RADEON X1800 ATI FireGL V7200 V7300 V7350 GigaFLOPS * Cypress ATI RADEON * Peak single-precision performance; For RV670, RV770 & Cypress divide by 5 for peak double-precision performance R580(+) ATI RADEON X19xx ATI FireStream R600 ATI RADEON HD 2900 ATI FireGL V7600 V8600 V8650 CTM GPGPU RV670 ATI RADEON HD 3800 ATI FireGL V7700 AMD FireStream 9170 RV770 ATI RADEON HD 4800 ATI FirePro V8700 AMD FireStream HD 5870 ALU 2.5 Stream SDK CAL+IL/Brook+ OpenCL 1.0 DirectX <+18% TDP Sep 05 Mar 06 Oct 06 Apr 07 Nov 07 Jun 08 Dec 08 Jul 09 13

AMD/ATI GPU AMD/ATI GPU 5-way VLIW Shader Processing Unit VLIW 5 ALU SPU(Stream Processing Unit) Stream Stream 5 Stream (T-Stream ) 4 Stream T-Stream Radeon HD5870 x: MULADD_e, R5.y, R2.x, T0.

14 AMD/ATI GPU AMD/ATI GPU 5-way VLIW Shader Processing Unit VLIW 5 ALU SPU(Stream Processing Unit) Stream Stream 5 Stream (T-Stream ) 4 Stream T-Stream Radeon HD5870 x: MULADD_e, R5.y, R2.x, T0.x VEC_120 y: MULADD_e R0.y, R5.y, R2.w, T1.w VEC_120 z: MULADD_e T1.z, R6.y, R2.x, T1.z VEC_210 w: MULADD_e, R7.y, R2.x, T0.w VEC_021 t: MULADD_e T0.z, R6.y, R2.z, T2.z VEC_120 Stream コア T-Stream コア 14 14

15 AMD/ATI GPU SIMD SIMD SIMD 16 SIMD SIMD SIMD RV Cypress 20 Cypress 15 15

16 GPU OpenCL GPU SIMD Stream OpenCL CU PE CU ID ID 2 ID gx, gy ID wx, wy ID sx, sy (gx, gy) = (wx * Sx+ sx, wy * Sy+ sy) Sx, Sy The OpenCL Specification, Version:

17 - CPU GPU CPU CPU GPU CAE GPGPU GPU Graphics Workloads Serial/Task-Parallel Workloads Other Highly Parallel Workloads 17

18 Disclaimer & Attribution DISCLAIMER The information presented in this document is for informational purposes only and may contain technical inaccuracies, omissions and typographical errors. The information contained herein is subject to change and may be rendered inaccurate for many reasons, including but not limited to product and roadmap changes, component and motherboard version changes, new model and/or product releases, product differences between differing manufacturers, software changes, BIOS flashes, firmware upgrades, or the like. AMD assumes no obligation to update or otherwise correct or revise this information. However, AMD reserves the right to revise this information and to make changes from time to time to the content hereof without obligation of AMD to notify any person of such revisions or changes. AMD MAKES NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE CONTENTS HEREOF AND ASSUMES NO RESPONSIBILITY FOR ANY INACCURACIES, ERRORS OR OMISSIONS THAT MAY APPEAR IN THIS INFORMATION. AMD SPECIFICALLY DISCLAIMS ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE. IN NO EVENT WILL AMD BE LIABLE TO ANY PERSON FOR ANY DIRECT, INDIRECT, SPECIAL OR OTHER CONSEQUENTIAL DAMAGES ARISING FROM THE USE OF ANY INFORMATION CONTAINED HEREIN, EVEN IF AMD IS EXPRESSLY ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. ATTRIBUTION 2010 Advanced Micro Devices, Inc. All rights reserved. AMD, the AMD Arrow logo, ATI, the ATI Logo, Radeon, FirePro, FireStream and combinations thereof are trademarks of Advanced Micro Devices, Inc. Microsoft, Windows, Windows Vista, and DirectX are registered trademarks of Microsoft Corporation in the United States and/or other jurisdictions. Other names are for informational purposes only and may be trademarks of their respective owners. OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos. 18

外部SQLソース入門

外部SQLソース入門 Introduction to External SQL Sources 外部 SQL ソース入門 3 ESS 3 ESS : 4 ESS : 4 5 ESS 5 Step 1:... 6 Step 2: DSN... 6 Step 3: FileMaker Pro... 6 Step 4: FileMaker Pro 1. 6 Step 5:... 6 Step 6: FileMaker Pro...