Increase OpenCL application performance with the new Intel® OpenCL SDK 1.5

By Arnon Peleg (Intel) (5 posts) on September 26, 2011 at 7:22 am

 

Intel released today an updated version of the Intel® OpenCL SDK for the CPU with significant new features and new performance improvements.

Available for free download at www.intel.com/go/opencl, the Intel® OpenCL SDK 1.5 makes it easy for you to design, build, debug, and profile OpenCL™ applications running on the CPU device, and is optimized for Intel® Core™ and Intel® Xeon® processors.

Design and Build:

The Intel OpenCL SDK 1.5 delivers performance improvements through full code generation using the Intel Advanced Vector Extensions (Intel® AVX). Intel AVX is a new 256-bit instruction set extension to SSE that was introduced with the 2nd Generation Intel® Core™ Processors. It is designed for applications that are Floating Point (FP) intensive.

Using the unique Implicit CPU Vectorization Module inside the Intel OpenCL SDK compiler, you can seamlessly utilize the full performance of Intel® Architecture by auto vectorization of OpenCL kernel code. For more information, check out this blog.

Another new tool in this release is the Intel® OpenCL SDK Offline Compiler for Linux*. With this standalone tool you can build kernels offline and analyze the code generation. In the assembly view window, for example, you can see how Intel® AVX code is generated.

For more performance features, look at the Tips and Tricks in writing OpenCL™ Code for CPU or the full optimization guide.

 

Debug:

I am especially excited about the new beta version of the Intel® OpenCL SDK Debugger, a Microsoft* Visual Studio* 2008 plug-in that enables you to set breakpoints in an instance of an OpenCL kernel and view kernel variables and memory.

Now you can debug OpenCL applications on the Intel platform!

 

Profile:

The latest introduction of Intel® Graphics Performance Analyzers (Intel® GPA) 4.2 and this release of the Intel® OpenCL SDK 1.5 have significantly improved your ability to profile OpenCL applications through the following views:

  • Device view. Enables you to see the distribution of the application’s OpenCL kernels and memory operations across the system’s software threads.
  • Context view.  Enables you to examine the flow of the OpenCL commands and their dependencies within the application’s context command queues.
  • API View. Enables you to measure the execution time of the application’s OpenCL API calls.

 

For the full list of new features, read the Intel® OpenCL SDK Release Notes.

To download the SDK and begin optimizing your OpenCL applications for Intel® CPUs, visit the Intel® OpenCL SDK community today.

I will be happy to read your feedback in this blog or in our OpenCL Support Forum.

*OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos.

Categories: Graphics & Media, Parallel Programming, Performance and Optimization
Tags: , , ,

For more complete information about compiler optimizations, see our Optimization Notice.

Comments (0)

Trackbacks (7)


Leave a comment  

To obtain technical support, please go to Software Support.
Name (required)*

Email (required; will not be displayed on this page)*

Your URL (optional)


Comment*