Achieving peak bandwidth on multi-socket systems
Let's say each CPU socket has 43 GB/s of bandwidth through its four memory channels. Let's say I have a dual socket system. A reduction operation should achieve performance of 86 GB/s, but it...
View ArticleCalculate GFLOPs
Hi, In Intel Vtune Amplifier profiler, there is no counter for how many instructions execute on Integrated GPUs.Instead, the profiler provide three metrics indicating the ratio of EU in state active,...
View ArticleCan't use intel GPU with OpenCL while AMD HD8790M is installed
In my DELL E6540 laptop I have a Intel HD4600 and an AMD HD8790M GPU.Before installing the AMD graphics driver I could use OpenCL with the Intel GPU (but obviously not the AMD GPU) as well as the Intel...
View ArticleProgram cannot build when setting device to "1"
Hello,I have a laptop with both intel haswell chip and intel HD 4400 GPUI have VS 2012 and the 2014 beta opencl sdk installed.I have taken the median filter sample and modified it to run my own...
View ArticleCan't install Intel OpenCL SDK 2013
Hi everyone,My system is windows 7 home premium, my GPU is Intel HD Graphics 4000, and I have updated the latest the driver. The VS version is Visual Sdudio 2012 Express. However, when I follow the...
View ArticleKernel Builder Crashes when trying to "Analyze" kernel. Build is fine
Hi I have this error when I try to analyze a kernel. The build is fine for both 32-bit and 64-bit Kernel Builder.What's the problem? Problem signature: Problem Event Name: APPCRASH Application...
View ArticleclBuildProgram crashes when targetting HD 4400 Intel GPU
Hello,I am working on an open source jpeg 2000 decoder:https://github.com/OpenCodec/ThousandthChickenI am on windows, using visual studio 2012 and the 2014 beta OpenCL sdk.My laptop has an intel CPU...
View ArticleOpencl: incorrect results on phi
This kernel runs on host cpu but produces wrong outputs when run Phi. What's the issue witht the Phi here? //KERNEL_SIMPLE __kernel void Convolve(const __global float * pInput,...
View ArticleI uninstalled sdk 2012, but can't install sdk 2013
I am having problems installing the 2013 version of the OpenCL SDK, on Windows 7 with VS 2010. I have a Xeon E5-26900 CPU(32-bit). I have installed 2012 version of the OpenCL SDK and it worked very...
View ArticleInconsistent cl_device_id order
I tried to create a program from source, build it and save its binary, and then create a program from binary. However, when I went to build it the second time I got the follow error: -42...
View ArticleIntel OpenCL SDK: clGetKernelWorkGroupInfo return value
Hi 1. I have a global work size of 1024 by 1024.2. I set the local work size to 16 by 16. 3. My CPU opnecl device has a maximum work-group-size of 8192.4. I call clEnqueueNDRangeKernel with the desired...
View ArticleclGetPlatformIDs returns -1001
I installed intel_sdk_for_ocl_applications_xe_2013_r3_sdk_3.2.1.16712_x64.tgz on my PC( Intel(R) Xeon(R) CPU E5-2680 0). But when I run opencl clGetPlatformIDs ,it returned -1001. My OS is SUSE Linux...
View ArticleclEnqueueCopyBufferRect crashes with access violation in cpu_device32.dll
I have several calls to clEnqueueCopyBufferRect to copy data between two device buffers.After a few calls, I get a crash with an access violation.I am using the 2014 beta sdk, targetting my Intel i7...
View ArticleAnnouncing new Beta version of the Intel® SDK for OpenCL™ Applications 2014
Dear OpenCL developers,I’m happy to announce the availability of the 2014 Beta version of the OpenCL SDK. Sign up and download now here.The Intel® SDK for OpenCL™ Applications 2014 Beta extends the...
View ArticleClCreateBuffer(| CL_MEM_USE_HOST_PTR): When does OpenCL framework transfer...
Hi Intel Xeon Phi OpenCL optimization guide suggests using Mapped buffers for data transfer between host and device memory. OpenCL spec also states that the technique is faster than having to write...
View ArticleHow dose clCreateBuffer( CL_MEM_USE_HOST_PTR) work when platform and device...
When my OpenCL platform and device are both CPU, would the behaviour (creating buffer using CL_MEM_USE_HOST_PTR )copy another uniform data or just link to the host_ptr address without copy?Dose any one...
View ArticleIntel OpenCL auto-vectorization switch
Hi, I want to disable the auto-vectorization of OpenCL kernel on Xeon Phi, is there any way to do that?
View ArticleRelative include paths don't work in clBuildProgram
I am using the 2014 beta SDK. I would like to pass "-I RELATIVE\\PATH\\TO\\FILES" in as a build parameters, but this doesn't work.I have to pass the absolute path instead. So, this forces me to hard...
View ArticleSome benchmarking experience with Phi: OpenCL vs OpenMP 2D Image Convolution
The performance of Xeon Phi benchmarked with 2D convolution in opnecl seems much better than an openmp implementation even with compiler-enabled vectorization. Openmp version was run in phi native...
View ArticleInfuriating Forum Format
I just wanted to say how annoying the format of this forum is! Every post's subject line begins with "new comment for blah blah blah" so it is hard to see the actual post title; every post I receive...
View Article