Visible to Intel only — GUID: GUID-435D7A7F-8831-4D6B-A970-533623EC8F64
Legal Information
Getting Help and Support
Introduction
Coding for the Intel® Processor Graphics
Platform-Level Considerations
Application-Level Optimizations
Optimizing OpenCL™ Usage with Intel® Processor Graphics
Check-list for OpenCL™ Optimizations
Performance Debugging
Using Multiple OpenCL™ Devices
Coding for the Intel® CPU OpenCL™ Device
OpenCL™ Kernel Development for Intel® CPU OpenCL™ device
Mapping Memory Objects
Using Buffers and Images Appropriately
Using Floating Point for Calculations
Using Compiler Options for Optimizations
Using Built-In Functions
Loading and Storing Data in Greatest Chunks
Applying Shared Local Memory
Using Specialization in Branching
Considering native_ and half_ Versions of Math Built-Ins
Using the Restrict Qualifier for Kernel Arguments
Avoiding Handling Edge Conditions in Kernels
Using Shared Context for Multiple OpenCL™ Devices
Sharing Resources Efficiently
Synchronization Caveats
Writing to a Shared Resource
Partitioning the Work
Keeping Kernel Sources the Same
Basic Frequency Considerations
Eliminating Device Starvation
Limitations of Shared Context with Respect to Extensions
Why Optimizing Kernel Code Is Important?
Avoid Spurious Operations in Kernel Code
Perform Initialization in a Separate Task
Use Preprocessor for Constants
Use Signed Integer Data Types
Use Row-Wise Data Accesses
Tips for Auto-Vectorization
Local Memory Usage
Avoid Extracting Vector Components
Task-Parallel Programming Model Hints
Visible to Intel only — GUID: GUID-435D7A7F-8831-4D6B-A970-533623EC8F64
Aligning Pointers to Microsoft DirectX* Buffers Upon Mapping
If your application utilizes resource sharing with Microsoft DirectX* 10 or 11 by use of mapping, use the CL_MEM_USE_HOST_PTR flag. Also note that mapping is less efficient than using the cl_khr_d3d10_sharing or cl_khr_d3d11_sharing.
Consider the general interoperability flow:
- Map a resource to CPU
- Create a buffer wrapping the memory
- Use the buffer by use of the OpenCL™ regular commands
- Upon competition of the OpenCL commands, the resource can be safely unmapped.
Now use the CL_MEM_USE_HOST_PTR flag to directly operate on the mapped memory and avoid data copying upon OpenCL buffer creation. This method requires properly aligned memory. See the “Mapping Memory Objects” section for more information.
NOTE:
Aligning might result in unprocessed data between original and aligned pointer.
If it is acceptable for your application, and/or the copying overhead is of concern, consider aligning of the pointer returned by the DirectX map call to comply with the "Mapping Memory Objects" section.
Another potential workaround is to allocate larger DirectX resource than it is required, so that you have some room for a safe alignment. This approach requires some additional application logic.