Visible to Intel only — GUID: GUID-561479B7-7D64-4EF8-8EDD-C7A9BE5BCD5A
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
See Also
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-561479B7-7D64-4EF8-8EDD-C7A9BE5BCD5A
Sharing Resources Efficiently
Objects, allocated at the context level, are shared between devices in the context. For example, buffers and images are effectively shared by default. Other resources that are shared automatically across all devices, include program and kernel objects.
NOTE:
Shared memory objects cannot be written concurrently by different command queues. Use explicit synchronization of the write access with OpenCL™ synchronization objects, such as events. Consider using sub-buffers, which enables you to simultaneously write to the non-overlapping regions.
You can also avoid implicit copying when you share data with the host, as explained in the “Mapping Memory Objects” section.
NOTE:
To avoid potential inefficiencies, especially associated with improper alignment, use 4k alignment for the host pointers in scenarios when the Intel® Graphics device is involved. Also align the allocation sizes to the cache line boundaries (64 bytes). Refer to the “Mapping Memory Objects” section for more details.