Visible to Intel only — GUID: GUID-A4E64E9C-405C-4049-B5C1-02EBF467CDDB
Visible to Intel only — GUID: GUID-A4E64E9C-405C-4049-B5C1-02EBF467CDDB
fopenmp-target-buffers, Qopenmp-target-buffers
Enables a way to overcome the problem where some OpenMP* offload SPIR-V* devices produce incorrect code when a target object is larger than 4GB. This feature is only available for ifx.
Syntax
Linux: |
-fopenmp-target-buffers=keyword |
macOS: |
None |
Windows: |
/Qopenmp-target-buffers:keyword |
Arguments
keyword |
Possible values are:
|
Default
default |
If you do not specify this option, the compiler may produce incorrect code on some OpenMP* offload SPIR-V* devices when a target object is larger than 4GB. |
Description
This option enables a way to overcome the problem where some OpenMP* offload SPIR-V* devices produce incorrect code when a target object is larger than 4GB (4294959104 bytes).
However, note that when -fopenmp-target-buffers=4GB (or /Qopenmp-target-buffers:4GB) is specified on Intel® GPUs, there may be a decrease in performance.
To use this option, you must also specify option -fopenmp-targets (Linux*) or /Qopenmp-targets (Windows*).
This option may have no effect for some OpenMP* offload SPIR-V* devices, and for OpenMP* offload targets different from SPIR*.
When OpenMP* offloading is enabled, this option only applies to device-specific compilation.
IDE Equivalent
Visual Studio: Code Generation > Specify buffer size for OpenMP offload kernel access limitations (ifx only)
Alternate Options
None