Intel® C++ Compiler 19.0 Developer Guide and Reference
Lets you specify the target architecture to use when offloading code. This option only applies when targeting Intel® Xeon Phi™ products.
Only available when targeting Intel® Xeon Phi™ products
Linux: | -qoffload-arch=arch |
macOS: | None |
Windows: | /Qoffload-arch:arch |
arch |
Specifies the target architecture. On Intel® 64 architecture targeting Intel® MIC Architecture, arch can be:
|
mic-avx512 |
On Intel® 64 architecture targeting Intel® MIC Architecture, the default arch is mic-avx512. |
This option lets you specify the target architecture to use when offloading code.
The following table shows the vISA version that each arch supports:
arch value |
Supported vISA |
---|---|
skylake |
vISA 3.2 |
broadwell |
vISA 3.1, vISA 3.2 |
haswell |
vISA 3.1 |
The following are various combinations and settings of command line options for offloading:
Offload: Intel® 64 architecture host and Intel® Xeon Phi™ product family x200 (formerly code name Knights Landing) target code generation:
icc … -qoffload-arch=mic-avx512 ! Linux icl … /Qoffload-arch=mic-avx512 ! Windows
Offload : Intel® Xeon Phi™ product family x200 (formerly code name Knights Landing) host and target code generation (for this host, target defaults to Intel® Xeon Phi™ product family x200 (formerly code name Knights Landing)):
icc … -xmic-avx512 ! Linux icl … /Qxmic-avx512 ! Windows
Native : Code Generation for Intel® Xeon Phi™ product family x200 (formerly code name Knights Landing):
icc … -xmic-avx512 –qno-offload ! Linux icl … /Qxmic-avx512 /Qno-offload ! Windows
Visual Studio: Code Generation [Intel C++] > Offload Target Architecture
Eclipse: Code Generation > Offload Target Architecture
Xcode: None
None