OpenMP Support

Clang fully supports OpenMP 4.5. Clang supports offloading to X86_64, AArch64, PPC64[LE] and has basic support for Cuda devices.

In addition, the LLVM OpenMP runtime libomp supports the OpenMP Tools Interface (OMPT) on x86, x86_64, AArch64, and PPC64 on Linux, Windows, and macOS.

For the list of supported features from OpenMP 5.0 see OpenMP implementation details.

General improvements

  • New collapse clause scheme to avoid expensive remainder operations. Compute loop index variables after collapsing a loop nest via the collapse clause by replacing the expensive remainder operation with multiplications and additions.
  • The default schedules for the distribute and for constructs in a parallel region and in SPMD mode have changed to ensure coalesced accesses. For the distribute construct, a static schedule is used with a chunk size equal to the number of threads per team (default value of threads or as specified by the thread_limit clause if present). For the for construct, the schedule is static with chunk size of one.
  • Simplified SPMD code generation for distribute parallel for when the new default schedules are applicable.

Directives execution modes

Clang code generation for target regions supports two modes: the SPMD and non-SPMD modes. Clang chooses one of these two modes automatically based on the way directives and clauses on those directives are used. The SPMD mode uses a simplified set of runtime functions thus increasing performance at the cost of supporting some OpenMP features. The non-SPMD mode is the most generic mode and supports all currently available OpenMP features. The compiler will always attempt to use the SPMD mode wherever possible. SPMD mode will not be used if:

  • The target region contains user code (other than OpenMP-specific directives) in between the target and the parallel directives.

Data-sharing modes

Clang supports two data-sharing models for Cuda devices: Generic and Cuda modes. The default mode is Generic. Cuda mode can give an additional performance and can be activated using the -fopenmp-cuda-mode flag. In Generic mode all local variables that can be shared in the parallel regions are stored in the global memory. In Cuda mode local variables are not shared between the threads and it is user responsibility to share the required data between the threads in the parallel regions.

Collapsed loop nest counter

When using the collapse clause on a loop nest the default behavior is to automatically extend the representation of the loop counter to 64 bits for the cases where the sizes of the collapsed loops are not known at compile time. To prevent this conservative choice and use at most 32 bits, compile your program with the -fopenmp-optimistic-collapse.

Features not supported or with limited support for Cuda devices

  • Cancellation constructs are not supported.
  • Doacross loop nest is not supported.
  • User-defined reductions are supported only for trivial types.
  • Nested parallelism: inner parallel regions are executed sequentially.
  • Static linking of libraries containing device code is not supported yet.
  • Automatic translation of math functions in target regions to device-specific math functions is not implemented yet.
  • Debug information for OpenMP target regions is supported, but sometimes it may be required to manually specify the address class of the inspected variables. In some cases the local variables are actually allocated in the global memory, but the debug info may be not aware of it.

OpenMP 5.0 Implementation Details

The following table provides a quick overview over various OpenMP 5.0 features and their implementation status. Please contact openmp-dev at for more information or if you want to help with the implementation.

Category Feature Status Reviews
loop extension support != in the canonical loop form done D54441
loop extension #pragma omp loop (directive) unclaimed  
loop extension collapse imperfectly nested loop done  
loop extension collapse non-rectangular nested loop done  
loop extension C++ range-base for loop done  
loop extension clause: nosimd for SIMD directives unclaimed  
loop extension inclusive scan extension (matching C++17 PSTL) unclaimed  
memory mangagement memory allocators done r341687,r357929
memory mangagement allocate directive and allocate clause done r355614,r335952
OMPD OMPD interfaces not upstream
OMPT OMPT interfaces mostly done  
thread affinity extension thread affinity extension done  
task extension taskloop reduction done  
task extension task affinity not upstream  
task extension clause: depend on the taskwait construct worked on  
task extension depend objects and detachable tasks worked on  
task extension mutexinoutset dependence-type for tasks done D53380,D57576
task extension combined taskloop constructs unclaimed  
task extension master taskloop done  
task extension parallel master taskloop done  
task extension master taskloop simd done  
task extension parallel master taskloop simd done  
SIMD extension atomic and critical constructs inside SIMD code unclaimed  
SIMD extension SIMD nontemporal unclaimed  
device extension infer target functions from initializers worked on  
device extension infer target variables from initializers worked on  
device extension OMP_TARGET_OFFLOAD environment variable done D50522
device extension support full ‘defaultmap’ functionality worked on  
device extension device specific functions unclaimed  
device extension clause: device_type done  
device extension clause: in_reduction unclaimed r308768
device extension omp_get_device_num() worked on D54342
device extension structure mapping of references unclaimed  
device extension nested target declare done D51378
device extension implicitly map ‘this’ (this[:1]) done D55982
device extension allow access to the reference count (omp_target_is_present) worked on  
device extension requires directive (unified shared memory) worked on  
device extension clause: unified_address, unified_shared_memory worked on D52625,D52359
device extension clause: reverse_offload unclaimed parts D52780
device extension clause: atomic_default_mem_order unclaimed parts D53513
device extension clause: dynamic_allocators unclaimed parts D53079
device extension user-defined mappers worked on D56326,D58638,D58523,D58074,D60972,D59474
device extension mapping lambda expression done D51107
device extension clause: use_device_addr for target data done  
device extension map(replicate) or map(local) when requires unified_shared_me worked on D55719,D55892
device extension teams construct on the host device worked on Clang part is done, r371553.
atomic extension hints for the atomic construct worked on D51233
base language C11 support unclaimed  
base language C++11/14/17 support unclaimed  
base language lambda support done  
misc extension array shaping unclaimed  
misc extension library shutdown (omp_pause_resource[_all]) unclaimed parts D55078
misc extension metadirectives unclaimed  
misc extension conditional modifier for lastprivate clause unclaimed  
misc extension user-defined function variants worked on D67294, D64095
misc extensions pointer/reference to pointer based array reductions unclaimed  
misc extensions prevent new type definitions in clauses unclaimed