FreeBSD manual

download PDF document: llvmopenmp15.1.pdf

LLVMOPENMP(1) LLVM/OpenMP LLVMOPENMP(1)
NAME llvmopenmp - LLVM/OpenMP 15.0.7
NOTE: This document is a work in progress and most of the expected content is not yet available. While you can expect changes, we always welcome feedback and additions. Please contact, e.g., through openmp-dev@lists.llvm.org.
OpenMP impacts various parts of the LLVM project, from the frontends (Clang and Flang), through middle-end optimizations, up to the multitude of available OpenMP runtimes.
A high-level overview of OpenMP in LLVM can be found here.
OPENMP IN LLVM --- DESIGN OVERVIEW Resources
o OpenMP Booth @ SC19: "OpenMP clang and flang Development" https://youtu.be/6yOa-hRi63M
LLVM/OpenMP Runtimes There are four distinct types of LLVM/OpenMP runtimes: the host runtime LLVM/OpenMP Host Runtime (libomp), the target offloading runtime LLVM/OpenMP Target Host Runtime (libomptarget), the target offloading plugin LLVM/OpenMP Target Host Runtime Plugins (libomptarget.rtl.XXXX), and finally the target device runtime LLVM/OpenMP Target Device Runtime (libomptarget-ARCH-SUBARCH.bc).
For general information on debugging OpenMP target offloading applications, see LIBOMPTARGET_INFO and Debugging
LLVM/OpenMP Host Runtime (libomp) An early (2015) design document for the LLVM/OpenMP host runtime, aka. libomp.so, is available as a pdf.
Environment Variables OMP_CANCELLATION Enables cancellation of the innermost enclosing region of the type specified. If set to true, the effects of the cancel construct and of cancellation points are enabled and cancellation is activated. If set to false, cancellation is disabled and the cancel construct and cancellation points are effectively ignored.
NOTE: Internal barrier code will work differently depending on whether cancellation is enabled. Barrier code should repeatedly check the global flag to figure out if cancellation has been triggered. If a thread observes cancellation, it should leave the barrier prematurely with the return value 1 (and may wake up other threads). Otherwise, it should leave the barrier with the return value 0.
Enables (true) or disables (false) cancellation of the innermost enclosing region of the type specified.
Default: false
Default: false
OMP_DEFAULT_DEVICE Sets the device that will be used in a target region. The OpenMP routine omp_set_default_device or a device clause in a parallel pragma can override this variable. If no device with the specified device number exists, the code is executed on the host. If this environment variable is not set, device number 0 is used.
OMP_DYNAMIC Enables (true) or disables (false) the dynamic adjustment of the number of threads. Default: false

OMP_MAX_ACTIVE_LEVELS The maximum number of levels of parallel nesting for the program. Default: 1

OMP_NESTED WARNING: Deprecated. Please use OMP_MAX_ACTIVE_LEVELS to control nested parallelism
Enables (true) or disables (false) nested parallelism. Default: false

OMP_NUM_THREADS Sets the maximum number of threads to use for OpenMP parallel regions if no other value is specified in the application.
The value can be a single integer, in which case it specifies the number of threads for all parallel regions. The value can also be a comma-separated list of integers, in which case each integer specifies the number of threads for a parallel region at that particular nesting level.
The first position in the list represents the outer-most parallel nesting level, the second position represents the next-inner parallel nesting level, and so on. At any level, the integer can be left out of the list. If the first integer in a list is left out, it implies the normal default value for threads is used at the outer-most level. If the integer is left out of any other level, the number of threads for that level is inherited from the previous level. Default: The number of processors visible to the operating system on which the program is executed. Syntax: OMP_NUM_THREADS=value[,value]* Example: OMP_NUM_THREADS=4,3

OMP_PLACES Specifies an explicit ordered list of places, either as an abstract name describing a set of places or as an explicit list of places described by non-negative numbers. An exclusion operator, !, can also be used to exclude the number or place immediately following the operator.
For explicit lists, an ordered list of places is specified with each o An explicit list of comma-separated non-negatives numbers Example: {0,2,4,6}
o An interval with notation <lower-bound>:<length>[:<stride>]. Example: {0:4:2}. When <stride> is omitted, a unit stride is assumed. The interval notation represents this set of numbers:
<lower-bound>, <lower-bound> + <stride>, ..., <lower-bound> + (<length> - 1) * <stride>
A place list can also be specified using the same interval notation: {place}:<length>[:<stride>]. This represents the list of length <length> places determined by the following:
{place}, {place} + <stride>, ..., {place} + (<length>-1)*<stride> Where given {place} and integer N, {place} + N = {place with every number offset by N} Example: {0,3,6}:4:1 represents {0,3,6}, {1,4,7}, {2,5,8}, {3,6,9}
Examples of explicit lists: These all represent the same set of places
OMP_PLACES="{0,1,2,3},{4,5,6,7},{8,9,10,11},{12,13,14,15}" OMP_PLACES="{0:4},{4:4},{8:4},{12:4}" OMP_PLACES="{0:4}:4:4"
NOTE: When specifying a place using a set of numbers, if any number cannot be mapped to a processor on the target platform, then that number is ignored within the place, but the rest of the place is kept intact. If all numbers within a place are invalid, then the entire place is removed from the place list, but the rest of place list is kept intact.
The abstract names listed below are understood by the run-time environment:
o threads: Each place corresponds to a single hardware thread.
o cores: Each place corresponds to a single core (having one or more hardware threads).
o sockets: Each place corresponds to a single socket (consisting of one or more cores).
o numa_domains: Each place corresponds to a single NUMA domain (consisting of one or more cores).
o ll_caches: Each place corresponds to a last-level cache (consisting of one or more cores).
The abstract name may be appended by a positive number in parentheses to denote the length of the place list to be created, that is abstract_name(num-places). If the optional number isn't specified, then the runtime will use all available resources of type abstract_name. When requesting fewer places than available on the system, the first available resources as determined by abstract_name are used. When requesting more places than available on the system, only the available resources are used.
Examples of abstract names:
as specifying KMP_AFFINITY=scatter. If disabled, this is the same as specifying KMP_AFFINITY=none.
Acceptable values: true, false, or a comma separated list, each element of which is one of the following values: master, close, spread, or primary.
Default: false
WARNING: master is deprecated. The semantics of master are the same as primary.
If set to false, the execution environment may move OpenMP threads between OpenMP places, thread affinity is disabled, and proc_bind clauses on parallel constructs are ignored. Otherwise, the execution environment should not move OpenMP threads between OpenMP places, thread affinity is enabled, and the initial thread is bound to the first place in the OpenMP place list.
If set to primary, all threads are bound to the same place as the primary thread.
If set to close, threads are bound to successive places, near where the primary thread is bound.
If set to spread, the primary thread's partition is subdivided and threads are bound to single place successive sub-partitions. Related environment variables: KMP_AFFINITY (overrides OMP_PROC_BIND).

OMP_SCHEDULE Sets the run-time schedule type and an optional chunk size. Default: static, no chunk size specified Syntax: OMP_SCHEDULE="kind[,chunk_size]"

OMP_STACKSIZE Sets the number of bytes to allocate for each OpenMP thread to use as the private stack for the thread. Recommended size is 16M.
Use the optional suffixes to specify byte units: B (bytes), K (Kilobytes), M (Megabytes), G (Gigabytes), or T (Terabytes) to specify the units. If you specify a value without a suffix, the byte unit is assumed to be K (Kilobytes).
This variable does not affect the native operating system threads created by the user program, or the thread executing the sequential part of an OpenMP program.
The kmp_{set,get}_stacksize_s() routines set/retrieve the value. The kmp_set_stacksize_s() routine must be called from sequential part, before first parallel region is created. Otherwise, calling kmp_set_stacksize_s() has no effect. Default:

o 32-bit architecture: 2M
program.
If this limit is reached and another native operating system thread encounters OpenMP API calls or constructs, the program can abort with an error message. If this limit is reached when an OpenMP parallel region begins, a one-time warning message might be generated indicating that the number of threads in the team was reduced, but the program will continue.
The omp_get_thread_limit() routine returns the value of the limit. Default: No enforced limit Related environment variable: KMP_ALL_THREADS (overrides OMP_THREAD_LIMIT).

OMP_WAIT_POLICY Decides whether threads spin (active) or yield (passive) while they are waiting. OMP_WAIT_POLICY=active is an alias for KMP_LIBRARY=turnaround, and OMP_WAIT_POLICY=passive is an alias for KMP_LIBRARY=throughput. Default: passive

NOTE: Although the default is passive, unless the user has explicitly set OMP_WAIT_POLICY, there is a small period of active spinning determined by KMP_BLOCKTIME.
KMP_AFFINITY (Windows, Linux) Enables run-time library to bind threads to physical processing units.
You must set this environment variable before the first parallel region, or certain API calls including omp_get_max_threads(), omp_get_num_procs() and any affinity API calls.
Syntax: KMP_AFFINITY=[<modifier>,...]<type>[,<permute>][,<offset>]
modifiers are optional strings consisting of a keyword and possibly a specifier
o respect (default) and norespect - determine whether to respect the original process affinity mask.
o verbose and noverbose (default) - determine whether to display affinity information.
o warnings (default) and nowarnings - determine whether to display warnings during affinity detection.
o reset and noreset (default) - determine whether to reset primary thread's affinity after outermost parallel region(s)
o granularity=<specifier> - takes the following specifiers thread, core (default), tile, socket, die, group (Windows only). The granularity describes the lowest topology levels that OpenMP threads are allowed to float within a topology map. For example, if granularity=core, then the OpenMP threads will be allowed to move between logical processors within a single core. If granularity=thread, then the OpenMP threads will be restricted to a single logical processor.
|<id_list> := | <proc_id> | | | | <proc_id>,<id_list> | +---------------+---------------------------+
Where each proc_id represents an operating system logical processor ID. For example, proclist=[3,0,{1,2},{0,3}] with OMP_NUM_THREADS=4 would place thread 0 on OS logical processor 3, thread 1 on OS logical processor 0, thread 2 on both OS logical processors 1 & 2, and thread 3 on OS logical processors 0 & 3.
type is the thread affinity policy to choose. Valid choices are none, balanced, compact, scatter, explicit, disabled
o type none (default) - Does not bind OpenMP threads to particular thread contexts; however, if the operating system supports affinity, the compiler still uses the OpenMP thread affinity interface to determine machine topology. Specify KMP_AFFINITY=verbose,none to list a machine topology map.
o type compact - Specifying compact assigns the OpenMP thread <n>+1 to a free thread context as close as possible to the thread context where the <n> OpenMP thread was placed. For example, in a topology map, the nearer a node is to the root, the more significance the node has when sorting the threads.
o type scatter - Specifying scatter distributes the threads as evenly as possible across the entire system. scatter is the opposite of compact; so the leaves of the node are most significant when sorting through the machine topology map.
o type balanced - Places threads on separate cores until all cores have at least one thread, similar to the scatter type. However, when the runtime must use multiple hardware thread contexts on the same core, the balanced type ensures that the OpenMP thread numbers are close to each other, which scatter does not do. This affinity type is supported on the CPU only for single socket systems.
o type explicit - Specifying explicit assigns OpenMP threads to a list of OS proc IDs that have been explicitly specified by using the proclist modifier, which is required for this affinity type.
o type disabled - Specifying disabled completely disables the thread affinity interfaces. This forces the OpenMP run-time library to behave as if the affinity interface was not supported by the operating system. This includes the low-level API interfaces such as kmp_set_affinity and kmp_get_affinity, which have no effect and will return a nonzero error code.
For both compact and scatter, permute and offset are allowed; however, if you specify only one integer, the runtime interprets the value as a permute specifier. Both permute and offset default to 0.
The permute specifier controls which levels are most significant when sorting the machine topology map. A value for permute forces the mappings to make the specified number of most significant levels of the sort the least significant, and it inverts the order of significance. The root node of the tree is not considered a separate level for the sort operations.
On Windows with multiple processor groups, the norespect affinity modifier is assumed when the process affinity mask equals a single processor group (which is default on Windows). Otherwise, the respect affinity modifier is used.
NOTE: On Windows with multiple processor groups, if the granularity is too coarse, it will be set to granularity=group. For example, if two processor groups exist across one socket, and granularity=socket the runtime will shift the granularity down to group since that is the largest granularity allowed by the OS.
KMP_ALL_THREADS Limits the number of simultaneously-executing threads in an OpenMP program. If this limit is reached and another native operating system thread encounters OpenMP API calls or constructs, then the program may abort with an error message. If this limit is reached at the time an OpenMP parallel region begins, a one-time warning message may be generated indicating that the number of threads in the team was reduced, but the program will continue execution. Default: No enforced limit. Related environment variable: OMP_THREAD_LIMIT (KMP_ALL_THREADS takes precedence)

KMP_BLOCKTIME Sets the time, in milliseconds, that a thread should wait, after completing the execution of a parallel region, before sleeping.
Use the optional character suffixes: s (seconds), m (minutes), h (hours), or d (days) to specify the units.
Specify infinite for an unlimited wait time. Default: 200 milliseconds Related Environment Variable: KMP_LIBRARY Example: KMP_BLOCKTIME=1s

KMP_CPUINFO_FILE Specifies an alternate file name for a file containing the machine topology description. The file must be in the same format as /proc/cpuinfo.
Default: None
KMP_DETERMINISTIC_REDUCTION Enables (true) or disables (false) the use of a specific ordering of the reduction operations for implementing the reduction clause for an OpenMP parallel region. This has the effect that, for a given number of threads, in a given parallel region, for a given data set and reduction operation, a floating point reduction done for an OpenMP reduction clause has a consistent floating point result from run to run, since round-off errors are identical. Default: false Example: KMP_DETERMINISTIC_REDUCTION=true

KMP_DYNAMIC_MODE Selects the method used to determine the number of threads to use for a parallel region when OMP_DYNAMIC=true. Possible values: (load_balance |
Default: load_balance (on all supported platforms)
KMP_HOT_TEAMS_MAX_LEVEL Sets the maximum nested level to which teams of threads will be hot.
NOTE: A hot team is a team of threads optimized for faster reuse by subsequent parallel regions. In a hot team, threads are kept ready for execution of the next parallel region, in contrast to the cold team, which is freed after each parallel region, with its threads going into a common pool of threads.
For values of 2 and above, nested parallelism should be enabled.
Default: 1
KMP_HOT_TEAMS_MODE Specifies the run-time behavior when the number of threads in a hot team is reduced. Possible values:
o 0 - Extra threads are freed and put into a common pool of threads.
o 1 - Extra threads are kept in the team in reserve, for faster reuse in subsequent parallel regions.
Default: 0
KMP_HW_SUBSET Specifies the subset of available hardware resources for the hardware topology hierarchy. The subset is specified in terms of number of units per upper layer unit starting from top layer downwards. E.g. the number of sockets (top layer units), cores per socket, and the threads per core, to use with an OpenMP application, as an alternative to writing complicated explicit affinity settings or a limiting process affinity mask. You can also specify an offset value to set which resources to use. When available, you can specify attributes to select different subsets of resources.
An extended syntax is available when KMP_TOPOLOGY_METHOD=hwloc. Depending on what resources are detected, you may be able to specify additional resources, such as NUMA domains and groups of hardware resources that share certain cache levels.
Basic syntax: [num_units|*]ID[@offset][:attribute] [,[num_units|*]ID[@offset][:attribute]...]
Supported unit IDs are not case-insensitive. S - socket num_units specifies the requested number of sockets.
D - die num_units specifies the requested number of dies per socket.
C - core num_units specifies the requested number of cores per die - if any - otherwise, per socket.
T - thread num_units specifies the requested number of HW threads per core. offset - (Optional) The number of units to skip.
attribute - (Optional) An attribute differentiating resources at a particular level. The attributes available to users are:
o Core type - On Intel architectures, this can be intel_atom or intel_core
o Core efficiency - This is specified as effnum where num is a number from 0 to the number of core efficiencies detected in the machine topology minus one. E.g., eff0. The greater the efficiency number the more performant the core. There may be more core efficiencies than core types and can be viewed by setting KMP_AFFINITY=verbose
NOTE: The hardware cache can be specified as a unit, e.g. L2 for L2 cache, or LL for last level cache.
Extended syntax when KMP_TOPOLOGY_METHOD=hwloc:
Additional IDs can be specified if detected. For example:
N - numa num_units specifies the requested number of NUMA nodes per upper layer unit, e.g. per socket.
TI - tile num_units specifies the requested number of tiles to use per upper layer unit, e.g. per NUMA node.
When any numa or tile units are specified in KMP_HW_SUBSET and the hwloc topology method is available, the KMP_TOPOLOGY_METHOD will be automatically set to hwloc, so there is no need to set it explicitly.
If you don't specify one or more types of resource, such as socket or thread, all available resources of that type are used.
The run-time library prints a warning, and the setting of KMP_HW_SUBSET is ignored if:
o a resource is specified, but detection of that resource is not supported by the chosen topology detection method and/or
o a resource is specified twice. An exception to this condition is if attributes differentiate the resource.
o attributes are used when not detected in the machine topology or conflict with each other.
This variable does not work if KMP_AFFINITY=disabled.
Default: If omitted, the default value is to use all the available hardware resources.
Examples:
o 2s,4c,2t: Use the first 2 sockets (s0 and s1), the first 4 cores on each socket (c0 - c3), and 2 threads per core.
o 2s@2,4c@8,2t: Skip the first 2 sockets (s0 and s1) and use 2 sockets (s2-s3), skip the first 8 cores (c0-c7) and use 4 cores on each
o 1s, 1d, 1n, 1c, 1t: Use 1 socket, 1 die, 1 NUMA node, 1 core, 1 thread - use HW thread as a result.
o 4c:intel_atom,5c:intel_core: Use all available sockets and use 4 Intel Atom(R) processor cores and 5 Intel(R) Core(TM) processor cores per socket.
o 2c:eff0@1,3c:eff1: Use all available sockets, skip the first core with efficiency 0 and use the next 2 cores with efficiency 0 and 3 cores with efficiency 1 per socket.
o 1s, 1c, 1t: Use 1 socket, 1 core, 1 thread. This may result in using single thread on a 3-layer topology architecture, or multiple threads on 4-layer or 5-layer architecture. Result may even be different on the same architecture, depending on KMP_TOPOLOGY_METHOD specified, as hwloc can often detect more topology layers than the default method used by the OpenMP run-time library.
o *c:eff1@3: Use all available sockets, skip the first three cores of efficiency 1, and then use the rest of the available cores of efficiency 1.
To see the result of the setting, you can specify verbose modifier in KMP_AFFINITY environment variable. The OpenMP run-time library will output to stderr the information about the discovered hardware topology before and after the KMP_HW_SUBSET setting was applied.
KMP_INHERIT_FP_CONTROL Enables (true) or disables (false) the copying of the floating-point control settings of the primary thread to the floating-point control settings of the OpenMP worker threads at the start of each parallel region.
Default: true
KMP_LIBRARY Selects the OpenMP run-time library execution mode. The values for this variable are serial, turnaround, or throughput. Default: throughput Related environment variable: KMP_BLOCKTIME and OMP_WAIT_POLICY

KMP_SETTINGS Enables (true) or disables (false) the printing of OpenMP run-time library environment variables during program execution. Two lists of variables are printed: user-defined environment variables settings and effective values of variables used by OpenMP run-time library.
Default: false
KMP_STACKSIZE Sets the number of bytes to allocate for each OpenMP thread to use as its private stack.
Recommended size is 16M.
Use the optional suffixes to specify byte units: B (bytes), K (Kilobytes), M (Megabytes), G (Gigabytes), or T (Terabytes) to specify the units. If you specify a value without a suffix, the byte unit is o 32-bit architectures: 2M
o 64-bit architectures: 4M
KMP_TOPOLOGY_METHOD Forces OpenMP to use a particular machine topology modeling method.
Possible values are:
o all - Let OpenMP choose which topology method is most appropriate based on the platform and possibly other environment variable settings.
o cpuid_leaf31 (x86 only) - Decodes the APIC identifiers as specified by leaf 31 of the cpuid instruction. The runtime will produce an error if the machine does not support leaf 31.
o cpuid_leaf11 (x86 only) - Decodes the APIC identifiers as specified by leaf 11 of the cpuid instruction. The runtime will produce an error if the machine does not support leaf 11.
o cpuid_leaf4 (x86 only) - Decodes the APIC identifiers as specified in leaf 4 of the cpuid instruction. The runtime will produce an error if the machine does not support leaf 4.
o cpuinfo - If KMP_CPUINFO_FILE is not specified, forces OpenMP to parse /proc/cpuinfo to determine the topology (Linux only). If KMP_CPUINFO_FILE is specified as described above, uses it (Windows or Linux).
o group - Models the machine as a 2-level map, with level 0 specifying the different processors in a group, and level 1 specifying the different groups (Windows 64-bit only).
NOTE: Support for group is now deprecated and will be removed in a future release. Use all instead.
o flat - Models the machine as a flat (linear) list of processors.
o hwloc - Models the machine as the Portable Hardware Locality (hwloc) library does. This model is the most detailed and includes, but is not limited to: numa domains, packages, cores, hardware threads, caches, and Windows processor groups. This method is only available if you have configured libomp to use hwloc during CMake configuration.
Default: all
KMP_VERSION Enables (true) or disables (false) the printing of OpenMP run-time library version information during program execution.
Default: false
KMP_WARNINGS Enables (true) or disables (false) displaying warnings from the OpenMP run-time library during program execution.
full list of supported environment variables is defined below.
o LIBOMPTARGET_DEBUG=<Num>
o LIBOMPTARGET_PROFILE=<Filename>
o LIBOMPTARGET_MEMORY_MANAGER_THRESHOLD=<Num>
o LIBOMPTARGET_INFO=<Num>
o LIBOMPTARGET_HEAP_SIZE=<Num>
o LIBOMPTARGET_STACK_SIZE=<Num>
o LIBOMPTARGET_SHARED_MEMORY_SIZE=<Num>
o LIBOMPTARGET_MAP_FORCE_ATOMIC=[TRUE/FALSE] (default TRUE)
LIBOMPTARGET_DEBUG LIBOMPTARGET_DEBUG controls whether or not debugging information will be displayed. This feature is only availible if libomptarget was built with -DOMPTARGET_DEBUG. The debugging output provided is intended for use by libomptarget developers. More user-friendly output is presented when using LIBOMPTARGET_INFO.
LIBOMPTARGET_PROFILE LIBOMPTARGET_PROFILE allows libomptarget to generate time profile output similar to Clang's -ftime-trace option. This generates a JSON file based on Chrome Tracing that can be viewed with chrome://tracing or the Speedscope App. Building this feature depends on the LLVM Support Library for time trace output. Using this library is enabled by default when building using the CMake option OPENMP_ENABLE_LIBOMPTARGET_PROFILING. The output will be saved to the filename specified by the environment variable. For multi-threaded applications, profiling in libomp is also needed. Setting the CMake option OPENMP_ENABLE_LIBOMP_PROFILING=ON to enable the feature. Note that this will turn libomp into a C++ library.
LIBOMPTARGET_MEMORY_MANAGER_THRESHOLD LIBOMPTARGET_MEMORY_MANAGER_THRESHOLD sets the threshold size for which the libomptarget memory manager will handle the allocation. Any allocations larger than this threshold will not use the memory manager and be freed after the device kernel exits. The default threshold value is 8KB. If LIBOMPTARGET_MEMORY_MANAGER_THRESHOLD is set to 0 the memory manager will be completely disabled.
LIBOMPTARGET_INFO LIBOMPTARGET_INFO allows the user to request different types of runtime information from libomptarget. LIBOMPTARGET_INFO uses a 32-bit field to enable or disable different types of information. This includes information about data-mappings and kernel execution. It is recommended to build your application with debugging information enabled, this will enable filenames and variable declarations in the information messages. OpenMP Debugging information is enabled at any level of debugging so a full debug runtime is not required. For minimal debugging information compile with -gline-tables-only, or compile with -g for full debug information. A full list of flags supported by LIBOMPTARGET_INFO is given below.

o Indicate when an entry is changed in the device mapping table: 0x08
o Print OpenMP kernel information from device plugins: 0x10
o Indicate when data is copied to and from the device: 0x20
Any combination of these flags can be used by setting the appropriate bits. For example, to enable printing all data active in an OpenMP target region along with CUDA information, run the following bash command.
$ env LIBOMPTARGET_INFO=$((0x1 | 0x10)) ./your-application
Or, to enable every flag run with every bit set.
$ env LIBOMPTARGET_INFO=-1 ./your-application
For example, given a small application implementing the ZAXPY BLAS routine, Libomptarget can provide useful information about data mappings and thread usages.
#include <complex>
using complex = std::complex<double>;
void zaxpy(complex *X, complex *Y, complex D, std::size_t N) { #pragma omp target teams distribute parallel for for (std::size_t i = 0; i < N; ++i) Y[i] = D * X[i] + Y[i]; }
int main() { const std::size_t N = 1024; complex X[N], Y[N], D; #pragma omp target data map(to:X[0 : N]) map(tofrom:Y[0 : N]) zaxpy(X, Y, D, N); }
Compiling this code targeting nvptx64 with all information enabled will provide the following output from the runtime library.
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O3 -gline-tables-only zaxpy.cpp -o zaxpy $ env LIBOMPTARGET_INFO=-1 ./zaxpy
Info: Entering OpenMP data region at zaxpy.cpp:14:1 with 2 arguments: Info: to(X[0:N])[16384] Info: tofrom(Y[0:N])[16384] Info: Creating new map entry with HstPtrBegin=0x00007fff0d259a40, TgtPtrBegin=0x00007fdba5800000, Size=16384, RefCount=1, Name=X[0:N] Info: Copying data from host to device, HstPtr=0x00007fff0d259a40, TgtPtr=0x00007fdba5800000, Size=16384, Name=X[0:N] Info: Creating new map entry with HstPtrBegin=0x00007fff0d255a40, TgtPtrBegin=0x00007fdba5804000, Size=16384, RefCount=1, Name=Y[0:N] Info: Copying data from host to device, HstPtr=0x00007fff0d255a40, TgtPtr=0x00007fdba5804000, Size=16384, Name=Y[0:N] Info: OpenMP Host-Device pointer mappings after block at zaxpy.cpp:14:1: Info: Host Ptr Target Ptr Size (B) RefCount Declaration Info: Mapping exists (implicit) with HstPtrBegin=0x00007fff0d255a40, TgtPtrBegin=0x00007fdba5804000, Size=0, RefCount=2 (incremented), Name=Y Info: Creating new map entry with HstPtrBegin=0x00007fff0d2559f0, TgtPtrBegin=0x00007fdba5808000, Size=16, RefCount=1, Name=D Info: Copying data from host to device, HstPtr=0x00007fff0d2559f0, TgtPtr=0x00007fdba5808000, Size=16, Name=D Info: Mapping exists (implicit) with HstPtrBegin=0x00007fff0d259a40, TgtPtrBegin=0x00007fdba5800000, Size=0, RefCount=2 (incremented), Name=X Info: Mapping exists with HstPtrBegin=0x00007fff0d255a40, TgtPtrBegin=0x00007fdba5804000, Size=0, RefCount=2 (update suppressed) Info: Mapping exists with HstPtrBegin=0x00007fff0d2559f0, TgtPtrBegin=0x00007fdba5808000, Size=16, RefCount=1 (update suppressed) Info: Mapping exists with HstPtrBegin=0x00007fff0d259a40, TgtPtrBegin=0x00007fdba5800000, Size=0, RefCount=2 (update suppressed) Info: Launching kernel __omp_offloading_10305_c08c86__Z5zaxpyPSt7complexIdES1_S0_m_l6 with 8 blocks and 128 threads in SPMD mode Info: Mapping exists with HstPtrBegin=0x00007fff0d259a40, TgtPtrBegin=0x00007fdba5800000, Size=0, RefCount=1 (decremented) Info: Mapping exists with HstPtrBegin=0x00007fff0d2559f0, TgtPtrBegin=0x00007fdba5808000, Size=16, RefCount=1 (deferred final decrement) Info: Copying data from device to host, TgtPtr=0x00007fdba5808000, HstPtr=0x00007fff0d2559f0, Size=16, Name=D Info: Mapping exists with HstPtrBegin=0x00007fff0d255a40, TgtPtrBegin=0x00007fdba5804000, Size=0, RefCount=1 (decremented) Info: Removing map entry with HstPtrBegin=0x00007fff0d2559f0, TgtPtrBegin=0x00007fdba5808000, Size=16, Name=D Info: OpenMP Host-Device pointer mappings after block at zaxpy.cpp:6:1: Info: Host Ptr Target Ptr Size (B) RefCount Declaration Info: 0x00007fff0d255a40 0x00007fdba5804000 16384 1 Y[0:N] at zaxpy.cpp:13:17 Info: 0x00007fff0d259a40 0x00007fdba5800000 16384 1 X[0:N] at zaxpy.cpp:13:11 Info: Exiting OpenMP data region at zaxpy.cpp:14:1 with 2 arguments: Info: to(X[0:N])[16384] Info: tofrom(Y[0:N])[16384] Info: Mapping exists with HstPtrBegin=0x00007fff0d255a40, TgtPtrBegin=0x00007fdba5804000, Size=16384, RefCount=1 (deferred final decrement) Info: Copying data from device to host, TgtPtr=0x00007fdba5804000, HstPtr=0x00007fff0d255a40, Size=16384, Name=Y[0:N] Info: Mapping exists with HstPtrBegin=0x00007fff0d259a40, TgtPtrBegin=0x00007fdba5800000, Size=16384, RefCount=1 (deferred final decrement) Info: Removing map entry with HstPtrBegin=0x00007fff0d255a40, TgtPtrBegin=0x00007fdba5804000, Size=16384, Name=Y[0:N] Info: Removing map entry with HstPtrBegin=0x00007fff0d259a40, TgtPtrBegin=0x00007fdba5800000, Size=16384, Name=X[0:N]
From this information, we can see the OpenMP kernel being launched on the CUDA device with enough threads and blocks for all 1024 iterations of the loop in simplified SPMD Mode. The information from the OpenMP data region shows the two arrays X and Y being copied from the host to the device. This creates an entry in the host-device mapping table associating the host pointers to the newly created device data. The data mappings in the OpenMP device kernel show the default mappings being used for all the variables used implicitly on the device. Because X and Y are already mapped in the device's table, no new entries are created. Additionally, the default mapping shows that D will be copied back from the device once the OpenMP device kernel region ends even though it isn't written to. Finally, at the end of the OpenMP data region the entries for X and Y are removed from the table.
The information level can be controlled at runtime using an internal extern foo();
int main() { __tgt_set_info_flag(0x10); #pragma omp target foo(); }
Errors: libomptarget provides error messages when the program fails inside the OpenMP target region. Common causes of failure could be an invalid pointer access, running out of device memory, or trying to offload when the device is busy. If the application was built with debugging symbols the error messages will additionally provide the source location of the OpenMP target region.
For example, consider the following code that implements a simple parallel reduction on the GPU. This code has a bug that causes it to fail in the offloading region.
#include <cstdio>
double sum(double *A, std::size_t N) { double sum = 0.0; #pragma omp target teams distribute parallel for reduction(+:sum) for (int i = 0; i < N; ++i) sum += A[i];
return sum; }
int main() { const int N = 1024; double A[N]; sum(A, N); }
If this code is compiled and run, there will be an error message indicating what is going wrong.
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O3 -gline-tables-only sum.cpp -o sum $ ./sum
CUDA error: an illegal memory access was encountered Libomptarget error: Copying data from device failed. Libomptarget error: Call to targetDataEnd failed, abort target. Libomptarget error: Failed to process data after launching the kernel. Libomptarget error: Consult https://openmp.llvm.org/design/Runtimes.html for debugging options. sum.cpp:5:1: Libomptarget error 1: failure of target construct while offloading is mandatory
This shows that there is an illegal memory access occuring inside the OpenMP target region once execution has moved to the CUDA device, suggesting a segmentation fault. This then causes a chain reaction of failures in libomptarget. Another message suggests using the LIBOMPTARGET_INFO environment variable as described in Environment Variables. If we do this it will print the sate of the host-target pointer mappings at the time of failure.
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O3 -gline-tables-only sum.cpp -o sum
is the sum variable that will be copied back from the device once the reduction has ended. There is no entry mapping the host array A to the device. In this situation, the compiler cannot determine the size of the array at compile time so it will simply assume that the pointer is mapped on the device already by default. The solution is to add an explicit map clause in the target region.
double sum(double *A, std::size_t N) { double sum = 0.0; #pragma omp target teams distribute parallel for reduction(+:sum) map(to:A[0 : N]) for (int i = 0; i < N; ++i) sum += A[i];
return sum; }
LIBOMPTARGET_STACK_SIZE This environment variable sets the stack size in bytes for the CUDA plugin. This can be used to increase or decrease the standard amount of memory reserved for each thread's stack.
LIBOMPTARGET_HEAP_SIZE This environment variable sets the amount of memory in bytes that can be allocated using malloc and free for the CUDA plugin. This is necessary for some applications that allocate too much memory either through the user or globalization.
LIBOMPTARGET_SHARED_MEMORY_SIZE This environment variable sets the amount of dynamic shared memory in bytes used by the kernel once it is launched. A pointer to the dynamic memory buffer can be accessed using the llvm_omp_target_dynamic_shared_alloc function. An example is shown in Dynamic Shared Memory.
OpenMP in LLVM --- Offloading Design OpenMP Target Offloading --- SPMD Mode OpenMP Target Offloading --- Generic Mode LIBOMPTARGET_MAP_FORCE_ATOMIC The OpenMP standard guarantees that map clauses are atomic. However, the this can have a drastic performance impact. Users that do not require atomic map clauses can disable them to potentially recover lost performance. As a consequence, users have to guarantee themselves that no two map clauses will concurrently map the same memory. If the memory is already mapped and the map clauses will only modify the reference counter from a non-zero count to another non-zero count, concurrent map clauses are supported regardless of this option. To disable forced atomic map clauses use "false"/"FALSE" as the value of the LIBOMPTARGET_MAP_FORCE_ATOMIC environment variable. The default behavior of LLVM 14 is to force atomic maps clauses, prior versions of LLVM did not.
LLVM/OpenMP Target Host Runtime Plugins (libomptarget.rtl.XXXX) Remote Offloading Plugin: The remote offloading plugin permits the execution of OpenMP target regions on devices in remote hosts in addition to the devices connected to the local host. All target devices on the remote host will be exposed to the application as if they were local devices, that is, the remote host CPU or its GPUs can be offloaded to with the appropriate device number. If the server is running on the same host, each device mapping during offloading. These are implemented via gRPC/protobuf so these libraries are required to build and use this plugin. The server must also have access to the necessary target-specific plugins in order to perform the offloading.
Due to the experimental nature of this plugin, the CMake variable LIBOMPTARGET_ENABLE_EXPERIMENTAL_REMOTE_PLUGIN must be set in order to build this plugin. For example, the rpc plugin is not designed to be thread-safe, the server cannot concurrently handle offloading from multiple applications at once (it is synchronous) and will terminate after a single execution. Note that openmp-offloading-server is unable to remote offload onto a remote host itself and will error out if this is attempted.
Remote offloading is configured via environment variables at runtime of the OpenMP application:
o LIBOMPTARGET_RPC_ADDRESS=<Address>:<Port>
o LIBOMPTARGET_RPC_ALLOCATOR_MAX=<NumBytes>
o LIBOMPTARGET_BLOCK_SIZE=<NumBytes>
o LIBOMPTARGET_RPC_LATENCY=<Seconds>
LIBOMPTARGET_RPC_ADDRESS The address and port at which the server is running. This needs to be set for the server and the application, the default is 0.0.0.0:50051. A single OpenMP executable can offload onto multiple remote hosts by setting this to comma-seperated values of the addresses.
LIBOMPTARGET_RPC_ALLOCATOR_MAX After allocating this size, the protobuf allocator will clear. This can be set for both endpoints.
LIBOMPTARGET_BLOCK_SIZE This is the maximum size of a single message while streaming data transfers between the two endpoints and can be set for both endpoints.
LIBOMPTARGET_RPC_LATENCY This is the maximum amount of time the client will wait for a response from the server.
LLVM/OpenMP Target Device Runtime (libomptarget-ARCH-SUBARCH.bc) The target device runtime is an LLVM bitcode library that implements OpenMP runtime functions on the target device. It is linked with the device code's LLVM IR during compilation.
Dynamic Shared Memory The target device runtime contains a pointer to the dynamic shared memory buffer. This pointer can be obtained using the llvm_omp_target_dynamic_shared_alloc extension. If this function is called from the host it will simply return a null pointer. In order to use this buffer the kernel must be launched with an adequate amount of dynamic shared memory allocated. Currently this is done using the LIBOMPTARGET_SHARED_MEMORY_SIZE environment variable. An example is given below.
void foo() { #pragma omp barrier if (omp_get_thread_num() == 1) x = *buf; } }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 shared.c $ env LIBOMPTARGET_SHARED_MEMORY_SIZE=256 ./shared
Debugging The device runtime supports debugging in the runtime itself. This is configured at compile-time using the flag -fopenmp-target-debug=<N> rather than using a separate debugging build. If debugging is not enabled, the debugging paths will be considered trivially dead and removed by the compiler with zero overhead. Debugging is enabled at runtime by running with the environment variable LIBOMPTARGET_DEVICE_RTL_DEBUG=<N> set. The number set is a 32-bit field used to selectively enable and disable different features. Currently, the following debugging features are supported.
o Enable debugging assertions in the device. 0x01
o Enable OpenMP runtime function traces in the device. 0x2
o Enable diagnosing common problems during offloading . 0x4
void copy(double *X, double *Y) { #pragma omp target teams distribute parallel for for (std::size_t i = 0; i < N; ++i) Y[i] = X[i]; }
Compiling this code targeting nvptx64 with debugging enabled will provide the following output from the device runtime library.
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -fopenmp-target-debug=3 $ env LIBOMPTARGET_DEVICE_RTL_DEBUG=3 ./zaxpy
Kernel.cpp:70: Thread 0 Entering int32_t __kmpc_target_init() Parallelism.cpp:196: Thread 0 Entering int32_t __kmpc_global_thread_num() Mapping.cpp:239: Thread 0 Entering uint32_t __kmpc_get_hardware_num_threads_in_block() Workshare.cpp:616: Thread 0 Entering void __kmpc_distribute_static_init_4() Parallelism.cpp:85: Thread 0 Entering void __kmpc_parallel_51() Parallelism.cpp:69: Thread 0 Entering <OpenMP Outlined Function> Workshare.cpp:575: Thread 0 Entering void __kmpc_for_static_init_4() Workshare.cpp:660: Thread 0 Entering void __kmpc_distribute_static_fini() Workshare.cpp:660: Thread 0 Entering void __kmpc_distribute_static_fini() Kernel.cpp:103: Thread 0 Entering void __kmpc_target_deinit()
OpenACC support is under development for both Flang and Clang. For this purpose, LLVM's OpenMP runtimes are being extended to serve as OpenACC runtimes. In some cases, Clang supports OpenMP extensions to make the additional functionality also available in OpenMP applications.
OPENACC SUPPORT OpenACC support is under development for both Flang and Clang. For this purpose, LLVM's OpenMP runtimes are being extended to serve as OpenACC runtimes. By default, Clang recognizes these extensions. The command-line option -fno-openmp-extensions can be specified to disable all OpenMP extensions, including those described in this section.
Motivation There are multiple benefits to exposing OpenACC functionality as LLVM OpenMP extensions:
o OpenMP applications can take advantage of the additional functionality.
o As LLVM's implementation of these extensions matures, it can serve as a basis for including these extensions in the OpenMP standard.
o Source-to-source translation from certain OpenACC features to OpenMP is otherwise impossible.
o Runtime tests can be written in terms of OpenMP instead of OpenACC or low-level runtime calls.
o More generally, there is a clean separation of concerns between OpenACC and OpenMP development in LLVM. That is, LLVM's OpenMP developers can discuss, modify, and debug LLVM's extended OpenMP implementation and test suite without directly considering OpenACC's language and execution model, which are handled by LLVM's OpenACC developers.
ompx_hold Map Type Modifier Example
#pragma omp target data map(ompx_hold, tofrom: x) // holds onto mapping of x throughout region { foo(); // might have map(delete: x) #pragma omp target map(present, alloc: x) // x is guaranteed to be present printf("%d\n", x); }
The ompx_hold map type modifier above specifies that the target data directive holds onto the mapping for x throughout the associated region regardless of any target exit data directives executed during the call to foo. Thus, the presence assertion for x at the enclosed target construct cannot fail.
Behavior
o Stated more generally, the ompx_hold map type modifier specifies that the associated data is not unmapped until the end of the construct. As usual, the standard OpenMP reference count for the data must also reach zero before the data is unmapped.
o If ompx_hold is specified for the same data on lexically or dynamically enclosed constructs, there is no additional effect as the data mapping is already held throughout their regions.
o The ompx_hold map type modifier is permitted to appear only on target constructs (and associated combined constructs) and target data constructs. It is not permitted to appear on target enter data or target exit data directives because there is no associated statement, so it is not meaningful to hold onto a mapping until the end of the struct even if other map clauses on the same directive specify other members without the ompx_hold map type modifier.
o ompx_hold support is not yet provided for defaultmap.
Implementation
o LLVM uses the term dynamic reference count for the standard OpenMP reference count for host/device data mappings.
o The ompx_hold map type modifier selects an alternate reference count, called the hold reference count.
o A mapping is removed only once both its reference counts reach zero.
o Because ompx_hold can appear only constructs, increments and decrements of the hold reference count are guaranteed to be balanced, so it is impossible to decrement it below zero.
o The dynamic reference count is used wherever ompx_hold is not specified (and possibly cannot be specified). Decrementing the dynamic reference count has no effect if it is already zero.
o The runtime determines that the ompx_hold map type modifier is in effect (see Behavior above) when the hold reference count is greater than zero.
Relationship with OpenACC OpenACC specifies two reference counts for tracking host/device data mappings. Which reference count is used to implement an OpenACC directive is determined by the nature of that directive, either dynamic or structured:
o The dynamic reference count is always used for enter data and exit data directives and corresponding OpenACC routines.
o The structured reference count is always used for data and compute constructs, which are similar to OpenMP's target data and target constructs.
Contrast with OpenMP, where the dynamic reference count is always used unless the application developer specifies an alternate behavior via our map type modifier extension. We chose the name hold for that map type modifier because, as demonstrated in the above example, hold concisely identifies the desired behavior from the application developer's perspective without referencing the implementation of that behavior.
The hold reference count is otherwise modeled after OpenACC's structured reference count. For example, calling acc_unmap_data, which is similar to omp_target_disassociate_ptr, is an error when the structured reference count is not zero.
While Flang and Clang obviously must implement the syntax and semantics for selecting OpenACC reference counts differently than for selecting OpenMP reference counts, the implementation is the same at the runtime level. That is, OpenACC's dynamic reference count is OpenMP's dynamic reference count, and OpenACC's structured reference count is our OpenMP hold reference count extension.
#pragma omp target teams map(tofrom:x) #pragma omp atomic update x++;
Relationship with OpenACC This extension is important when translating OpenACC to OpenMP because OpenACC does not have the same restriction for its corresponding constructs. For example, the following is conforming OpenACC:
#pragma acc parallel copy(x) #pragma acc atomic update x++;
LLVM, since version 11 (12 Oct 2020), has an OpenMP-Aware optimization pass as well as the ability to perform "scalar optimizations" across OpenMP region boundaries.
In-depth discussion of the topic can be found here.
OPENMP OPTIMIZATIONS IN LLVM LLVM, since version 11 (12 Oct 2020), has an OpenMP-Aware optimization pass as well as the ability to perform "scalar optimizations" across OpenMP region boundaries.
OpenMP-Aware Optimizations LLVM, since version 11 (12 Oct 2020), supports an OpenMP-Aware optimization pass. This optimization pass will attempt to optimize the module with OpenMP-specific domain-knowledge. This pass is enabled by default at high optimization levels (O2 / O3) if compiling with OpenMP support enabled.
OpenMPOpt
o OpenMP Runtime Call Deduplication
o Globalization
OpenMPOpt contains several OpenMP-Aware optimizations. This pass is run early on the entire Module, and later on the entire call graph. Most optimizations done by OpenMPOpt support remarks. Optimization remarks can be enabled by compiling with the following flags.
$ clang -Rpass=openmp-opt -Rpass-missed=openmp-opt -Rpass-analysis=openmp-opt
OpenMP Runtime Call Deduplication The OpenMP runtime library contains several functions used to implement features of the OpenMP standard. Several of the runtime calls are constant within a parallel region. A common optimization is to replace invariant code with a single reference, but in this case the compiler will only see an opaque call into the runtime library. To get around this, OpenMPOpt maintains a list of OpenMP runtime functions that are constant and will manually deduplicate them.
Globalization The OpenMP standard requires that data can be shared between different threads. This requirement poses a unique challenge when offloading to GPU accelerators. Data cannot be shared between the threads in a GPU by default, in order to do this it must either be placed in global or
void use(void *) { }
void foo() { int x; use(&x); }
int main() { #pragma omp target parallel foo(); }
In many cases, this transformation is not actually necessary but still carries a significant performance penalty. Because of this, OpenMPOpt can perform and inter-procedural optimization and scan each known usage of the globalized variable and determine if it is potentially captured and shared by another thread. If it is not actually captured, it can safely be moved back to fast register memory.
Another case is memory that is intentionally shared between the threads, but is shared from one thread to all the others. Such variables can be moved to shared memory when compiled without needing to go through the runtime library. This allows for users to confidently declare shared memory on the device without needing to use custom OpenMP allocators or rely on the runtime.
static void share(void *);
static void foo() { int x[64]; #pragma omp parallel share(x); }
int main() { #pragma omp target foo(); }
These optimizations can have very large performance implications. Both of these optimizations rely heavily on inter-procedural analysis. Because of this, offloading applications should ideally be contained in a single translation unit and functions should not be externally visible unless needed. OpenMPOpt will inform the user if any globalization calls remain if remarks are enabled. This should be treated as a defect in the program.
Resources
o 2021 OpenMP Webinar: "A Compiler's View of OpenMP" https://youtu.be/eIMpgez61r4
o 2020 LLVM Developers' Meeting: "(OpenMP) Parallelism-Aware Optimizations" https://youtu.be/gtxWkeLCxmU
o 2019 EuroLLVM Developers' Meeting: "Compiler Optimizations for (OpenMP) Target Offloading to GPUs" https://youtu.be/3AbS82C3X30
Inter-procedural Fixpoint Iteration Framework" https://youtu.be/CzWkc_JcfS0
LLVM has an elaborate ecosystem around analysis and optimization remarks issues during compilation. The remarks can be enabled from the clang frontend [1] [2] in various formats [3] [4] to be used by tools, i.a., opt-viewer or llvm-opt-report (dated).
The OpenMP optimizations in LLVM have been developed with remark support as a priority. For a list of OpenMP specific remarks and more information on them, please refer to OpenMP Optimization Remarks.
o [1] https://clang.llvm.org/docs/UsersManual.html#options-to-emit-optimization-reports
o [2] https://clang.llvm.org/docs/ClangCommandLineReference.html#diagnostic-flags
o [3] https://clang.llvm.org/docs/ClangCommandLineReference.html#cmdoption-clang-foptimization-record-file
o [4] https://clang.llvm.org/docs/ClangCommandLineReference.html#cmdoption-clang1-fsave-optimization-record
OPENMP OPTIMIZATION REMARKS The OpenMP-Aware optimization pass is able to generate compiler remarks for performed and missed optimisations. To emit them, pass these options to the Clang invocation: -Rpass=openmp-opt -Rpass-analysis=openmp-opt -Rpass-missed=openmp-opt. For more information and features of the remark system, consult the clang documentation:
o Clang options to emit optimization reports
o Clang diagnostic and remark flags
o The -foptimization-record-file flag and the -fsave-optimization-record flag
OpenMP Remarks Potentially unknown OpenMP target region caller [OMP100] A function remark that indicates the function, when compiled for a GPU, is potentially called from outside the translation unit. Note that a remark is only issued if we tried to perform an optimization which would require us to know all callers on the GPU.
To facilitate OpenMP semantics on GPUs we provide a runtime mechanism through which the code that makes up the body of a parallel region is shared with the threads in the team. Generally we use the address of the outlined parallel region to identify the code that needs to be executed. If we know all target regions that reach the parallel region we can avoid this function pointer passing scheme and often improve the register usage on the GPU. However, If a parallel region on the GPU is in a function with external linkage we may not know all callers statically. If there are outside callers within target regions, this remark is to be ignored. If there are no such callers, users can modify the linkage and thereby help optimization with a static or __attribute__((internal)) function annotation. If changing the linkage is impossible, e.g., because there are outside callers on the host, one An analysis remark that indicates that a parallel region has unknown calls.
Parallel region is not called from a unique kernel. Will not attempt to rewrite the state machine. [OMP102] This analysis remark indicates that a given parallel region is called by multiple kernels. This prevents the compiler from optimizing it to a single kernel and rewrite the state machine.
Moving globalized variable to the stack. [OMP110] This optimization remark indicates that a globalized variable was moved back to thread-local stack memory on the device. This occurs when the optimization pass can determine that a globalized variable cannot possibly be shared between threads and globalization was ultimately unnecessary. Using stack memory is the best-case scenario for data globalization as the variable can now be stored in fast register files on the device. This optimization requires full visibility of each variable.
Globalization typically occurs when a pointer to a thread-local variable escapes the current scope. The compiler needs to be pessimistic and assume that the pointer could be shared between multiple threads according to the OpenMP standard. This is expensive on target offloading devices that do not allow threads to share data by default. Instead, this data must be moved to memory that can be shared, such as shared or global memory. This optimization moves the data back from shared or global memory to thread-local stack memory if the data is not actually shared between the threads.
Examples A trivial example of globalization occurring can be seen with this example. The compiler sees that a pointer to the thread-local variable x escapes the current scope and must globalize it even though it is not actually necessary. Fortunately, this optimization can undo this by looking at its usage.
void use(int *x) { }
void foo() { int x; use(&x); }
int main() { #pragma omp target parallel foo(); }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 omp110.cpp -O1 -Rpass=openmp-opt omp110.cpp:6:7: remark: Moving globalized variable to the stack. [OMP110] int x; ^
A less trivial example can be seen using C++'s complex numbers. In this case the overloaded arithmetic operators cause pointers to the complex numbers to escape the current scope, but they can again be removed once the usage is visible.
#include <complex> }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 omp110.cpp -O1 -Rpass=openmp-opt In file included from omp110.cpp:1: In file included from /usr/bin/clang/lib/clang/13.0.0/include/openmp_wrappers/complex:27: /usr/include/c++/8/complex:328:20: remark: Moving globalized variable to the stack. [OMP110] complex<_Tp> __r = __x; ^ /usr/include/c++/8/complex:388:20: remark: Moving globalized variable to the stack. [OMP110] complex<_Tp> __r = __x; ^
Diagnostic Scope OpenMP target offloading optimization remark.
Replaced globalized variable with X bytes of shared memory. [OMP111] This optimization occurs when a globalized variable's data is shared between multiple threads, but requires a constant amount of memory that can be determined at compile time. This is the case when only a single thread creates the memory and is then shared between every thread. The memory can then be pushed to a static buffer of shared memory on the device. This optimization allows users to declare shared memory on the device without using OpenMP's custom allocators.
Globalization occurs when a pointer to a thread-local variable escapes the current scope. If a single thread is known to be responsible for creating and sharing the data it can instead be mapped directly to the device's shared memory. Checking if only a single thread can execute an instruction requires that the parent functions have internal linkage. Otherwise, an external caller could invalidate this analysis but having multiple threads call that function. The optimization pass will make internal copies of each function to use for this reason, but it is still recommended to mark them as internal using keywords like static whenever possible.
Example This optimization should apply to any variable declared in an OpenMP target region that is then shared with every thread in a parallel region. This allows the user to declare shared memory without using custom allocators. A simple stencil calculation shows how this can be used.
void stencil(int M, int N, double *X, double *Y) { #pragma omp target teams distribute collapse(2) \ map(to : X [0:M * N]) map(tofrom : Y [0:M * N]) for (int i0 = 0; i0 < M; i0 += MC) { for (int j0 = 0; j0 < N; j0 += NC) { double sX[MC][NC];
#pragma omp parallel for collapse(2) shared(sX) default(firstprivate) for (int i1 = 0; i1 < MC; ++i1) for (int j1 = 0; j1 < NC; ++j1) sX[i1][j1] = X[(i0 + i1) * N + (j0 + j1)];
#pragma omp parallel for collapse(2) shared(sX) default(firstprivate) for (int i1 = 1; i1 < MC - 1; ++i1) for (int j1 = 1; j1 < NC - 1; ++j1) Y[(i0 + i1) * N + j0 * j1] = (sX[i1 + 1][j1] + sX[i1 - 1][j1] + sX[i1][j1 + 1] + sX[i1][j1 - 1] +
double sX[MC][NC]; ^
The default mapping for variables captured in an OpenMP parallel region is shared. This means taking a pointer to the object which will ultimately result in globalization that will be mapped to shared memory when it could have been placed in registers. To avoid this, make sure each variable that can be copied into the region is marked firstprivate either explicitly or using the OpenMP 5.1 feature default(firstprivate).
Diagnostic Scope OpenMP target offloading optimization remark.
Found thread data sharing on the GPU. Expect degraded performance due to data globalization. [OMP112] This missed remark indicates that a globalized value was found on the target device that was not either replaced with stack memory by OMP110 or shared memory by OMP111. Globalization that has not been removed will need to be handled by the runtime and will significantly impact performance.
The OpenMP standard requires that threads are able to share their data between each-other. However, this is not true by default when offloading to a target device such as a GPU. Threads on a GPU cannot shared their data unless it is first placed in global or shared memory. In order to create standards complaint code, the Clang compiler will globalize any variables that could potentially be shared between the threads. In the majority of cases, globalized variables can either be returns to a thread-local stack, or pushed to shared memory. However, in a few cases it is necessary and will cause a performance penalty.
Examples This example shows legitimate data sharing on the device. It is a convoluted example, but is completely complaint with the OpenMP standard. If globalization was not added this would result in different results on different target devices.
#include <omp.h> #include <cstdio>
#pragma omp declare target static int *p; #pragma omp end declare target
void foo() { int x = omp_get_thread_num(); if (omp_get_thread_num() == 1) p = &x;
#pragma omp barrier
printf ("Thread %d: %d\n", omp_get_thread_num(), *p); }
int main() { #pragma omp target parallel foo(); }
A less convoluted example globalization that cannot be removed occurs when calling functions that aren't visible from the current translation unit.
extern void use(int *x);
void foo() { int x; use(&x); }
int main() { #pragma omp target parallel foo(); }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O1 -Rpass-missed=openmp-opt omp112.cpp omp112.cpp:4:7: remark: Found thread data sharing on the GPU. Expect degraded performance due to data globalization. [OMP112] [-Rpass-missed=openmp-opt] int x; ^
Diagnostic Scope OpenMP target offloading missed remark.
Could not move globalized variable to the stack. Variable is potentially captured in call. Mark parameter as __attribute__((noescape)) to override. [OMP113] This missed remark indicates that a globalized value could not be moved to the stack because it is potentially captured by a call to a function we cannot analyze. In order for a globalized variable to be moved to the stack, copies to its pointer cannot be stored. Otherwise it is considered captured and could potentially be shared between the threads. This can be overridden using a parameter level attribute as suggested in the remark text.
Globalization will occur when a pointer to a thread-local variable escapes the current scope. In most cases it can be determined that the variable cannot be shared if a copy of its pointer is never made. However, this remark indicates a copy of the pointer is present or that sharing is possible because it is used outside the current translation unit.
Examples If a pointer to a thread-local variable is passed to a function not visible in the current translation unit we need to assume a copy is made of it that can be shared between the threads. This prevents OMP110 from triggering, which will result in a performance penalty when executing on the target device.
extern void use(int *x);
void foo() { int x; use(&x); }
int main() { #pragma omp target parallel int x; ^
As the remark suggests, this behaviour can be overridden using the noescape attribute. This tells the compiler that no reference to the object the pointer points to that is derived from the parameter value will survive after the function returns. The user is responsible for verifying that this assertion is correct.
extern void use(__attribute__((noescape)) int *x);
void foo() { int x; use(&x); }
int main() { #pragma omp target parallel foo(); }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O2 -Rpass=openmp-opt omp113.cpp missed.cpp:4:7: remark: Moving globalized variable to the stack. [OMP110] int x; ^
Diagnostic Scope OpenMP target offloading missed remark.
Transformed generic-mode kernel to SPMD-mode [OMP120] This optimization remark indicates that the execution strategy for the OpenMP target offloading kernel was changed. Generic-mode kernels are executed by a single thread that schedules parallel worker threads using a state machine. This code transformation can move a kernel that was initially generated in generic mode to SPMD-mode where all threads are active at the same time with no state machine. This execution strategy is closer to how the threads are actually executed on a GPU target. This is only possible if the instructions previously executed by a single thread have no side-effects or can be guarded. If the instructions have no side-effects they are simply recomputed by each thread.
Generic-mode is often considerably slower than SPMD-mode because of the extra overhead required to separately schedule worker threads and pass data between them.This optimization allows users to use generic-mode semantics while achieving the performance of SPMD-mode. This can be helpful when defining shared memory between the threads using OMP111.
Examples Normally, any kernel that contains split OpenMP target and parallel regions will be executed in generic-mode. Sometimes it is easier to use generic-mode semantics to define shared memory, or more tightly control the distribution of the threads. This shows a naive matrix-matrix multiplication that contains code that will need to be guarded.
void matmul(int M, int N, int K, double *A, double *B, double *C) { #pragma omp target teams distribute collapse(2) \ map(to:A[0: M*K]) map(to:B[0: K*N]) map(tofrom:C[0 : M*N]) for (int i = 0; i < M; i++) { C[i*N + j] = sum; } } }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -fopenmp-version=51 -O2 -Rpass=openmp-opt omp120.cpp omp120.cpp:6:14: remark: Replaced globalized variable with 8 bytes of shared memory. [OMP111] double sum = 0.0; ^ omp120.cpp:2:1: remark: Transformed generic-mode kernel to SPMD-mode. [OMP120] #pragma omp target teams distribute collapse(2) \ ^
This requires guarding the store to the shared variable sum and the store to the matrix C. This can be thought of as generating the code below.
void matmul(int M, int N, int K, double *A, double *B, double *C) { #pragma omp target teams distribute collapse(2) \ map(to:A[0: M*K]) map(to:B[0: K*N]) map(tofrom:C[0 : M*N]) for (int i = 0; i < M; i++) { for (int j = 0; j < N; j++) { double sum; #pragma omp parallel default(firstprivate) shared(sum) { #pragma omp barrier if (omp_get_thread_num() == 0) sum = 0.0; #pragma omp barrier
#pragma omp for reduction(+:sum) for (int k = 0; k < K; k++) sum += A[i*K + k] * B[k*N + j];
#pragma omp barrier if (omp_get_thread_num() == 0) C[i*N + j] = sum; #pragma omp barrier } } } }
Diagnostic Scope OpenMP target offloading optimization remark.
Value has potential side effects preventing SPMD-mode execution. Add __attribute__((assume("ompx_spmd_amenable"))) to the called function to override. [OMP121] This analysis remarks indicates that a potential side-effect that cannot be guarded prevents the target region from executing in SPMD-mode. SPMD-mode requires that each thread is active inside the region. Any instruction that cannot be either recomputed by each thread independently or guarded and executed by a single thread prevents the region from executing in SPMD-mode.
This remark will attempt to print out the instructions preventing the region from being executed in SPMD-mode. Calls to functions outside the current translation unit will prevent this transformation from extern int work();
void use(int x);
void foo() { #pragma omp target teams { int x = work(); #pragma omp parallel use(x);
} }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O2 -Rpass-analysis=openmp-opt omp121.cpp omp121.cpp:8:13: remark: Value has potential side effects preventing SPMD-mode execution. Add `__attribute__((assume("ompx_spmd_amenable")))` to the called function to override. [OMP121] int x = work(); ^
As the remark suggests, the problem is caused by the unknown call to the external function work. This can be overridden by asserting that it does not contain any code that prevents SPMD-mode execution.
__attribute__((assume("ompx_spmd_amenable"))) extern int work();
void use(int x);
void foo() { #pragma omp target teams { int x = work(); #pragma omp parallel use(x);
} }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O2 -Rpass=openmp-opt omp121.cpp omp121.cpp:6:1: remark: Transformed generic-mode kernel to SPMD-mode. [OMP120] #pragma omp target teams ^
Diagnostic Scope OpenMP target offloading analysis remark.
Removing unused state machine from generic-mode kernel. [OMP130] This optimization remark indicates that an unused state machine was removed from a target region. This occurs when there are no parallel regions inside of a target construct. Normally, a state machine is required to schedule the threads inside of a parallel region. If there are no parallel regions, the state machine is unnecessary because there is only a single thread active at any time.
Examples This optimization should occur on any target region that does not contain any parallel work.
omp130.cpp:2:1: remark: Removing unused state machine from generic-mode kernel. [OMP130] #pragma omp target teams distribute map(tofrom: X[0:N]) map(tofrom: Y[0:N]) ^
Diagnostic Scope OpenMP target offloading optimization remark.
Rewriting generic-mode kernel with a customized state machine. [OMP131] This optimization remark indicates that a generic-mode kernel on the device was specialized for the given target region. When offloading in generic-mode, a state machine is required to schedule the work between the parallel worker threads. This optimization specializes the state machine in cases where there is a known number of parallel regions inside the kernel. A much simpler state machine can be used if it is known that there is no nested parallelism and the number of regions to schedule is a static amount.
Examples This optimization should occur on any generic-mode kernel that has visibility on all parallel regions, but cannot be moved to SPMD-mode and has no nested parallelism.
#pragma omp declare target int TID; #pragma omp end declare target
void foo() { #pragma omp target { TID = omp_get_thread_num(); #pragma omp parallel { work(); } } }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O2 -Rpass=openmp-opt omp131.cpp omp131.cpp:8:1: remark: Rewriting generic-mode kernel with a customized state machine. [OMP131] #pragma omp target ^
Diagnostic Scope OpenMP target offloading optimization remark.
Generic-mode kernel is executed with a customized state machine that requires a fallback. [OMP132] This analysis remark indicates that a state machine rewrite occurred, but could not be done fully because of unknown calls to functions that may contain parallel regions. The state machine handles scheduling work between parallel worker threads on the device when operating in generic-mode. If there are unknown parallel regions it prevents the optimization from fully rewriting the state machine.
Examples This will occur for any generic-mode kernel that may contain unknown parallel regions. This is typically coupled with the OMP133 remark.
extern void setup(); work(); } } }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O2 -Rpass-analysis=openmp-opt omp132.cpp omp133.cpp:4:1: remark: Generic-mode kernel is executed with a customized state machine that requires a fallback. [OMP132] #pragma omp target ^
Diagnostic Scope OpenMP target offloading analysis remark.
Call may contain unknown parallel regions. Use __attribute__((assume("omp_no_parallelism"))) to override. [OMP133] This analysis remark identifies calls that prevented OMP131 from providing the generic-mode kernel with a fully specialized state machine. This remark will identify each call that may contain unknown parallel regions that caused the kernel to require a fallback.
Examples This will occur for any generic-mode kernel that may contain unknown parallel regions. This is typically coupled with the OMP132 remark.
extern void setup();
void foo() { #pragma omp target { setup(); #pragma omp parallel { work(); } } }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O2 -Rpass-analysis=openmp-opt omp133.cpp omp133.cpp:6:5: remark: Call may contain unknown parallel regions. Use `__attribute__((assume("omp_no_parallelism")))` to override. [OMP133] setup(); ^
The remark suggests marking the function with the assumption that it contains no parallel regions. If this is done then the kernel will be rewritten with a fully specialized state machine.
__attribute__((assume("omp_no_parallelism"))) extern void setup();

void foo() { #pragma omp target { setup(); #pragma omp parallel { work(); }
Diagnostic Scope OpenMP target offloading analysis remark.
Could not internalize function. Some optimizations may not be possible. [OMP140] This analysis remark indicates that function internalization failed for the given function. Internalization occurs when a call to a function that ordinarily has external visibility is replaced with a call to a copy of that function with only internal visibility. This allows the compiler to make strong static assertions about the context a function is called in. Without internalization this analysis would always be invalidated by the possibility of someone calling the function in a different context outside of the current translation unit. This is necessary for optimizations like OMP111 and OMP120. If a function failed to be internalized it most likely has linkage that cannot be copied. Internalization is currently only enabled by default for OpenMP target offloading.
Examples This will occur for any function declaration that has incompatible linkage.
__attribute__((weak)) void setup();
void foo() { #pragma omp target { setup(); #pragma omp parallel { work(); } } }
$ clang++ -fopenmp -fopenmp-targets=nvptx64 -O1 -Rpass-analysis=openmp-opt omp140.cpp omp140.cpp:1:1: remark: Could not internalize function. Some optimizations may not be possible. [OMP140] __attribute__((weak)) void setup() { ^
Diagnostic Scope OpenMP analysis remark.
Parallel region merged with parallel region at <location>. [OMP150] This optimization remark indicates that a parallel region was merged with others into a single parallel region. Parallel region merging fuses consecutive parallel regions to reduce the team activation overhead of forking and increases the scope of possible OpenMP-specific optimizations within merged parallel regions. This optimization can also guard sequential code between two parallel regions if applicable.
Example This optimization should apply to any compatible and consecutive parallel regions. In this case the sequential region between the parallel regions will be guarded so it is only executed by a single thread in the new merged region.
parallel_work(); }
$ clang++ -fopenmp -O2 -Rpass=openmp-opt -mllvm -openmp-opt-enable-merging omp150.cpp omp150.cpp:2:1: remark: Parallel region merged with parallel region at merge.cpp:7:1. [OMP150] #pragma omp parallel ^
Diagnostic Scope OpenMP optimization remark.
Removing parallel region with no side-effects. [OMP160] This optimization remark indicates that a parallel region was deleted because it was not found to have any side-effects. This can occur if the region does not write any of its results to memory visible outside the region. This optimization is necessary because the barrier between sequential and parallel code typically prevents dead code elimination from completely removing the region. Otherwise there will still be overhead to fork and merge the threads with no work done.
Example This optimization occurs whenever a parallel region was not found to have any side-effects. This can occur if the parallel region only reads memory or is simply empty.
void foo() { #pragma omp parallel { } #pragma omp parallel { int x = 1; } } }
$ clang++ -fopenmp -O2 -Rpass=openmp-opt omp160.cpp omp160.cpp:4:1: remark: Removing parallel region with no side-effects. [OMP160] [-Rpass=openmp-opt] #pragma omp parallel ^ delete.cpp:2:1: remark: Removing parallel region with no side-effects. [OMP160] [-Rpass=openmp-opt] #pragma omp parallel ^ ^
Diagnostic Scope OpenMP optimization remark.
OpenMP runtime call <call> deduplicated. [OMP170] This optimization remark indicates that a call to an OpenMP runtime call was replaced with the result of an existing one. This occurs when the compiler knows that the result of a runtime call is immutable. Removing duplicate calls is done by replacing all calls to that function with the result of the first call. This cannot be done automatically by the compiler because the implementations of the OpenMP runtime calls live in a separate library the compiler cannot see.
Example This optimization will trigger for known OpenMP runtime calls whose return value will not change.
void foo(int N) { }
$ clang -fopenmp -O2 -Rpass=openmp-opt omp170.c ompi170.c:2:26: remark: OpenMP runtime call omp_get_thread_limit deduplicated. [OMP170] double *A = malloc(N * omp_get_thread_limit()); ^
Diagnostic Scope OpenMP optimization remark.
Replacing OpenMP runtime call <call> with <value>. This optimization remark indicates that analysis determined an OpenMP runtime calls can be replaced with a constant value. This can occur when an OpenMP runtime call that queried some internal state was found to always return a single value after analysis.
Example This optimization will trigger for most target regions to simplify the runtime once certain constants are known. This will trigger for internal runtime functions so it requires enabling verbose remarks with -openmp-opt-verbose-remarks (prefixed with -mllvm for use with clang).
void foo() { #pragma omp target parallel { } }
$ clang test.c -fopenmp -fopenmp-targets=nvptx64 -O1 -Rpass=openmp-opt \ -mllvm -openmp-opt-verbose-remarks remark: Replacing runtime call __kmpc_is_spmd_exec_mode with 1. [OMP180] [-Rpass=openmp-opt] remark: Replacing runtime call __kmpc_is_spmd_exec_mode with 1. [OMP180] [-Rpass=openmp-opt] remark: Replacing runtime call __kmpc_parallel_level with 1. [OMP180] [-Rpass=openmp-opt] remark: Replacing runtime call __kmpc_parallel_level with 1. [OMP180] [-Rpass=openmp-opt]
Diagnostic Scope OpenMP optimization remark.
Redundant barrier eliminated. (device only) This optimization remark indicates that analysis determined an aligned barrier in the device code to be redundant. This can occur when state updates that have been synchronized by the barrier were eliminated too. See also "Co-Designing an OpenMP GPU Runtime and Optimizations for Near-Zero Overhead Execution", IPDPS'22.
Example This optimization will trigger for most target regions if state initialization was removed as a consequence of "state forwarding". This will trigger for internal runtime functions so it requires enabling verbose remarks with -openmp-opt-verbose-remarks (prefixed with -mllvm for use with clang).
Diagnostic Scope OpenMP optimization remark.
+-------------------+------------------+-----------------------------------------------+ |Diagnostics Number | Diagnostics Kind | Diagnostics Description | +-------------------+------------------+-----------------------------------------------+ |OMP100 | Analysis | Potentially unknown OpenMP target region | | | | caller. | | | | machine. | +-------------------+------------------+-----------------------------------------------+ |OMP110 | Optimization | Moving globalized variable to the stack. | +-------------------+------------------+-----------------------------------------------+ |OMP111 | Optimization | Replaced globalized variable with X bytes of | | | | shared memory. | +-------------------+------------------+-----------------------------------------------+ |OMP112 | Missed | Found thread data sharing on the GPU. Expect | | | | degraded performance due to data | | | | globalization. | +-------------------+------------------+-----------------------------------------------+ |OMP113 | Missed | Could not move globalized variable to the | | | | stack. Variable is potentially captured in | | | | call. Mark parameter as | | | | __attribute__((noescape)) to override. | +-------------------+------------------+-----------------------------------------------+ |OMP120 | Optimization | Transformed generic-mode kernel to SPMD-mode. | +-------------------+------------------+-----------------------------------------------+ |OMP121 | Analysis | Value has potential side effects preventing | | | | SPMD-mode execution. Add | | | | __attribute__((assume("ompx_spmd_amenable"))) | | | | to the called function to override. | +-------------------+------------------+-----------------------------------------------+ |OMP130 | Optimization | Removing unused state machine from | | | | generic-mode kernel. | +-------------------+------------------+-----------------------------------------------+ |OMP131 | Optimization | Rewriting generic-mode kernel with a | | | | customized state machine. | +-------------------+------------------+-----------------------------------------------+ |OMP132 | Analysis | Generic-mode kernel is executed with a | | | | customized state machine that requires a | | | | fallback. | +-------------------+------------------+-----------------------------------------------+ |OMP133 | Analysis | Call may contain unknown parallel regions. | | | | Use | | | | __attribute__((assume("omp_no_parallelism"))) | | | | to override. | +-------------------+------------------+-----------------------------------------------+ |OMP140 | Analysis | Could not internalize function. Some | | | | optimizations may not be possible. | +-------------------+------------------+-----------------------------------------------+ |OMP150 | Optimization | Parallel region merged with parallel region | | | | at <location>. | +-------------------+------------------+-----------------------------------------------+ |OMP160 | Optimization | Removing parallel region with no | | | | side-effects. | +-------------------+------------------+-----------------------------------------------+ |OMP170 | Optimization | OpenMP runtime call <call> deduplicated. | +-------------------+------------------+-----------------------------------------------+ |OMP180 | Optimization | Replacing OpenMP runtime call <call> with | | | | <value>. | +-------------------+------------------+-----------------------------------------------+ |OMP190 | Optimization | Redundant barrier eliminated. (device only) | +-------------------+------------------+-----------------------------------------------+
Dealing with OpenMP can be complicated. For help with the setup of an OpenMP (offload) capable compiler toolchain, its usage, and common problems, consult the Support and FAQ page.

Calls OpenMP in LLVM Technical Call
o Development updates on OpenMP (and OpenACC) in the LLVM Project, including Clang, optimization, and runtime work.
o Join OpenMP in LLVM Technical Call.
o Time: Weekly call on every Wednesday 7:00 AM Pacific time.
o Meeting minutes are here.
o Status tracking page.
OpenMP in Flang Technical Call
o Development updates on OpenMP and OpenACC in the Flang Project.
o Join OpenMP in Flang Technical Call
o Time: Weekly call on every Thursdays 8:00 AM Pacific time.
o Meeting minutes are here.
o Status tracking page.
FAQ NOTE: The FAQ is a work in progress and most of the expected content is not yet available. While you can expect changes, we always welcome feedback and additions. Please contact, e.g., through openmp-dev@lists.llvm.org.
Q: How to contribute a patch to the webpage or any other part? All patches go through the regular LLVM review process.
Q: How to build an OpenMP GPU offload capable compiler? To build an effective OpenMP offload capable compiler, only one extra CMake option, LLVM_ENABLE_RUNTIMES="openmp", is needed when building LLVM (Generic information about building LLVM is available here.). Make sure all backends that are targeted by OpenMP to be enabled. By default, Clang will be built with all backends enabled. When building with LLVM_ENABLE_RUNTIMES="openmp" OpenMP should not be enabled in LLVM_ENABLE_PROJECTS because it is enabled by default.
For Nvidia offload, please see Q: How to build an OpenMP NVidia offload capable compiler?. For AMDGPU offload, please see Q: How to build an OpenMP AMDGPU offload capable compiler?.
NOTE: The compiler that generates the offload code should be the same (version) as the compiler that builds the OpenMP device runtimes. The OpenMP host runtime can be built by a different compiler.
Q: How to build an OpenMP NVidia offload capable compiler? The Cuda SDK is required on the machine that will execute the openmp application.
compute capacity of your GPU, e.g., 75.
Q: How to build an OpenMP AMDGPU offload capable compiler? A subset of the ROCm toolchain is required to build the LLVM toolchain and to execute the openmp application. Either install ROCm somewhere that cmake's find_package can locate it, or build the required subcomponents ROCt and ROCr from source.
The two components used are ROCT-Thunk-Interface, roct, and ROCR-Runtime, rocr. Roct is the userspace part of the linux driver. It calls into the driver which ships with the linux kernel. It is an implementation detail of Rocr from OpenMP's perspective. Rocr is an implementation of HSA.
SOURCE_DIR=same-as-llvm-source # e.g. the checkout of llvm-project, next to openmp BUILD_DIR=somewhere INSTALL_PREFIX=same-as-llvm-install
cd $SOURCE_DIR git clone git@github.com:RadeonOpenCompute/ROCT-Thunk-Interface.git -b roc-4.2.x \ --single-branch git clone git@github.com:RadeonOpenCompute/ROCR-Runtime.git -b rocm-4.2.x \ --single-branch
cd $BUILD_DIR && mkdir roct && cd roct cmake $SOURCE_DIR/ROCT-Thunk-Interface/ -DCMAKE_INSTALL_PREFIX=$INSTALL_PREFIX \ -DCMAKE_BUILD_TYPE=Release -DBUILD_SHARED_LIBS=OFF make && make install
cd $BUILD_DIR && mkdir rocr && cd rocr cmake $SOURCE_DIR/ROCR-Runtime/src -DIMAGE_SUPPORT=OFF \ -DCMAKE_INSTALL_PREFIX=$INSTALL_PREFIX -DCMAKE_BUILD_TYPE=Release \ -DBUILD_SHARED_LIBS=ON make && make install
IMAGE_SUPPORT requires building rocr with clang and is not used by openmp.
Provided cmake's find_package can find the ROCR-Runtime package, LLVM will build a tool bin/amdgpu-arch which will print a string like gfx906 when run if it recognises a GPU on the local system. LLVM will also build a shared library, libomptarget.rtl.amdgpu.so, which is linked against rocr.
With those libraries installed, then LLVM build and installed, try:
clang -O2 -fopenmp -fopenmp-targets=amdgcn-amd-amdhsa example.c -o example && ./example
Q: What are the known limitations of OpenMP AMDGPU offload? LD_LIBRARY_PATH or rpath/runpath are required to find libomp.so and libomptarget.so
There is no libc. That is, malloc and printf do not exist. Libm is implemented in terms of the rocm device library, which will be searched for if linking with '-lm'.
Some versions of the driver for the radeon vii (gfx906) will error unless the environment variable 'export HSA_IGNORE_SRAMECC_MISREPORT=1' is set.
o libomp.so (or similar), the host openmp runtime
o libomptarget.so, the target-agnostic target offloading openmp runtime
o plugins loaded by libomptarget.so:
o libomptarget.rtl.amdgpu.so
o libomptarget.rtl.cuda.so
o libomptarget.rtl.x86_64.so
o libomptarget.rtl.ve.so
o and others
o dependencies of those plugins, e.g. cuda/rocr for nvptx/amdgpu
The compiled executable is dynamically linked against a host runtime, e.g. libomp.so, and against the target offloading runtime, libomptarget.so. These are found like any other dynamic library, by setting rpath or runpath on the executable, by setting LD_LIBRARY_PATH, or by adding them to the system search.
libomptarget.so has rpath or runpath (whichever the system default is) set to $ORIGIN, and the plugins are located next to it, so it will find the plugins without any environment variables set. If LD_LIBRARY_PATH is set, whether it overrides which plugin is found depends on whether your system treats -Wl,-rpath as RPATH or RUNPATH.
The plugins will try to find their dependencies in plugin-dependent fashion.
The cuda plugin is dynamically linked against libcuda if cmake found it at compiler build time. Otherwise it will attempt to dlopen libcuda.so. It does not have rpath set.
The amdgpu plugin is linked against ROCr if cmake found it at compiler build time. Otherwise it will attempt to dlopen libhsa-runtime64.so. It has rpath set to $ORIGIN, so installing libhsa-runtime64.so in the same directory is a way to locate it without environment variables.
In addition to those, there is a compiler runtime library called deviceRTL. This is compiled from mostly common code into an architecture specific bitcode library, e.g. libomptarget-nvptx-sm_70.bc.
Clang and the deviceRTL need to match closely as the interface between them changes frequently. Using both from the same monorepo checkout is strongly recommended.
Unlike the host side which lets environment variables select components, the deviceRTL that is located in the clang lib directory is preferred. Only if it is absent, the LIBRARY_PATH environment variable is searched to find a bitcode file with the right name. This can be overridden by passing a clang flag, --libomptarget-nvptx-bc-path or --libomptarget-amdgcn-bc-path. That can specify a directory or an exact bitcode file to use. For now, the answer is most likely no. Please see Q: How to build an OpenMP GPU offload capable compiler?.
Q: Does Clang support <math.h> and <complex.h> operations in OpenMP target on GPUs? Yes, LLVM/Clang allows math functions and complex arithmetic inside of OpenMP target regions that are compiled for GPUs.
Clang provides a set of wrapper headers that are found first when math.h and complex.h, for C, cmath and complex, for C++, or similar headers are included by the application. These wrappers will eventually include the system version of the corresponding header file after setting up a target device specific environment. The fact that the system header is included is important because they differ based on the architecture and operating system and may contain preprocessor, variable, and function definitions that need to be available in the target region regardless of the targeted device architecture. However, various functions may require specialized device versions, e.g., sin, and others are only available on certain devices, e.g., __umul64hi. To provide "native" support for math and complex on the respective architecture, Clang will wrap the "native" math functions, e.g., as provided by the device vendor, in an OpenMP begin/end declare variant. These functions will then be picked up instead of the host versions while host only variables and function definitions are still available. Complex arithmetic and functions are support through a similar mechanism. It is worth noting that this support requires extensions to the OpenMP begin/end declare variant context selector that are exposed through LLVM/Clang to the user as well.
Q: What is a way to debug errors from mapping memory to a target device? An experimental way to debug these errors is to use remote process offloading. By using libomptarget.rtl.rpc.so and openmp-offloading-server, it is possible to explicitly perform memory transfers between processes on the host CPU and run sanitizers while doing so in order to catch these errors.
Q: Why does my application say "Named symbol not found" and abort when I run it? This is most likely caused by trying to use OpenMP offloading with static libraries. Static libraries do not contain any device code, so when the runtime attempts to execute the target region it will not be found and you will get an an error like this.
CUDA error: Loading '__omp_offloading_fd02_3231c15__Z3foov_l2' Failed CUDA error: named symbol not found Libomptarget error: Unable to generate entries table for device id 0.
Currently, the only solution is to change how the application is built and avoid the use of static libraries.
Q: Can I use dynamically linked libraries with OpenMP offloading? Dynamically linked libraries can be only used if there is no device code split between the library and application. Anything declared on the device inside the shared library will not be visible to the application when it's linked.
Q: How to build an OpenMP offload capable compiler with an outdated host compiler? Enabling the OpenMP runtime will perform a two-stage build for you. If like to use in the second stage.
Q: How can I include OpenMP offloading support in my CMake project? Currently, there is an experimental CMake find module for OpenMP target offloading provided by LLVM. It will attempt to find OpenMP target offloading support for your compiler. The flags necessary for OpenMP target offloading will be loaded into the OpenMPTarget::OpenMPTarget_<device> target or the OpenMPTarget_<device>_FLAGS variable if successful. Currently supported devices are AMDGPU and NVPTX.
To use this module, simply add the path to CMake's current module path and call find_package. The module will be installed with your OpenMP installation by default. Including OpenMP offloading support in an application should now only require a few additions.
cmake_minimum_required(VERSION 3.13.4) project(offloadTest VERSION 1.0 LANGUAGES CXX)
list(APPEND CMAKE_MODULE_PATH "${PATH_TO_OPENMP_INSTALL}/lib/cmake/openmp")
find_package(OpenMPTarget REQUIRED NVPTX)
add_executable(offload) target_link_libraries(offload PRIVATE OpenMPTarget::OpenMPTarget_NVPTX) target_sources(offload PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/src/Main.cpp)
Using this module requires at least CMake version 3.13.4. Supported languages are C and C++ with Fortran support planned in the future. Compiler support is best for Clang but this module should work for other compiler vendors such as IBM, GNU.
Q: What does 'Stack size for entry function cannot be statically determined' mean? This is a warning that the Nvidia tools will sometimes emit if the offloading region is too complex. Normally, the CUDA tools attempt to statically determine how much stack memory each thread. This way when the kernel is launched each thread will have as much memory as it needs. If the control flow of the kernel is too complex, containing recursive calls or nested parallelism, this analysis can fail. If this warning is triggered it means that the kernel may run out of stack memory during execution and crash. The environment variable LIBOMPTARGET_STACK_SIZE can be used to increase the stack size if this occurs.
The current (in-progress) release notes can be found here while release notes for releases, starting with LLVM 12, will be available on the Download Page.
OPENMP 15.0.0 RELEASE NOTES WARNING: These are in-progress notes for the upcoming LLVM 15.0.0 release. Release notes for previous releases can be found on the Download Page.
Introduction This document contains the release notes for the OpenMP runtime, release 15.0.0. Here we describe the status of OpenMP, including major improvements from the previous release. All OpenMP releases may be
COPYRIGHT 2013-2023, LLVM/OpenMP
15 December 15, 2023 LLVMOPENMP(1)