This patch introduces the logic implementing hierarchical scheduling.
First and foremost, hierarchical scheduling is off by default
To enable, use -DLIBOMP_USE_HIER_SCHED=On during CMake's configure stage.
This work is based off if the IWOMP paper:
"Workstealing and Nested Parallelism in SMP Systems"
Hierarchical scheduling is the layering of OpenMP schedules for different layers
of the memory hierarchy. One can have multiple layers between the threads and
the global iterations space. The threads will go up the hierarchy to grab
iterations, using possibly a different schedule & chunk for each layer.
[ Global iteration space (0-999) ]
(use static)
[ L1 | L1 | L1 | L1 ]
(use dynamic,1)
[ T0 T1 | T2 T3 | T4 T5 | T6 T7 ]
In the example shown above, there are 8 threads and 4 L1 caches begin targeted.
If the topology indicates that there are two threads per core, then two
consecutive threads will share the data of one L1 cache unit. This example
would have the iteration space (0-999) split statically across the four L1
caches (so the first L1 would get (0-249), the second would get (250-499), etc).
Then the threads will use a dynamic,1 schedule to grab iterations from the L1
cache units. There are currently four supported layers: L1, L2, L3, NUMA
OMP_SCHEDULE can now read a hierarchical schedule with this syntax:
OMP_SCHEDULE='EXPERIMENTAL LAYER,SCHED[,CHUNK][:LAYER,SCHED[,CHUNK]...]:SCHED,CHUNK
And OMP_SCHEDULE can still read the normal SCHED,CHUNK syntax from before
I've kept most of the hierarchical scheduling logic inside kmp_dispatch_hier.h
to try to keep it separate from the rest of the code.
Differential Revision: https://reviews.llvm.org/D47962
llvm-svn: 336571
Summary:
Currently Cuda plugin supports loading of the single image, though we
may have the executable with the several images, if it has target
regions inside of the dynamically loaded library. Patch allows to load
multiple images.
Reviewers: grokos
Subscribers: guansong, openmp-commits, kkwli0
Differential Revision: https://reviews.llvm.org/D49036
llvm-svn: 336569
This patch reorganizes the loop scheduling code in order to allow hierarchical
scheduling to use it more effectively. In particular, the goal of this patch
is to separate the algorithmic parts of the scheduling from the thread
logistics code.
Moves declarations & structures to kmp_dispatch.h for easier access in
other files. Extracts the algorithmic part of __kmp_dispatch_init() and
__kmp_dispatch_next() into __kmp_dispatch_init_algorithm() and
__kmp_dispatch_next_algorithm(). The thread bookkeeping logic is still kept in
__kmp_dispatch_init() and __kmp_dispatch_next(). This is done because the
hierarchical scheduler needs to access the scheduling logic without the
bookkeeping logic. To prepare for new pointer in dispatch_private_info_t, a
new flags variable is created which stores the ordered and nomerge flags instead
of them being in two separate variables. This will keep the
dispatch_private_info_t structure the same size.
Differential Revision: https://reviews.llvm.org/D47961
llvm-svn: 336568
These are preliminary changes that attempt to use C++11 Atomics in the runtime.
We are expecting better portability with this change across architectures/OSes.
Here is the summary of the changes.
Most variables that need synchronization operation were converted to generic
atomic variables (std::atomic<T>). Variables that are updated with combined CAS
are packed into a single atomic variable, and partial read/write is done
through unpacking/packing
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D47903
llvm-svn: 336563
The flag "--no-as-needed" is not recognized by the linker on macOS making the following tests fail:
ompt/loadtool/tool_available/tool_available.c
ompt/loadtool/tool_not_available/tool_not_available.c
This patch removes this flag for macOS and adds it only for Linux and Windows.
I tested it on Ubuntu 16.04 and macOS HighSierra, with Clang/LLVM 6.0.1 and OpenMP trunk.
This solution was also discussed in the OpenMP-dev mailing list.
Patch provided by Simone Atzeni
Differential Revision: https://reviews.llvm.org/D48888
llvm-svn: 336327
The testcase potentially fails when a thread is reused.
The added synchronization makes sure this does not happen.
Patch provided by Simon Convent
Differential Revision: https://reviews.llvm.org/D48932
llvm-svn: 336326
When compiling with icc, there is a problem with reenter frame addresses in
parallel_begin callbacks in the interoperability.c testcase. (The address is
not available. thus NULL)
Using alloca() forces availability of the frame pointer.
Patch provided by Simon Convent
Differential Revision: https://reviews.llvm.org/D48282
llvm-svn: 336088
Several runtime entry points have not been tested from non-OpenMP threads. This
adds tests to an existing testcase. While at it, the testcase was reformatted
Patch provided by Simon Convent
Differential Revision: https://reviews.llvm.org/D48124
llvm-svn: 336087
Especially the thread_end callback has not been tested before.
This adds a testcase for nested and non-nested threads.
Patch provided by Simon Convent
Differential Revision: https://reviews.llvm.org/D47824
llvm-svn: 336086
The current implementation always provides the thread-num for the current
parallel region. This patch fixes the behavior for ancestor levels >0.
Differential Revision: https://reviews.llvm.org/D46533
llvm-svn: 336085
Summary:
Patch fixes several problems in the implementation of NVPTX RTL.
1. Detection of the last iteration for loops with static scheduling, no chunks.
2. Fixes reductions for the serialized parallel constructs.
3. Fixes handling of the barriers.
Reviewers: grokos
Reviewed By: grokos
Subscribers: Hahnfeld, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D48480
llvm-svn: 335469
Upcoming changes to FileCheck will modify CHECK-DAG to not match
overlapping regions of the input. This test was found to be affected
because it expects to find four threads to invoke events of type
ompt_event_implicit_task_begin. It turns out this is wrong because
OMP_THREAD_LIMIT is set to 2, so there are only two threads. The
rest of the test got it right so it went unnoticed until now.
(Rewrite test and apply clang-format to it as discussed in the past.)
Differential Revision: https://reviews.llvm.org/D47119
llvm-svn: 333361
The generic entry points for static loop scheduling previously
hardcoded that the runtime was initialized. This can be wrong if
the compiler analyzes that the runtime is not needed and calls
the init functions accordingly.
This didn't affect clang-ykt because they have entry points for
different combinations of SPMD x Runtime not needed. I didn't do
measurements yet but with inlining we might get away with always
calling the generic interface and letting compiler and runtime
figure out the rest.
In any case, a correct runtime is always better than having
functions that may only be called if previous calls passed in
a specific set of arguments!
Differential Revision: https://reviews.llvm.org/D47131
llvm-svn: 333285
Introduce OPENMP_INSTALL_LIBDIR and use in all install() commands.
This also fixes installation of libomptarget-nvptx that previously
didn't honor {OPENMP,LLVM}_LIBDIR_SUFFIX.
Differential Revision: https://reviews.llvm.org/D47130
llvm-svn: 333284
The existing implementation of the dynamic scheduling
breaks the contract introduced by the original openmp
runtime and, thus, is incorrect. Patch fixes it and
introduces correct dynamic scheduling model.
Thanks to Alexey Bataev for submitting this patch.
Differential Revision: https://reviews.llvm.org/D47333
llvm-svn: 333225
We already know where the CUDA SDK is, so there is no point in
letting Clang search for it again and possibly finding no or
a different installation.
--cuda-path is supported since the beginning of CUDA support in
Clang, so making this required doesn't impose additional restrictions.
Differential Revision: https://reviews.llvm.org/D46930
llvm-svn: 332495
Move all logic related to selecting the bitcode compiler and linker
into a new file and dynamically test required compiler flags. This
also adds -fcuda-rdc for Clang trunk as previously attempted in D44992
which fixes the build.
As a result this change also enables building the library by default
if all prerequisites are met.
Differential Revision: https://reviews.llvm.org/D46901
llvm-svn: 332494
Summary: Add function to the NVPTX libomptarget library that will return true if the current target region is being executed in SPMD mode.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D46840
llvm-svn: 332360
implicit_task_end callbacks in nested parallel regions did not always give the
correct thread_num, since the inner parallel region may have already been
finalized.
Now, the thread_num is stored at the beginning of the implicit task and
retrieved at the end, whenever necessary.
A testcase was added as well.
Differential Revision: https://reviews.llvm.org/D46260
llvm-svn: 331632
The api_calls_misc.c testcase tests the following api calls:
ompt_get_callback()
ompt_get_state()
ompt_enumerate_states()
ompt_enumerate_mutex_impls()
These have not been tested previously.
The api_calls.c testcase has been renamed to api_calls_places.c because it only tests api calls that are related to places.
Differential Revision: https://reviews.llvm.org/D42523
llvm-svn: 331631
Summary:
Enable the device side debug messages at compile time, use env var to control at runtime.
To achieve this, an environment data block is passed to the device lib when it is loaded.
By default, the message is off, to enable it, a user need to set LIBOMPDEVICE_DEBUG=1.
Reviewers: grokos
Reviewed By: grokos
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D46210
llvm-svn: 331550
Summary: Minor printf format correction. NVCC ignore those. Clang will give warning on these if debug is enabled.
Reviewers: grokos
Reviewed By: grokos
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D45528
llvm-svn: 330944
Summary: The LIBOMPTARGET_NVPTX_DEBUG flag is inconsistent between using nvcc to generate .a file and clang to generate .bc file. Sync the two setting so we can get debug messages from the bc file path as well.
Reviewers: grokos
Subscribers: Hahnfeld, openmp-commits, mgorny
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D45530
llvm-svn: 330477
Currently, the affinity API reports garbage for the initial place list and any
thread's place lists when using KMP_AFFINITY=none|compact|scatter.
This patch does two things:
for KMP_AFFINITY=none, Creates a one entry table for the places, this way, the
initial place list is just a single place with all the proc ids in it. We also
set the initial place of any thread to 0 instead of KMP_PLACE_ALL so that the
thread reports that single place (place 0) instead of garbage (-1) when using
the affinity API.
When non-OMP_PROC_BIND affinity is used
(including KMP_AFFINITY=compact|scatter), a thread's place list is populated
correctly. We assume that each thread is assigned to a single place. This is
implemented in two of the affinity API functions
Differential Revision: https://reviews.llvm.org/D45527
llvm-svn: 330283
This patch introduces GOMP_taskloop to our API. It adds GOMP_4.5 to our
version symbols. Being a wrapper around __kmpc_taskloop, the function
creates a task with the loop bounds properly nested in the shareds so that
the GOMP task thunk will work properly. Also, the firstprivate copy constructors
are properly handled using the __kmp_gomp_task_dup() auxiliary function.
Currently, only linear spawning of tasks is supported
for the GOMP_taskloop interface.
Differential Revision: https://reviews.llvm.org/D45327
llvm-svn: 330282
Summary:
This one line change is to remove this warning message
"warning: integer conversion resulted in a change of sign"
Reviewers: grokos
Reviewed By: grokos
Subscribers: openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D45415
llvm-svn: 329713
This change removes the unnecessary lock operation on __kmp_initz_lock inside
the __kmp_atfork_child() function for Linux; the lock variable is initialized
in the same function later.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D44949
llvm-svn: 328900
The summarizeStats.py script processes raw data provided by the
instrumented (stats-gathering) OpenMP* runtime library. It provides:
1) A radar chart which plots counters as frequency (per GigaTick) of use within
the program. The frequencies are plotted as log10, however values less than
one are kept as it is and represented in red color. This was done to help
visualize the differences better.
2) Pie charts separating total time as compute and non-compute. The compute and
non-compute times have their own pie charts showing the constructs that
contributed to them. The percentages listed are with respect to the total
time.
3) '.csv' file with percentage of time spent within the different constructs.
The script can be used as:
$ python $PATH_TO_SCRIPT/summarizeStats.py instrumented1.csv instrumented2.csv
Patch by Taru Doodi
Differential Revision: https://reviews.llvm.org/D41838
llvm-svn: 328568
Summary: The global stack initialization function may be called multiple times. The initialization of the shared memory slots should only happen when the function is called for the first time for a given warp master thread.
Reviewers: grokos, carlo.bertolli, ABataev, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44754
llvm-svn: 328148
Summary: The check for the master warp must take into consideration the actual number of warps: the master warp is equal to the last active warp not necessarily WARPSIZE - 1.
Reviewers: grokos, carlo.bertolli, ABataev, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44537
llvm-svn: 328146
Summary:
This patch allows worker to have a global memory stack managed by the runtime. This patch is needed for completeness and consistency with the globalization policy: if a worker-side variable escapes the current context it then needs to be globalized.
Until now, only the master thread was allowed to have such a stack. These global values can now potentially be shared amongst workers if the semantics of the OpenMP program require it.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44487
llvm-svn: 328144
Added settings code to read OMP_TARGET_OFFLOAD environment variable. Added
target-offload-var ICV as __kmp_target_offload, set via OMP_TARGET_OFFLOAD,
if available, otherwise defaulting to DEFAULT. Valid values for the ICV are
specified as enum values {0,1,2} for disabled, default, and mandatory. An
internal API access function __kmpc_get_target_offload is provided.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D44577
llvm-svn: 328046
Summary:
Allow the runtime to use the existing shared memory statically allocated slots.
When a variable is globalized, the underlying memory can be either shared or global memory (both have block-wide visibility). In this case, we allow that the storage to use a limited amount of shared memory that has been statically allocated already. Only if shared memory doesn't prove to be enough do we then invoke malloc() to create a new global memory slot.
Reviewers: ABataev, carlo.bertolli, grokos, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44486
llvm-svn: 327639
Summary: To save on calls to malloc, this patch enables the re-use of pre-allocated global memory slots.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44470
llvm-svn: 327637
Summary:
This patch adds support for the sharing of variables from the master thread of a team to the worker threads of the team.
The runtime uses a stack structure implemented as a doubly-linked list of slots with each slot having the exact same size as the size requested. This implementation leverages existing data structures. The runtime functions are added as separate functions to avoid interfering with the current interface.
Limitations to be addressed in future patches:
- This current patch only employs global memory. In a future patch we will enable to usage for shared memory as an optimization.
- Allow the allocation of several requested sizes in the same slot.
Reviewers: ABataev, grokos, caomhin, carlo.bertolli
Reviewed By: grokos
Subscribers: Hahnfeld, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44260
llvm-svn: 327440
Summary: To make the two parts of the union have the same size, the size of vect needs to be increased by 16 bits.
Reviewers: grokos, carlo.bertolli, caomhin, ABataev
Reviewed By: grokos, ABataev
Subscribers: fedor.sergeev, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D44254
llvm-svn: 327040
Summary:
This patch reverts the changes to libomptarget that were coupled with the changes to Clang code gen for data sharing using shared memory. A similar patch exists for Clang: D43625
Shared memory is meant to be used as an optimization on top of a more general scheme. So far we didn't have a global memory implementation ready so shared memory was a solution which applied to the current level of OpenMP complexity supported by trunk on GPU devices (due to the missing NVPTX backend patch this functionality has never been exercised). Now that we have a global memory solution this patch is "in the way" and needs to be removed (for now). This patch (or an equivalent version of it) will be put out for review once the global memory scheme is in place.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: grokos
Subscribers: Hahnfeld, guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D43626
llvm-svn: 326950
We have to ensure that the runtime is initialized _before_ waiting
for the two started threads to guarantee that the master threads
post their ompt_event_thread_begin before the worker threads. This
is not guaranteed in the parallel region where one worker thread
could start before the other master thread has invoked the callback.
The problem did not happen with Clang becauses the generated code
calls __kmpc_global_thread_num() and cashes its result for functions
that contain OpenMP pragmas.
Differential Revision: https://reviews.llvm.org/D43882
llvm-svn: 326435
The thread_num parameter of ompt_get_task_info() was not being used previously,
but need to be set.
The print_task_type() function (form the task-types.c testcase) was merged into
the print_ids() function (in callback.h). Testing of ompt_get_task_info() was
added to the task-types.c testcase. It was not tested extensively previously.
Differential Revision: https://reviews.llvm.org/D42472
llvm-svn: 326338
The main change of this patch is to insert {{.*}} in current_address=[[RETURN_ADDRESS_END]].
This is needed to match any of the alternatively printed addresses.
Additionally, clang-format is applied to the two tests.
Differential Revision: https://reviews.llvm.org/D43115
llvm-svn: 326312
This is required to be NULL for implicit barriers at the end of a
parallel region. Noticed in review of D43191.
Differential Revision: https://reviews.llvm.org/D43308
llvm-svn: 325922
The compiler inlines the user code in the task. Check for that case at
runtime by comparing the frame addresses and print the expected exit
address.
Also showcase how I think the OMPT tests could be reformatted to match
LLVM's code style. In my opinion it would be great to that kind of change
to all tests that need to be touched for whatever reason...
Differential Revision: https://reviews.llvm.org/D43191
llvm-svn: 325921
Test whether OMPT-callbacks for two threads that initiate a parallel region are correct.
Differential Revision: https://reviews.llvm.org/D41942
llvm-svn: 325423
Only use ompt_ functions when testing OMPT in api_calls testcase.
Add size parameter to print_list.
Fix small bug in implementation of ompt_get_partition_place_nums(): return correct length.
Differential Revision: https://reviews.llvm.org/D42162
llvm-svn: 325422
GlobalISel doesn't yet implement blockaddress and falls back to
SelectionDAG. This results in additional branch instruction to
the next basic block which breaks the OMPT tests.
Disable GlobalISel for now when compiling the tests because fixing
them is not easily possible. See http://llvm.org/PR36313 for full
discussion history.
Differential Revision: https://reviews.llvm.org/D43195
llvm-svn: 325218
This affects all outlined functions, not just tasks! Only show warning
when using Clang 5.0 or later.
Differential Revision: https://reviews.llvm.org/D43190
llvm-svn: 325131
Summary:
Different NVIDIA GPUs support different compute capabilities. To enable the inlining of runtime functions and the best performance on different generations of NVIDIA GPUs, a bc library for each compute capability needs to be compiled. The same compiler build will then be usable in conjunction with multiple generations of NVIDIA GPUs.
To differentiate between versions of the same bc lib, the output file name will contain the compute capability ID.
Depends on D14254
Reviewers: Hahnfeld, hfinkel, carlo.bertolli, caomhin, ABataev, grokos
Reviewed By: Hahnfeld, grokos
Subscribers: guansong, mgorny, openmp-commits
Differential Revision: https://reviews.llvm.org/D41724
llvm-svn: 324904
Tests the search for tools as defined in the spec. The OMP_TOOL_LIBRARIES
environment variable contains paths to the following files(in that order)
-to a nonexisting file
-to a shared library that does not have a ompt_start_tool function
-to a shared library that has an ompt_start_tool implementation returning NULL
-to a shared library that has an ompt_start_tool implementation returning a
pointer to a valid instance of ompt_start_tool_result_t
The expected result is that the last tool gets active and can print in the
thread-begin callback.
Differential Revision: https://reviews.llvm.org/D42166
llvm-svn: 324588
Add a testcase that checks wheter the runtime can handle an ompt_start_tool
method that returns NULL indicating that no tool shall be loaded.
All tool_available testcases need a separate folder to avoid file conflicts for
the generated tools.
Differential Revision: https://reviews.llvm.org/D41904
llvm-svn: 324587
Summary: This patch extends the libomptarget functionality in patch D14254 with support for the data sharing scheme for supporting implicitly shared variables. The runtime therefore maintains a list of references to shared variables.
Reviewers: carlo.bertolli, ABataev, Hahnfeld, grokos, caomhin, hfinkel
Reviewed By: Hahnfeld, grokos
Subscribers: guansong, llvm-commits, openmp-commits
Differential Revision: https://reviews.llvm.org/D41485
llvm-svn: 324495
If tool initialization returns 0, OMPT should not be active. The current
implementation provided some callback invocations in this case.
Differential Revision: https://reviews.llvm.org/D42709
llvm-svn: 324320
https://reviews.llvm.org/D42757
The method ThreadsInTeam is used to determine the number of threads to be used in a parallel region under SPMD mode (see line 127 of supporti.h in libomptarget/deviceRTLs/nvptx/src/). This patch fixes the corresponding debug print upon initialization of the kernel in SPMD mode.
llvm-svn: 323978
That's what we really need to link the CUDA plugin against,
not the CUDA runtime API in CUDA_LIBRARIES! While the latter
comes with the CUDA SDK, the Driver API is installed with
the kernel driver and there is at most one per system. As
fallback we can use the stubs library distributed with the
CUDA SDK for linking.
Differential Revision: https://reviews.llvm.org/D42643
llvm-svn: 323787
Use equivalents for the last calls to the Runtime API. Remove
stray assert in case of an error found during review, we should
only return OFFLOAD_FAIL.
Differential Revision: https://reviews.llvm.org/D42686
llvm-svn: 323786
This patch implements the device runtime library whose interface is used in the code generation for OpenMP offloading devices.
Currently there is a single device RTL written in CUDA meant to CUDA enabled GPUs.
The interface is a variation of the kmpc interface that includes some extra calls to do thread and storage management that only make sense for a GPU target.
Differential revision: https://reviews.llvm.org/D14254
llvm-svn: 323649
Use fuzzy return addresses in lock testcases so that these
testcases can also be run using the Intel Compiler.
Patch by Simon Convent!
Differential Revision: https://reviews.llvm.org/D41896
llvm-svn: 323529
Add Workaround for Intel Compiler Bug with Case#: 03138964
A critical region within a nested task causes a segfault in icc 14-18:
int main()
{
#pragma omp parallel num_threads(2)
#pragma omp master
#pragma omp task
#pragma omp task
#pragma omp critical
printf("test\n");
}
When the critical region is in a separate function, the segault does not occur.
So we add noinline to make sure that the function call stays there.
Differential Revision: https://reviews.llvm.org/D41182
llvm-svn: 322622
The defintion is not part of the spec and thus should not have the prefix
"ompt_" but rather a prefix that indicates that this is implementation
specific.
Differential Revision: https://reviews.llvm.org/D41166
llvm-svn: 322621
When the current thread is not an (initialized) OpenMP thread, the runtime
entry points return values that correspond to "not available" or similar
Differential Revision: https://reviews.llvm.org/D41167
llvm-svn: 322620
If user requested affinity with granularity=tile we need to either use HWLOC
or ignore the request. The change allows user to not specify
KMP_TOPOLOGY_METHOD=hwloc and choose it automatically instead.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D40905
llvm-svn: 322205
This change simplifies __kmp_expand_threads to take a single argument.
Previously, it allowed two arguments and had logic to decide on different
potential expansion sizes. However, no calls to __kmp_expand_threads in the
runtime make use of this extra logic. Thus the extra argument and logic is
removed here.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D41836
llvm-svn: 322204
This change improves stability of the runtime when the application forks child
processes. Acquiring/releasing __kmp_initz_lock and __kmp_forkjoin_lock in the
atfork handlers insures that the actual fork does not occur while those two
locks are held, and __kmp_itt_reset() reverts the itt's global state to the
initial state which also initializes the mutex stored in the global state.
Some missing initialization code was also inserted in the child's atfork handler.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D41462
llvm-svn: 322202
This field is defined as kmp_int32, so we should use neither
pointers to kmp_int64 nor 64 bit atomic instructions.
(Found while testing on a Raspberry Pi, 32 bit ARM)
Differential Revision: https://reviews.llvm.org/D41656
llvm-svn: 321964
This patch enables OMPT by default if version 50 or later is built and the config says, that OMPT will be supported.
Differential Revision: https://reviews.llvm.org/D41508
llvm-svn: 321675
We now have several options that apply for both libraries and they
shouldn't be documented in multiple files. When already merging
the two Build_With_CMake.txt documents, convert them to
reStructuredText which is used for all of LLVM's documentation.
Differential Revision: https://reviews.llvm.org/D40920
llvm-svn: 321481
As for normal task creation, the task frame addresses need to be stored
for the encountering task.
Differential Revision: https://reviews.llvm.org/D41165
llvm-svn: 321421
The compiler warns that _BSD_SOURCE is deprecated and _DEFAULT_SOURCE should
be used instead. We keep _BSD_SOURCE for older compilers, that don't know
about _DEFAULT_SOURCE.
The linker drops the tool when linking, since there is no visible need for
the library. So we need to tell the linker, that the tool should be linked
anyway.
Differential Revision: https://reviews.llvm.org/D41499
llvm-svn: 321362
The format string for hints only prints the second argument (string) and drops
the first argument (hint id). Depending on how you read the POSIX text for
printf, this could be valid. But for practical reason, i.e., unpacking the
va_list passed to printf based on the formating information, it makes sense
to fix the implementation and not pass the id for hint.
Failing testcases were:
misc_bugs/teams-reduction.c
ompt/parallel/not_enough_threads.c
Differential Revision: https://reviews.llvm.org/D41504
llvm-svn: 321361
This function is defined in OpenMP-TR6 section 4.1.5.1.6
The functions was not implemented yet.
Since ompt-functions can only be called after the runtime was initialized and
has loaded a tool, it can assume the runtime to be initialized. In contrast
to omp_get_num_procs which needs to check whether the runtime is initialized.
Differential Revision: https://reviews.llvm.org/D40949
llvm-svn: 321269
This revision fixes failing testcases with parallel for loops and the gomp
interface. The return address needs to be stored at entry to runtime.
The storage is cleared on usage, so we need to update the storage before
calling again internal functions, that will trigger event callbacks.
Differential Revision: https://reviews.llvm.org/D41181
llvm-svn: 321265
We use the bitmap ompt_enabled thoughout the runtime, to avoid loading the
vector of callback functions when testing if specific code should be executed.
Before invoking an event callback function, the pointer is tested for NULL.
This revision resets the corresponding bit in ompt_enabled to 0 if
NULL is passed in set_callback.
Differential Revision: https://reviews.llvm.org/D41171
llvm-svn: 321264
Clang 5 or higher adds an intermediate function call in certain cases when
compiling with debug flag. This revision updates the testcases to work
correctly.
Differential Revision: https://reviews.llvm.org/D40595
llvm-svn: 321263
Reasons for expected failures are mainly bugs when using lables in OpenMP regions
or missing support of some OpenMP features.
For some worksharing clauses, support to distinguish the kind of workshare was
added just recently.
If an issue was fixed in a minor release version of a compiler, we flag the
test as unsupported for this compiler version to avoid false positives.
Same for fixes that where backported to older compiler versions.
Differential Revision: https://reviews.llvm.org/D40384
llvm-svn: 321262
There are two /proc/cpuinfo layots in use for AArch64: old and new.
The old one has all 'processor : n' lines in one section, hence
checking for duplications does not make sense.
Differential Revision: https://reviews.llvm.org/D41000
llvm-svn: 320593
All architectures except x86_64 used the linear barrier implementation
by default which doesn't give good performance for a larger number
of threads.
Improvements for PARALLEL overhead (EPCC) with this patch on a Power8
system (2 sockets x 10 cores x 8 threads, OMP_PLACES=cores)
20 threads: 4.55us -> 3.49us
40 threads: 8.84us -> 4.06us
80 threads: 19.18us -> 4.74us
160 threads: 54.22us -> 6.73us
Differential Revision: https://reviews.llvm.org/D40358
llvm-svn: 320152
To make thread affinity work according to the OpenMP spec, the
runtime needs information about the hardware topology. On Linux
the default way is to parse /proc/cpuinfo which contains this
information for x86 machines but (at least) not for AArch64 and
Power architectures.
Fortunately, there is a different code path which is able to get
that data from sysfs. The needed patch has landed in 2006 for
Linux 2.6.16 which is safe to assume nowadays (even RHEL 5 had
a kernel version derived from 2.6.18, and we are now at RHEL 7!).
Differential Revision: https://reviews.llvm.org/D40357
llvm-svn: 320151
Otherwise I see hangs in the omp_single_copyprivate test when
compiling in release mode. With the debug assertions, I get a
failure `head > 0 && tail > 0`.
Differential Revision: https://reviews.llvm.org/D40722
llvm-svn: 320150
This last of four patches adds a new file for the interface
functions that Clang uses during code generation. The only
change except simply moving the current code is renaming the
function CheckDeviceAndCtors() and using the correct type for
64bit device ids.
Differential Revision: https://reviews.llvm.org/D40801
llvm-svn: 319972
This third patch moves the implementation of the user-facing
OpenMP API functions into its own file. For now, the code is
only moved, no cleanups applied yet.
Differential Revision: https://reviews.llvm.org/D40800
llvm-svn: 319971
This is the second patch to split the current monolithic
implementation into separate files. Note that this change
doesn't cleanup the code yet.
Differential Revision: https://reviews.llvm.org/D40799
llvm-svn: 319970
This is the first of four patches to split the target agnostic
library into multiple (smaller) files. It only moves the code
to separate implementation files and does no cleanup (yet) except
removing unneeded headers.
Differential Revision: https://reviews.llvm.org/D40798
llvm-svn: 319969
Future patches will add (private) header files in src/ that should
not be visible to plugins, so move the "public" ones to a new
include/ directory. This is still internal in a sense that the
contained files won't be installed for the user.
Similarly, the target agnostic offloading library should be built
directly in src/. The parent directory is responsible for finding
dependencies and including all subdirectories.
Differential Revision: https://reviews.llvm.org/D40797
llvm-svn: 319968
Redundant extra verbose output of binding to full mask in case
affinity=balanced or OMP_PLACES=<any> or OMP_PROC_BIND=<any>
Differential Revision: https://reviews.llvm.org/D40624
llvm-svn: 319960
This change is a trivial fix for enums that removes specification of "last" or
"upper" values, or other boundary values. This simplifies the code in places,
and results in never needing to update the "upper" values again.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D40804
llvm-svn: 319957
The runtime will use the global kmp_critical_name as a lock and
tries to atomically store a pointer in there. This will fail
if the global is only aligned by 4 bytes, the size of one int32_t
element. Use a union to ensure the global is aligned to the size
of a pointer on the current platform.
llvm-svn: 319811
__kmpc_reduce_nowait() correctly swapped the teams for reductions
in a teams construct. Apply the same logic to __kmpc_reduce() and
__kmpc_reduce_end().
Differential Revision: https://reviews.llvm.org/D40753
llvm-svn: 319788
Perform a nested CMake invocation to avoid writing our own parser
for compiler versions when we are not testing the in-tree compiler.
Use the extracted information to mark a test as unsupported that
hangs with Clang prior to version 4.0.1 and restrict tests for
libomptarget to Clang version 6.0.0 and later.
Differential Revision: https://reviews.llvm.org/D40083
llvm-svn: 319448
This change makes kmp_r_sched_t type into a union for simpler
comparisons and assignments
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D40374
llvm-svn: 319379
kmp_aligned_malloc() always returned NULL on Windows (stub library only)
that may cause Fortran application crash. With this change all memory
allocation functions were fixed to use aligned{m,re,rec}alloc() to
allocate/reallocate memory. To deallocate that memory _aligned_free() is
used in kmp_free().
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D40296
llvm-svn: 319375
Added two warnings:
1) Before building the topology map check if tiles are requested but the
topo method is not hwloc;
2) After building the topology map check if tiles are requested but not
detected by the library.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D40340
llvm-svn: 319374
Fortran array elements made default integer in OMP_GET_PLACE_PROC_IDS and
OMP_GET_PARTITION_PLACE_NUMS subroutines, otherwise call to them produces
incorrect result.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D40356
llvm-svn: 319372
The code for the two OpenMP runtime libraries was very similar.
Move to common CMake file that is included and provides a simple
interface for adding testsuites. Also add a common check-openmp
target that runs all testsuites that have been registered.
Note that this renames all test options to the common OPENMP
namespace, for example OPENMP_TEST_C_COMPILER instead of
LIBOMP_TEST_COMPILER and so on.
Differential Revision: https://reviews.llvm.org/D40082
llvm-svn: 319343
These are needed by both libraries, so we can do that in a
common namespace and unify configuration parameters.
Also make sure that the user isn't requesting libomptarget
if the library cannot be built on the system. Issue an error
in that case.
Differential Revision: https://reviews.llvm.org/D40081
llvm-svn: 319342
As a first step, this allows us to generalize the detection of
standalone builds and make it fully compatible when building in
llvm/runtimes/ which automatically sets OPENMP_STANDLONE_BUILD.
Differential Revision: https://reviews.llvm.org/D40080
llvm-svn: 319341
Summary:
We want to automatically copy the appropriate mailing list
for review requests to the openmp repository.
For context, see the proposal and discussion here:
http://lists.llvm.org/pipermail/cfe-dev/2017-November/056032.html
Similar to D40179, I set up a new Diffusion repository with callsign
"OMP" for OpenMP:
https://reviews.llvm.org/source/openmp/
This explicitly updates openmp's .arcconfig to point to the new
OMP repository in Diffusion, which will let us use Herald rule H272
to automatically subscribe openmp-commits to review requests.
Reviewers: hans, grokos, Hahnfeld
Reviewed By: grokos
Subscribers: sammccall, klimek, openmp-commits
Differential Revision: https://reviews.llvm.org/D40499
llvm-svn: 319254
Power has a weak consistency model so we need memory barriers to
make writes (both from runtime and from user code) available for
all threads.
Differential Revision: https://reviews.llvm.org/D40175
llvm-svn: 318848
We have just fixed the codegen of omp_is_initial_device() to reliably work
when offloading to the same device, see commit r316001. This fixes the
failing tests that were the reason why we disabled the library for 5.0.
Differential Revision: https://reviews.llvm.org/D39052
llvm-svn: 318847
This is the libomptarget-side patch which changes the __tgt_* API function signatures in preparation for the new map interface.
Changes are: Device IDs 32bits --> 64bits, Flags 32bits --> 64bits
Differential revision: https://reviews.llvm.org/D40313
llvm-svn: 318790
These tests were failing rarely on my MacBook when there was some
activity in the background. Read: one of a thousand executions?
* sections.c missed the sorting based on thread ids. This worked
as long as the master thread finished its section before the
worker thread started the second one but failed if the master
thread was put to sleep by the OS.
* The checks in single.c assumed that the master thread executes
the single region which works most of the time because it is
usually faster than the newly spawned worker thread.
Differential Revision: https://reviews.llvm.org/D39853
llvm-svn: 318527
The testsuite directory is not used or updated and confuses new users to the
OpenMP project. These tests were rewritten using the lit format and put under
the runtime/test directory. This patch removes the entire testsuite/ directory.
Differential Revision: https://reviews.llvm.org/D39767
llvm-svn: 318056
Traditionally, the library had a weak symbol for ompt_start_tool()
that served as fallback and disabled OMPT if called. Tools could
provide their own version and replace the default implementation
to register callbacks and lookup functions. This mechanism has
worked reasonably well on Linux systems where this interface was
initially developed.
On Darwin / Mac OS X the situation is a bit more complicated and
the weak symbol doesn't work out-of-the-box. In my tests, the
library with the tool needed to link against the OpenMP runtime
to make the process work. This would effectively mean that a tool
needed to choose a runtime library whereas one design goal of the
interface was to allow tools that are agnostic of the runtime.
The solution is to use dlsym() with the argument RTLD_DEFAULT so
that static implementations of ompt_start_tool() are found in the
main executable. This works because the linker on Mac OS X includes
all symbols of an executable in the global symbol table by default.
To use the same code path on Linux, the application would need to
be built with -Wl,--export-dynamic. To avoid this restriction, we
continue to use weak symbols on Linux systems as before.
Finally this patch extends the existing test to cover all possible
ways of initializing the tool as described by the standard. It
also fixes ompt_finalize() to not call omp_get_thread_num() when
the library is shut down which resulted in hangs on Darwin.
The changes have been tested on Linux to make sure that it passes
the current tests as well as the newly extended one.
Differential Revision: https://reviews.llvm.org/D39801
llvm-svn: 317980
In standalone build, plugins where previously built in their
subdirectory in plugins/ and tests couldn't find them.
Differential Revision: https://reviews.llvm.org/D39920
llvm-svn: 317979
For up-to-date compilers, this assertion is reasonable, but it breaks
compatibility with the typical compiler installed on most systems.
This patch changes the default value to what we had when there was no
compiler support. A warning about the outdated compiler is printed during
runtime, when this point is reached.
Differential Revision: https://reviews.llvm.org/D39890
llvm-svn: 317928
In these places the const attribute seems correct and doesn't
need any other change, so let's do it.
Differential Revision: https://reviews.llvm.org/D39756
llvm-svn: 317798
Allocated memory is typically not 'const' if it needs to be freed.
This patch removes around 50 wrong const attributes, modifies the
corresponding functions and finally gets rid of some const_casts.
These have especially been strange for __kmp_str_fname_free() that
added a 'const' to call __kmp_str_free() which removed it again.
Two minor cleanups that I performed in this process:
* __kmp_tool_libraries now lives in kmp_settings.cpp as it is
used nowhere else.
* __kmp_msg_empty was removed as it was never used and Clang
now complained that it was assigned a string literal that
is 'const char *'.
Differential Revision: https://reviews.llvm.org/D39755
llvm-svn: 317797
If a parallel region is cancelled, execution resumes at the end
of the structured block. That is why this test cannot use the
"normal" macros that print right after inserting the label.
Instead it previously printed the addresses before the pragma
and swapped the checks compared to the other tests.
However, this does not work because FileChecks '*' is greedy
so that RETURN_ADDRESS always matched the second address. This
makes the test fail when an "overflow" occurrs and the first
address matches the value of codeptr_ra.
I discovered this on my MacBook but I'm unable to reproduce the
failure with the current version. Nevertheless we should fix this
problem to avoid that this test fails later after an unrelated change.
Differential Revision: https://reviews.llvm.org/D39708
llvm-svn: 317787
Return addresses are determined based on the address of a label
that is inserted directly after a pragma / API call. In some cases
the tests can assume a known number of instructions between the
addresses. However, the instructions and their encoded lengths
depend on the target that the test is compiled on.
Firstly, this patch refactors the macro print_current_address() to
allow such target dependent modifications and adds information for
the observed instructions on POWER. Secondly, it adapts the related
macro print_fuzzy_address() to reuse much of "hacky" code and fixes
the used formatting strings in the printf() call. Finally, it also
adds documentation about how these macros are intended to work.
Differential Revision: https://reviews.llvm.org/D39699
llvm-svn: 317786
1) Get rid of xaliasify, xexpand and xversionify for KMP_EXPAND_NAME and
KMP_VERSION_SYMBOL. KMP_VERSION_SYMBOL is a combination of xaliasify and
xversionify.
2) Put all attribute and __declspec definitions in kmp_os.h
Differential Revision: https://reviews.llvm.org/D39516
llvm-svn: 317636
The TR6 document is expected to be publically released around November 15.
This patch does not implement OMPT for libomptarget.
Patch by Simon Convent and Joachim Protze
Differential Revision: https://reviews.llvm.org/D39182
llvm-svn: 317436
This is part of the renaming of data types from OpenMP TR4 to TR6
Patch by Simon Convent
Differential Revision: https://reviews.llvm.org/D39326
llvm-svn: 317435
The TR6 document is expected to be publically released around November 15.
This patch does not implement OMPT for libomptarget.
Patch by Simon Convent and Joachim Protze
Differential Revision: https://reviews.llvm.org/D39182
llvm-svn: 317339
This is part of the renaming of data types from OpenMP TR4 to TR6
Patch by Simon Convent
Differential Revision: https://reviews.llvm.org/D39326
llvm-svn: 317338
This was never used in the upstream compiler and was responsible
for some problems with reductions in the clang-ykt fork.
Differential Revision: https://reviews.llvm.org/D39553
llvm-svn: 317214
This is a partial fix for bug 34050.
This prevents callers of omp_set_lock (which does not hold __kmp_global_lock)
from ever seeing an uninitialized version of __kmp_i_lock_table.table.
It does not solve a use-after-free race condition if omp_set_lock obtains a
pointer to __kmp_i_lock_table.table before it is updated and then attempts to
dereference afterwards. That race is far less likely and can be handled in a
separate patch.
The unit test usually segfaults on the current trunk revision. It passes with
the patch.
Patch by Adam Azarchs
Differential Revision: https://reviews.llvm.org/D39439
llvm-svn: 317115
The code is tested to work with latest clang, GNU and Intel compiler. The implementation
is optimized for low overhead when no tool is attached shifting the cost to execution with
tool attached.
This patch does not implement OMPT for libomptarget.
Patch by Simon Convent and Joachim Protze
Differential Revision: https://reviews.llvm.org/D38185
llvm-svn: 317085
Replacing call to __kmp_msg(kmp_ms_fatal,...) with __kmp_fatal(...) caused an
issue when incomplete message is displayed in case an error message is followed
by another message (e.g. by a hint messa)ge. This is because __kmp_fatal()
passes incomplete list of arguments to __kmp_msg().
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D39248
llvm-svn: 316623
The problem is due to the runtime's threadprivate cleanup code which tries to
access data that was already destroyed by one of the root threads.
__kmp_init_gtid is used as a checker here since it is set to false before actual
resource cleanup is done in __kmp_cleanup().
Patch by Hansang Bae
llvm-svn: 316452
If both KMP_HW_SUBSET and KMP_PLACE_THREADS are set and KMP_PLACE_THREADS gets
parsed first, then the current environment variable parser rejects both and
neither get used. This patch uses the rivals mechanism that is used for other
environment variable groups (e.g., KMP_STACKSIZE, GOMP_STACKSIZE, OMP_STACKSIZE).
If both are set, then it tells the user that it is ignoring KMP_PLACE_THREADS in
favor of KMP_HW_SUBSET. The message about deprecating KMP_PLACE_THREADS when it
is set is still printed regardless.
Differential Revision: https://reviews.llvm.org/D38292
llvm-svn: 315091
Removes semicolons after if {} blocks, function definitions, etc.
I was able to apply the large OMPT patch cleanly on top of this one
with no conflicts.
llvm-svn: 314340
Minor code cleanup of Klocwork issues. Fatal messages are given no return
attribute. Define and use KMP_NORETURN to work for multiple C++ versions.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D37275
llvm-svn: 312538
Cleanup code to remove BUILD_TV and unused code bracketed by it.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D36011
llvm-svn: 311114
Disable default debugging dumps for libomptarget and plugins and move dumps
under control of environment variable LIBOMPTARGET_DEBUG=<integer>. Dumps
are enabled when LIBOMPTARGET_DEBUG is set to a positive integer value.
Debugging dumps are available only in debug build; release build does not
support it.
Differential Revision: https://reviews.llvm.org/D33227
llvm-svn: 310841
This change improves the way threads are spread across cores
when OMP_PROC_BIND=spread is set and no unusual affinity masks are in use.
Differential Revision: https://reviews.llvm.org/D36510
llvm-svn: 310670
We use symbol versioning for GNU-compatibility but libgomp has versioned symbols
only in the shared library but not in the static.
Moreover, version symbols in the static library can cause an error at link time.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D36225
llvm-svn: 309877
The original locations can be reached without initializing the lock variable
(td_deque_lock), so it is potentially unsafe. It is guaranteed that the lock
is initialized if the deque (td_deque) is not NULL, and lock functions can be
safely called.
Patch by Hansang Bae
Differential Revision: https://reviews.llvm.org/D36017
llvm-svn: 309875
This change adds a new environment variable, KMP_TEAMS_THREAD_LIMIT, which is
used to set a new global variable, __kmp_teams_max_nth, which is checked when
determining the size and quantity of teams that will be created in the teams
construct. Specifically, it is a limit on the total number of threads in a given
teams construct. It differentiates the limits for the teams construct from the
limits for regular parallel regions (KMP_DEVICE_THREAD_LIMIT/__kmp_max_nth and
OMP_THREAD_LIMIT/__kmp_cg_max_nth). When each individual team is formed, it is
still subject to those limits. After the clauses to the teams construct are
parsed and calculated, we check to make sure we are within this limit, and if
not, reduce num_threads per team and/or number of teams, accordingly. The
default value is set to the number of available processors on the system.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D36009
llvm-svn: 309874
This change fixes the implementation of OMP_THREAD_LIMIT. The implementation of
this previously was not restricted to a contention group (but it should be,
according to the spec), and this is fixed here. A field is added to root thread
to store a counter of the threads in the contention group. An extra check is
added when reserving threads for a parallel region that checks this variable and
compares to threadlimit-var, which is implemented as a new global variable,
kmp_cg_max_nth. Associated settings changes were also made, and clean up of
comments that referred to OMP_THREAD_LIMIT, but should refer to the new
KMP_DEVICE_THREAD_LIMIT (added in an earlier patch).
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D35912
llvm-svn: 309319
This change drops in KMP_DEVICE_THREAD_LIMIT to replace KMP_MAX_THREADS. It's
possible there will eventually be a OMP_DEVICE_THREAD_LIMIT, and we need
something to distinguish from OMP_THREAD_LIMIT, which is currently implemented
incorrectly (the fix for that will be added soon in a separate patch).
KMP_ALL_THREADS is deprecated here, but we can keep the "all" option on
KMP_DEVICE_THREAD_LIMIT to support that functionality. KMP_DEVICE_THREAD_LIMIT
now has priority over its deprecated rival KMP_ALL_THREADS. I also cleaned up
some comments that incorrectly referred to non-existent kmp_max_threads variable
instead of kmp_max_nth.
I've left the name of where this setting eventually ends up as
__kmp_max_nth, for now.
This change does not change much in the way of functionality. It does NOT change
OMP_THREAD_LIMIT. It's just cleaning up and setting up for that.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D35860
llvm-svn: 309168
Introduce OPENMP_ENABLE_LIBOMPTARGET which defaults to OFF at the moment.
libomptarget is not yet ready for prime time:
- Offloading to NVIDIA GPUs is not completed yet (compiler, device RTL)
- The generic ELF plugin for offloading to the host (meant for testing)
uses a single instance of the OpenMP runtime (libomp). That is why
omp_is_initial_device() returns 1 which makes the tests fail.
Because of these reasons, we want to disable building (and testing!)
for release 5.0.
See https://bugs.llvm.org/show_bug.cgi?id=33859
Differential Revision: https://reviews.llvm.org/D35719
llvm-svn: 309115
Removed unused __kmp_env_* variables. Also clangified other people's code.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D35808
llvm-svn: 309000
Summary:
The kmp_os.h header is defining the `PAGE_SIZE` macro unconditionally,
even while it is only used directly after its definition, for the
Windows implementation of the `KMP_GET_PAGE_SIZE()` macro.
On at least FreeBSD, but likely all other BSDs too, this macro conflicts
with the one defined in system headers, so remove it, since nothing else
uses it. Make all Unixes use `getpagesize()` instead, and use
`GetSystemInfo()` for the Windows case.
Reviewers: jlpeyton, jcownie, emaste, AndreyChurbanov
Reviewed By: AndreyChurbanov
Subscribers: AndreyChurbanov, hfinkel, zturner
Differential Revision: https://reviews.llvm.org/D35072
llvm-svn: 308355
Summary:
Taskloop implementation is extended by using recursive task scheduling.
Envirable KMP_TASKLOOP_MIN_TASKS added as a manual threshold for the user
to switch from recursive to linear tasks scheduling.
Details:
* The calculations for the loop parameters are moved from __kmp_taskloop_linear
upper level
* Initial calculation is done in the __kmpc_taskloop, further range splitting
is done in the __kmp_taskloop_recur.
* Added threshold to switch from recursive to linear tasks scheduling;
* One half of split range is scheduled as an internal task which just moves
sub-range parameters to the stealing thread that continues recursive
scheduling (if number of tasks still enough), the other half is processed
recursively;
* Internal task duplication routine fixed to assign parent task, that was not
needed when all tasks were scheduled by same thread, but is needed now.
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D35273
llvm-svn: 308338
The internal details of this setting are not meant to be user visible and only create confusion.
Differential Revision: https://reviews.llvm.org/D35269
llvm-svn: 308189
Changes are: got all atomics to accept volatile pointers that allowed
to simplify many type conversions. Windows specific code fixed correspondingly.
Differential Revision: https://reviews.llvm.org/D35417
llvm-svn: 308164
The first bit is actually the "untied" flag. That is why the condition was
wrong and has to be inverted to set the flag correctly.
Found and initial patch by Simon Convent!
llvm-svn: 307899
Summary:
On Unix, a .S file is normally an assembly source which must be
preprocessed with a C preprocessor, while a .s file is "plain" assembly.
The former is handled by the compiler driver (cc), the latter is
directly passed to the assembler binary (as).
Because z_Linux_asm.s is supposed to be preprocessed, rename it to .S,
so it can be automatically picked up correctly by build systems.
Reviewers: AndreyChurbanov, emaste, jlpeyton
Reviewed By: AndreyChurbanov
Subscribers: mgorny, openmp-commits
Differential Revision: https://reviews.llvm.org/D35171
llvm-svn: 307680
While importing libomp into the FreeBSD base system we encountered
Clang warnings that "'register' storage class specifier is deprecated
and incompatible with C++1z [-Wdeprecated-register]".
Differential Revision: https://reviews.llvm.org/D35124
llvm-svn: 307441
Changes are: replaced C-style casts with cons_cast and reinterpret_cast;
type of several counters changed to signed; type of parameters of 32-bit and
64-bit AND and OR intrinsics changes to unsigned; changed files formatted
using clang-format version 3.8.1.
Differential Revision: https://reviews.llvm.org/D34759
llvm-svn: 307020
I've found it very difficult to get test/parallel/omp_nested.c to pass
consistently across my build environments. The problem is that it creates N^2
threads (it is testing nested parallel regions), and that often exceeds the
thread limits on systems with many cores. We do raise the process limits in
lit, and that often helps, but if running lit with a smaller number of threads
or on a system where we're otherwise resource constrained, this particular test
tends to fail (because the runtime cannot create a sufficient number of
threads).
This seems to work: if the maximum number of threads is more than some small
number, then cap the number of threads used for the parallel region. The choice
of 4 here is somewhat arbitrary.
Differential Revision: https://reviews.llvm.org/D32033
llvm-svn: 306357
Summary: On BSDs, there is no `libdl.so`, and functions like `dlopen`
are implemented in the main C library instead. Use the `CMAKE_DL_LIBS`
variable instead of hardcoding a dependency on the `dl` library.
Reviewers: grokos, joerg, emaste
Reviewed By: emaste
Subscribers: jlpeyton, mgorny, openmp-commits
Differential Revision: https://reviews.llvm.org/D34632
llvm-svn: 306319
Reset affinity to none (false for proc-bind-var) so that threads in the child
processes are not bound tightly, unless the user explicitly sets this in
KMP_AFFINITY/OMP_PROC_BIND, in child processes. This can improve
performance for scripting languages which fork for parallelism like Python's
multiprocessing module.
Differential Revision: https://reviews.llvm.org/D34154
llvm-svn: 305513
If OpenMP is initialized before fork()-ing occurs and affinity is set to
something like compact, then the master thread will be pinned to a single HW
thread/core after initialization. If the master (or any other thread) then
forks N processes, all N processes will then be pinned to that same single HW
thread/core. To reset the affinity for the new child process, the atfork
handler for the child process can call kmp_set_thread_affinity_mask_initial()
to reset its affinity to the initial affinity of the application before it
re-initializes libomp. The parent process will not be affected and still
keeps its affinity setting.
Differential Revision: https://reviews.llvm.org/D34118
llvm-svn: 305306
Fix static initializers to use the proper unlocked value for the poll
field of the tas and futex locks.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D33794
llvm-svn: 304828
Some code was restructured to move it under KMP_DEBUG. The rest is
formatting changes to fix some things broken by clang-format
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D33744
llvm-svn: 304438
With these settings, the create_hwloc_map() method was being called causing an
assert(). After some consideration, it was determined that disabling affinity
explicitly should just disable hwloc as well. i.e., KMP_AFFINITY overrides
KMP_TOPOLOGY_METHOD. This lets the user know that the Hwloc mechanism is being
ignored when KMP_AFFINITY=disabled.
Differential Revision: https://reviews.llvm.org/D33208
llvm-svn: 304344
This change checks if the initial affinity mask is equal to exactly one
Windows processor group's affinity mask. If it is, then the code does not
respect the initial affinity mask and uses the entire machine instead.
The reasoning behind this is that, by default, Windows assigns exactly one
processor group as the initial affinity mask even when there are multiple
Windows processor groups available. User's typically want to use the whole
machine, so we ignore this special case and use the entire machine.
If the initial affinity mask is a proper subset of one group, or spans multiple
groups, then the initial affinity mask is respected since we can assume that the
operating system did not assign this initial affinity mask. This change only
affects machines with multiple processor groups
Differential Revision: https://reviews.llvm.org/D33210
llvm-svn: 304343
An assert() was being tripped when KMP_AFFINITY=respect + Multiple Processor
Groups. Let __kmp_affinity_create_proc_group_map() function be able to create
address2os object which contains a single group by deleting restriction that
process affinity mask must span multiple groups.
llvm-svn: 303101
This patch contains the clang-format and cleanup of the entire code base. Some
of clang-formats changes made the code look worse in places. A best effort was
made to resolve the bulk of these problems, but many remain. Most of the
problems were mangling line-breaks and tabbing of comments.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D32659
llvm-svn: 302929
This patch chagnes the plugin interface so that:
1) future plugins can take advantage of systems with shared CPU/device storage
2) instead of using base addresses, target regions are launched by providing target addresseds and base offsets explicitly.
Differential revision: https://reviews.llvm.org/D33028
llvm-svn: 302663
Older Hwloc libraries (< 1.10.0) don't offer the HWLOC_OBJ_NUMANODE nor
HWLOC_OBJ_PACKAGE types. Instead they are named HWLOC_OBJ_NODE and
HWLOC_OBJ_SOCKET instead. This patch just defines the newer names based on
the older names when using an older Hwloc.
Differential Revision: https://reviews.llvm.org/D32496
llvm-svn: 301349
Without this fix cancellation status for parallel, sections and for persists
across construct boundaries.
Differential Revision: https://reviews.llvm.org/D31419
llvm-svn: 299434
Jonathan Peyton