This is an archive of the discontinued LLVM Phabricator instance.

Differential D106374

[OpenMP][Documentation] Fix hyperlink location
ClosedPublic

Authored by jhuber6 on Jul 20 2021, 9:07 AM.

Details

Reviewers
jdoerfert
Commits
rG6242f9b96697: [OpenMP][Documentation] Fix hyperlink location
Summary

Fixes the documentation hyperlinks not showing the header.

Diff Detail

Event Timeline

jhuber6 created this revision.Jul 20 2021, 9:07 AM
Herald added subscribers: guansong, yaxunl. · View Herald TranscriptJul 20 2021, 9:07 AM
jhuber6 requested review of this revision.Jul 20 2021, 9:07 AM
Herald added a project: Restricted Project. · View Herald TranscriptJul 20 2021, 9:07 AM
Herald added subscribers: openmp-commits, sstefan1. · View Herald Transcript
jdoerfert accepted this revision.Jul 20 2021, 9:41 AM

LG, feel free to submit things like this w/o review. Post commit review for such changes is sufficient, especially if this has been discussed.

This revision is now accepted and ready to land.Jul 20 2021, 9:41 AM
This revision was landed with ongoing or failed builds.Jul 20 2021, 9:42 AM
Closed by commit rG6242f9b96697: [OpenMP][Documentation] Fix hyperlink location (authored by jhuber6). · Explain Why
This revision was automatically updated to reflect the committed changes.
jhuber6 added a commit: rG6242f9b96697: [OpenMP][Documentation] Fix hyperlink location.
Harbormaster completed remote builds in B115123: Diff 360166.Jul 20 2021, 10:08 AM

Revision Contents

PathSize
openmp/
docs/
remarks/
6 lines
4 lines
4 lines
4 lines
4 lines
4 lines
4 lines
4 lines
4 lines
4 lines
4 lines
4 lines
4 lines
4 lines

Diff 360178

openmp/docs/remarks/OMP100.rst

Potentially unknown OpenMP target region caller `[OMP100]`
==========================================================
.. _omp100: .. _omp100:
.. _omp_no_external_caller_in_target_region: .. _omp_no_external_caller_in_target_region:
Potentially unknown OpenMP target region caller `[OMP100]`
==========================================================
A function remark that indicates the function, when compiled for a GPU, is 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 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 only issued if we tried to perform an optimization which would require us to
know all callers on the GPU. know all callers on the GPU.
To facilitate OpenMP semantics on GPUs we provide a runtime mechanism through 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 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 threads in the team. Generally we use the address of the outlined parallel
Show All 12 Lines

openmp/docs/remarks/OMP101.rst

.. _omp101:
Parallel region is used in unknown / unexpected ways. Will not attempt to rewrite the state machine. [OMP101] Parallel region is used in unknown / unexpected ways. Will not attempt to rewrite the state machine. [OMP101]
============================================================================================================= =============================================================================================================
.. _omp101:
An analysis remark that indicates that a parallel region has unknown calls. An analysis remark that indicates that a parallel region has unknown calls.

openmp/docs/remarks/OMP102.rst

.. _omp102:
Parallel region is not called from a unique kernel. Will not attempt to rewrite the state machine. [OMP102] Parallel region is not called from a unique kernel. Will not attempt to rewrite the state machine. [OMP102]
=========================================================================================================== ===========================================================================================================
.. _omp102:
This analysis remark indicates that a given parallel region is called by This analysis remark indicates that a given parallel region is called by
multiple kernels. This prevents the compiler from optimizing it to a single multiple kernels. This prevents the compiler from optimizing it to a single
kernel and rewrite the state machine. kernel and rewrite the state machine.

openmp/docs/remarks/OMP110.rst

.. _omp110:
Moving globalized variable to the stack. [OMP110] Moving globalized variable to the stack. [OMP110]
================================================= =================================================
.. _omp110:
This optimization remark indicates that a globalized variable was moved back to 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 thread-local stack memory on the device. This occurs when the optimization pass
can determine that a globalized variable is not possibly be shared between can determine that a globalized variable is not possibly be shared between
threads and globalization was unnecessary. Using stack memory is the best-case threads and globalization was unnecessary. Using stack memory is the best-case
scenario for data globalization as the variable can now be stored in fast 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 register files on the device. This optimization requires full visibility of each
variable. variable.
▲ Show 20 Lines • Show All 70 Lines • Show Last 20 Lines

openmp/docs/remarks/OMP111.rst

.. _omp111:
Replaced globalized variable with X bytes of shared memory. [OMP111] Replaced globalized variable with X bytes of shared memory. [OMP111]
==================================================================== ====================================================================
.. _omp111:
This optimization occurs when a globalized variable's data is shared between This optimization occurs when a globalized variable's data is shared between
multiple threads, but requires a static amount of memory that can be determined multiple threads, but requires a static amount of memory that can be determined
at compile time. This is the case when only a single thread creates the memory at compile time. This is the case when only a single thread creates the memory
and then shares is between every thread. The memory can then be pushed to a and then shares is between every thread. The memory can then be pushed to a
static buffer of shared memory on the device. This optimization allows users to static buffer of shared memory on the device. This optimization allows users to
declare shared memory on the device without using OpenMP's allocators. declare shared memory on the device without using OpenMP's allocators.
Globalization normally occurs when a pointer to a thread-local variables escapes Globalization normally occurs when a pointer to a thread-local variables escapes
▲ Show 20 Lines • Show All 53 Lines • Show Last 20 Lines

openmp/docs/remarks/OMP112.rst

.. _omp112:
Found thread data sharing on the GPU. Expect degraded performance due to data globalization. [OMP112] Found thread data sharing on the GPU. Expect degraded performance due to data globalization. [OMP112]
===================================================================================================== =====================================================================================================
.. _omp112:
This missed remark indicates that a globalized value was found on the target This missed remark indicates that a globalized value was found on the target
device that was not either replaced with stack memory by :ref:`OMP110 <omp110>` device that was not either replaced with stack memory by :ref:`OMP110 <omp110>`
or shared memory by :ref:`OMP111 <omp111>`. Globalization that has not been or shared memory by :ref:`OMP111 <omp111>`. Globalization that has not been
removed will need to be handled by the runtime and will significantly hurt removed will need to be handled by the runtime and will significantly hurt
performance. performance.
The OpenMP standard expects that threads can always share their data between The OpenMP standard expects that threads can always share their data between
each-other. However, this is not true by default when offloading to a target each-other. However, this is not true by default when offloading to a target
▲ Show 20 Lines • Show All 76 Lines • Show Last 20 Lines

openmp/docs/remarks/OMP113.rst

.. _omp113:
Could not move globalized variable to the stack. Variable is potentially captured in call. Mark parameter as `__attribute__((noescape))` to override. [OMP113] Could not move globalized variable to the stack. Variable is potentially captured in call. Mark parameter as `__attribute__((noescape))` to override. [OMP113]
============================================================================================================================================================== ==============================================================================================================================================================
.. _omp113:
This missed remark indicates that a globalized value could not be moved to the 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 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 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 its pointer cannot be stored. Otherwise it is considered captured and could
potentially be shared between the threads. This can be overridden using a potentially be shared between the threads. This can be overridden using a
parameter level attribute as suggested in the remark text. parameter level attribute as suggested in the remark text.
Globalization will occur when a pointer to a thread-local variable escapes Globalization will occur when a pointer to a thread-local variable escapes
▲ Show 20 Lines • Show All 68 Lines • Show Last 20 Lines

openmp/docs/remarks/OMP120.rst

.. _omp120:
Transformed generic-mode kernel to SPMD-mode [OMP120] Transformed generic-mode kernel to SPMD-mode [OMP120]
===================================================== =====================================================
.. _omp120:
This optimization remark indicates that the execution strategy for the OpenMP This optimization remark indicates that the execution strategy for the OpenMP
target offloading kernel was changed. Generic-mode kernels execute by using a target offloading kernel was changed. Generic-mode kernels execute by using a
single thread to schedule parallel worker threads using a state machine. This single thread to schedule parallel worker threads using a state machine. This
code transformation can move a kernel that was initially generated in generic 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 mode to SPMD-mode where all threads are active at the same time with no state
machine. This is only possible if the instructions previously executed by a machine. 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 single thread have no side-effects or can be guarded. If the instructions have
no side-effects they are simply recomputed by each thread. no side-effects they are simply recomputed by each thread.
▲ Show 20 Lines • Show All 80 Lines • Show Last 20 Lines

openmp/docs/remarks/OMP121.rst

.. _omp121:
Value has potential side effects preventing SPMD-mode execution. Add `__attribute__((assume(\"ompx_spmd_amenable\")))` to the called function to override. [OMP121] Value has potential side effects preventing SPMD-mode execution. Add `__attribute__((assume(\"ompx_spmd_amenable\")))` to the called function to override. [OMP121]
=================================================================================================================================================================== ===================================================================================================================================================================
.. _omp121:
This analysis remarks indicates that a potential side-effect that cannot be This analysis remarks indicates that a potential side-effect that cannot be
guarded prevents the target region from executing in SPMD-mode. SPMD-mode guarded prevents the target region from executing in SPMD-mode. SPMD-mode
requires that each thread is active inside the region. Any instruction that requires that each thread is active inside the region. Any instruction that
cannot be either recomputed by each thread independently or guarded and executed cannot be either recomputed by each thread independently or guarded and executed
by a single thread prevents the region from executing in SPMD-mode. by a single thread prevents the region from executing in SPMD-mode.
This remark will attempt to print out the instructions preventing the region This remark will attempt to print out the instructions preventing the region
from being executed in SPMD-mode. Calls to functions outside the current from being executed in SPMD-mode. Calls to functions outside the current
▲ Show 20 Lines • Show All 67 Lines • Show Last 20 Lines

openmp/docs/remarks/OMP130.rst

.. _omp130:
Removing unused state machine from generic-mode kernel. [OMP130] Removing unused state machine from generic-mode kernel. [OMP130]
================================================================ ================================================================
.. _omp130:
This optimization remark indicates that an unused state machine was removed from 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 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 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 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. is unnecessary because there is only a single thread active at any time.
Examples Examples
-------- --------
Show All 23 Lines

openmp/docs/remarks/OMP140.rst

.. _omp140:
Could not internalize function. Some optimizations may not be possible. [OMP140] Could not internalize function. Some optimizations may not be possible. [OMP140]
==================================================================================================================== ====================================================================================================================
.. _omp140:
This analysis remark indicates that function internalization failed for the This analysis remark indicates that function internalization failed for the
given function. Internalization occurs when a call to a function that ordinarily 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 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 only internal visibility. This allows the compiler to make strong static
assertions about the context a function is called in. Without internalization assertions about the context a function is called in. Without internalization
this analysis would always be invalidated by the possibility of someone calling this analysis would always be invalidated by the possibility of someone calling
the function in a different context outside of the current translation unit. the function in a different context outside of the current translation unit.
This is necessary for optimizations like :ref:`OMP111 <omp111>` and :ref:`OMP120 This is necessary for optimizations like :ref:`OMP111 <omp111>` and :ref:`OMP120
Show All 36 Lines

openmp/docs/remarks/OMP150.rst

.. _omp150:
Parallel region merged with parallel region at <location>. [OMP150] Parallel region merged with parallel region at <location>. [OMP150]
=================================================================== ===================================================================
.. _omp150:
This optimization remark indicates that a parallel region was merged with others This optimization remark indicates that a parallel region was merged with others
into a single parallel region. Parallel region merging fuses consecutive into a single parallel region. Parallel region merging fuses consecutive
parallel regions to reduce the team activation overhead of forking and increases parallel regions to reduce the team activation overhead of forking and increases
the scope of possible OpenMP-specific optimizations within merged parallel the scope of possible OpenMP-specific optimizations within merged parallel
regions. This optimization can also guard sequential code between two parallel regions. This optimization can also guard sequential code between two parallel
regions if applicable. regions if applicable.
Example Example
Show All 29 Lines

openmp/docs/remarks/OMP160.rst

.. _omp160:
Removing parallel region with no side-effects. [OMP160] Removing parallel region with no side-effects. [OMP160]
======================================================= =======================================================
.. _omp160:
This optimization remark indicates that a parallel region was deleted because it 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 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 write any of its results to memory visible outside the region. This optimization
is necessary because the barrier between sequential and parallel code typically is necessary because the barrier between sequential and parallel code typically
prevents dead code elimination from completely removing the region. Otherwise prevents dead code elimination from completely removing the region. Otherwise
there will still be overhead to fork and merge the threads. there will still be overhead to fork and merge the threads.
Example Example
Show All 31 Lines

openmp/docs/remarks/OMP170.rst

.. _omp170:
OpenMP runtime call <call> deduplicated. [OMP170] OpenMP runtime call <call> deduplicated. [OMP170]
==================================================================== ====================================================================
.. _omp170:
This optimization remark indicates that a call to an OpenMP runtime call was 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 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 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 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 cannot be done automatically by the compiler because the implementations of the
OpenMP runtime calls live in a separate library the compiler cannot see. OpenMP runtime calls live in a separate library the compiler cannot see.
Example Example
Show All 28 Lines