This path defines the newly added `__kmpc_disitrute_static_init`
functions in the device runtime library. These functions are currently
exact copies of the current worksharing method but can be tuned later.
Depends on D110429
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D110430
This is a follow-up of D110029, which uses bitset to indicate execution mode. This patches makes the changes in the function call.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D110279
Parallel regions are outlined as functions with capture variables explicitly generated as distinct parameters in the function's argument list. That complicates the fork_call interface in the OpenMP runtime: (1) the fork_call is variadic since there is a variable number of arguments to forward to the outlined function, (2) wrapping/unwrapping arguments happens in the OpenMP runtime, which is sub-optimal, has been a source of ABI bugs, and has a hardcoded limit (16) in the number of arguments, (3) forwarded arguments must cast to pointer types, which complicates debugging. This patch avoids those issues by aggregating captured arguments in a struct to pass to the fork_call.
Reviewed By: jdoerfert, jhuber6
Differential Revision: https://reviews.llvm.org/D102107
This patch adds fields for the device number and number of devices into
the device environment struct and debugging values.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D110004
In some build configurations, the target we depend on is not available for declaring the build dependency.
We only need to declare the build dependency, if the build target is available in the same build.
Fixes the issue raised in https://reviews.llvm.org/D107156#2969862
This patch should go into release/13 together with D108404
Differential Revision: https://reviews.llvm.org/D108868
Lets wavefront size be 32 for amdgpu openmp, as well as 64.
Fixes up as little as possible to pass that through the libraries. This change
is end to end, as opposed to updating clang/devicertl/plugin separately. It can
be broken up for review/commit if preferred. Posting as-is so that others with
a gfx10 can try it out. It works roughly as well as gfx9 for me, but there are
probably bugs remaining as well as the todo: for letting grid values vary more.
Reviewed By: ronlieb
Differential Revision: https://reviews.llvm.org/D108708
Lets the amdgpu plugin write to omptarget_device_environment
to enable debugging. Intend to use in the near future to record the
wavesize that a given deviceRTL was compiled with for running on hardware
that supports 32 or 64.
Patch sets all the attributes that are useful. Notably .data means the variable
is set by writing to host memory before copying to the GPU instead of launching
a kernel to update the image. Can simplify the plugin slightly to drop the
code for patching after load if this is used consistently.
NFC on nvptx, cuda plugin seems to work fine without any annotations.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D108698
The use of `$<TARGET_FILE:clang>` was adapted too broadly from D101265.
Fixes llvm.org/PR51579
Also see discussion in D108534.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D108640
Compiled nvptx devicertl as freestanding, breaking the
dependency on host glibc and gcc-multilibs. Thus build it by default.
Comes at the cost of #defining out printf. Tried mapping it onto
__builtin_printf but that gets transformed back to printf instead
of hitting the cuda/openmp lowering transform.
Printf could be preserved by one of:
- dropping all the standard headers and ffreestanding
- providing a header only printf implementation
- changing the compiler handling of printf
Reviewed By: grokos
Differential Revision: https://reviews.llvm.org/D108349
Add include path to the cmakefiles and set the target_impl enums
from the llvm constants instead of copying the values.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D108391
D107156 and D107320 are not sufficient when OpenMP is built as llvm runtime
(LLVM_ENABLE_RUNTIMES=openmp) because dependencies only work within the same
cmake instance.
We could limit the dependency to cases where libomptarget/plugins are really
built. But compared to the whole llvm project, building openmp runtime is
negligible and postponing the build of OpenMP runtime after the dependencies
are ready seems reasonable.
The direct dependency introduced in D107156 and D107320 is necessary for the
case where OpenMP is built as llvm project (LLVM_ENABLE_PROJECTS=openmp).
Differential Revision: https://reviews.llvm.org/D108404
Use uint64_t for lanemask on all GPU architectures at the interface
with clang. Updates tests. The deviceRTL is always linked as IR so the zext
and trunc introduced for wave32 architectures will fold after inlining.
Simplification partly motivated by amdgpu gfx10 which will be wave32 and
is awkward to express in the current arch-dependant typedef interface.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D108317
The device runtime contains several calls to `__kmpc_get_hardware_num_threads_in_block`
and `__kmpc_get_hardware_num_blocks`. If the thread_limit and the num_teams are constant,
these calls can be folded to the constant value.
In this patch we use the already introduced `AAFoldRuntimeCall` and the `NumTeams` and
`NumThreads` kernel attributes (to be introduced in a different patch) to fold these functions.
The code checks all the kernels, and if their attributes match, the functions are folded.
In the future we will explore specializing for multiple values of NumThreads and NumTeams.
Depends on D106390
Reviewed By: jdoerfert, JonChesterfield
Differential Revision: https://reviews.llvm.org/D106033
Similar to D105787, this patch tries to fold `__kmpc_parallel_level` if possible.
Note that `__kmpc_parallel_level` doesn't take activeness into consideration,
based on current `deviceRTLs`, its return value can be such as 0, 1, 2, instead
of 0, 129, 130, etc. that also indicate activeness.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106154
We build `deviceRTLs` with `-O1` by default, which also triggers OpenMPOpt. When
the info cache is created, some attributes are removed. As a result, although we
mark a few functions `noinline`, they are still inlined when the bitcode library
is generated. This can cause an issue in middle end optimization.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106710
Unrolling this loop provides better performance in practice because it is
executed on the device and is likely to be very small.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D106692
With D106496 we can make the globalization fallback stack much simpler
and this version doesn't seem to experience the spurious failures and
deadlocks we have seen before.
Differential Revision: https://reviews.llvm.org/D106576
These functions should follow the camel case convention. These are really easy to change
and are needed for D106033.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D106390
Function internalization can sometimes occur in situations where we want to
keep the call sites intact. This patch adds an option to disable function
internalization and prevents the device runtime from being internalized while
creating the bitcode library.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106438
This patch introduces `__kmpc_is_generic_main_thread_id` which splits the old
comparison into its own runtime function. The purpose of this is so we can fold
this part independently, so when both this and `is_spmd_mode` are folded the
final function will be folded as well.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106437
This patch changes `__kmpc_free_shared` to take an additional argument
corresponding to the associated allocation's size. This makes it easier to
implement the allocator in the runtime.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106496
The patch exposes the libomptarget runtime function that gets the hardware thread id through the kmpc API. This is to be used in SPMDization for checking the thread id to execute regions by a single thread in a block.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106323
In `deviceRTLs`, the parallel level is stored in a shared variable of type `uint8_t`.
`__kmpc_parallel_level` currently returns a 16-bit interger. This patch first
changes the return type of the function to `uint8_t`, same as the shared variable,
and then corrects function type which was updated in D105955.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106384
Currently the NPVTX work function is marked volatile. This prevents some
optimizations from using this value.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106310
Currently when we compile the project in debug mode, `-g` will not be added to
compilation flag. The bc files generated in different mode are of different size.
When using GPU debuggers like `cuda-gdb`, it is expected to provide more info
with a debug version of bc lib.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D106229
Parallel regions are outlined as functions with capture variables explicitly generated as distinct parameters in the function's argument list. That complicates the fork_call interface in the OpenMP runtime: (1) the fork_call is variadic since there is a variable number of arguments to forward to the outlined function, (2) wrapping/unwrapping arguments happens in the OpenMP runtime, which is sub-optimal, has been a source of ABI bugs, and has a hardcoded limit (16) in the number of arguments, (3) forwarded arguments must cast to pointer types, which complicates debugging. This patch avoids those issues by aggregating captured arguments in a struct to pass to the fork_call.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D102107
There are two places in current deviceRTLs where it computes parallel level explicitly,
which is basically the functionality of `__kmpc_parallel_level`. Starting from
D105787, we plan to introduce a series of function call folding based on information
that can be deducted during compilation time. Computation of parallel level is
the next target. This patch makes steps for the optimization.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D105955
Simplifies control flow to allow store/load forwarding
This change folds two basic blocks into one, leaving a single store to parallelLevel.
This is a step towards spmd kernels with sufficiently aggressive inlining folding
the loads from parallelLevel and thus discarding the nested parallel handling
when it is unused.
Transform:
```
int threadId = GetThreadIdInBlock();
if (threadId == 0) {
parallelLevel[0] = expr;
} else if (GetLaneId() == 0) {
parallelLevel[GetWarpId()] = expr;
}
// =>
if (GetLaneId() == 0) {
parallelLevel[GetWarpId()] = expr;
}
// because
unsigned GetLaneId() { return GetThreadIdInBlock() & (WARPSIZE - 1);}
// so whenever threadId == 0, GetLaneId() is also 0.
```
That replaces a store in two distinct basic blocks with as single store.
A more aggressive follow up is possible if the threads in the warp/wave
race to write the same value to the same address. This is not done as
part of this change.
```
if (GetLaneId() == 0) {
parallelLevel[GetWarpId()] = expr;
}
// =>
parallelLevel[GetWarpId()] = expr;
// because
unsigned GetWarpId() { return GetThreadIdInBlock() / WARPSIZE; }
// so GetWarpId will index the same element for every thread in the warp
// and, because expr is lane-invariant in this case, every lane stores the
// same value to this unique address
```
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D105699
In order to fold calls based on high-level knowledge and control flow
tracking it helps to expose the information as a runtime call. The
logic: `!SPMD && getTID() == getMasterTID()` was used in various places
and is now encapsulated in `__kmpc_is_generic_main_thread`. As part of
this rewrite we replaced eager computation of arguments with on-demand
computation, especially helpful if the calls can be folded and arguments
don't need to be computed consequently.
Differential Revision: https://reviews.llvm.org/D105768
In order to avoid malloc/free, up to NUM_SHARED_VARIABLES_IN_SHARED_MEM
(=64) variables are communicated in dedicated shared memory instead. The
simplification does avoid the need for an "init" and requires "deinit"
only if we ever communicate more than NUM_SHARED_VARIABLES_IN_SHARED_MEM
variables.
Differential Revision: https://reviews.llvm.org/D105767
We had multiple functions to determine the execution mode (SPMD/Generic)
and runtime status (initialized/uninitialized) but that just increased
complexity without a real benefit. Especially with D102307 in mind it
is helpful to reduce the dependence on the `ident_t` flags.
Differential Revision: https://reviews.llvm.org/D105586
In the spirit of TRegions [0], this patch provides a simpler and uniform
interface for a kernel to set up the device runtime. The OMPIRBuilder is
used for reuse in Flang. A custom state machine will be generated in the
follow up patch.
The "surplus" threads of the "master warp" will not exit early anymore
so we need to use non-aligned barriers. The new runtime will not have an
extra warp but also require these non-aligned barriers.
[0] https://link.springer.com/chapter/10.1007/978-3-030-28596-8_11
This was in parts extracted from D59319.
Reviewed By: ABataev, JonChesterfield
Differential Revision: https://reviews.llvm.org/D101976
Broke check-clang, see https://reviews.llvm.org/D102307#2869065
Ran `git revert -n ebbe149a6f08535ede848a531a601ae6591cfbc5..269416d41908bb670f67af689155d5ab8eea689a`
We had multiple functions to determine the execution mode (SPMD/Generic)
and runtime status (initialized/uninitialized) but that just increased
complexity without a real benefit. Especially with D102307 in mind it
is helpful to reduce the dependence on the `ident_t` flags.
Differential Revision: https://reviews.llvm.org/D105586
In the spirit of TRegions [0], this patch provides a simpler and uniform
interface for a kernel to set up the device runtime. The OMPIRBuilder is
used for reuse in Flang. A custom state machine will be generated in the
follow up patch.
The "surplus" threads of the "master warp" will not exit early anymore
so we need to use non-aligned barriers. The new runtime will not have an
extra warp but also require these non-aligned barriers.
[0] https://link.springer.com/chapter/10.1007/978-3-030-28596-8_11
This was in parts extracted from D59319.
Reviewed By: ABataev, JonChesterfield
Differential Revision: https://reviews.llvm.org/D101976
In our ongoing work, we are using `AbstractAttributor` to deduct execution model
of device functions, and potententially remove unnecessary function calls to
`__kmpc_is_spmd_exec_mode`. In current device runtime, we have mixed use of
`isSPMDMode` and `__kmpc_is_spmd_exec_mode`, but in fact in `__kmpc_is_spmd_exec_mode`
it simply calls `isSPMDMode`. Since all functions starting with `__kmpc` is C
function, which doesn't have things like name mangling. It is more optimization
friendly. In this patch, we simply replaced all calls to `isSPMDMode` with
`__kmpc_is_spmd_exec_mode` to pave the way for the optimization.
Reviewed By: JonChesterfield
Differential Revision: https://reviews.llvm.org/D105211
[libomptarget][amdgpu] Build openmp for two more targets
The 4800U APU is a gfx902 and the MI100 accelerator is a gfx908.
Both numbers are listed in ROCT topology.c
Reviewed By: jhuber6
Differential Revision: https://reviews.llvm.org/D104922
Currently the runtime implementation of `__kmpc_alloc_shared` is extremely slow because it allocated memory for each thread individually. This patch adds a small buffer for the threads to share data and will greatly improve performance for builds where all globalization could not be optimized out. If the shared buffer is full, then memory will not only be allocated per-warp rather than per-thread.
Depends on D97680
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D104666
Summary:
This patch introduces the new globalization runtime to be used by D97680. These
runtime calls will replace the __kmpc_data_sharing_push_stack and
__kmpc_data_sharing_pop_stack functions.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D102532
[libomptarget][amdgpu] Mark alloc, free weak to facilitate local experimentation
There are a lot of different ways we might implement the devicertl local alloc
and free functions. Via host, local buffers (stack or arena), specialising per
kernel etc. It is not yet clear what the right design is. This change makes the
alloc and free functions weak, so one can override them from local tests while
comparing options.
Not strictly necessary, as a comparable patch can be applied locally each time,
but would be convenient for out of tree dev. Plan would be to drop the weak
attribute at the same time as introducing a working allocator to trunk.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D102499
[libomptarget][amdgpu] Fix truncation error for partial wavefront
The partial barrier implementation involves one wavefront resetting and N-1
waiting. This change future proofs against launching with a number of threads
that is not a multiple of the wavefront size.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D102407
Joseph Huber