This patch makes `num_teams` and `thread_limit` mandatory for bare
kernels,
similar to a reguar kernel language that when launching a kernel, the
grid size
has to be set explicitly.
Summary:
Libomptarget supports JIT by treating an LLVM-IR file as a regular input
image. The handling here used a global map to keep track of triples once
it was parsed. This was done to same time, however this created a global
constructor as well as an extra mutex to handle it. This patch removes
the use of this map.
Instead, we simply use the file magic to perform a quick check if the
input image is valid bitcode. If not, we then create a lazy module. This
should roughly equivalent to the old handling that create an IR symbol
table. Here we can prevent the module from materializing everything but
the single triple metadata we read in later.
Summary:
Currently, we don't emit any of the AMDGPU specific flags and options
when printing in GNU-mode (the default). This patch adds extra code to
handle outputting these values to match the output from GNU's `readelf`
when used on an AMDGPU image.
This reverts commit a01307a6ee788fc6ac2e09e58f0f52e5666def86 and its
follow-up fix 32d5221ec4810dd723ccebaabbda1df5d3b4cfcf.
It caused unexpected warnings emitted for nested designators in C.
This patch removes assumptions that std::array's iterators are raw
pointers in the source code and in our test suite. While this is true
right now, this doesn't have to be true and ion the future we might want
to enable bounded iterators in std::array, which would require this
change.
This is a pre-requisite for landing #74482
Fix mapping of structs to device.
The following example fails:
```
#include <stdio.h>
#include <stdlib.h>
struct Descriptor {
int *datum;
long int x;
int xi;
long int arr[1][30];
};
int main() {
Descriptor dat = Descriptor();
dat.datum = (int *)malloc(sizeof(int)*10);
dat.xi = 3;
dat.arr[0][0] = 1;
#pragma omp target enter data map(to: dat.datum[:10]) map(to: dat)
#pragma omp target
{
dat.xi = 4;
dat.datum[dat.arr[0][0]] = dat.xi;
}
#pragma omp target exit data map(from: dat)
return 0;
}
```
This is a rework of the previous attempt:
https://github.com/llvm/llvm-project/pull/72410
Some tools with a specified target arch, but no full triple default to
the host triple. On macos hosts, this would then force using macho on
targets that didn't expect it, resulting in assertions.
We should also probably emit explicit errors if the object format is
specified on targets which don't handle it.
Otherwise we may replace undef with poison.
Note that a lot of tests regressing here already have variants
that use poison instead of undef (often in a separate
inseltpoison file), which is why I'm not adjusting them to the
new pattern.
This means that getRegInterval no longer depends on the MCInstrDesc, so
it could be simplified further to take just a MachineOperand or just a
physical register. NFCI.
The core implementation of the dataflow anlysis framework is
interpocedural by design. While this offers better analysis precision,
it also comes with additional cost as it takes longer for the analysis
to reach the fixpoint state. Add a configuration mechanism to the
dataflow solver to control whether it operates inteprocedurally or not
to offer clients a choice.
As a positive side effect, this change also adds hooks for explicitly
processing external/opaque function calls in the dataflow analyses,
e.g., based off of attributes present in the the function declaration or
call operation such as alias scopes and modref available in the LLVM
dialect.
This change should not affect existing analyses and the default solver
configuration remains interprocedural.
Co-authored-by: Jacob Peng <jacobmpeng@gmail.com>
The CheckInteger routine called from TableGen-generated selection logic
uses getSExtValue - which will abort if the underlying APInt does not
fit into an int64_t.
This case is now triggered by the SystemZ back-end since i128 is a legal
type on certain machines. While we do not have any regular instructions
that take 128-bit immediates (like most other platforms), there are
patterns in the .td files that recognize an i128 "xor ..., -1" as a
"not".
These patterns cause code to be generated that calls the CheckInteger
routine on some i128-valued integer, which may trigger the assert.
Fix by using trySExtValue instead.
Fixes https://github.com/llvm/llvm-project/issues/75710
Previously "yamlize" overload for validatedMappingTraits was unconditionally calling "MappingTraits<T>::validate" even if "MappingContextTraits<T, Context>" was passed to it.
Therefore compilation failed when specifying "MappingContextTraits<T,Context>::validate()"
The specialisation will not be valid when ConstantInt gains native
support for vector types.
This is largely a mechanical change but with extra attention paid to constant
folding, InstCombineVectorOps.cpp, LoopFlatten.cpp and Verifier.cpp to
remove the need to call `getIntegerType()`.
Co-authored-by: Nikita Popov <github@npopov.com>
LLVM intrinsics `get_fpmode`, `set_fpmode` and `reset_fpmode` operate
control modes, the bits of FP environment that affect FP operations. On
ARM these bits are in FPSCR together with the status bits. The
implementation of these intrinsics produces code close to that of
functions `fegetmode` and `fesetmode` from GLIBC.
Pull request: https://github.com/llvm/llvm-project/pull/74054
In line with updated shufflevector semantics, this represents the
poison elements rather than undef elements now. This commit is a
pure rename, without any logic changes.
This was caught by coverity, reported as: `dead_error_condition`.
Since the conditional revolves around `CF`, it is guaranteed to be null
in the else clause, hence making the second part of the statement
redundant.
Add intrinsics of the form:
svboolx2_t svwhile<cond>_b{8,16,32,64}_[{s,u}64]_x2([u]int64_t, [u]int64_t);
and their overloaded variants as specified in
https://github.com/ARM-software/acle/pull/257
Check if workshare loop without loop body is lowered correctly i.e.:
1) null pointer is passed to OpenMP device RTL function as a
parameter which denotes loop function body aggregated parameters
2) Outlined loop function body has only one parameter - loop counter
Due to the way the f string was written, the text ended up with 4 spaces
at the start. 4 space indent in Markdown means plain text, which is not
what we intend here.
Binutils 2.36 had a somewhat controversial change in how the
--preprocessor option was handled in GNU windres; previously, the option
was interpreted as a part of the command string, potentially containing
multiple arguments (which even was hinted at in the documentation).
In Binutils 2.36, this was changed to interpret the --preprocessor
argument as one argument (possibly containing spaces) pointing at the
preprocessor executable.
The existing behaviour where implicit arguments like -E -xc -DRC_INVOKED
are dropped if --preprocessor is specified, was kept.
This was a breaking change for some users of GNU windres, see
https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=21c33bcbe36377abf01614fb1b9be439a3b6de20,
https://sourceware.org/bugzilla/show_bug.cgi?id=27594, and
https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=5edb8e3f5ad8d74a83fc0df7f6e4514eed0aa77f.
As multiple years have passed since, the behaviour change seems to be
here to stay, and any users of the previous form of the option have been
forced to avoid this construct. Thus update llvm-windres to match the
new way Binutils of handling this option.
One construct for specifying the path to the preprocessor, which works
both before and after binutils 2.36 (and this change in llvm-windres) is
to specify options like this:
--preprocessor path/to/executable --preprocessor-arg -E
--preprocessor-arg -xc -DRC_INVOKED
With commit https://reviews.llvm.org/D152366 I introduced functionality
that permitted the hoisting of runtime memory checks from a vectorised
inner loop to the preheader of the next outer-most loop. This is useful
for benchmarks like SPEC2017's x264 where the inner loop is vectorised
and only has a small trip count. In such cases the runtime memory checks
become expensive and since the checks never fail in the case of x264 it
makes sense to do this. However, this behaviour was controlled by the
flag -hoist-runtime-checks which was off by default.
This patch enables this flag by default for all targets, since I believe
this is a generally beneficial thing to do. I have tested this with
SPEC2017 and I see 2.3% and 2.6% improvements with x264 on neoverse-v1
and neoverse-n1, respectively. Similarly, I saw slight improvements in
the overall geomean on both machines. The only other notable changes
were a 1% drop in the roms benchmark, which was compensated for by a 1%
improvement in fotonik3d.
This PR adds a warning that's emitted when a non-streaming or
non-streaming-compatible builtin is called in an unsuitable function.
Uses work by Kerry McLaughlin.
This is a re-upload of #74064 and fixes a compile time increase.
So far, if there was a chain of record type prvalues,
`getResultObjectLocation()` would assign a different result object
location to
each one. This makes no sense, of course, as all of these prvalues end
up
initializing the same result object.
This patch fixes this by propagating storage locations up through the
entire
chain of prvalues.
The new implementation also has the desirable effect of making it
possible to
make `getResultObjectLocation()` const, which seems appropriate given
that,
logically, it is just an accessor.
This takes the code from D144103 and extends it to maxloc, to allow the
simplifyMinMaxlocReduction method to work with both min and max
intrinsics by switching condition and limit/initial value.
