The builtins we were using to implement __clc_is(finite|inf|nan|normal)
-- __builtin_isfinite, etc. -- don't take vector types so we were
previously scalarizing. The __builtin_isfpclass builtin does take vector
types and thus allows us to keep things in vectors.
There is no change in codegen to the scalar versions of any of these
builtins.
This commit moves the implementation of the copysign builtin to the CLC
library.
It simultaneously optimizes it for vector types by avoiding
scalarization. It does so by using the __builtin_elementwise_copysign
clang builtins, which can handle vector types.
It also fixes a bug in the half/fp16 implementation of the builtin. This
version was using an incorrect mask (0x7FFFF instead of 0x7FFF) and was
thus preserving the original sign bit, rather than masking it out.
By using the vector reduction buitins we can avoid scalarization.
Targets that don't support vector reductions will scalarize later on
anyway. The vector reduction builtins should be well-enough supported by
the middle-end to be a generic solution.
This produces conceptually equivalent code: all vector elements are
OR'd/AND'd together and the final scalar is bit-shifted and masked to
produce the final result.
The 'normalize' builtin uses 'all' so its code has similarly improved in
places.
These are similar to 347fb208, but these builtins are expressed in terms
of other builtins. The LLVM IR generated features the same fcmp ord/uno
comparisons as before, but consistently in vector form.
Clang knows how to perform relational operations on OpenCL vectors, so
we don't need to use the Clang builtins. The builtins we were using
didn't support vector types, so we were previously scalarizing.
This commit generates the same LLVM fcmp operations as before, just
without the scalarization.
There were two implementations of this - one that implemented nextafter
in software, and another that called a clang builtin. No in-tree targets
called the builtin, so all targets build the software version. The
builtin version has been removed, and the software version has been
renamed to be the "default".
This commit also optimizes nextafter, to avoid scalarization as much as
possible. Note however that the (CLC) relational builtins still
scalarize; those will be optimized in a separate commit.
Since nextafter is used by some convert_type builtins, the diff to IR
codegen is not limited to the builtin itself.
Having the fp16 pragmas enabled in the header file is risky. The macros
defined by that header don't (and can't) include the pragmas that make
fp16 types themselves legal, and another header may disable the fp16
pragma before the macro's use.
The safest thing to do is the use of pragmas surrounding each use of the
macro in the implementation files. This pattern is also far more common
across the codebase.
The half variants were missing. The integer bitselect builtins weren't
going through __clc_bitselect due to an oversight when the CLC version
was introduced.
This was missed during the introduction of select. This also unifies the
various .inc files used for each, as they were essentially identical.
The __clc_select function is now also built for SPIR-V targets.
All targets build `__clc_mad` -- even SPIR-V targets -- since it
compiles to the optimal `llvm.fmuladd` intrinsic. There is no change to
the bytecode generated for non-SPIR-V targets.
The `mix` builtin, which is implemented as a wrapper around `mad`, is
left as an OpenCL-layer wrapper of `__clc_mad`. I don't know if it's
worth having a specific CLC version of `mix`.
The changes to the other CLC files/functions are moving uses of `mad` to
`__clc_mad`, and reformatting. There is an additional instance of
`trunc` becoming `__clc_trunc`, which was missed before.
Missing half variants were also added.
The builtins are now consistently emitted in vector form (i.e., with a
splat of the literal to the appropriate vector size).
This commit moves the implementation of the smoothstep function to the
CLC library, whilst optimizing the codegen.
This commit also adds support for 'half' versions of smoothstep, which
were previously missing.
The CLC smoothstep implementation now keeps everything in vectors,
rather than recursively splitting vectors by half down to the scalar
base form. This should result in more optimal codegen across the board.
This commit also removes some non-standard overloads of smoothstep with
mixed types, such as 'double smoothstep(float, float, float)'. There
aren't any mixed-(element )type versions of smoothstep as far as I can
see:
gentype smoothstep(gentype edge0, gentype edge1, gentype x)
gentypef smoothstep(float edge0, float edge1, gentypef x)
gentyped smoothstep(double edge0, double edge1, gentyped x)
gentypeh smoothstep(half edge0, half edge1, gentypeh x)
The CLC library only defines the first type, for simplicity; the OpenCL
layer is responsible for handling the scalar/scalar/vector forms. Note
that the scalar/scalar/vector forms now splat the scalars to the vector
type, rather than recursively split vectors as before. The macro that
used to 'vectorize' smoothstep in this way has been moved out of the
shared clcmacro.h header as it was only used for the smoothstep builtin.
Note that the CLC clamp function is now built for both SPIR-V targets.
This is to help build the CLC smoothstep function for the Mesa SPIR-V
target.
There's no automatic way of checking these headers are self-contained.
Instead of including these common files many times across the whole
codebase, we can include them in the generic `gentype.inc` and
`floatn.inc` files which are included by most CLC headers.
These functions all map to the corresponding LLVM intrinsics, but the
vector intrinsics weren't being generated. The intrinsic mapping from
CLC vector function to vector intrinsic was working correctly, but the
mapping from OpenCL builtin to CLC function was suboptimally recursively
splitting vectors in halves.
For example, with this change, `ceil(float16)` calls `llvm.ceil.v16f32`
directly once optimizations are applied.
Now also, instead of generating LLVM intrinsics through `__asm` we now
call clang elementwise builtins for each CLC builtin. This should be a
more standard way of achieving the same result
The CLC versions of each of these builtins are also now built and
enabled for SPIR-V targets. The LLVM -> SPIR-V translator maps the
intrinsics to the appropriate OpExtInst, so there should be no
difference in semantics, despite the newly introduced indirection from
OpenCL builtin through the CLC builtin to the intrinsic.
The AMDGPU targets make use of the same `_CLC_DEFINE_UNARY_BUILTIN`
macro to override `sqrt`, so those functions also appear more optimal
with this change, calling the vector `llvm.sqrt.vXf32` intrinsics
directly.
The OpenCL relational functions now call their CLC counterparts, and the
CLC relational functions are defined identically to how the OpenCL
functions were defined.
As usual, clspv and spir-v targets bypass these.
No observable changes to any libclc target (measured with llvm-diff).
These functions are all mapped to LLVM intrinsics.
The clspv and spirv targets don't declare or define any of these CLC
functions, and instead map these to their corresponding OpenCL symbols.
These functions are "shared" between integer and floating-point types,
hence the directory name. They are used in several CLC internal
functions such as __clc_ldexp.
Note that clspv and spirv targets don't want to define these functions,
so pre-processor macros replace calls to __clc_min with regular min, for
example. This means they can use as much of the generic CLC source files
as possible, but where CLC functions would usually call out to an
external __clc_min symbol, they call out to an external min symbol. Then
they opt out of defining __clc_min itself in their CLC builtins library.
Preprocessor definitions for these targets have also been changed
somewhat: what used to be CLC_SPIRV (the 32-bit target) is now
CLC_SPIRV32, and CLC_SPIRV now represents either CLC_SPIRV32 or
CLC_SPIRV64. Same goes for CLC_CLSPV.
There are no differences (measured with llvm-diff) in any of the final
builtins libraries for nvptx, amdgpu, or clspv. Neither are there
differences in the SPIR-V targets' LLVM IR before it's actually lowered
to SPIR-V.
Some libclc builtins currently use internal builtins prefixed with
'__clc_' for various reasons, e.g., to avoid naming clashes.
This commit formalizes this concept by starting to isolate the
definitions of these internal clc builtins into a separate
self-contained bytecode library, which is linked into each target's
libclc OpenCL builtins before optimization takes place.
The goal of this step is to allow additional libraries of builtins
that provide entry points (or bindings) that are not written in OpenCL C
but still wish to expose OpenCL-compatible builtins. By moving the
implementations into a separate self-contained library, entry points can
share as much code as possible without going through OpenCL C.
The overall structure of the internal clc library is similar to the
current OpenCL structure, with SOURCES files and targets being able to
override the definitions of builtins as needed. The idea is that the
OpenCL builtins will begin to need fewer target-specific overrides, as
those will slowly move over to the clc builtins instead.
Another advantage of having a separate bytecode library with the CLC
implementations is that we can internalize the symbols when linking it
(separately), whereas currently the CLC symbols make it into the final
builtins library (and perhaps even the final compiled binary).
This patch starts of with 'dot' as it's relatively self-contained, as
opposed to most of the maths builtins which tend to pull in other
builtins.
We can also start to clang-format the builtins as we go, which should
help to modernize the codebase.
