DPC++ daily 2022-08-04

[SYCL][FPGA] Remove support for intel::fpga_pipeline attribute (#6519)

Support for new FPGA attribute called [[intel::fpga_pipeline(N)]]
on https://github.com/intel/llvm/pull/6254 is no longer needed.

This patch removes the support from frontend.

Signed-off-by: Soumi Manna <soumi.manna@intel.com>

DPC++ daily 2022-08-02

[SYCL] Address tests failures after #6469 #6506 (#6509)

Just a workaround in sycl/test/regression/fsycl-host-compiler-win.cpp
while the proper fix is being investigated/prepared.

DPC++ daily 2022-08-01

LLVM and SPIRV-LLVM-Translator pulldown (WW32)

LLVM: llvm/llvm-project@c9737b6f1818056de3a69e43150d501ac0ee2851
SPIRV-LLVM-Translator: KhronosGroup/SPIRV-LLVM-Translator@1b8a00741caafac50de84f1f860b78e702722585

DPC++ daily 2022-07-30

[BuildBot] Uplift GPU RT version for Linux CI Process (#6472)

Uplift GPU RT version for Linux to 22.29.23750

Signed-off-by: bb-sycl <bb-sycl@intel.com>

DPC++ daily 2022-07-29

[SYCL][GDB] Fix op[] when called with typedef argument (#6459)

The implementation of 'index' did not strip typedefs from the argument
passed to it, when passed as a single number.  Instead, TYPE_CODE_INT was
expected.

This leads to failures when calling 'accessor[arg]' using an arg that is
either a typedef or something like size_t, which are considered
TYPE_CODE_TYPEDEF inside GDB.  The check fails and the function
continues on and fails even though the argument could have been used in
the int cast.

Typedef stripping was added to the if condition to fix this.

Signed-off-by: Nils-Christian Kempke <nils-christian.kempke@intel.com>

DPC++ daily 2022-07-28

[NFCI][SYCL] Refactor reduction implementation (#6445)

The core part of it is adding the user's reduction variable type as an
explicit template parameter and modifying reduction to store just that,
instead of having three members (RWAcc/DWAcc/UsmPtr) of which only one
is used to store it.

That, in turn, made the whole default_reduction_algorithm redundant as
all it was used for was to propagate that exact information in an
obscure way.

Once done, it enabled some further simplifications:

Simplifying some methods to use compile-time "if constexpr" instead
of checking the type of the user reduction variable in runtime
(i.e., previous "if (MRWAcc)" check).

Making reduction_impl_algo::reduction_impl_algo ctor generic and
having a single instance of it, eliminating three version for
different types of user's reduction variable.

Unify reduction_impl::reduction_impl ctors accessing RW/DW accessor
eliminating code duplication

create make_reduction helper function to make creation of
reduction_impl require less boilerplate code by using template type
parameters deduction (couldn't use CTAD as some of the params can't
be deduced and there is no such thing as partial CTAD).

Simplify reduction_impl creation further by benefiting from
accessor's CTAD.

DPC++ daily 2022-07-26

LLVM and SPIRV-LLVM-Translator pulldown (WW31)

LLVM: llvm/llvm-project@46e6dd84b778dd2e30368183fec265beceecdd0f
SPIRV-LLVM-Translator: KhronosGroup/SPIRV-LLVM-Translator@c8cd2f6895091f84823d94a661753c0a9765a7a7

DPC++ daily 2022-07-25

[SYCL] Handle nullptr arguments in level_zero (#6464)

* [SYCL] Handle nullptr arguments in level_zero

If specialization constant is not used on device, then the buffer for
it is not created on device side. In this case the host part passes nullptr
for the corresponding kernel operand holding the reference to the buffer
with constants.
This patch fixes the level_zero arguments handling to avoid SegFault
in the plugin.

Signed-off-by: Vyacheslav N Klochkov <vyacheslav.n.klochkov@intel.com>

DPC++ daily 2022-07-23

[SYCL] Fix reductions regressions after #6343 (#6460)

Three pieces:
  * Access to private static handler::withAuxHandler
  * Pow2WG -> HasUniformWG
  * Out accessor must be the last created (after other temps)

DPC++ daily 2022-07-22

[SYCL] Add DPCPP_ENABLE_OPAQUE_POINTERS option (#6378)

This option controls the default behavior of LLVM projects regarding
opaque pointers. The default value is OFF as most of the SYCL compiler
components are not ready to handle opaque pointers.

This change adds a new nightly job testing SYCL compiler with enabled
opaque pointers by default.