Jeff Hammond

Results 414 comments of Jeff Hammond

@jan-wassenberg Indeed, I got stuck on the VPU count aspect of e.g. Silver. Sorry about that. Giving access to the documented base and turbo frequencies for non-AVX, AVX, AVX-512 would...

> Unfortunately resize will default construct items in memory sequentially, which will zero storage for types that are trivially constructible. Sorry, but why is this bad? Because initialization faults pages...

This is the trivial if you are impatient w.r.t. the pull request. ```patch diff --git a/src/lcals/FIRST_MIN-OMP.cpp b/src/lcals/FIRST_MIN-OMP.cpp index efe8f944..a03ae084 100644 --- a/src/lcals/FIRST_MIN-OMP.cpp +++ b/src/lcals/FIRST_MIN-OMP.cpp @@ -12,6 +12,10 @@ #include +#if...

This is not similar to other implementations but is modernC++errific implementation that uses lambdas a lot 😄 ``` typedef std::array Reduce_type; Reduce_type result = std::transform_reduce( counting_iterator(ibegin), counting_iterator(iend), Reduce_type{m_vsum_init,m_vmin_init,m_vmax_init}, [=](Reduce_type a,...

@bader I'm not sure about the `cuda` label. This is specific to ARM, not CUDA.

The MCVE is below: ``` jrhammon@xavier-agx:~/ISYCL$ cd /home/jrhammon/ISYCL/build/tools/libdevice && \ /home/jrhammon/ISYCL/build/bin/clang-12 -S -fsycl-device-only -fno-sycl-use-bitcode \ -Wno-sycl-strict -Wno-undefined-internal -sycl-std=2017 \ /home/jrhammon/ISYCL/libdevice/fallback-cmath.cpp \ -o /home/jrhammon/ISYCL/build/lib/libsycl-fallback-cmath.spv error: unknown target ABI 'aapcs' ```

I don't know. Haven't had time to retest.

I thought https://github.com/devinamatthews/tblis supported TBB but I can't tell anymore. You might try that.

Multiple OpenMP runtimes in a single process is not a legal use case for OpenMP. Nothing is the specification required it to work and there are good reasons why it...