Karthikeyan

@albert-github Sorry. I thought, `@return` overrides it. It doesn't. It actually adds extra "Returns" section. But doxygen does not throw a warning. But doxygen throws a warning for `multiple @param...

@gpucibot merge

Benchmark comparisons. [BENCH_JSONs.zip](https://github.com/rapidsai/cudf/files/9031713/BENCH_JSONs.zip) commit ff63c0a745382d79eaf0ae45931e7167d9763b88 (HEAD -> branch-22.08) vs This PR branch. ``` ./_deps/benchmark-src/tools/compare.py benchmarks branch-22.08/CSV_READER_BENCH.json enh-test_float_parse/CSV_READER_BENCH.json Comparing branch-22.08/CSV_READER_BENCH.json to enh-test_float_parse/CSV_READER_BENCH.json Benchmark Time CPU Time Old Time New CPU Old...

rerun tests

style check cmake-format.............................................................Failed

The main reason for creating this iterator is to provide a generalized output iterator. For input iterator, we can create `counting_iterator` and use a `make_transform_iterator` to translate it to access...

@jrhemstad In https://github.com/NVIDIA/cccl/issues/792 godbolt example, it still uses `f.begin()` as starting point for the output iterator. This output iterator is transformed using projections. Instead of `f.begin()`, can we use a...

deleted copy ctor declaration of cuda atomic? Will it stop working in Linux after the fix?

Right now, `device_uvector

Usage of (present) cmath functions in Windows nvcc 11.2 causes error. `nvcc -std=c++17 cmath_test.cu --expt-relaxed-constexpr` ``` cmath_test.cu(4): error: calling a __host__ function("exp ") from a __global__ function("kernel") is not allowed...