10 Commits

Author	SHA1	Message	Date
Hans Dembinski	92a873c746	Faster linearize (#230) ... * introduce offset for faster linearization of non-growing axes * added traits::is_inclusive and constexpr bool inclusive() methods for builtin axes * bug fixes for fill method when weight array is used with non-inclusive axis and when growing axes are used * bug fix of axis::options::test(...) * coverage tested with gcc-8, updated CONTRIBUTING.md with coverage info	2019-10-06 23:03:45 +02:00
Hans Dembinski	016532f011	Histogram fill method (#54) ... huge refactor, various speed improvements, potential for further improvements and parallelization	2019-08-19 23:55:43 +02:00
Hans Dembinski	5f3f9d8fd7	Benchmark improvements and scripts to research history ... * allow benchmarks to run in isolation, new scripts to run historic benchmarks * make sure variant is never trivial in tests and benchmarks	2019-06-02 19:46:57 +02:00
Hans Dembinski	237bf234c5	simplications to linearize and benchmarking	2019-05-30 19:36:02 +02:00
Hans Dembinski	c4c6d90c21	improved benchmarks, turn off asserts	2019-05-30 13:49:26 +02:00
Hans Dembinski	a1c76d2b0a	generate random numbers outside benchmark loop (#195)	2019-05-29 22:26:11 +02:00
Hans Dembinski	4d383fae1b	enable "unsafe" math optimizations in benchmarks	2019-05-28 23:15:44 +02:00
Hans Dembinski	ecd142080d	speed improvements for 1d and 2d histograms ... * rename get_size to axes_rank and move axes buffer to axes.hpp * added fill experiments * faster and simpler fill_impl * faster specializations for linearize_value	2019-05-26 23:51:15 +02:00
Hans Dembinski	cf9520bdfa	make benchmarks compile with -O3 -march=native -fno-exceptions and -fno-rtti	2019-05-15 21:53:32 +02:00
Hans Dembinski	e75aa6b210	Google benchmark and benchmark improvements ... * fetching google benchmark with cmake * use new thread-safe accumulator in histogram_parallel_filling * update boost, root and gsl benchmarks to use google benchmark	2019-05-03 22:06:01 +02:00