AlphaC Posted March 9, 2018 Posted March 9, 2018 (edited) Please consider adding the real-world based benchmarks used in specwpc (https://www.spec.org/gwpg/wpc.static/wpc-v2.1-info.html) which scale insanely with cores: Rodinia (pre-euler 3D) --- this is representative of Computational Fluid dynamics packages , see https://www.cs.virginia.edu/~skadron/wiki/rodinia/index.php/Rodinia:Accelerating_Compute-Intensive_Applications_with_Accelerators WPCcfd (based on OpenFoam) --- this is representative of Computational Fluid dynamics packages CalculiX ---- this is representative of Finite Element Analysis packages Poisson's Equation ---- partial differentiation solver Monte Carlo , Black Scholes, Binomial --- for financial analysts Luxrender --- maybe , it's a render benchmark but I think Corona benchmark would be more widely used ( https://corona-renderer.com/benchmark/all) FFTW --- anyone using Fast Fourier Transforms will make use of this (http://www.fftw.org/) single precision AVX. octave --- representative of MATLAB workloads to a certain extent (it uses .m files) python --- because python is used in the real world Kirchoff migration --- representative of geophysics workloads Just about nobody uses Superpi or wprime in daily use. The practical applications of that many digits of pi is extremely low. Also nobody buys a 10+ core CPU to run Superpi. x264 and x265 benchmarks are semi-useful as are Cinebench and Blender but there ought to be less focus on Digital Content Creation benchmarks and more scientific/engineering ones. I don't mind if no points are given but it would be helpful to people that use their CPUs for more than "just gaming" to have a better picture of the performance in real workloads. Just a thought. See also: https://www.anandtech.com/show/11839/intel-core-i9-7980xe-and-core-i9-7960x-review/7 http://www.tomshardware.co.uk/amd-ryzen-threadripper-1950x-cpu,review-33976-12.html Edited March 9, 2018 by AlphaC Quote
AlphaC Posted April 8, 2018 Author Posted April 8, 2018 New suggestions: Blender barbershop interior , including Blender team's results : https://code.blender.org/2017/11/cycles-benchmark-amd-update-new-benchmark-file/ Paraview benchmark https://www.paraview.org/ParaView/Doc/Nightly/www/py-doc/paraview.benchmark.html# Quote
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.