Basics

• Questions on a number of code acceleration techniques

Benchmarking

• Don't do timing measurements in global scope.
• Do NOT use code_llvm as your first tool to identify performance issue.
• Compare Speed Python3 vs Julia [on hold]
• How efficient are push!() and append!() methods in Julia?
• Benchmarking Julia on a PDE: the Kuramoto-Sivashinksy equation

Arch D. Robison's Workshop Videos

• JuliaCon 2016 | Introduction to Writing High Performance Julia
• JuliaCon 2015 | Workshop on writing high performance Julia

Type Stability

• Why type instability?
• Type stability when arguments are functions
• Julia: Efficiency of Multiple Parameters

Function Barrier

• Bikeshedding mapslices
• Introduce a function barrier to mapslices #22229
• dealing with types in kwargs in Julia

Allocation

• Generic n-dimensional loop without allocations
• splatting penalty
• Splat on a vector uses allocations and not on a tuple
• Why is a manual in-place addition so much faster than += and .+= on range-indexed arrays?
• Julia - get real part of complex array
• Why is Julia allocating so much memory?
• Julia - nested loops are consuming a lot of memory

  Related Issues

• GC support for stack allocated objects #12205
• Want non-allocating array views #14955
• WIP/RFC unbox more immutables #18632

Constant

• Compile-time constants
• Does declaring a variable to be constant do anything at all?

Keyword Arguments

• Performance of functions with keyword arguments

  Related Issue

• WIP: add NamedTuples #22194

Splatting Penalty

• Call-site splatting optimization #13359
• avoid splatting penalty in chkstride1 #16416
• Fast indexing of arrays

Spatial Locality

• Boxes and Registers
• TiledIteration.jl

Memoization

• wiki
• Caching function values
• Memoize.jl

@pure Function

• @pure macro
• Optimizing offsetof() down to a constant value?

  Related Issues

• Pure function notation #414
• implement pure function annotations #13555
• constant-propagation for @pure functions #14324

Parallel Accelerator

• An introduction to ParallelAccelerator.jl
• JuliaCon 2016 | A Tour of ParallelAccelerator.jl | Lindsey Kuper
• Hai Liu - Parallelizing Julia with a Non-invasive DSL

Case-study

• Julia vs R vs Python
• Kron vs scalar product speed difference. python code faster?
• Speed up this code game
• Machine code discrepancy between Julia and C
• Why are calls to this function slow, compared to Python+Numpy?
• Zeroing elements of matrix A depending on matrix B. Fast?
• Julia - the way of kings (generator performance)