Punchline: rewrote the code to look almost identical to C++, hand-held the compiler by adding @-marks to disable safety checks, forced SIMD codegen and fastmath on.
End result: code that is uglier and still much slower than C++. Kind of a shame.
Very interesting post and I think this exposes the limitations of the Julia compiler. Note that an old version of the compiler is used (1.0.3 from 2019).
One could say that we can almost replicate the semantic of a C++ program, but writing in Julia. For example we can remove bounds checks in arrays or remove hidden memory allocations.
But the goal of a language for numerical computing is capturing the mathematical formulas using high level constructs closer to the original representation while compiling to efficient code.
Domain scientists want to play with the math and the formulas, not doing common subexpression elimination in their programs. Just curious to see how it evolves
Making Julia as Fast as C++ (2019)
(flow.byu.edu)28 points by d_tr 6 May 2026 | 13 comments
Comments
End result: code that is uglier and still much slower than C++. Kind of a shame.
One could say that we can almost replicate the semantic of a C++ program, but writing in Julia. For example we can remove bounds checks in arrays or remove hidden memory allocations.
But the goal of a language for numerical computing is capturing the mathematical formulas using high level constructs closer to the original representation while compiling to efficient code.
Domain scientists want to play with the math and the formulas, not doing common subexpression elimination in their programs. Just curious to see how it evolves