Python: Difference between revisions
→Performance
m (→Applications) |
|||
| Line 393: | Line 393: | ||
== Performance == | == Performance == | ||
The derivation of CPython that is available via <code>nixpkgs</code> only contains optimizations that do not harm reproducibility. Link-Time-Optimization (LTO) is only enabled on 64-bit Linux systems, while Profile Guided Optimization (PGO) is currently disabled. See [https://docs.python.org/3/using/configure.html#performance-options Configuring Python 3.1.3. Performance options] | The derivation of CPython that is available via <code>nixpkgs</code> only contains optimizations that do not harm reproducibility. Link-Time-Optimization (LTO) is only enabled on 64-bit Linux systems, while Profile Guided Optimization (PGO) is currently disabled. See [https://docs.python.org/3/using/configure.html#performance-options Configuring Python 3.1.3. Performance options] | ||
Additionally, when | Additionally, when compiling something within <code>nix-shell</code> or a derivation security hardening flags are passed to the compiler by default which may have a small performance impact. | ||
At the time of writing certain optimizations cause Python wheels to be non-reproducible and increase install times. For a detailed overview of the trials and tribulations of discovering such performance regressions see [https://discourse.nixos.org/t/why-is-the-nix-compiled-python-slower/18717 Why is the nix-compiled Python slower?]. | |||
=== Regression === | === Regression === | ||
With the <code>nixpkgs</code> version of Python you can expect anywhere from a 30-40% regression on synthetic benchmarks. For example: | With the <code>nixpkgs</code> version of Python you can expect anywhere from a 30-40% regression on synthetic benchmarks. For example: | ||