ochafik

# Context: https://github.com/openscad/openscad/pull/3636

Notice the N=3 vs. N=5 below. All test on a MacBook Pro w/ 2.6GHz Core i7.

# Baseline
time openscad spheres_single_for_loop.scad  -o out.stl -DN=3 -Doverlap=true # 52s
time openscad spheres_nested_for_loops.scad -o out.stl -DN=3 -Doverlap=true # 45s
time openscad spheres_single_for_loop.scad  -o out.stl -DN=5 -Doverlap=true # 3min32
time openscad spheres_nested_for_loops.scad -o out.stl -DN=5 -Doverlap=true # 2min41

ochafik

#!/bin/bash
# Copyright 2021 Google LLC.
# SPDX-License-Identifier: Apache-2.0
#
# Author: Olivier Chafik (http://ochafik.com)
#
# Usage:
#   - Checkout my optimization integration branch (doesn't include multithreaded changes):
#     git clone https://github.com/ochafik/openscad openscad-ochafik
#     cd openscad-ochafik

ochafik

// Copyright 2021 Google LLC.
// SPDX-License-Identifier: Apache-2.0
//
// Author: Olivier Chafik (http://ochafik.com)
//
// This is extracted from my little chain/scalemail project shared in
// https://gist.github.com/ochafik/6d19e3b131153305b44ea275d5d930e3
//
// It is simpler to render, this example is crafted to exercice the fast-union
// feature I'm proposing in https://github.com/openscad/openscad/pull/3636

ochafik

ochafik

// Context: ochafik@'s attempts to speed up OpenSCAD
// fast-union feature: https://github.com/openscad/openscad/pull/3636
// lazier union features: https://github.com/openscad/openscad/pull/3637
//
// (and parallelization attempts, not sent as PRs yet)
//
// fast-union, lazy-union, push-transforms-down-unions, lazy-module and flatten-children all help here!
// openscad --enable=fast-union cube-with-half-spheres-dents.scad -o cube-with-half-spheres-dents.stl

top_transform = true;

ochafik

Here are some micro benchmarks that demonstrate big gains of OpenSCAD rendering with --enable=fast-csg (a feature I developed in this pull request).

Note: some (but not all) models making heavy use of minkowski may be slower with the feature enabled, as CGALUtils::applyMinkowski is forcing conversions to the old nef format. I hope to address that + publish benchmarks involving minkowski operations soon!*

Note2: these figures need updating, lots of changes since last time I tested these files! Most notably, I'm now proactively falling back to Nef computations in many cases that could have been handled by fast PMP operations on surface meshes, just by fear of hitting corner cases that might corrupt the data. Makes things slower :-(

• uncacheable-grid.scad (see below): 35sec -> 1.5sec (31x faster)
• chainmail.scad (see below): 24min 15sec -> 4min 27sec (5x faster)
• spheres_nested_for_loops.scad (see below): 44sec

ochafik

// Copyright 2021 Google LLC.
// SPDX-License-Identifier: Apache-2.0
//
/*
This model exercices a few experimental features I've been brewing.
It's not a theoretical case tho: I was trying to print a chainmail with
variable tiles when OpenSCAD's rendering scalability challenges hit me hard :-S
(guess I fell down the rabbit hole of 3d rendering optimization at the
temporary expense of my 3d printing projects, oh well!)

ochafik

mkdir build
( cd build && cmake .. -DCMAKE_BUILD_TYPE=Release -DEXPERIMENTAL=1 )
( cd build && time make -j12 )
( cd build/tests && time ctest -j12 -E 'edges_view-options-tests|nef3_broken' )
( cd build/tests && time ctest -j12 -R 'fastcsg.*' )

ochafik

import 'dart:js';
import 'dart:html';

main() {
  printJsObject(o) =>
    (JsObject(context['Function'], ['console.log(arguments[0])']) as JsFunction).apply([o]);

  printJsObject(DivElement());
  // prints <div></div>
  

ochafik

import 'dart:js';
import 'dart:html';

main() {
  printJsObject(document.body);
  var o = IntersectionObserver((entries, observer) {
    printJsObject(observer);
    printJsObject(JsObject.fromBrowserObject(observer));
    for (final entry in entries) {
      printJsObject(entry);