This patch implements the vector types (i.e:`float2`) by making heavy
usage of templating. All vector functions are now outside of the vector
classes (inside the `blender::math` namespace) and are not vector size
dependent for the most part.
In the ongoing effort to make shaders less GL centric, we are aiming
to share more code between GLSL and C++ to avoid code duplication.
####Motivations:
- We are aiming to share UBO and SSBO structures between GLSL and C++.
This means we will use many of the existing vector types and others
we currently don't have (uintX, intX). All these variations were
asking for many more code duplication.
- Deduplicate existing code which is duplicated for each vector size.
- We also want to share small functions. Which means that vector
functions should be static and not in the class namespace.
- Reduce friction to use these types in new projects due to their
incompleteness.
- The current state of the `BLI_(float|double|mpq)(2|3|4).hh` is a
bit of a let down. Most clases are incomplete, out of sync with each
others with different codestyles, and some functions that should be
static are not (i.e: `float3::reflect()`).
####Upsides:
- Still support `.x, .y, .z, .w` for readability.
- Compact, readable and easilly extendable.
- All of the vector functions are available for all the vectors types
and can be restricted to certain types. Also template specialization
let us define exception for special class (like mpq).
- With optimization ON, the compiler unroll the loops and performance
is the same.
####Downsides:
- Might impact debugability. Though I would arge that the bugs are
rarelly caused by the vector class itself (since the operations are
quite trivial) but by the type conversions.
- Might impact compile time. I did not saw a significant impact since
the usage is not really widespread.
- Functions needs to be rewritten to support arbitrary vector length.
For instance, one can't call `len_squared_v3v3` in
`math::length_squared()` and call it a day.
- Type cast does not work with the template version of the `math::`
vector functions. Meaning you need to manually cast `float *` and
`(float *)[3]` to `float3` for the function calls.
i.e: `math::distance_squared(float3(nearest.co), positions[i]);`
- Some parts might loose in readability:
`float3::dot(v1.normalized(), v2.normalized())`
becoming
`math::dot(math::normalize(v1), math::normalize(v2))`
But I propose, when appropriate, to use
`using namespace blender::math;` on function local or file scope to
increase readability.
`dot(normalize(v1), normalize(v2))`
####Consideration:
- Include back `.length()` method. It is quite handy and is more C++
oriented.
- I considered the GLM library as a candidate for replacement. It felt
like too much for what we need and would be difficult to extend / modify
to our needs.
- I used Macros to reduce code in operators declaration and potential
copy paste bugs. This could reduce debugability and could be reverted.
- This touches `delaunay_2d.cc` and the intersection code. I would like
to know @howardt opinion on the matter.
- The `noexcept` on the copy constructor of `mpq(2|3)` is being removed.
But according to @JacquesLucke it is not a real problem for now.
I would like to give a huge thanks to @JacquesLucke who helped during this
and pushed me to reduce the duplication further.
Reviewed By: brecht, sergey, JacquesLucke
Differential Revision: https://developer.blender.org/D13791
This patch implements the vector types (i.e:`float2`) by making heavy
usage of templating. All vector functions are now outside of the vector
classes (inside the `blender::math` namespace) and are not vector size
dependent for the most part.
In the ongoing effort to make shaders less GL centric, we are aiming
to share more code between GLSL and C++ to avoid code duplication.
####Motivations:
- We are aiming to share UBO and SSBO structures between GLSL and C++.
This means we will use many of the existing vector types and others
we currently don't have (uintX, intX). All these variations were
asking for many more code duplication.
- Deduplicate existing code which is duplicated for each vector size.
- We also want to share small functions. Which means that vector
functions should be static and not in the class namespace.
- Reduce friction to use these types in new projects due to their
incompleteness.
- The current state of the `BLI_(float|double|mpq)(2|3|4).hh` is a
bit of a let down. Most clases are incomplete, out of sync with each
others with different codestyles, and some functions that should be
static are not (i.e: `float3::reflect()`).
####Upsides:
- Still support `.x, .y, .z, .w` for readability.
- Compact, readable and easilly extendable.
- All of the vector functions are available for all the vectors types
and can be restricted to certain types. Also template specialization
let us define exception for special class (like mpq).
- With optimization ON, the compiler unroll the loops and performance
is the same.
####Downsides:
- Might impact debugability. Though I would arge that the bugs are
rarelly caused by the vector class itself (since the operations are
quite trivial) but by the type conversions.
- Might impact compile time. I did not saw a significant impact since
the usage is not really widespread.
- Functions needs to be rewritten to support arbitrary vector length.
For instance, one can't call `len_squared_v3v3` in
`math::length_squared()` and call it a day.
- Type cast does not work with the template version of the `math::`
vector functions. Meaning you need to manually cast `float *` and
`(float *)[3]` to `float3` for the function calls.
i.e: `math::distance_squared(float3(nearest.co), positions[i]);`
- Some parts might loose in readability:
`float3::dot(v1.normalized(), v2.normalized())`
becoming
`math::dot(math::normalize(v1), math::normalize(v2))`
But I propose, when appropriate, to use
`using namespace blender::math;` on function local or file scope to
increase readability.
`dot(normalize(v1), normalize(v2))`
####Consideration:
- Include back `.length()` method. It is quite handy and is more C++
oriented.
- I considered the GLM library as a candidate for replacement. It felt
like too much for what we need and would be difficult to extend / modify
to our needs.
- I used Macros to reduce code in operators declaration and potential
copy paste bugs. This could reduce debugability and could be reverted.
- This touches `delaunay_2d.cc` and the intersection code. I would like
to know @howardt opinion on the matter.
- The `noexcept` on the copy constructor of `mpq(2|3)` is being removed.
But according to @JacquesLucke it is not a real problem for now.
I would like to give a huge thanks to @JacquesLucke who helped during this
and pushed me to reduce the duplication further.
Reviewed By: brecht, sergey, JacquesLucke
Differential Revision: https://developer.blender.org/D13791
This patch implements the vector types (i.e:float2) by making heavy
usage of templating. All vector functions are now outside of the vector
classes (inside the blender::math namespace) and are not vector size
dependent for the most part.
In the ongoing effort to make shaders less GL centric, we are aiming
to share more code between GLSL and C++ to avoid code duplication.
Motivations:
- We are aiming to share UBO and SSBO structures between GLSL and C++.
This means we will use many of the existing vector types and others we
currently don't have (uintX, intX). All these variations were asking
for many more code duplication.
- Deduplicate existing code which is duplicated for each vector size.
- We also want to share small functions. Which means that vector functions
should be static and not in the class namespace.
- Reduce friction to use these types in new projects due to their
incompleteness.
- The current state of the BLI_(float|double|mpq)(2|3|4).hh is a bit of a
let down. Most clases are incomplete, out of sync with each others with
different codestyles, and some functions that should be static are not
(i.e: float3::reflect()).
Upsides:
- Still support .x, .y, .z, .w for readability.
- Compact, readable and easilly extendable.
- All of the vector functions are available for all the vectors types and
can be restricted to certain types. Also template specialization let us
define exception for special class (like mpq).
- With optimization ON, the compiler unroll the loops and performance is
the same.
Downsides:
- Might impact debugability. Though I would arge that the bugs are rarelly
caused by the vector class itself (since the operations are quite trivial)
but by the type conversions.
- Might impact compile time. I did not saw a significant impact since the
usage is not really widespread.
- Functions needs to be rewritten to support arbitrary vector length. For
instance, one can't call len_squared_v3v3 in math::length_squared() and
call it a day.
- Type cast does not work with the template version of the math:: vector
functions. Meaning you need to manually cast float * and (float *)[3] to
float3 for the function calls.
i.e: math::distance_squared(float3(nearest.co), positions[i]);
- Some parts might loose in readability:
float3::dot(v1.normalized(), v2.normalized())
becoming
math::dot(math::normalize(v1), math::normalize(v2))
But I propose, when appropriate, to use
using namespace blender::math; on function local or file scope to
increase readability. dot(normalize(v1), normalize(v2))
Consideration:
- Include back .length() method. It is quite handy and is more C++
oriented.
- I considered the GLM library as a candidate for replacement.
It felt like too much for what we need and would be difficult to
extend / modify to our needs.
- I used Macros to reduce code in operators declaration and potential
copy paste bugs. This could reduce debugability and could be reverted.
- This touches delaunay_2d.cc and the intersection code. I would like to
know @Howard Trickey (howardt) opinion on the matter.
- The noexcept on the copy constructor of mpq(2|3) is being removed.
But according to @Jacques Lucke (JacquesLucke) it is not a real problem
for now.
I would like to give a huge thanks to @Jacques Lucke (JacquesLucke) who
helped during this and pushed me to reduce the duplication further.
Reviewed By: brecht, sergey, JacquesLucke
Differential Revision: http://developer.blender.org/D13791
- Added space below non doc-string comments to make it clear
these aren't comments for the symbols directly below them.
- Use doxy sections for some headers.
- Minor improvements to doc-strings.
Ref T92709
A parallel loop to create the interesection meshes for each triangle
meant that with parallelism, the output order of the created meshes
could vary with each execution. Keep the parallelism for doing the CDTs
for interesection, but move the extraction of the new faces into a
serial loop afterwards, for repeatability.
The problem is that the fast triangulator (based on polyfill)
sometimes makes degenerate triangles. Commit 8115f0c5bd91f had
a check for degenerate triangles but it wasn't thorough enough.
This commit uses a more thorough (and pessimistic) test for
degenerate triangles, using the exact triangulator in those cases.
Using part of a patch from Erik Abrahamsson, this replaces the
use of linked lists for original id tracking by Sets.
I had thought that the lists were unlikely to grow to more than
a few elements, but when the mesh has a lot of holes (whose
original ids go *outside* the hole, and therefore, most of the
mesh), this assumption can be very wrong.
On a Text regression test, the time went from 11.67s to 0.16s
with this fix. I also tested to make sure that Boolean didn't
slow down with this, and found it actually had a very slight speedup.
Using Sets exposed a dependency on the ordering of the items
in the id lists, luckily caught by a mesh intersect regression test,
so fixed that.
From patch D11780 from Erik Abrahamsson.
It parallelizes making the vertices, destruction of map entries,
finding if the result is PWN, finding triangle adjacencies,
and finding the ambient cell.
The latter needs a parallel_reduce from tbb, so added one into
BLI_task.hh so that if WITH_TBB is false, the code will still work.
On Erik's 6-core machine, the elapsed time went from 17.5s to 11.8s
(33% faster) on an intersection of two spheres with 3.1M faces.
On Howard's 24-core machine, the elapsed time went from 18.7s to 10.8s
for the same test.
Went back to using Blender's polyfill for triangulation, which is much
faster (time for a 3.1M face boolean went from 103s to 48s).
Had to put in detection for the case that needs the exact triangulator
(bug T86805), and also a fix for non-convex quads (bug T89330).
The problem was an optimization I put in to triangulate quads.
It was wrong if the quad, after projecting onto a 2d plane, was
not convex. Handling quads the same as other faces fixes the bug.
Unfortunately, this will slow down Exact Boolean when the input has
many quads (the usual case, of course).
Will attempt to fix that with a later change, but for now, this
at least restores correctness.
Use bounding box tests quickly tell that two components cannot
have a containment relation between each other. This change
cut about 0.6s off a test with 25 big icospheres.
This is from patch D11432 from Erik Abrahamsson. He found that
in some mpq3 functions called frequently from loops, passing in
buffers for termporary mpq3 values can save substantial time.
On my machine, his example in that patch went from 9.48s to 7.50s
for the boolean part of the calculation. On his machine, a running
time went from 17s to 10.3s.
There was a quadratic algorithm extracting triangles from a coplanar
cluster. This is now linear.
Also found and fixed a bug in the same area related to the triangulator
added recently: it didn't get the right correspondence between new
edges and original edges.
This removes a lot of unnecessary code that is generated by
the compiler automatically.
In very few cases, a defaulted destructor in a .cc file is
still necessary, because of forward declarations in the header.
I removed some defaulted virtual destructors, because they are not
necessary, when the parent class has a virtual destructor already.
Defaulted constructors are only necessary when there is another
constructor, but the class should still be default constructible.
Differential Revision: https://developer.blender.org/D10911
The existing hash function didn't work well with Set's method of
masking to the lower bits, because many verts have zeros in the
lower bits.
Also, replaced VectorSet with Set for Vert deduping.
Two problems were fixed. One, the code for dissolving vertices
left a face around if dissolving a vertex would leave less than
three vertices. Instead, the face should be deleted.
Two, with transformations like "rotate 180 degrees", this should
be no problem with exact, but the current transformation matrix
has very small non-zero entries where it shouldn't. Cleaning the
transformation matrix makes it more likely that user expectations
about coplanar faces will be fulfilled.
The code for determining coplanar clusters had a bug where it would
miss some triangles. The fix for now is to just put triangles in
the cluster if their bounding boxes overlap. This works but maybe
makes clusters bigger then they have to be. I'll follow this commit
with work on making the CDT routine faster when using exact arithmetic.
Also removed a lot of unused code, and added some new intersect
performance tests.
The code that decided to use a faster double version of plane
side testing forgot to take an absolute value, so half the time
the exact code was being used when it was unnecessary.
The code that decided to use a faster double version of plane
side testing forgot to take an absolute value, so half the time
the exact code was being used when it was unnecessary.
The code to detect non-trivial coplanar intersection sometimes
falsely said there wasn't one. This caused some coplanar intersections
to be missed. Also added a test for this case.
Instead of calculating exact normals for all faces, just do it
for those that potentially intersect. A big improvement for dense
meshes that only intersect in relatively few places.
The code that found coplanar clusters was not updating a bounding box.
Also, code that was detecting non-trivial coplanar intersects was
slightly wrong, but that would not have caused any functional problems.
