- 22 10月, 2020 6 次提交
-
-
由 Matt Pharr 提交于
Issue #47.
-
由 Matt Pharr 提交于
Possible fix for #63.
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
activeTransformBits wasn't being reset to AllTransformBits, like it was before. (Bug introduced in 7164c954.)
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
-
- 19 10月, 2020 2 次提交
-
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
-
- 18 10月, 2020 2 次提交
-
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
(They are essentially useless now that texture and media definitions do not obey attribute scope any more.)
-
- 17 10月, 2020 1 次提交
-
-
由 Matt Pharr 提交于
No functional changes.
-
- 16 10月, 2020 3 次提交
-
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
-
- 15 10月, 2020 6 次提交
-
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
Some of the TypePack capabilities have been improved, in preparation for forthcoming cleanups in the TaggedPointer implementation.
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
If a surface is either reflective or transmissive but not both, offset the reference point passed to Light::SampleLi() so that it it is on the same side of the surface as the outgoing ray will be. Fixes issue #59.
-
由 Matt Pharr 提交于
Replace piecewise-linear LUT in LineartoSRGB() with rational polynomial approximation from enoki. Also, use a rational polynomial approximation in place of std::pow() for the general SRGBToLinear() function.
-
- 14 10月, 2020 2 次提交
-
-
由 Matt Pharr 提交于
Actually initialize RGBSpectrum::rgb member variable in the constructor. Remove now-unused ParameterType() and ParameterString() methods.
-
由 Matt Pharr 提交于
-
- 13 10月, 2020 6 次提交
-
-
由 Matt Pharr 提交于
RGBSigmoidPolynomial: use an exact inverse mapping for uniform colors
-
由 Wenzel Jakob 提交于
-
由 Matt Pharr 提交于
This had broken the GPU build, since we need to be able to call its size() method in GPU code.
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
Calling GetMesh() was a bad idea, since the underlying vector could be updated concurrently as other bilinear patch meshes were created in parallel, which in turn could lead to the vector being resized and the constructor reading freed memory.
-
由 Matt Pharr 提交于
No meaningful changes, though in FastExp(), underflow and overflow are now handled more rigorously.
-
- 12 10月, 2020 1 次提交
-
-
由 Matt Pharr 提交于
-
- 10 10月, 2020 5 次提交
-
-
由 Matt Pharr 提交于
The rounded floating-point math intrinsics are used on the GPU now, which gives tighter bounds thanks to less unnecessary conservative rounding.
-
由 Matt Pharr 提交于
No change to functionality, but the Interval class is no longer templated on the Float type. This added a whole bunch of complexity that wasn't actually being used.
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
(i.e., don't sample it when MIS is happening.) Also, fix a bug in UniformInfiniteLight::Phi() (though no one ever calls it...)
-
由 Matt Pharr 提交于
Rework the compilation to PTX and embedding
-
- 09 10月, 2020 1 次提交
-
-
由 Matt Pharr 提交于
-
- 08 10月, 2020 5 次提交
-
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
-
由 Matt Pharr 提交于
(Be consistent with changes in #55.)
-
由 Pierre Moreau 提交于
Since we have a CMake target, we can do scoped include directories. Remove duplicate compile flags, and also add compile definitions which enables PBRT to compile on Windows for Pascal-architecture cards. Compared to commit a685e020, this also * changes the dependency to be directly on the library rather than some of its object file; * adds a missing dependency on several generated header files; * avoids passing `/D _CRT_SECURE_NO_WARNINGS` to nvcc; * removes compiler flags that are part of the regular CUDA flags.
-