Kévin Petit

Depends on https://github.com/google/clspv/issues/321 - [ ] Basic type support, constants and operators - [ ] Conversions - [ ] Math built-ins - [ ] Common built-ins - [ ] Geometric...

Depends on https://github.com/google/clspv/issues/276 - [ ] Basic type support, constants and operators - [ ] Conversions - [ ] Math built-ins - [ ] Common built-ins - [ ] Geometric...

So others can easily run conformance tests.

Support for physical addressing

3

## Motivation - Full support for OpenCL C requires physical addressing. While clspv has done a great job of enabling many OpenCL C programs without the use of physical addressing,...

acosh built-in function more accurate with -cl-native-math than without

4

#715 regressed the acosh bruteforce CTS test (`-w` wimpy, `-1` scalar) for me (NVidia GeForce GTX 1050 Ti, Linux driver 460.56). Passing `-cl-native-math` to clspv fixes the issue. This could...

Support double-precision floating point

2

clspv generally doesn't have support for double-precision floating point. The following at least are missing: - [x] support double literals (see #157) - [x] support double constants - [ ]...

Cross-kernel influence in binding allocation

2

Clvk currently assumes that the highest binding number in a descriptor set is less than the number of resources that require a binding. This no longer seems to be true...