Error Compiling

[half-potato]

/home/amai/ParticleNeRF/src/testbed.cu:179:23: error: no match for ‘operator=’ (operand types are ‘ngp::Testbed::Volume’ and ‘<brace-enclosed initializer list>’)
  179 |         m_volume = {};
      |                       ^
/home/amai/ParticleNeRF/include/neural-graphics-primitives/testbed.h:852:15: note: candidate: ‘ngp::Testbed::Volume& ngp::Testbed::Volume::operator=(ngp::Testbed::Volume&&)’
  852 |         struct Volume {
      |               ^~~~~~
/home/amai/ParticleNeRF/include/neural-graphics-primitives/testbed.h:852:15: note:   no known conversion for argument 1 from ‘<brace-enclosed initializer list>’ to ‘ngp::Testbed::Volume&&’
/home/amai/ParticleNeRF/include/particles/particles.h: In instantiation of ‘void particle::Particles<T, POS_DIMS>::init_positions_grid(tcnn::default_rng_t&, uint32_t) [with T = __half; unsigned int POS_DIMS = 3; tcnn::default_rng_t = tcnn::pcg32; uint32_t = unsigned int]’:
/home/amai/ParticleNeRF/include/particles/particles.h:356:1:   required from ‘void particle::Particles<T, POS_DIMS>::init_random(tcnn::default_rng_t&, uint32_t, uint32_t) [with T = __half; unsigned int POS_DIMS = 3; tcnn::default_rng_t = tcnn::pcg32; uint32_t = unsigned int]’
/home/amai/ParticleNeRF/include/particles/particle_encoding.h:275:36:   required from ‘particle::ParticleEncodingTemplated<T, POS_DIMS, N_FEATURE_DIMS>::ParticleEncodingTemplated(const particle::json&) [with T = __half; unsigned int POS_DIMS = 3; unsigned int N_FEATURE_DIMS = 2; particle::json = nlohmann::basic_json<>]’
/home/amai/ParticleNeRF/include/particles/particle_encoding.h:643:17:   required from ‘particle::ParticleEncoding<T>* particle::create_particle_encoding(uint32_t, const json&) [with T = __half; uint32_t = unsigned int; json = nlohmann::basic_json<>]’
/home/amai/ParticleNeRF/src/testbed.cu:3772:92:   required from here
/home/amai/ParticleNeRF/include/particles/particles.h:369:26: warning: narrowing conversion of ‘res’ from ‘uint32_t’ {aka ‘unsigned int’} to ‘Eigen::Matrix<int, 3, 1>::Scalar’ {aka ‘int’} [-Wnarrowing]
  369 |         Eigen::Vector3i res3d = {res, res, res};
      |                          ^~~
/home/amai/ParticleNeRF/include/particles/particles.h:369:31: warning: narrowing conversion of ‘res’ from ‘uint32_t’ {aka ‘unsigned int’} to ‘Eigen::Matrix<int, 3, 1>::Scalar’ {aka ‘int’} [-Wnarrowing]
  369 |         Eigen::Vector3i res3d = {res, res, res};
      |                               ^~~
/home/amai/ParticleNeRF/include/particles/particles.h:369:36: warning: narrowing conversion of ‘res’ from ‘uint32_t’ {aka ‘unsigned int’} to ‘Eigen::Matrix<int, 3, 1>::Scalar’ {aka ‘int’} [-Wnarrowing]
  369 |         Eigen::Vector3i res3d = {res, res, res};
      |                                    ^~~
/home/amai/ParticleNeRF/include/particles/particles.h: In instantiation of ‘void particle::Particles<T, POS_DIMS>::init_positions_grid(tcnn::default_rng_t&, uint32_t) [with T = __half; unsigned int POS_DIMS = 2; tcnn::default_rng_t = tcnn::pcg32; uint32_t = unsigned int]’:
/home/amai/ParticleNeRF/include/particles/particles.h:356:1:   required from ‘void particle::Particles<T, POS_DIMS>::init_random(tcnn::default_rng_t&, uint32_t, uint32_t) [with T = __half; unsigned int POS_DIMS = 2; tcnn::default_rng_t = tcnn::pcg32; uint32_t = unsigned int]’
/home/amai/ParticleNeRF/include/particles/particle_encoding.h:275:36:   required from ‘particle::ParticleEncodingTemplated<T, POS_DIMS, N_FEATURE_DIMS>::ParticleEncodingTemplated(const particle::json&) [with T = __half; unsigned int POS_DIMS = 2; unsigned int N_FEATURE_DIMS = 2; particle::json = nlohmann::basic_json<>]’
/home/amai/ParticleNeRF/include/particles/particle_encoding.h:654:17:   required from ‘particle::ParticleEncoding<T>* particle::create_particle_encoding(uint32_t, const json&) [with T = __half; uint32_t = unsigned int; json = nlohmann::basic_json<>]’
/home/amai/ParticleNeRF/src/testbed.cu:3772:92:   required from here
/home/amai/ParticleNeRF/include/particles/particles.h:369:26: warning: narrowing conversion of ‘res’ from ‘uint32_t’ {aka ‘unsigned int’} to ‘Eigen::Matrix<int, 3, 1>::Scalar’ {aka ‘int’} [-Wnarrowing]
  369 |         Eigen::Vector3i res3d = {res, res, res};
      |                          ^~~
/home/amai/ParticleNeRF/include/particles/particles.h:369:31: warning: narrowing conversion of ‘res’ from ‘uint32_t’ {aka ‘unsigned int’} to ‘Eigen::Matrix<int, 3, 1>::Scalar’ {aka ‘int’} [-Wnarrowing]
  369 |         Eigen::Vector3i res3d = {res, res, res};
      |                               ^~~
/home/amai/ParticleNeRF/include/particles/particles.h:369:36: warning: narrowing conversion of ‘res’ from ‘uint32_t’ {aka ‘unsigned int’} to ‘Eigen::Matrix<int, 3, 1>::Scalar’ {aka ‘int’} [-Wnarrowing]
  369 |         Eigen::Vector3i res3d = {res, res, res};

This is most likely caused by gcc 12.2.1 and CUDA 12.1, but I'm interested in updating the code to ensure compatibility with these newer compilers. Not entirely sure how to fix this issue though.

[jc211]

Hey there, that code is unchanged from InstantNGP. I'm not sure what compilers they support. Maybe you could just go

m_volume = Volume()

instead of

m_volume = {}

[half-potato]

That fixed it, along with pybind/pybind11#4606. Thanks.