RayD Dr.Jit

PyPI License

RayD is a Dr.Jit-native GPU library for differentiable ray geometry and multipath simulation primitives with OptiX and pure-CUDA ray-tracing backends.

pip install rayd-drjit

RayD is not a full renderer. It exposes low-level scene, ray, edge, visibility, and reflection-path queries for building custom renderers, RF simulators, acoustic tools, and inverse-design systems without adopting a material-light-integrator framework.

Release compatibility

Release wheels cover Linux x86-64 and Windows x86-64 on CPython 3.10-3.14. They are built with CUDA 12.8 and contain native code for sm_70, sm_75, sm_80, sm_86, sm_87, sm_89, sm_90, sm_100, sm_101, and sm_120, plus compute_120 PTX for forward compatibility. sm_87 is present for Orin-compatible builds, but the published x86-64 wheels themselves are not aarch64 packages; Jetson currently requires a native/source build.

Ray-tracing backend selection

Scene() uses OptiX when the current CUDA device/context accepts an OptiX context and otherwise selects the pure-CUDA triangle backend plus the Dr.Jit/CUDA edge BVH. Use trace_backend="optix" or "cuda" and edge_bvh_backend="optix" or "drjit" to require a backend. scene.capabilities() reports the resolved choice. An explicit OptiX request fails when unavailable, and operational pipeline, allocation, or CUDA failures are not converted into a fallback. RAYD_DISABLE_OPTIX=1 forces capability discovery to report OptiX unavailable.

RayD is Dr.Jit-native and does not depend on PyTorch. Because its nanobind extension uses the CPython ABI, CI builds one wheel per Python version rather than one Python-independent wheel. The complete build matrix and release configuration are documented in CI_BUILD_MATRIX.md.

Scope

RayD focuses on geometry and wave-propagation primitives:

  • differentiable ray-mesh intersection
  • scene-level GPU acceleration through OptiX or a pure-CUDA BVH
  • nearest-edge queries through a scene-global edge BVH
  • primary-edge sampling support for edge-based gradient terms
  • segment visibility and multipath reflection primitives
  • Dr.Jit arrays and Dr.Jit autodiff as the public Python API

RayD intentionally does not provide:

  • a material or BSDF system
  • emitters
  • integrators
  • scene loading
  • image I/O
  • a complete path-tracing runtime
  • alternate Python frontend wrappers

Why RayD?

RayD is for users who need fast differentiable geometry queries, but do not want a full rendering framework.

Mitsuba is excellent for physically based rendering, but it can be too high-level when the main workload is RF propagation, acoustics, sonar, visibility analysis, or custom wave simulation. In those settings, direct control over ray-scene queries, edge queries, reflection paths, and geometry gradients is often more useful than a complete renderer.

RayD keeps the API surface small: meshes, scenes, rays, intersections, edges, and multipath query results.

Core API

  • Mesh: triangle geometry, transforms, UVs, and edge topology
  • Scene: a container of meshes plus OptiX acceleration structures
  • scene.intersect(ray): differentiable ray-mesh intersection
  • scene.shadow_test(ray): occlusion testing
  • scene.nearest_edge(query): nearest-edge queries for points and rays
  • scene.set_edge_mask(mask) / scene.edge_mask(): scene-global filtering for secondary-edge queries
  • scene.trace_reflections(...): specular reflection-path tracing
  • scene.visible(...): batched segment visibility queries
  • scene.trace_refl_epc(...): equivalent-path correction primitives for reflection paths
  • scene.accum_dfr_direct(...) / scene.accum_dfr(...): native diffraction grid accumulation
  • scene.trace_dfr_paths(...): compact native first-order diffraction path export

Feature Overview

Each core query, what it computes, its input/output, and how it behaves under Dr.Jit autodiff (AD):

API What it computes Input -> Output AD
scene.intersect(ray) Closest-hit ray-mesh intersection rays -> Intersection (t, point, normal, uv, ids) AD
scene.shadow_test(ray) Any-hit occlusion test rays -> boolean mask Boolean
scene.nearest_edge(point) Nearest scene edge to each point points -> NearestPointEdge (distance, points, edge id) AD
scene.nearest_edge(ray) Nearest scene edge to each ray (segment on [0, tmax] when tmax is finite) rays -> NearestRayEdge AD
scene.nearest_edges(point, k) k nearest scene edges per point (k <= 16) points -> NearestEdgesTopK AD
scene.visible(start, end) Mutual visibility between two segment endpoints endpoints -> SegmentVisibility Boolean
scene.visible_pair / visible_chain / visible_edge Shared-origin pair, polyline-chain, and edge-sample visibility segments -> segment/chain/edge visibility Boolean
scene.trace_reflections(ray, max_bounces) Specular reflection paths with image sources rays -> ReflectionChain (hit points, normals, image sources, ids) AD
scene.trace_refl_epc(ray, receiver, ...) Equivalent-path-correction reflection toward a receiver rays + receiver -> ReflEpc Detached
scene.trace_refl_epc_field(tx, receiver, ...) EPC trace returning the complex reflected field tx position + receiver + ReflEpcFieldOptions(AD) -> ReflEpcField (complex E-field) Native AD
scene.accumulate_reflections(...) Accumulate reflected field/power onto a grid rays + grid + material -> AccumResult Native AD
scene.accum_dfr_direct(...) Native direct/Keller/suffix diffraction accumulation onto a grid DfrStates + DfrGrid + DfrMaterial -> DfrAccum Native AD for direct, Keller, and suffix-reflection order-1, including suffix reflector mesh-vertex gradients; detached for unsupported AD strategies
scene.accum_dfr(...) Native order-2/3 diffraction-chain accumulation onto a grid initial/recursive DfrStates + grid + material -> DfrAccum Native AD for direct, Keller, and suffix-reflection order-2/3 chains, including suffix reflector mesh-vertex gradients
scene.trace_dfr_paths(...) Compact first-order diffraction path export tx/rx + DfrStates + DfrMaterial -> DfrPaths AD

AD legend:

  • AD - differentiable geometry: geometric outputs carry Dr.Jit gradients with respect to mesh vertices and transforms; the discrete hit/edge/path selection runs detached.
  • Native AD - the native RayD multipath implementation preserves Dr.Jit gradients for continuous inputs passed as AD arrays and supplies RayD-side derivative code. Diffraction grid accumulation uses CUDA custom-op JVP/VJP over a fixed OptiX forward tape for direct, Keller, and suffix-reflection order-1 and order-2/3 chain accumulation. Discrete visibility/path-selection decisions remain detached.
  • Boolean - returns occlusion/visibility booleans and is not differentiable.
  • Detached - native fast path: runs detached only and rejects AD inputs.

Trace vs. accum:

  • trace_* APIs enumerate geometric or field paths and return per-ray/per-path records. They do not reduce contributions into receiver grids.
  • accum_* APIs launch native accumulation kernels that reduce contributions into aggregate outputs, usually grid cells, with atomic or tiled writes. They are the high-throughput path for radiomap-style workloads when individual path records are not needed.

Minimal Example

The example below traces one ray against one triangle and backpropagates the hit distance to the vertex positions.

import drjit as dr
import rayd.drjit as rd


mesh = rd.Mesh(
    dr.cuda.Array3f([0.0, 1.0, 0.0],
                    [0.0, 0.0, 1.0],
                    [0.0, 0.0, 0.0]),
    dr.cuda.Array3i([0], [1], [2]),
)

verts = dr.cuda.ad.Array3f(
    [0.0, 1.0, 0.0],
    [0.0, 0.0, 1.0],
    [0.0, 0.0, 0.0],
)
dr.enable_grad(verts)
mesh.vertex_positions = verts

scene = rd.Scene()
scene.add_mesh(mesh)
scene.build()

ray = rd.RayAD(
    dr.cuda.ad.Array3f([0.25], [0.25], [-1.0]),
    dr.cuda.ad.Array3f([0.0], [0.0], [1.0]),
)

its = scene.intersect(ray)
loss = dr.sum(its.t)
dr.backward(loss)

print("t =", its.t)
print("grad z =", dr.grad(verts)[2])

Edge Queries

RayD provides a scene-level edge acceleration structure for nearest-edge and edge-sampling workloads.

This is useful for:

  • edge sampling
  • nearest-edge queries
  • visibility-boundary terms
  • geometric diffraction models

Scene.set_edge_mask(mask) filters the secondary-edge BVH in scene-global edge index space. It does not modify scene.edge_info(), scene.edge_topology(), or scene.mesh_edge_offsets().

Multipath Queries

RayD includes low-level reflection and visibility primitives for custom wave simulators:

  • trace_reflections(...) for specular reflection chains
  • visible(...), visible_pair(...), visible_chain(...), and visible_edge(...) for segment and polyline-chain visibility
  • trace_refl_epc(...) and trace_refl_epc_field(...) for equivalent-path correction workflows
  • accumulate_reflections(...) for reflection grid accumulation workloads
  • accum_dfr_direct(...), accum_dfr(...), trace_dfr_paths(...) for native diffraction kernels

These APIs expose primitives, not a complete propagation simulator. Callers own the source model, receiver model, material policy, objective, and optimization loop.

Naming rule: use Dfr for diffraction (DfrStates, DfrAccum, accum_dfr_direct), Refl for reflection-specific short names, keep Epc in equivalent-path-correction APIs, and reserve AD for automatic differentiation. See API_NAMING_STANDARD.md.

Examples

Performance

The chart below was generated on March 25, 2026 on an NVIDIA GeForce RTX 5080 and AMD Ryzen 7 9800X3D, comparing RayD (0.1.2) against Mitsuba 3.8.0 with the cuda_ad_rgb variant.

Raw benchmark data is stored in docs/performance_benchmark.json.

  • RayD is consistently faster on static forward and static gradient workloads across all three scene sizes.
  • Dynamic reduced forward reaches parity or better from the medium scene onward, and dynamic full is effectively tied on the largest case.
  • On the largest 192x192 mesh / 384x384 ray benchmark, RayD vs Mitsuba average latency in milliseconds is: static full 0.162 vs 0.190, static reduced 0.124 vs 0.224, dynamic full 0.741 vs 0.740, dynamic reduced 0.689 vs 0.714, gradient static 0.411 vs 0.757, gradient dynamic 1.324 vs 1.413.
  • Correctness stayed aligned throughout the sweep: forward mismatch counts remained 0, and the largest static gradient discrepancy was 9.54e-7.

RayD vs Mitsuba performance benchmark

Device Selection

RayD follows Dr.Jit's current-thread CUDA device selection. Choose a GPU before constructing RayD resources:

import rayd.drjit as rd

rd.set_device(0)

Existing RayD scenes, OptiX pipelines, and BVHs should not be reused across device switches in the same process.

Building Locally

RayD is a Python package with a C++/CUDA extension.

You need Python >=3.10,<3.15, CUDA Toolkit >=11.0, CMake, a C++17 compiler, drjit==1.3.1, nanobind==2.9.2, and scikit-build-core.

On Windows, use Visual Studio 2022 with Desktop C++ tools. On Linux, use GCC or Clang with C++17 support.

Recommended environment:

conda create -n myenv python=3.10 -y
conda activate myenv
python -m pip install -U pip setuptools wheel
python -m pip install cmake scikit-build-core nanobind==2.9.2
python -m pip install drjit==1.3.1

Install from the repository root:

python -m pip install .

For editable development builds:

python -m pip install --no-build-isolation -ve .

Repository Layout

Testing

python -m unittest tests.drjit.test_geometry -v
python -m unittest tests.drjit.test_visibility_topk -v
python -m unittest tests.drjit.test_reflection_epc -v
python -m unittest tests.drjit.test_reflection_accumulation -v
python -m unittest tests.test_project_metadata -v

The default development environment used by this repository is:

conda activate witwin3

Credits

RayD is developed with reference to:

Citation

@inproceedings{chen2026rfdt,
  title     = {Physically Accurate Differentiable Inverse Rendering
               for Radio Frequency Digital Twin},
  author    = {Chen, Xingyu and Zhang, Xinyu and Zheng, Kai and
               Fang, Xinmin and Li, Tzu-Mao and Lu, Chris Xiaoxuan
               and Li, Zhengxiong},
  booktitle = {Proceedings of the 32nd Annual International Conference
               on Mobile Computing and Networking (MobiCom)},
  year      = {2026},
  doi       = {10.1145/3795866.3796686},
  publisher = {ACM},
  address   = {Austin, TX, USA},
}

License

BSD 3-Clause. See LICENSE.

Key dates

PyPI data
Data sourced directly from PyPI's database.
  • Released:
Newer release available (0.8.0)

1 maintainer

PyPI data
Data sourced directly from PyPI's database.
Avatar for Asixa from gravatar.com Asixa

Credits

Author: Xingyu Chen

License

BSD-3-Clause

Requires

Python <3.15, >=3.10

Report project as malware

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

rayd_drjit-0.7.0-cp314-cp314-win_amd64.whl (26.2 MB view details)

Uploaded CPython 3.14Windows x86-64

rayd_drjit-0.7.0-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (24.6 MB view details)

Uploaded CPython 3.14manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

rayd_drjit-0.7.0-cp313-cp313-win_amd64.whl (26.3 MB view details)

Uploaded CPython 3.13Windows x86-64

rayd_drjit-0.7.0-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (24.6 MB view details)

Uploaded CPython 3.13manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

rayd_drjit-0.7.0-cp312-cp312-win_amd64.whl (26.3 MB view details)

Uploaded CPython 3.12Windows x86-64

rayd_drjit-0.7.0-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (24.6 MB view details)

Uploaded CPython 3.12manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

rayd_drjit-0.7.0-cp311-cp311-win_amd64.whl (26.3 MB view details)

Uploaded CPython 3.11Windows x86-64

rayd_drjit-0.7.0-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (24.6 MB view details)

Uploaded CPython 3.11manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

rayd_drjit-0.7.0-cp310-cp310-win_amd64.whl (26.3 MB view details)

Uploaded CPython 3.10Windows x86-64

rayd_drjit-0.7.0-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (24.6 MB view details)

Uploaded CPython 3.10manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

File details

Details for the file rayd_drjit-0.7.0-cp314-cp314-win_amd64.whl.

File metadata

  • Download URL: rayd_drjit-0.7.0-cp314-cp314-win_amd64.whl
  • Upload date:
  • Size: 26.2 MB
  • Tags: CPython 3.14, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.14

File hashes

Hashes for rayd_drjit-0.7.0-cp314-cp314-win_amd64.whl
Algorithm Hash digest
SHA256 9ff47fc9e5bd1f335c68d8aa5770f7e1ebf2750b774473a5a8964cc8e9779d41
MD5 411afae6a0f9d4e38618aad3775947c8
BLAKE2b-256 29b66aa51fa8484ccae1984bdf57997717cef22a50a173060061c1a7c713c2fd

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp314-cp314-win_amd64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file rayd_drjit-0.7.0-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for rayd_drjit-0.7.0-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 8ed3e375fef7100328245205c412f390f633dd8025eac67210a021fa8756b5ce
MD5 653b0448ca1f44f37cb0bd35f94137d8
BLAKE2b-256 941b01d890f17d6cbbcf8e9d1a5097d410d2976d41dfde6f203edab2a287ce20

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file rayd_drjit-0.7.0-cp313-cp313-win_amd64.whl.

File metadata

  • Download URL: rayd_drjit-0.7.0-cp313-cp313-win_amd64.whl
  • Upload date:
  • Size: 26.3 MB
  • Tags: CPython 3.13, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.14

File hashes

Hashes for rayd_drjit-0.7.0-cp313-cp313-win_amd64.whl
Algorithm Hash digest
SHA256 11263d277935251d5588583f9a6a770977291c9aa91a10ca2a626b5fad35d79f
MD5 a152d5c9923c33ea96f09a1e4f92b8c1
BLAKE2b-256 739e59a5967c49b53d1eb61b2cacc8a5965a8b0edd00a46dee63a60982366818

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp313-cp313-win_amd64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file rayd_drjit-0.7.0-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for rayd_drjit-0.7.0-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 4af14e450eb35931c8631835864ec4799dfa6207076292dbb9ab252968342204
MD5 154bfe64177a79fa418ca47b6c378a30
BLAKE2b-256 99a68652d9992f389f6f3d3677a1179b0915715b49bb29b1e88fcc87f35c9698

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file rayd_drjit-0.7.0-cp312-cp312-win_amd64.whl.

File metadata

  • Download URL: rayd_drjit-0.7.0-cp312-cp312-win_amd64.whl
  • Upload date:
  • Size: 26.3 MB
  • Tags: CPython 3.12, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.14

File hashes

Hashes for rayd_drjit-0.7.0-cp312-cp312-win_amd64.whl
Algorithm Hash digest
SHA256 8b8b053515a4d06a01fb6064e859d2347b89043a3fded46e2158a81a79c9ed27
MD5 4462d8318b1e37ec726959aaf527857f
BLAKE2b-256 7fc9f0615ba000a70d439541f1135867998fe750c0bb8bb8cef5f9a2010cad66

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp312-cp312-win_amd64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file rayd_drjit-0.7.0-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for rayd_drjit-0.7.0-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 eb8b66fd27d755cac94fb506e1a388bae6c16b2d6da73c179a752c2e5a779d48
MD5 4a1aa01f6b5f691363a1706b2a45ad46
BLAKE2b-256 9c93dccfad470c7588b9d62b64c992ab56a52d2757cc9ca349b66eaabc993b7e

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file rayd_drjit-0.7.0-cp311-cp311-win_amd64.whl.

File metadata

  • Download URL: rayd_drjit-0.7.0-cp311-cp311-win_amd64.whl
  • Upload date:
  • Size: 26.3 MB
  • Tags: CPython 3.11, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.14

File hashes

Hashes for rayd_drjit-0.7.0-cp311-cp311-win_amd64.whl
Algorithm Hash digest
SHA256 d07d0ccccd28336c371085dfd49a2500341b4a06acdb294630a8c3db4f2c24b4
MD5 f8c87992c62e5f2b663d912fccc1efca
BLAKE2b-256 4d5a4ec12ac43649339e5925f72e10abd23bc696944f9b99f86a3e17277e9f5c

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp311-cp311-win_amd64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file rayd_drjit-0.7.0-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for rayd_drjit-0.7.0-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 1b0aa0803ee3e428e2212c1b8ad2485a3b7ea8398c712b3c306f3152d2614a9b
MD5 0e56bfa2c1958094bff3a1b8c203efa3
BLAKE2b-256 b6427b9b704baabb7ba21f66851022533a5301d72dc8b532eed070b3b1c71b46

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file rayd_drjit-0.7.0-cp310-cp310-win_amd64.whl.

File metadata

  • Download URL: rayd_drjit-0.7.0-cp310-cp310-win_amd64.whl
  • Upload date:
  • Size: 26.3 MB
  • Tags: CPython 3.10, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.14

File hashes

Hashes for rayd_drjit-0.7.0-cp310-cp310-win_amd64.whl
Algorithm Hash digest
SHA256 a2c397547d9113031f5ce008001147788a9a42de50a9f39881292bcf638bcd28
MD5 178a00644860a086ecec887b42dccb2a
BLAKE2b-256 68db42cb2c50de9e583bc570ea3f274fcf9391faf94ae8521196d6794610a503

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp310-cp310-win_amd64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file rayd_drjit-0.7.0-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for rayd_drjit-0.7.0-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 aff51be3158c8690efb4605cdbde53a6c63cc3164796662951460231d761ecf2
MD5 192af1fae0e742c20ba9b127d3b8a43a
BLAKE2b-256 51b9199cf979a32e8ef8f7a5a67cff8d0bc0555262ffcfc7436f9dc1bf4014a3

See more details on using hashes here.

Provenance

The following attestation bundles were made for rayd_drjit-0.7.0-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl:

Publisher: pypi.yml on Asixa/RayD

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Release history Release notifications | RSS feed

0.8.0

10 files

This release

0.7.0 This release

10 files

0.6.0

10 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page