Skip to content

Navigation Menu

Appearance settings

View all features
- BY COMPANY SIZE
  Enterprises
  Small and medium teams
  Startups
  Nonprofits
- BY USE CASE
  App Modernization
  DevSecOps
  DevOps
  CI/CD
  View all use cases
- BY INDUSTRY
  Healthcare
  Financial services
  Manufacturing
  Government
  View all industries
View all solutions
- EXPLORE BY TOPIC
  AI
  Software Development
  DevOps
  Security
  View all topics
- EXPLORE BY TYPE
  Customer stories
  Events & webinars
  Ebooks & reports
  Business insights
  GitHub Skills
- SUPPORT & SERVICES
  Documentation
  Customer support
  Community forum
  Trust center
  Partners
View all resources
- COMMUNITY
  GitHub SponsorsFund open source developers
- PROGRAMS
  Security Lab
  Maintainer Community
  Accelerator
  GitHub Stars
  Archive Program
- REPOSITORIES
  Topics
  Trending
  Collections
- ENTERPRISE SOLUTIONS
  Enterprise platformAI-powered developer platform
- AVAILABLE ADD-ONS
  GitHub Advanced SecurityEnterprise-grade security features
  Copilot for BusinessEnterprise-grade AI features
  Premium SupportEnterprise-grade 24/7 support
Pricing

Search code, repositories, users, issues, pull requests...

Search

Clear

Search syntax tips

Provide feedback

We read every piece of feedback, and take your input very seriously.

Include my email address so I can be contacted

Saved searches

Use saved searches to filter your results more quickly

Name

Query

To see all available qualifiers, see our documentation.

Appearance settings

You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session.

Dismiss alert

alicevision / AliceVision Public

Notifications You must be signed in to change notification settings
Fork 874
Star 3.4k

Code
Issues 16
Pull requests 35
Discussions
Actions
Projects
Wiki
Security and quality
Insights

Additional navigation options

Code
Issues
Pull requests
Discussions
Actions
Projects
Wiki
Security and quality
Insights

Dev/port cuda to sycl#2077

Open

SFN-eu wants to merge 63 commits intoalicevision:developalicevision/AliceVision:developfrom

SFN-eu:dev/port-cuda-to-syclSFN-eu/AliceVision:dev/port-cuda-to-syclCopy head branch name to clipboard

Conversation Commits63 (63)Checks Files changed

Open

Dev/port cuda to sycl#2077
SFN-eu wants to merge 63 commits intoalicevision:developfrom
SFN-eu:dev/port-cuda-to-sycl

Commits

Commits on Apr 22, 2026

Add AdaptiveCpp to src/CMakeLists.txt dependency checks
sfn
committed
Split off directory for SYCL port and add necessary cmake plumbing
sfn
committed
Port computeOnMultiGPUs (CUDA) to computeOnMultiDevices (SYCL)
sfn
committed
Replace CUDA's float2 and float3 with sycl::vec
sfn
committed
Port cuda/host/memory.hpp to sycl/memory.hpp
sfn
committed
Port Refine.hpp and Sgm.hpp to SYCL (but not their respective .cpp files)
sfn
committed
Port DeviceMipmapImage.hpp and DeviceCache.hpp to SYCL (but not their respective .cpp files)
sfn
committed
Remove default sycl::event constructors to make queue dependencies explicit
sfn
committed
Port DepthMapEstimator to SYCL (both .cpp and .hpp)
sfn
committed
Various bugfixes
sfn
committed
Replace explicit sycl::vec<float, n> syntax with sycl::floatn shorthand
sfn
committed
Port depthMapUtils to SYCL (both .cpp and .hpp)
sfn
committed
Port NormalMapEstimator to SYCL (both .cpp and .hpp) and fix bugs in sycl/memory.hpp while at it
sfn
committed
Port deviceDepthSimilarityMap.hpp to SYCL (but not cuda host code or kernels)
sfn
committed
Port Refine to SYCL (both .cpp and .hpp)
sfn
committed
Port deviceSimilarityVolume.hpp to SYCL (but not cuda host code or kernels)
sfn
committed
Port Sgm to SYCL (both .cpp and .hpp)
sfn
committed
Port volumeIO to SYCL (both .cpp and .hpp)
sfn
committed
Cleanup: remove unnecessary cmake define, drop requirement of in-order queues, be more precise in the device aspects we require
sfn
committed
Add files and CMake plumbing to port deviceSimilarityVolume Kernels
sfn
committed
Port utility headers (buffer, color, SimStat, PatchPattern, CameraParams) and adjust existing progress to match
sfn
committed
Rework how CameraParams and mipmap image sampling works
sfn
committed
Port custom PatchPattern as a reference singleton
sfn
committed
Implement lookup, nearest sampling and linear sampling for SyclDeviceMemoryPitched in memory.hpp
sfn
committed
Replace INFINITY macro with the more correct std::numeric_limits<float>::infinity()
sfn
committed
Create minimal shell "accessor" objects for copying over to device, with full trilinear interpolation
sfn
committed
Rework memory layout of multidimensional arrays to allow easy use of memcpy and memset on Y slices of volumes
sfn
committed
Pass queues by reference
sfn
committed
Port deviceSimilarityVolume.cpp, including all device code and interface with rest of project
sfn
committed
Port depthMapEstimantion and depthMapFiltering to be compatible with SYCL as well as CUDA
sfn
committed
Various bugfixes 2: debug statement boogaloo
sfn
committed
Poort DeviceCache.cpp to SYCL
sfn
committed
Port DeviceMipmapImage.cpp to SYCL
sfn
committed
Finish device implementation of DeviceMipmapImage, along fixing some
sfn
committed
Fix mismatched symbols between deviceSimilarityVolume.cpp and deviceSimilarityVolue.hpp
sfn
committed
Misc formatting cleanup and bug fixes
sfn
committed
Rework construction, storage and destruction of Sgm and Refine objects to reduce (device) memory usage and reallocation
sfn
committed
Swap device cache from using std::map (ordered red-black tree, logarithmic time) into using std::unordered_map (hashmap, ammortized linear time)
sfn
committed
Consolidate all multi-dimensional array coordinate to address calculation in sycl/buffer.hpp
sfn
committed
Fix bug in DepthMapEstimator.cpp that could result in dereferencing a non-existant compute object if a tile was skipped
sfn
committed
Port deviceDepthSimilarityMap to SYCL
sfn
committed
Fix miscelanous memory access and allocation issues
sfn
committed
Collection of minor correctness and performance improvements (with a big impact)
sfn
committed
General cleanup while tracking down an issue (which will be fixed in the next commit)
sfn
committed
Fix mipmap images
sfn
committed
General cleanup
sfn
committed
Performance improvements
sfn
committed
Correctness fixes. The port now produces equivalent results to the original CUDA implementation, and further changes should be checked against regressions.
sfn
committed
Fix compile error after rebasing
sfn
committed
Fix CUDA build
sfn
committed
Correctness fixes
sfn
committed
Rework build and library system to allow building cuda and sycl backend support at the same time
sfn
committed
Fix build to conform to upstream changes in depthmap
sfn
committed
Fix minor glitches due to having previous versions of aliceVision installed elsewhere on my system
sfn
committed
Properly break out of switch statement when selecting backend
sfn
committed
Performance: seperate kernels for computation with custom patch patterns to speed up the more likely case of a default patch
sfn
committed
Performance: optimize MipmapImageAccess struct
sfn
committed
Refactor DepthMapEstimator to simplify logic with the use of omp async tasks
sfn
committed
Performance (minor): replace loop with std::fill
sfn
committed
Performance: condense memory access of DeviceMipmapImage
sfn
committed
Performance: minor tweaks
sfn
committed
Fix: ensure DeviceCache is threadsafe by using locks to protect modification of containers
sfn
committed
Generalize per-device logic embedded in DeviceCache to it's own PerDevice class
sfn
committed

Footer

© 2026 GitHub, Inc.

Footer navigation

Terms
Privacy
Security
Status
Community
Docs
Contact

You can’t perform that action at this time.