Ios - #401
Open
Leonhardmaster2 wants to merge 9 commits into
Open
Conversation
…s, and implement missing examples - Add complete Doxygen comments to hydrology headers `drainage_basin.hpp` and `drainage_basin_cell_based.hpp`. - Correct references and links to examples and output images across library headers (e.g. `statistics.hpp`, `transform.hpp`, `synthesis.hpp`, and `string_utils.hpp`). - Generate and implement 21 missing example folders and source files matching header Doxygen `@include` tags. - Fix API compilation issues in example files (updated `hmap::Texture` to `hmap::Tensor`, correct `hmap::Array` constructors, and added `init_opencl` where required). - Optimize the `non_parametric_sampling` example settings to speed up compilation validation and execution. - Run all examples and copy generated images to `docs/images/` for clean Doxygen output compilation.
The MSVC branch passed /Od via add_compile_options, which lands after the per-config flags on the cl command line. MSVC takes the last optimization flag, so /Od overrode Release's /O2 (cl warns D9025) and every Windows build shipped the entire library compiled without optimization. Drop /Od (each config's default now applies: /O2 Release, /Od Debug) and add /fp:fast outside Debug to match the -ffast-math the GCC/Clang builds already use. Measured on a Windows 11 machine (RTX 5060, 6 cores), Hesiod headless batch of the island bootstrap at 4096x4096: 47s before, 17.4s after (2.7x), matching equivalent-hardware Linux times.
…ptimization fix(cmake): stop disabling MSVC optimizations in Release builds
Bumps external/CLWrapper to include recent changes: - Removed runtime kernel argument validation step. - Renamed benchmark and unit test targets to prevent build conflicts. - Disabled unit tests by default.
… kernel loading Refactors `init_opencl` in `gpu_opencl.cpp` to: - Use native `clwrapper::KernelManager` API (`clear_sources`, `set_build_options`, `add_kernel`, and `build_program`) instead of string accumulation. - Introduce a local lambda helper `add(...)` to clean up kernel loading readability and reduce boilerplate around each `#include`.
- fix(cmake): match AppleClang so macOS gets -O3/-ffast-math (STREQUAL 'Clang' silently skipped Apple's compiler, leaving macOS builds unoptimized; same bug class as the MSVC /Od fix in 7237678); guard optimization flags per-config and drop deprecated -Ofast - perf(cmake): add -mcpu=apple-m1 tuning for arm64 macOS - fix(cmake): OpenMP detection on macOS via Homebrew libomp hints; OpenMP becomes optional on macOS (still required elsewhere), conditional OpenMP::OpenMP_CXX link, multi-token flag splitting - fix(cmake): detect keg-only opencl-headers/opencl-clhpp-headers and rebind stale Homebrew assimp libz.tbd path to the active SDK zlib - perf(threading): distributed_tile_loop now uses dynamic work-pull scheduling (atomic task counter) instead of static strided partitioning; new recommended_thread_count() sizes the pool after Apple Silicon performance cores (hw.perflevel0.physicalcpu) - fix(virtual_array): guard RamTileStorage lazy tile creation with a mutex (data race under VA_DISTRIBUTED) and make clone() lock-aware - perf(openmp): init_openmp default changed from hardcoded 8 threads to omp_get_num_procs() - docs/readme: macOS build instructions; ci: add macOS arm64 job Measured on Apple M3 (2048^2 CPU kernels): 13.3x total speedup (hydraulic_diffusion 42x, gradient_norm 21x, noise_fbm 6.6x); ex_virtual_array 15/15 runtime checks pass, OpenMP 5.1 active, OpenCL Apple M3 GPU verified.
Brings the INOUT/create_imagef/get_imagef Run API required by the GPU-resident flow_simulation rewrite.
Member
|
Hi, thanks for the PR! A couple of requests:
Thanks! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.