Mirrored_Repos/DarkflameServer
1
0
Fork 0
mirror of https://github.com/DarkflameUniverse/DarkflameServer.git synced 2026-10-22 12:42:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(ugc): ray_backend=hiprt traces on the GPU with HIPRT and Orochi (optional build)

Browse Source 
HIPRT (MIT) behind the CMake option DLU_HIPRT (off): its headers come from its
SDK (HIPRT_ROOT, else ROCm's /opt/rocm) and are copied next to the servers; its
library is loaded when first used (hiprtew), as HIP or CUDA are by Orochi
(MIT, fetched pinned by hash; CUDA when its toolkit is found). The trace
kernels (nearest hit skipping the triangle a ray leaves, any hit) are compiled
the first time and kept in cache/hiprt. One GPU context for the process
(hiprt_device picks the GPU); the workers take turns on it. When HIPRT, the
GPU or a scene's upload fails, Embree is used instead, and the UGC server logs
why at start.

For a GPU the rays go in batches (UgcRays::Scene gets batch queries; the CPU
backends answer them a ray at a time):
- hidden faces: with a batch backend the paths are traced side by side, a
  bounce at a time (the path code split into Start, Scatter and Bounce, the one
  by one tracing unchanged); the same paths with the same random numbers, so
  the same triangles are decided (tested with builtin side by side)
- the occlusion bake and the denoised icons' traced occlusion always ask in
  batches (the same rays, the same results)

Its symbols are hidden: the servers export theirs (-rdynamic), and HIPRT's
library, which has an Orochi of its own, would otherwise call ours.

Check: configure with -DDLU_HIPRT=ON on a machine with ROCm (or HIPRT's SDK)
and an AMD RDNA or NVIDIA GPU; UgcServer --make-model x.lxfml out hiprt; the
UGC tests (hits, hidden faces and occlusion against builtin); a build without
it leaves everything as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Aaron Kimbrell
2026-09-29 11:57:37 -05:00
parent 2e2e8153e2
commit 07dbf30dc8
13 changed files with 719 additions and 92 deletions
Download Patch File Download Diff File Expand all files Collapse all files

31
thirdparty/CMakeLists.txt vendored
View File

@@ -208,3 +208,34 @@ if(DLU_OIDN)
endif()
message(STATUS "Open Image Denoise ${OpenImageDenoise_VERSION}: the UGC server can denoise icons")
endif()
# HIPRT (MIT), the UGC server's ray_backend=hiprt on the GPU: optional, off by default. Needs HIPRT's SDK (its headers;
# HIPRT_ROOT, else ROCm's /opt/rocm), whose library is loaded at run time (hiprtew), as HIP or CUDA are by Orochi (MIT,
# fetched here; CUDA too when its toolkit is found). The headers are copied next to the servers: the GPU kernels are
# compiled from them when the UGC server first uses the GPU.
option(DLU_HIPRT "Build the UGC server with HIPRT (the ray_backend=hiprt setting, on the GPU)" OFF)
if(DLU_HIPRT)
find_path(DLU_HIPRT_INCLUDE_DIR hiprt/hiprtew.h PATHS "$ENV{HIPRT_ROOT}/include" "${HIPRT_ROOT}/include" "/opt/rocm/include" REQUIRED)
FetchContent_Declare(orochi
URL https://github.com/GPUOpen-LibrariesAndSDKs/Orochi/archive/dbaf54bb61007d7ce87802b9458df54eeb574fe1.tar.gz
URL_HASH SHA256=13d5b946aceadc07b09633e5cb2917433209f562a596e76181325d04cd0d4173
DOWNLOAD_EXTRACT_TIMESTAMP TRUE)
FetchContent_MakeAvailable(orochi) # no CMakeLists.txt of its own: only downloaded
add_library(dlu_orochi STATIC "${orochi_SOURCE_DIR}/Orochi/Orochi.cpp" "${orochi_SOURCE_DIR}/contrib/hipew/src/hipew.cpp")
target_include_directories(dlu_orochi PUBLIC "${orochi_SOURCE_DIR}")
target_link_libraries(dlu_orochi PUBLIC ${CMAKE_DL_LIBS})
# The servers export their symbols (-rdynamic): HIPRT's library has an Orochi of its own, whose names ours mustn't take
set_target_properties(dlu_orochi PROPERTIES CXX_VISIBILITY_PRESET hidden VISIBILITY_INLINES_HIDDEN ON)
find_package(CUDAToolkit QUIET)
if(CUDAToolkit_FOUND)
target_sources(dlu_orochi PRIVATE "${orochi_SOURCE_DIR}/contrib/cuew/src/cuew.cpp")
target_compile_definitions(dlu_orochi PUBLIC OROCHI_ENABLE_CUEW)
target_include_directories(dlu_orochi PUBLIC ${CUDAToolkit_INCLUDE_DIRS})
endif()
file(COPY "${DLU_HIPRT_INCLUDE_DIR}/hiprt" DESTINATION "${CMAKE_BINARY_DIR}/hiprt/include")
if(CUDAToolkit_FOUND)
message(STATUS "HIPRT headers in ${DLU_HIPRT_INCLUDE_DIR}: the UGC server can trace rays on AMD (HIP) and NVIDIA (CUDA) GPUs")
else()
message(STATUS "HIPRT headers in ${DLU_HIPRT_INCLUDE_DIR}: the UGC server can trace rays on AMD GPUs (HIP; no CUDA toolkit found)")
endif()
endif()