Skip to content

111 batch running features gpu runner#142

Draft
jmaack24 wants to merge 52 commits into
developfrom
111-batch-running-features---gpu-runner
Draft

111 batch running features gpu runner#142
jmaack24 wants to merge 52 commits into
developfrom
111-batch-running-features---gpu-runner

Conversation

@jmaack24
Copy link
Copy Markdown
Member

@jmaack24 jmaack24 commented May 15, 2026

Implement/do the following for the OptixRunner:

  • Improve performance of SolTraceSystem::get_buffer_results
  • Add user specified batch size controlling number of rays traced per optixLaunch
  • Add logic to estimate batch size to fit on available device memory

jmaack24 added 30 commits March 27, 2026 16:13
@jmaack24 jmaack24 linked an issue May 15, 2026 that may be closed by this pull request
3 tasks
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

@coveralls
Copy link
Copy Markdown

coveralls commented May 19, 2026

Coverage Report for CI Build 26186442341

Coverage decreased (-0.02%) to 86.244%

Details

  • Coverage decreased (-0.02%) from the base build.
  • Patch coverage: 6 uncovered changes across 3 files (29 of 35 lines covered, 82.86%).
  • 4 coverage regressions across 1 file.

Uncovered Changes

File Changed Covered %
coretrace/simulation_runner/native_runner/native_runner.hpp 4 0 0.0%
coretrace/simulation_runner/embree_runner/embree_runner.cpp 1 0 0.0%
coretrace/simulation_runner/native_runner/native_runner.cpp 2 1 50.0%

Coverage Regressions

4 previously-covered lines in 1 file lost coverage.

File Lines Losing Coverage Coverage
coretrace/simulation_runner/embree_runner/trace_embree.cpp 4 92.73%

Coverage Stats

Coverage Status
Relevant Lines: 7553
Covered Lines: 6514
Line Coverage: 86.24%
Coverage Strength: 2961185.24 hits per line

💛 - Coveralls

@jmaack24 jmaack24 requested a review from Copilot May 19, 2026 15:14
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 38 out of 38 changed files in this pull request and generated 8 comments.

Comment thread coretrace/simulation_runner/optix_runner/OptixCSP/src/core/soltrace_system.cpp Outdated
Comment on lines +156 to +158
std::vector<HitRecord> m_hit_records;

// Global ray index (ray_offset + local_index) for each logical hit ray in m_hit_records.
Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added error check in OptixRunner::setup_parameters to ensure that ray ids are less than uint32_t max value.

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Turns out this is needed for some of the round robin testing. Working on a fix.

Comment thread coretrace/CMakeLists.txt Outdated
Comment thread coretrace/simulation_runner/embree_runner/ftz_daz.hpp Outdated
Comment thread coretrace/simdriver/main.cpp Outdated
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 38 out of 38 changed files in this pull request and generated 4 comments.

Comment thread coretrace/simulation_runner/optix_runner/OptixCSP/src/core/soltrace_system.cpp Outdated
Comment on lines +19 to +30
# define SOLTRACE_SET_FTZ_DAZ() do { \
uint64_t _fpcr = _ReadStatusReg(ARM64_FPCR); \
_fpcr |= (1ULL << 24); \
_WriteStatusReg(ARM64_FPCR, _fpcr); \
} while(0)
# else
/* GCC / Clang on ARM64 */
# define SOLTRACE_SET_FTZ_DAZ() do { \
uint64_t _fpcr; \
__asm__ __volatile__("mrs %0, fpcr" : "=r"(_fpcr)); \
_fpcr |= (1ULL << 24); \
__asm__ __volatile__("msr fpcr, %0" : : "r"(_fpcr)); \
Comment thread coretrace/simulation_runner/simulation_runner.hpp
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Batch running features - GPU Runner

3 participants