byeo@hermes:/mnt/nvme0n1/byeo/projects/traccc/traccc_build_7911e66ca08e8e7c65732eb2f53652ca5edf0f81$ compute-sanitizer ./bin/traccc_test_cuda --gtest_filter=*Kalman*Telescope*/0*
========= COMPUTE-SANITIZER
Note: Google Test filter = *Kalman*Telescope*/0*
[==========] Running 1 test from 1 test suite.
[----------] Global test environment set-up.
[----------] 1 test from CUDAKalmanFitTelescopeValidation/KalmanFittingTelescopeTests
[ RUN ] CUDAKalmanFitTelescopeValidation/KalmanFittingTelescopeTests.Run/0
WARNING: No entries in volume finder
Detector check: OK
Using hit time
========= Invalid __local__ read of size 4 bytes
========= at void traccc::cuda::kernels::fit<traccc::kalman_fitter<detray::rk_stepper<covfie::field_view<covfie::backend::constant<covfie::vector::vector_d<float, (unsigned long)3>, covfie::vector::vector_d<float, (unsigned long)3>>>, algebra::plugin::array<float>, detray::constrained_step<float>, detray::stepper_rk_policy<float>, detray::stepping::void_inspector>, detray::navigator<const detray::detector<detray::default_metadata<algebra::plugin::array<float>>, detray::container_types<vecmem::device_vector, vecmem::jagged_device_vector, detray::dmap>>, (unsigned long)10, detray::navigation::void_inspector, detray::intersection2D<detray::surface_descriptor<detray::detail::typed_index<detray::default_metadata<algebra::plugin::array<float>>::mask_ids, unsigned int, unsigned int, (unsigned int)4026531840, (unsigned int)268435455>, detray::detail::typed_index<detray::default_metadata<algebra::plugin::array<float>>::material_ids, unsigned int, unsigned int, (unsigned int)4026531840, (unsigned int)268435455>, unsigned int, unsigned short>, algebra::plugin::array<float>, (bool)0>>>>(T1::config_type, traccc::device::fit_payload<T1>)+0x247e0
========= by thread (0,0,0) in block (0,0,0)
========= Address 0xffedb6 is misaligned
========= Saved host backtrace up to driver entry point at kernel launch time
========= Host Frame: [0x2ef36f]
========= in /usr/lib/x86_64-linux-gnu/libcuda.so.1
========= Host Frame: [0x15803]
========= in /usr/local/cuda-12.4/targets/x86_64-linux/lib/libcudart.so.12
========= Host Frame:cudaLaunchKernel [0x75230]
========= in /usr/local/cuda-12.4/targets/x86_64-linux/lib/libcudart.so.12
========= Host Frame:_ZNK6traccc4cuda17fitting_algorithmINS_13kalman_fitterIN6detray10rk_stepperIN6covfie10field_viewINS5_7backend8constantINS5_6vector8vector_dIfLm3EEESB_EEEEN7algebra6plugin5arrayIfEENS3_16constrained_stepIfEENS3_17stepper_rk_policyIfEENS3_8stepping14void_inspectorEEENS3_9navigatorIKNS3_8detectorINS3_16default_metadataISH_EENS
The fitting failure is Initially reported by @tsulaiav and I found that it started failing since commit 7911e66(or #950)
@niermann999 Could you take a look as detray v0.95.0 is contributed by you?
We also need to run
compute-sanitizerin the CI to avoid the same situation