Errors encountered while compiling Chrono's dem_module

35 views
Skip to first unread message

Zero

unread,
Jan 8, 2026, 6:00:45 AM (4 days ago) Jan 8
to ProjectChrono
1>F:\GitHub\chrono\src\chrono_dem/cuda/ChDem_SMC.cuh(237): error : namespace "cub" has no member "Inequality"
1>      Block_Discontinuity(temp_storage_disc).FlagHeads(head_flags, SDsTouched, cub::Inequality());
1>                                                                                    ^
1>
1>1 error detected in the compilation of "F:/GitHub/chrono/src/chrono_dem/cuda/ChDem_SMC_trimesh.cu".
1>ChDem_SMC_trimesh.cu

The above errors appeared when compiling dem_module.

"In Chrono::DEM's dem_module there are still uses of the old CUB APIs (CUDA 10/11), such as cub::Inequality. After CUDA 12/13, CUB was merged into the CUDA C++ Core Libraries (CCCL) and many tag-type comparators were removed, e.g. cub::Inequality, cub::Equality, cub::Less, cub::Greater, etc."

"I'm using CUDA( version 13.1).Should this part be updated?"

Radu Serban

unread,
Jan 9, 2026, 5:37:32 AM (3 days ago) Jan 9
to ProjectChrono

Indeed, the GPU-based Chrono modules are not updated to work with CUDA 13 yet.  Such an update will have to include not only Chrono::DEM, but also Chrono::FSI_SPH and Chrono::Sensor. Furthermore, we may want to provide some level of backward compatibility, as not all users are able to update CUDA at any time.

 

For now, you will have to use an older version of CUDA. Version 12.9 works fine; the cub functions you mention are still present in that CUDA release (albeit marked as deprecated).

 

--Radu

--
You received this message because you are subscribed to the Google Groups "ProjectChrono" group.
To unsubscribe from this group and stop receiving emails from it, send an email to projectchron...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/projectchrono/9106777c-fb96-4865-9bc7-d46498339ed6n%40googlegroups.com.

Zero

unread,
Jan 11, 2026, 7:55:08 AM (19 hours ago) Jan 11
to ProjectChrono
Hi,dear

Thank you for your suggestion.I'll try CUDA version 12.9.

Reply all
Reply to author
Forward
0 new messages