-
Dear hypre team,
I am using BoomerAMG via PETSc as a preconditioner for the PETSc GMRES solver. When using the ParILUK as a smoother, I am noticing a steady increase in memory usage and eventuall…
-
As discussed yesterday with @tjhei we (@jdannberg, @RanpengLi and myself) are investigating a problem that looks like a MPI dead lock inside the GMG preconditioner. We are still working to break it d…
-
**Using**
PC OS: Ubuntu 22.04
PC ROS Distro: Humble
Raspberry Pi OS: Ubuntu 22.04
Raspberry Pi ROS Distro: Humble
Create® 3 firmware version: H.1.2
**Expected behaviour**
Follow the [tutoria…
-
**Describe the bug**
When you factorize a GPU matrix, the output is on a CPU. I believe
[https://github.com/JuliaGPU/CUDA.jl/blob/6d2e9ab2f2a1bb1df7791bb30178e0fe956940a3/lib/cusolver/linalg.jl#L4…
-
MRtrix3 can take a long time to build. I often like to use a Windows laptop (i5 10th generation quad-core processor) to test changes on MSYS2 and compile times are in the 20-25-minute ballpark. The si…
-
While reviewing the code for the nystrom_kernel_svd in Eigenpro3 `beta = torch.from_numpy(kmat).diag().max()` and top_eigensystem `beta = n * (scaled_kmat.diag() - (E[:,:q].pow(2)*(L[:q]-L[q])).sum(-1…
-
In most examples of MFEM, the homogeneous dirichlet boundary is used. But I want to set nonhomogeneous Dirichlet boundary in a time-dependent solver. I hope to know what I should do. And is there any …
-
I've added broken tags but systest indicates this is passing and enforced in CI in both verilated models and on the fpga
-
At the moment we can only have one paradiag object, because of how we handle appctx (and maybe other reasons, if so please tell me about them!).
In Firedrake's NonlinearVariationalSolver, the appctx …
-
This is no longer dependent on "Hermes3 in 2D in NESO". Plan is to get Dave Moxey contributing directly to NESO.
- reuse as much code between 2D and 3D