CONQUEST-release
CONQUEST-release copied to clipboard
OrderN
Full public release of large scale and linear scaling DFT code CONQUEST
* De-duplicating code in exx_kernel_default.f90 * Adding omp threading to `m_kern_exx_eri(_gto)` * Adding reframe run-only test for exx benchmarks
The CI doesn't test this build option. Should it?
# Description In the branch [f-exx-opt](https://github.com/OrderN/CONQUEST-release/tree/f-exx-opt) the Conquest_out results appear to be dependent on the compiler/libraries and the number of MPI ranks you use. # Compiler and library versions and...
Based on our meeting with @lionelalexandre on April 12, we need to investigate how much we can reduce code duplication in the subroutines https://github.com/OrderN/CONQUEST-release/blob/c00a071c18a257d4e708abaede7bb4769b80ccbd/src/exx_kernel_default.f90#L921 https://github.com/OrderN/CONQUEST-release/blob/c00a071c18a257d4e708abaede7bb4769b80ccbd/src/exx_kernel_default.f90#L1268 https://github.com/OrderN/CONQUEST-release/blob/c00a071c18a257d4e708abaede7bb4769b80ccbd/src/exx_kernel_default.f90#L1649 https://github.com/OrderN/CONQUEST-release/blob/c00a071c18a257d4e708abaede7bb4769b80ccbd/src/exx_kernel_default.f90#L2019 and do some...
There's possibly a bug in the MPI communication which appears when running on one process. Collecting hints in this issue In `test_004` of `f-exx-opt` we notice a difference in the...
The weak scaling test in [benchmarks](https://github.com/OrderN/CONQUEST-release/tree/develop/benchmarks/matrix_multiply) can now be run automatically (tested on Archer2), using ReFrame and Spack. The spack package is [upstreamed to spack](https://packages.spack.io/package.html?name=conquest), the ReFrame test is currently...
The backtracking line minimiser applies an Armijo criterion to decide whether a change in energy is acceptable. Under some circumstances this can lead to a failure to perform a line...
Cosma is a HPC system in the University of Durham https://www.durham.ac.uk/departments/academic/physics/cosma/ It has a wide range of profiling tools available due to the performance analysis course ran by Tobias Weinzierl
Fixes #265
In https://github.com/OrderN/CONQUEST-release/blob/f424a27d322011e6c131293ba204a87bc50e0472/src/fft_fftw3.f90#L56-L58 in fftx_exec_wrapper there is a loop over columns. It should be possible to thread it trivially. If @davidbowler can provide a test case where this becomes a bottleneck,...