quda
quda copied to clipboard
lattice
QUDA is a library for performing calculations in lattice QCD on GPUs.
It seems like bicgstab does not have a parameter `maxResIncrease` like the CG so if one uses reliable updates with an unreachable tolerance it will iterate until it hits max...
When I try to create a two-flavour twisted clover (non-degnerate) operator, it seems that this is no longer supported. I thought this was implemented during the Juelich hackathon?
We have been experimenting with using complex b[s] and c[s] which appears to give similarly good approximation to Shamir compared to Mobius with only real coefficient. Supporting zMobius would either...
On systems such as Summit, while there is peer-to-peer access across the node, the performance between the hemispheres is significantly less than within the hemisphere. And moreover, improved throughput can...
We need to add multi-rhs support to the rewritten dslash kernels. This should be done for all dslash kernels and likely use the composite framework as was done with the...
This variable isn't needed anymore with the dslash kernel rewrite using accessors. We can nuke it once the old staggered dslash kernels are removed.
Describe building QUDA with cmake, getting cmake if not available and currently available options in cmake. Particularly MPI detection.
A major optimization that can be explored is to see if amortizing the gauge field loads by acting on multiple spinors simultaneously will lead a large speed up. This is...
Look into providing imported targets and use of cmake for building QMP / QIO with the QUDA_DOWNLOAD_USQCD. This is a follow-up for #995