-
With the advent of unified memory address spaces, there seems to be little point in keeping around separate cpu and cuda fields, since a cpu field can be mapped into the device memory address space. …
-
I believe MILC supports this, but Chroma does not (at least the interface doesn't). We need to support this in Chroma as well. Guochun, what do you do for MILC that allows for this: is it just a cas…
-
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
-
As it stands, this combination triggers an error. After PR #873 gets merged, this can be triggered in `release/1.0.x` as of August 9, 2019 by making the following modifications to `staggered_eigensolv…
-
When the inversion source has support on only even sites, convergence seems to be OK with a tiny heavy-quark residual. But when a source has support on only odd sites, the residual decreases at a gl…
-
(Bug report and suggested fix from Joel Giedt)
Below is what I did with the ordering. The issue is that the array of 4d spinors is
a full parity spinor, with the even parity part in the first half o…
-
The gauge fixing with FFTs, computeGaugeFixingFFTQuda(...), does not support multi-GPUs.
The multi-GPU gauge fixing support is only available wwith the gauge fixing with overrelaxation, computeGaugeFi…
-
While not directly required for deletion of the legacy dslash kernels, we should fully document the new QUDA kernel framework, with the appropriate wiki pages. For example, we should have pages for:
…
-
@Marcogarofalo @pittlerf @urbach @robfre21
I guess this is one of the things that we need to work on to provide a QUDA implementation of the BSM3 operator. The tests and sanity tests of the CPU ve…
-
Program hangs without error after output file prints "MAKING PATH TABLES" with QUDA_ENABLE_TUNING set to 0. When I enable tuning and set the cache file location with QUDA_RESOURCE_PATH, the program ha…