-
The multi-GPU Wilson solver is presently broken for certain communications partitions. I have tracked this to commit b62515dfb67e55dcaf7f42f655e398450f6c439d which is when the merge of the non-deg tw…
-
At present, to properly run an application built with QUDA over QMP, it's necessary to specify "-geom Px Py Pz Pt" on the command-line. This is awkward in cases where the application has built-in log…
-
This was already mentioned as part of another issue, which has since been closed. Basically, most of the necessary code modifications have already been implemented in a side branch. That code requires…
-
I had been having problems with half-precision in the staggered DD solver. I now realise the problem arose because I had switched from QDP-ordered host gauge fields to MILC-ordered gauge fields in the…
-
I'm guessing that this was introduced by Mike's comms merge. With an MPI build, running "dslash_test --partition 8" gives
```
ERROR: (CUDA) invalid resource handle (rank 0, host blast, dslash_quda.c…
-
I just noticed this problem on Blue Waters yesterday when I was testing the MPI build. staggered_dslash_test and staggered_invert_test run fine on 4 GPUs, but hang in tests involving 8 and 16 GPUs. Th…
-
Currently, MPI communicators are created on demand. This will not play nicely with peer-2-peer MPI between devices since the cost to set up the handshaking is significant. Thus all MPI communication…
-
The entire library should be enclosed in a namespace "quda". Namespaces have already been used sporadically in the library, this needs to be made uniform throughout.
-
In a recent master branch we observed the following behaviour at JLab in the clover solver:
- i) The first solve succeeds
- ii) The second solve appears to converge to the wrong answer
This was teste…
bjoo updated
11 years ago
-
Not too troublesome, since you can get it to work by using QUDA_WILSON_LINKS instead. Still, it's inconsistent, and caused me a couple of hours of pain.