Speaking of VMs, I was wondering how the performance of VMs was. During the Pentathlon I had used VMs to sandbag. Afterward, I ran a few Rosetta tasks in a Vmware Player VM (host: Win10, guest: Win10 on an E5-2680), one in an ESXi VM (guest: Win10 on an E5-2670) and ran a few on Win10 on both the E5-2680 and i7-4960X. CPU times seemed to be all about the same. It's hard to tell with the variance of tasks. I had let the VMs run out and just ran physical during the event. I was a little worried because the VMs in VMware Player would say 1000 ops/s in the BOINC benchmarks, where the others would say like 13000+.
I had a few hundred queued up to release, and they are just sitting in the transfer pile.
Not too much lost time for me, and I generally don't get too worked up over hiccups in projects. I crunched through thousands of bad MilkyWay WUs, to help clear their servers when they had a bad round of tasks a few years ago. lol
Na, these should provide high GPU util as long as there is a CPU to feed it. No need for any kind of swan_sync environment variable any more. I think there's still a 24hr bonus so def don't want to run over that to get a bit higher util. These are much smaller than the Long tasks GPUGrid used to have.
Still NV/Cuda only and there has been no CPU work for awhile.
A forum community dedicated to overclocking enthusiasts and testing the limits of computing. Come join the discussion about computing, builds, collections, displays, models, styles, scales, specifications, reviews, accessories, classifieds, and more!