With docker allowed to take 4 cores of my i7-4770, it runs at a little over 7 Sol/s. If I boot into ubuntu and run with 8 threads I get around 11 IIRC.
How exactly did you do that? My docker container only reaches .8sol/s, and it only sees one processor for come reason (echo $(nproc) == 1)
Also, I tried a configuration that’s really similar to yours and my container starts validating all of the transactions every time it starts. Yours should do the same. Where you able to solve this?
I tried to use persistent data, but it doesn’t seem to work on my machine because of an issue between the db system zcashd uses and the virtualbox based docker volumes (it should work all right with the native docker, which doesn’t run on my machine, unfortunately, because of the processor not supporting Apple HyperVisor technology, as I understand it).
Well, I was using native docker. So all the persistent data under .zcash was stored in the host file system (see the -v flag to the docker run command above) and every time I start the container it can pick up where it left off. I suppose native docker probably also explains the CPU count difference. .8 sounds pretty low though, are you using the tromp solver?
Yes and I’m getting 4sol/s on a virtualbox vm (without docker). As I understand, you’re not using any additional flag to configure docker to use all of the processors, correct?
Native docker has a little menu bar thing that lets you configure how much RAM/CPU to give to docker (doesn’t appear ot be a hard reservation, just a limit). I adjusted that to let docker use 4 cores.