There is a bug found by @penguin on v0.14.2. It seems to affect all CPU miners.
I've compiled a Window binary with this patch and ran it for 16 hours, there is no improvement in terms of hashing power, but it did find a block for me.
I posted the Binary on Github if anyone is interested.
If I understand the bug, you wouldn't see a change in reported hash power, but the unpatched version is overstating its effective hashpower.
With N threads running (on N cores), the effective hash power of the unpatched version would be (1/N) times the reported number. With the patched version, we're actually getting the reported hash power.
Thank you.!
If I understand the bug, you wouldn't see a change in reported hash power, but the unpatched version is overstating its effective hashpower.
With N threads running (on N cores), the effective hash power of the unpatched version would be (1/N) times the reported number. With the patched version, we're actually getting the reported hash power.