2019-04-22 18:10:08
Is the windows miner fixed yet?

Fable
2019-04-22 19:27:05
what needs fixing? people have gotten comparable hashrates on windows

Rotonen
2019-04-22 19:33:30
with multiple devices on the same OS?

Fable
2019-04-22 19:33:43
it wasnt working with more than one m2 device before when I was trying it

Fable
2019-04-22 19:34:25
I tried it with 3 nvme drives on windows, hashrate was fine, just slightly lower than on linux

mjay
2019-04-22 19:36:30
z200 or z300? windows 10?

Fable
2019-04-22 19:37:16
z200?

mjay
2019-04-22 19:37:26
it was x370 and windows 8.1 I think

mjay
2019-04-22 21:53:06
just make a raid of them

Rotonen
2019-04-22 21:53:21
pooling resources over multiple disks is not in the problem scope of a miner

Rotonen
2019-04-22 21:53:39
if two instances of poolminer doesn't perform well, I would suspect a hardware bottleneck of some sort

Fireduck
2019-04-22 21:53:53
or do a local arktika cluster, if you also have multiple NUMA nodes

Rotonen
2019-04-22 21:55:00
@Fireduck users have been having weird issues on windows since the beginning and i've been curious as to what kills the hashrate for many - i'd not peg it at windows, but rather at how the disks are set up in the system

Rotonen
2019-04-22 21:57:02
There is an issue where windows prefers multiple file channels while linux doesn't care

Fireduck
2019-04-22 21:57:18
I should revisit that and make sure I am doing multiple channels on windows correctly

Fireduck
2019-04-22 21:58:11
I suspect the linux supports a direct read_at_offset while windows is needing a lock, seek, then read, then unlock

Fireduck
2019-04-22 22:02:22
that means there could be a range of hardware combinations which hit that wall before hitting a hardware limit

Rotonen
2019-04-22 22:11:43
in other news, optane continues to be meh for snowblossom
https://www.anandtech.com/print/14249/the-intel-optane-memory-h10-review-two-ssds-in-one The Intel Optane Memory H10 Review: QLC and Optane In One SSD

Rotonen