11:48:22
sech1:
https://gist.github.com/SChernykh/6058ecf01c929883b9d19c7eeadc8809 needs fresh tests
11:48:40
sech1:
We've added one more tweak to v2 which made it much more efficient and faster
21:26:55
tevador:
sech1: v2 looks good. Can you update #274 or open a new PR?
21:27:33
sech1:
Yes, I will open a new PR
21:27:47
sech1:
Still need to deal with ppc64be intrinsics, at least test it
21:27:54
sech1:
Maybe it works already, maybe not
21:28:32
sech1:
fallback intrinsics work fine, CI passes on s390x (which is big endian)
21:29:03
sech1:
So I'll open a new PR hopefully next week
21:30:25
tevador:
I don't think ppc64be is a blocker, as long as the portable code works on BE.
21:32:53
tevador:
Btw, any idea how to get the CI to run on this PR? https://github.com/tevador/RandomX/pull/294
21:34:55
sech1:
If you can push commits to it, just add the c-cpp.yml file (and add 'cyjan-fix-ios' to the list of branches in the begginning of the file)
21:35:27
sech1:
btw I don't think that PR is necessary
21:35:41
sech1:
CI runs on M1 instances too and it passes tests
21:36:27
sech1:
294 is only useful for weird compilers (non-clang). clang supports __builtin___clear_cache intrinsic
21:37:28
sech1:
If ppc64be is not a blocker, then I'll make a PR as soon as I get more hashrate tests for design_v2.md.
21:45:18
tevador:
I think #294 is for building on iOS.
21:47:56
tevador:
We have an open issue for failing build on iOS: https://github.com/tevador/RandomX/issues/153
22:04:50
sech1:
CI doesn't have an iOS build though
22:05:00
sech1:
I don't know how to test it without an iOS device
22:35:43
DataHoarder:
You don't build on iOS but for iOS afaik
22:35:58
DataHoarder:
Via xcode(?)
23:11:25
trampakoulas:
is this the best command "xmrig --cpu-max-threads-hint=20 --randomx-no-numa " for lenovo server with 4 cpu and 386gb ram
23:15:47
DataHoarder:
#monero-mining