I have updated monero-project/monero #6111/files at last, randomx functional tests. If anyone has ideas for other special cases, let me know.

getting this on master and release: paste.centos.org/view/raw/2489e9fc

gots my blockchain erkns!

and then this

im trying to solo mine from xmrig to daemon

gingeropolous: (i think that...) it's trying to allocate huge pages and can't

you can increase vm.nr_hugepages sysctl, or make it stop spamming with `set_log -stacktrace`. there is an environment variable to tell it to not try allocating with huge pages, but it didn't do anything for me :(

MONERO_RANDOMX_UMASK=1

well changing the user of the task to root didn't fix things

ah, ok. xmrig knows how to do it itself, but the daemon doesn't ... ?

if you run xmrig as root i think it'll try and enable them for you

it should be safe to ignore the messages in your daemon, at least i did until my log file got to be a few gigabytes in size lol

Did you monkey with the log settings or is there a bug in the rotation ?

i didn't monkey with log settings

That was for asymptotically.

(beyond set_log -stacktrace)

It should not grow beyond ~100 MB.

i think i only used --log-file and --log-level, then i made the system logrotate sort it out

ah. sorry. well, increasing huge pages to 1024 didn't do anything

OK, maybe --log-file disables rotation.

gingeropolous: you might need more. for 1 dataset and 24 mining threads i need 1193

Does anyone have a testnet node I can use for IBD? Can't even get sync to start, guess the seed nodes are down/overwhelmed

s/specify/use

fort3hlulz: sorry my seednode was down for maintenance, I've just brought it back up

just give it a bit to catch up

fort3hlulz: Height: 1487376/1487376 (100.0%) on testnet

should be fine now

yup, just started syncing a bit ago!

Thanks.

Is stagenet used at all? I'm setting up seeding nodes for both just to give a bit more backbone to those networks.

Stagenet is used, yes.

sweet, thanks

Is something special with keccak_256 as implemented in the Monero code? I have luigi's JavaScript implementation on xmr.llcoins.net/addresstests.html calculating different hashes than a C# implementation I grabed somewhere which gives the same as this here: emn178.github.io/online-tools/keccak_256.html

It also gives a different value for the empty string: c5d2460186f7233c927e7db2dcc703c0e500b653ca82273b7bfad8045d85a470 instead of c5d2460186f7233c927e7db2dcc703c0e500b653ca82273b7bfad8045d85a470

(according to en.wikipedia.org/w/index.php?title=…amples_of_SHA-3_and_Keccak_variants)

Monero has one of the original pre-standard submissions. NIST changed a couple things IIRC, padding and... something else.

So "Keccak" really means "one of the submissions as amended".

Yeah, I tried to play with the padding, to no avail. So maybe it's "something else" :)

yeah it's something else

I saw where Luigi got its code, maybe a file compare will lead me into the right direction

There have been occasional ideas to move to standard SHA3, but we'd still need to haul around the old keccak code anyway

shrug

Anyway, if somebody has a pointer laying around somewhere which version it is exactly, I would be grateful.

src/crypto/keccak.c :P

Try monero.stackexchange.com.

Alright, thanks. Maybe I will land there after an attempt on my own.

Hmmm, that comment in that keccak.c file sound promising: "A baseline Keccak (3rd round) implementation"

it's just padding I think

yeah just padding changes

your two empty string messages are the same too...

my code gives the same values as emn178

I think it's the same code anyway so that would make sense

var KECCAK_PADDING = [1, 256, 65536, 16777216];

var PADDING = [6, 1536, 393216, 100663296];

padding is just x6 looks like