07:06:52 <Inge-> Threadripper PRO 3995WX 8 channel memory, 64 cores
12:31:29 <hyc> threadripper pro sounds like just an EPyc server chip
13:36:50 <Inge-> Pretty much
13:36:55 <Inge-> but faster base clock I think
13:37:06 <Inge-> so could be some improvements over the Zen2 Epyc
13:37:22 <Inge-> Also no MSRP. Only an OEM part - so you need to go to Lenovo to get a workstatino currently
16:47:18 <gethh> hyc you're right it is exactly 7702P with 280W TDP
16:47:59 <sech1> https://xmr.nanopool.org/account/45aeGq9Pn1gHCQnWsJG7EhQWTvydsUtDJ9XHY8H5QHuXWy7FZhdFGqpPbtZiE6cg5oP9Avhkz8MzKdTPav8ojXVjDDHPMiB
16:48:02 <sech1> 350 MH/s
16:48:19 <sech1> https://xmr.nanopool.org/account/84HxQcbxuhtWS2x9as1TtzNKxXMoQY2jQh8mDEd2xFxhGWS74mdqqWuZf9JuSKLRmnbc7VpuxKqB1jRKe734ahpF6tdLwdk
16:48:22 <sech1> 97 MH/s
17:36:01 <gethh> insane
17:37:17 <tevador> 350 MH/s is what we estimated for a top10 supercomputer
17:37:38 <tevador> top5 even
17:37:52 <sech1> those 2 are probably the same entity
17:37:57 <sech1> so 450 MH/s
17:38:26 <scoobybejesus> nearly a third of network hashrate then
17:39:39 <tevador> that's suspiciously close to the estimate for Archer2
17:44:47 <sech1> same nonce patterns, average worker speed 1400 h/s
17:51:10 <jwinterm> aws instances or something?
17:53:05 <gethh> no fucking clue, but could be
17:53:42 <gethh> ramp up is very fast usually, so it's not botnet 100%
17:55:16 <tevador> 320k workers -> ~10k AMD Epyc CPUs, 2 cores per worker
17:59:52 <gethh> you numbers are right sir
18:02:46 <gethh> acquisition cost  ~90mln $
18:02:49 <jwinterm> why would they split workers on same cpu tho?
18:13:26 <tevador> https://www.nextplatform.com/2019/10/18/amd-cpus-will-power-uks-next-generation-archer2-supercomputer/
18:13:32 <tevador> maybe?
18:15:25 <tevador> either that or Amazon/Google/MS have started mining; this is too big to be hijacked VMs or free credits
18:41:31 <gethh> or state actor fucking around :)
19:15:01 <hyc> Amazon has certainly ramped up AMD in their own infrastructure
19:19:44 <hyc> hmm. can't find the reference I was thinking of
19:20:07 <hyc> but AMD's secure memory encryption is probably winning a lot of business right now
19:39:08 <sech1> At least Github is doing something https://i.imgur.com/qpAkhIR.png
19:40:55 <hyc> interesting. what was moneroocean doing that tripped their mining check?
19:42:46 <gethh> having wrong name
19:42:47 <gethh> hehe
19:45:07 <sech1> They were building their fork of XMRig using Github CI
20:09:50 <asymptotically> github ci blocks all connections to moneroocean but it seems like they didn't bother blocking any other pools
20:14:16 <xmrpow> I rly dont get why they arent mining at their own daemon... If these guys can have 400k per month  they could easily hire sb to configure it properly.
20:16:44 <xmrpow> Everybody can simply track their activities.