17:56:06 <hyc> thinking again about a cluster of cheap cores for a mining ASIC
17:56:36 <hyc> a bunch of cheap ARM Cortex-A53s still don't have enough on-chip cache to compete with a Ryzen
17:57:15 <hyc> that would be where to focus and innovate. putting a bunch of cores onto a chip, that don't need to coordinate with each other
17:57:21 <hyc> as tightly as a standard multiprocessor chip
17:57:46 <hyc> they can share read-only access to one chunk of pool RAM
17:58:01 <hyc> and can each have their own private caches for scratchpad
17:58:30 <hyc> not having to implement inter-core cache sharing protocols should simplify their work
17:58:53 <hyc> it would basically be a bunch of single-core CPUs sharing a chip package, as opposed to a multi-core CPU chip
17:59:49 <hyc> they could have a few shared memory ports for command/control, to start/stop computation, pass params. wouldn't need much else.
19:09:34 <derpy_bridge> <[discord] Kayla#5718>: and a shared nvme for pxeboot UwU
21:25:25 <tevador> hyc: that would most likely require an ARM architectural license, which is currently held only by Applied Micro, Broadcom, Cavium, Huawei, Nvidia, AMD, Qualcomm, Samsung and Apple
22:03:50 <needbrrrrrrr90> https://www.cryptunit.com/newsid/157-top-coins-to-mine-with-cpu-randomx-april-2020
22:04:00 <needbrrrrrrr90> didn't realize there were that many coins that had adopted rx
22:08:32 <jwinterm> but only one coin adopted it first
22:09:00 <jwinterm> I thought there was gonna be 157 coins
22:09:13 <needbrrrrrrr90> you didnt adopt randomx first tho
22:09:16 <needbrrrrrrr90> you adopted randomwow
22:09:22 <jwinterm> tru
22:10:42 <tevador> all apart from Monero are laughably insecure
22:10:59 <jwinterm> tru
22:10:59 <tevador> I mean, 200 KH/s?
22:11:28 <jwinterm> even monero is not exactly immutable at this level
22:11:40 <jwinterm> 1200 Mh/s is maybe a million CPUs?
22:12:29 <jwinterm> BredoLab is by far the largest recorded botnet to date, as it combined the resources of over 30 million computers around the world.
22:13:02 <jwinterm> and supercomputers and aws and asymptotically's secret stash of server racks
22:14:16 <tevador> 30M computer botnet sounds powerful on paper, but why aren't they mining monero then? are there more profitable things to do? I doubt it
22:15:23 <jwinterm> I've heard that profitable isn't necessarily easy to define
22:15:59 <jwinterm> I think mining is an easy way to make people realize their computer is messed up (CPU fans spinning all the time)
22:16:32 <tevador> or the botnet is really stable enough for mining or the hardware is not up to the task
22:16:34 <jwinterm> so maybe would be more profitable for a week, but longer term if you decimate your botnet you can't do other botnet stuff like ddos or whatever
22:19:38 <selsta> tevador: do you know if any AV software adapted randomx-sniffer?
22:20:49 <tevador> no idea, but anecdotal evidence says probably not
22:29:56 <tevador> I wonder why they removed Monero from here: https://www.crypto51.app/
22:34:22 <jwinterm> randomx not on nicehash?
22:34:32 <jwinterm> the whole thing probably runs on nicehash api it looks like
22:35:38 <tevador> it is on nicehash AFAIK
22:36:06 <tevador> https://www.nicehash.com/algorithm/randomxmonero 47 MH/s
22:36:23 <jwinterm> yea, just got there
22:38:17 <asymptotically> jwinterm: you can balance a lot of compute power on the rafters of your garage :D
22:39:11 <jwinterm> https://github.com/tdickman/crypto51/issues/55
23:24:56 <needbrrrrrrr90> Monero was removed because we solved 51% attacks
23:25:08 <needbrrrrrrr90> Duh
23:25:48 <jwinterm> dash has an issue on there basically saying that
23:25:53 <jwinterm> decred more or less as well