16:01:19 I suppose other pool software could use this rpc too, also for memory savings if they needed it 16:31:27 the glory of opensource, if they want to implement it, its all there for them to do so :) 16:32:50 was looking at that arduino coin that can be mined on nanos etc. Then realised you could just GPU mine it, but they move you onto an extreme difficulty net 16:33:21 then I looked a bit deeper, and the dev's control everything, its more centralised than the earth's core. 17:38:05 "then I looked a bit deeper..." <- Nice 18:18:18 yeah, the dev's could just turn around and say "Hey, you're mining too quick, thats against the rules, closing your account and seizing your coins" 20:06:31 someone went bonkers on RandomX memory: https://github.com/GiverofMemory/RandomX#randomxb 22:10:42 wow. we're totally repeating 2014 22:11:21 that was pebllecoin, it took cryptonight and made the memory requirement like 8gigs or something 22:32:17 sounds like a good way to waste a lot of perfectly good hardware. like chia :P 22:32:43 inb4 people claim randomx has only lasted this long because of the global chip shortage 22:33:59 oh please, will take a lot more than readily available silicon to take out RandomX :P 22:34:41 Those Russian ARM boxes might be a nice way to mine, but it'll be years before we can get our hands on them 22:41:34 Fun that the light mode is as much as the normal mode of RandomX 22:42:37 These are small numbers, 1 TiB RAM or bust 22:45:26 Their whole premise is wrong, no GPUs on the market can mine RandomX profitably. 22:46:46 technically speaking, no tech on the market can mine RandomX profitably unless you get lucky as fek solo mining 22:47:14 ROI's are really long - especially with the chip shortage throwing the investment costs through the roof 22:47:41 * pauliouk says that, while the Ryzen 5 3600 build next to him is about 8 months from ROI. 22:53:35 if you already have the hardware, you can mine profitably with a CPU 22:54:07 GPU mining RandomX makes no sense even with stolen electricity 22:56:20 true. I doubt you'd even be able to mine enough to cover the wear and tear of the GPU 22:57:18 the reference to "AI type graphics cards" was particularly laughable, since AI workloads are optimized for 8 bit ints 23:05:41 And small floats 23:06:19 tbh some CPUs on EPYC line are getting close to GPU level performance while being a CPU on their own 23:06:35 Insane L3, and also number of cores 23:07:20 I remember the joke of “oh I can boot Linux on this 12 MiB cache” but what do we have now, half a GiB of L3 cachev 23:07:51 :P 23:08:32 aaaah to have a shipping contained filled with pre-build server boards, loaded with DDR5 memory and MilanX chips... 23:08:38 It is becoming similar sized to GPU memory, with complex NUMA designs, and all kinds of hardware accelerations 23:09:18 So yeah just do raytracing might be about as quick :’) 23:10:18 will I finally be able to run games from this millennia? :D 23:13:12 the on-chip caches certainly make coldboot easier. we used to have to write bootloaders that fit in 256 bytes of cache