09:14:53 There is specialized hardware for CN/R, I'm sure now 09:15:39 Sumokoin nonce distribution chart: https://i.imgur.com/C9eFmD6.png 09:15:58 They stayed on CN/R, red line marks when Monero forked to RandomX 09:25:05 hm? the nonce distribution looks more uniform after the fork 09:25:55 well - one obvious division in the nonces, but the two halves have no obvious streaks 09:26:45 what happened between blocks 270000 and 280000? 09:26:55 it's not any of existing GPU miner software for sure 09:27:11 270-280k - it forked to CN/R 09:27:55 neither any proxies produce such nonce pattern, so remnant botnets mining through proxies are also out of the equation 09:29:59 so you're suggesting that CNR ASICs left Monero and jumped onto Sumo after the randomX hardfork? 09:30:23 yes 09:30:33 probably FPGAs 09:30:46 https://whattomine.com/coins/196-sumo-cryptonightr?utf8=%E2%9C%93&hr=2400&p=200.0&fee=0.0&cost=0.1&hcost=0.0&commit=Calculate 09:31:00 Vega GPU spends 2 times more on power than it earns from mining 09:31:43 got a hashrate chart for them around that time? 09:32:05 hmm 09:32:21 I can add it to the picture 09:32:59 I have full "get_block_headers_range" json RPC output from sumo daemon 09:36:31 that 2:1 cost is pretty nasty. So are Sumo miners complaining? 09:37:28 moneromooo: so scratch my suggestion to use CN/R for wallet KDF 09:42:53 with difficulty chart added: https://i.imgur.com/mtSBS1p.png 09:47:13 any explanation of the pattern between 275000 and 355000? 09:48:32 this is normal CN/R pattern 09:48:40 GPU miners split nonces that way 09:48:50 depending on the amount of GPUs in a rig 09:49:03 hence horizontal lines 09:53:10 I'm guessing nothing was visible in the Monero nonce distribution? 09:54:00 it looks like the FPGAs are only ~30 MH/s in total 09:54:17 just looked at top 2 sumo pools. Each of them has ~10 MH/s miner 09:54:28 so 2 individual miners account for 70% of Sumo hashrate 09:54:58 Monero nonce distribution looked fine, at least on this scale of FPGA mining 09:55:05 too subtle to notice 09:56:16 maybe it's even a single miner who spread hashrate between 2 pools 10:11:11 why did difficulty drop so sharply at CN/R introduction? were CN ASIC miners on there before too? 10:13:10 they were on cn/heavy before 10:13:22 so I guess cn/heavy also had some specialized hardware 10:13:27 ah 10:16:45 so many CN tweaks have been made - and all of them ASIC'd 10:17:18 that's what you get with static algorithms ... 10:17:47 personally I'd prefer to see more uniform nonce distribution, and less of these streaks 10:19:48 what's xmrig's nonce pattern look like? also streaky? 10:26:59 This is Monero's nonces: https://i.imgur.com/qiZoooA.png 10:27:10 last 2 forks are shown 10:27:47 RandomX pattern is dominated by proxies 10:27:54 so 256 horizontal lines 10:28:03 and a bit thicker line at the bottom - individual workers 10:51:15 so it sounds to me like ASICs are doing the right thing - starting at randomly selected nonces 10:51:31 and GPU miners are being too regimented 10:52:39 I wonder if that means they're using a true hardware RNG 11:00:35 "they" = asics in that last sentence, hyc ? 11:47:20 https://twitter.com/Inno_Miner/status/1217830543963770880 11:47:31 another Grin ASIC cancelled 12:19:45 Inge-: yes 12:22:05 tevador what do they mean "GPU ASIC" ? 12:22:22 so basically they're saying their chosen foundry partner was incompetent? 12:22:29 TSMC I presume 12:23:32 or their design was too exotic for the foundry 12:24:08 I guess they call it a GPU ASIC since Grin is mined by GPUs 12:24:23 otherwise it has nothing to do with GPUs 12:28:22 you would think the foundry guys would assess up front whether they can handle the design or not 12:38:20 in any case, RandomX ASIC looks even less likely to be designed 12:43:11 it amazes me after all these years of bitcoin asics screwing over people, that there were enough suckers to buy into another asic preorder 12:55:37 well, they're all getting full refunds, so I wouldn't call 'em suckers unless they don't get their money back 12:57:15 I wonder who the actual foundry is, since TSMC clearly have no problem producing GPUs and CPUs, high complexity stuff 12:58:09 I thought that mining ASIC is even easier to design than CPU/GPU. 13:54:45 ravencoin are realizing that tweaks don't work: https://www.reddit.com/r/Ravencoin/comments/epkqpe/looks_loke_we_hit_a_peak_in_difficulty_again/ 13:58:58 "Go to the ravencoin discord, ask admins to join algo room, and be part of the discussion regarding next algo" 14:24:52 sech1, can you share code that transform nonce.txt into image? 14:26:29 I can promise that will not use it for profit. 14:27:00 http://paste.debian.net/1126527/ 14:27:49 and the code that creates txt files from RPC json data: http://paste.debian.net/1126528/ 14:28:04 first one is for gnuplot 17:31:38 sweet, that's what I came to ask for! 21:38:37 cohcho: pool PR is great 21:38:41 thanks 22:23:09 total miners 28564 22:53:49 I've just got what they change in xmrig 22:54:03 (they == botnets) 22:57:08 b'donate.v2.xmrig.com' replaced with b'127.0.0.1\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00'