02:51:23 well decentralization is never the easiest or most practical solution to a problem 02:55:28 51% unknown? 02:55:40 total miners: 49438 03:02:40 gingeropolous: minexmr website has issues 03:04:33 ah, it dang ol dropped off the thing didnt it 03:04:48 thanks selsta 13:26:37 got my M1 Macbook pro. 13:26:51 /randomx-benchmark --mine --init 8 --jit --secure --nonces 100000 --threads 8 13:27:06 Memory initialized in 4.8494 s 13:27:16 Performance: 2320.18 hashes per second 13:30:36 any #s for power draw? 13:32:32 hmmm. 13:32:42 I have a power meter around here somewhere, haven't used it in ages 13:32:48 dunno if I canfind it 13:33:02 <\x> hyc put it on the database 13:33:09 <\x> https://xmrig.com/benchmark 13:33:42 <\x> if youre willing do a single thread run, im interested in that too, itll take time though 13:37:39 470.573H/s 1000 nonces 1 thread 13:37:56 no, in xmrig 13:38:04 ./xmrig --bench=1M -t 1 --submit 13:38:11 ah 13:38:15 haven't compiled that yet 13:38:24 911.174H 2 threads 2000 nonces 13:38:54 1254.48 3 threads 3000 nonces 13:39:15 <\x> ehemm sech1 13:39:20 1713.2 4 threads 4000 nonces 13:39:20 <\x> ./xmrig --bench=1M -t 1 --submit --user=hyc 13:39:36 <\x> for the maximum basedness 13:39:55 1972.44 5 threads 5000 nonces 13:40:09 so there are only 4 high-perf cores, so perf tails off here 13:40:11 <\x> neway hyc im gonna wait for the results on xmrig's db 13:40:45 2140.84 6 threads 6000 nonces 13:41:11 2234.52 7 threads 7000 nonces 13:41:50 2335.6 8 threads 8000 nonces 13:41:55 I'll get xmrig now 13:42:32 Interesting that hashrate increases from 6 to 8 threads 13:42:51 Apple M1 has 4 fast cores with 12 MB cache + 4 efficient cores with 4 MB cache 13:42:57 so 6 threads should be optimal 13:45:10 <\x> 4 threads on monero, 4 threads on wownero 13:45:20 <\x> whatchu think sech1? 13:45:42 maybe 13:46:00 but if you look at threads here: https://xmrig.com/benchmark/58FCwe 13:46:08 all threads have the same hashrate, weird 13:49:34 xmrig --bench wants a config.json 13:57:03 <\x> ./xmrig --bench=1M -t 1 --submit --user=hyc 13:57:06 <\x> try this 13:57:16 <\x> ./xmrig --bench=1M --submit --user=hyc 13:57:21 <\x> --user is optional 13:57:33 <\x> but itll be nice to see on the leaderboards eh 13:57:52 got it going 13:57:58 487.2H/s so far 13:58:01 5.9% 13:58:10 <\x> yeah it needs to hash 1 million :p 13:58:29 <\x> verification on xmrig's side also takes time 13:58:42 <\x> he verifies 1 thread with like 800 h/s cpu 13:58:50 basically ~same as randomx-benchmark 13:58:58 <\x> 1t bench isnt the same as nT bench on xmrig 13:59:38 <\x> sech1 did special sauce on it to prevent cheating on 1t since 1t xmrig benchmarking is like kinda competetive 14:00:13 eh, --submit doesn't work 14:00:21 <\x> extreme memory setups, some cpus even had to do it on cold, the 5800X there is on cold, -60C 14:00:23 benchmark failed: "Unknown CPU." 14:00:30 <\x> sech1 ^^ 14:02:15 this needs an update then: https://github.com/xmrig/xmrig/commit/8faef28e7da5cff809778269ca62a328aa05fed0 14:03:03 what does that mean, Apple M1 on Linux? this is running MacOSX 14:06:52 hmm. I built without hwloc before. just added hwloc, that seems to make it happier 14:07:10 always build with hwloc for benchmark 14:07:58 still 487H/s for 1T 14:09:44 <\x> damn sech1 14:09:50 <\x> xmrig benchmarks is uhh 14:09:55 <\x> i mean a lot uses it now 14:10:04 <\x> i thought itll just be a meme feature lmao 14:10:36 benchmark_XX.cmd in xmrig folder probably helps 14:11:35 <\x> it also got added on ddr4 oc guide 14:11:54 lol that's excellent 14:11:59 and makes perfect sense 14:12:17 <\x> https://github.com/integralfx/MemTestHelper/blob/master/DDR4%20OC%20Guide.md 14:12:21 <\x> ctrl f xmrig 14:12:49 I won't be satisfied until Anandtech adds it to their CPU benchmark suite 14:13:11 <\x> he benched it last time but he did it wrong 14:13:15 have people sugested that to them already? 14:13:23 how can he do it wrong 14:13:33 <\x> he didnt run as admin 14:13:42 <\x> cpus untuned too 14:13:44 meh. I'm not running as admin right now 14:13:47 doh 14:13:56 <\x> on x86 running as admin matters a lot 14:14:02 <\x> need wrmsr 14:14:06 must be a windows thing 14:14:10 <\x> nah 14:14:14 <\x> wrmsr 14:14:21 <\x> need to change a few msrs 14:14:55 <\x> i think we he ran it at first hugepages didnt kick in too 14:15:06 no hugepages on maxos 14:15:13 but IIRC M1 uses 64KB pages 14:15:51 <\x> https://www.youtube.com/watch?v=YJWnpn7p1hc 14:16:03 will find out when I start testing LMDB 14:19:11 <\x> ah yeah about that sech1 14:19:22 <\x> i think hwbot is willing to add xmrig on it now 14:19:33 <\x> i talked to them last time with my alt there 14:19:49 <\x> benchmate also willing to add it once in hwbot 14:19:55 hm. hashrate just took a dive. 269.8 14:20:45 <\x> warning, image dumps 14:20:47 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834795875368566794/unknown.png 14:20:57 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834795916728860683/unknown.png 14:21:26 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834796041663545437/unknown.png 14:21:38 <\x> neway, sech1 hwbot forum or email them 14:22:11 <\x> im not doing it man, you two (xmrig, sech1) are the devs for it, i think you two have to "plzbro" them 14:22:52 who are you, not an xmrig project guy? 14:23:00 <\x> benchmate dev, mat also seems to be willing to add it 14:23:11 <\x> hyc? me? im basically just a loiterer around here 14:23:16 ah 14:23:24 hashrate has recovered to 458 14:23:35 <\x> though im the one who shilled randomx stressing to overclockers 14:23:36 still down from where it was. no idea what changed 14:23:58 <\x> which is not actually a meme on zen2, its actually the best way to stress amd's fclk 14:23:59 Getting hot and throttling ? 14:24:15 machine doesn't feel much warmer 14:24:26 seems to have no fan. no fan noise 14:24:33 <\x> neway sech1, yeah, i wanna see the day xmrig is on hwbot 🙂 14:24:38 I remember reading it doesn't have fans and needs throttling to do so. 14:24:44 hyc: it does have a fan 14:24:49 the Pro at least 14:25:01 13" macbook pro. dead silent 14:25:32 if I put my ear on the touchpad I can hear the ocean 14:26:09 <\x> getting it on hwbot is huge 14:26:24 <\x> anandtech's cpu suite is meh compared to hwbot 14:27:00 <\x> hwbot is more for competetive oc and world records though so idk if you dont see it as hella based 14:30:25 <\x> seems somebody also already asked about it that isnt me back then sech1 14:30:35 <\x> though he asked benchmate dev to just look at it 14:30:44 <\x> benchmate is not gonna dd it if its not on hwbot first 14:30:47 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834798390964650005/unknown.png 14:31:14 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834798507209916446/unknown.png 14:35:53 heading out for a bit. bench should finish before I get back 14:40:24 <\x> sech1: benchmate kinda solves the offline benchmark issue, you can have a look on what benchmate is 14:40:36 <\x> https://benchmate.org/ 14:41:03 <\x> itll also dumb down a lot of things, like adding extra tweaks, core affinities etc 14:41:27 4:24 PM if I put my ear on the touchpad I can hear the ocean 14:41:28 the new iCean *ba dum ssst* 15:24:14 benchmark submitted 15:24:16 https://xmrig.com/benchmark/2AYxHc 15:27:07 starting -t 2 931.2H/s 15:27:21 I suppose it will throttle at some point as well 15:53:47 <\x> well if you really have the time for it then sure :p 15:53:59 <\x> i only wanted to see 1t and nT 15:54:02 there was no sign of throttling in the intermediate results 15:54:09 but it posted a final result of 743.4 15:54:21 <\x> but yeah if youre doing a brake-in its not bad :p 15:54:25 https://xmrig.com/benchmark/65pY8a 15:55:05 <\x> https://xmrig.com/benchmark/recent 15:55:30 <\x> its faster than my core2duo machine 15:55:51 <\x> https://xmrig.com/benchmark?cpu=Intel%28R%29+Core%28TM%292+Duo+CPU+P8600+%40+2.40GHz 15:56:04 <\x> dual core penryn 2.4GHz 15:56:14 that's a pretty old chip anyway 15:56:22 <\x> yeah 15:56:39 <\x> the only reason i keep it is cuz its memed 15:56:52 <\x> its the last thinkpad that can boot and run without intel ME 15:57:12 <\x> i did a few mods on it over the years 15:57:41 <\x> coreboot with no iME, nvme m.2 ssd adapted to mini pcie, newest wifi card 15:57:49 not bad 15:58:07 <\x> https://boards.4channel.org/g/thread/81119687#p81140428 15:58:15 <\x> https://i.4cdn.org/g/1618370485501.jpg 15:58:28 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834820460289064970/1618370682730.png 15:58:29 I still have my Intel quad-core dinosaur qx9300 15:58:31 <\x> nvme on pcie 1.0 x1 15:58:35 no nvme 15:58:45 <\x> sata2 is faster 15:58:51 <\x> but not in 4k random read/write 15:58:56 <\x> nvme wins that 15:59:14 yeah most times random I/O is more important 16:03:43 <\x> you can put nvme on most of the laptops of that era 16:03:48 <\x> issue is making it bootable 16:04:03 <\x> you can chainload it though like via grub or clover 16:04:17 <\x> or if coreboot is available then its easy peasy 16:04:42 it's a Dell Precision M4400 - built like a tank 16:04:54 with a beautiful WLED 1920x1200 screen 16:05:03 that's the reason I still have it 16:05:11 <\x> thats highend for that era 16:05:12 but it maxed out at 8GB RAM, so a bit painful to use 16:05:52 <\x> yeah 16:06:13 <\x> fairly slow ram too like ddr2-800 / ddr3-1066 16:06:20 yeah ddr2 16:06:24 <\x> you can raise fsb at best but needs hw modifications 16:06:51 <\x> yeah you can adapt an nvme on this one 16:06:55 <\x> looking at the board 16:07:27 mini-pci slot? 16:07:33 <\x> mini pcie 16:07:44 <\x> you need to buy a 2230/2242 m.2 ssd 16:07:53 yeah those are hardly worth the trouble 16:07:57 <\x> then an msata to m.2, its wired the same 16:08:39 <\x> if it has expresscard, then that can be used too 16:08:54 <\x> expresscard to nvme, itll sit flush on the device 16:09:12 yeah has expresscard 16:09:22 <\x> https://item.taobao.com/item.htm?id=627243558992 16:09:32 I have a USB3 card in there right now 16:09:36 <\x> youre chinese roght, youll be able to read reviews 16:09:52 <\x> https://www.aliexpress.com/item/1005002023877404.html 16:09:57 <\x> see both of them 16:11:16 [M&T3makes sense 16:11:31 <\x> expresscard is just pcie 16:11:33 <\x> heh 16:11:35 so basically this is just a connector converter 16:11:39 <\x> yep 16:11:52 <\x> after this you just need to figure out bootability 16:12:01 <\x> grub/clover will do 16:12:37 <\x> though you can always just use it as any other drive 16:12:45 <\x> maybe mount / on it 16:14:55 <\x> for upgrades, you can still get a core 2 quad extreme for it, but ehh who cares, but yeah its doable 16:15:29 right, just don't care. no hw AES 16:15:59 3 threads https://xmrig.com/benchmark/4hCkpY 16:20:39 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834826042320879626/unknown.png 16:20:47 <\x> a 2242 will fit on the wwan slot 16:20:53 <\x> needs to be adapted though 16:22:50 <\x> lel wrong image 16:22:51 <\x> here 16:22:54 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834826607549349898/unknown.png 16:42:42 hm yeah I think my wwan slot is still empty 16:44:43 https://www.brighttalk.com/webcast/663/479465 16:44:54 smart NICs to accelerate blockchain ops ? 16:46:15 4 threads 1722.8h/s https://www.brighttalk.com/webcast/663/479465 16:46:22 <\x> the adapter is uhh "msata to ngff" its wired the same 16:46:26 doh. https://xmrig.com/benchmark/6VHsmS 16:46:29 <\x> https://lewd.pics/p/O9eM.jpg 16:46:55 single slot supports both msata and nvme? 16:46:55 <\x> its like 2~3$ 16:47:04 <\x> nah 16:47:07 <\x> its like this 16:47:23 <\x> before sandy bridge intel offered pcie + usb on mini pcie slots 16:47:42 <\x> with sandy and ivy bridge intel offered msata+usb and pcie+usb on mini pcie slots 16:48:15 <\x> i dont know any of pre sandy bridge laptops that offers msata 16:50:19 <\x> https://www.ebay.com/itm/284248748727 16:51:05 <\x> you might need to remove bios whitelisting though 16:51:17 <\x> somebody likely has modded it for the m4400 16:55:06 <\x> https://lewd.pics/p/zNHw.jpg 16:58:56 <\x> you can check if it has a bios whitelist by putting any unsupported wifi card on the wwan slot 16:59:13 <\x> if it complains at boot then bios needs to be modded 17:02:50 <\x> "No wifi cards are whitelisted." 17:02:57 <\x> according to techinferno thread 17:03:38 <\x> so.. i guess youre good to go but maybe do more research if you really want to put an nvme on a c2q laptop :p 17:08:14 <\x> http://forum.notebookreview.com/threads/precision-m4400-owners-lounge-part-2.425222/page-57#post-9756067 17:11:53 heh 17:12:43 5 threads 2004.6H/s https://xmrig.com/benchmark/Y3qiM 17:25:31 <\x> hyc this is much more based though https://forum.51nb.com/thread-1997148-1-1.html 17:28:11 hyc: still no fan during all this benchmarking? 17:29:08 <\x> lol 17:29:36 <\x> the first hours of getting the machine is the time he benchmarked thread by thread damn 17:30:00 <\x> now the next meme is finding the best core with affinities 17:30:08 <\x> though thats too much lmao 17:30:29 lol 17:30:53 6 threads 2179.9h/s https://xmrig.com/benchmark/5yLsWg 17:30:54 <\x> it does matter though, but yeah this is too much 17:31:10 and yes, running benchmarks is all I do with any newly acquired machine 17:31:22 <\x> like the top 10900kf on xmrig benchmark site binned cores for it 17:31:39 <\x> since some cores are physically closer to the memory / memory controller 17:31:52 I used to mess with thread affinity for openldap benchmarking too but yeah it's a bit too tedious 17:32:11 but at the same time it can yield big payoffs 17:32:44 <\x> theres two guys that were competetively benching their 10900k's back then, theyre out of core clock, out of memory clock, the other guy basically won just by binned cores 17:32:59 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834844244962312282/unknown.png 17:33:01 <\x> 8 h/s 17:33:04 <\x> lmao 17:33:52 what's the diff between K and KF ? 17:33:59 <\x> F has no igpu 17:34:10 <\x> same die but igpou disabled 17:34:56 aha. so more thermal headroom :P 17:35:12 <\x> the 10900k guy also disabled his igpu 17:35:37 is that an equivalent disabling tho... totally power gated, or not 17:35:49 <\x> im not sure 17:37:58 <\x> both did bin cores it looks like 17:38:00 <\x> https://xmrig.com/benchmark/35F6jB 17:38:04 <\x> https://xmrig.com/benchmark/4p7YoP 17:38:13 <\x> see threads, affinity 17:39:16 <\x> https://cdn.discordapp.com/attachments/725316395428085831/834845824587399188/unknown.png 17:39:19 <\x> he binned with superpi 17:39:25 <\x> the 10900kf guy 17:39:50 <\x> core 4 is the best core for him so he affined to that 17:40:51 7 threads 2273.6 https://xmrig.com/benchmark/3waJhc 17:48:40 <\x> 8 threads done too 17:48:46 <\x> core binning time lmao 17:48:54 <\x> waste hours and hours binning cores 17:49:03 8 threads 2382.5 https://xmrig.com/benchmark/rskzp 17:49:10 yeah I don't care enough to bin individual cores 17:50:51 ooo. someone benched an Opteron 6386 17:55:45 ok I was wrong. M1 is using 16KB pages, not 64KB 17:55:54 ... moved on to LMDB testing now 17:56:06 which is boring, because it all Just Works 17:57:20 <\x> sech1: you emailed hwbot yet? 17:57:26 <\x> maaan, fucking do it plez