09:46:18 <mechanic41turk[m> <sech1> "oh, there's even "release-static..." <- So, do I just do, `make -D NO_AES=ON` ?
09:46:54 <mechanic41turk[m> I see inside the Makefile, there is the release-static-linux-armv8
09:46:54 <mechanic41turk[m> https://github.com/monero-project/monero/blob/master/Makefile#L133
10:08:00 <mechanic41turk[m> or do I add `-D NO_AES=ON` to the line 135 which seems to list all the specific options targeted for the linux-armv8 building.
10:15:54 <sech1> line 135, in the cmake command
10:16:19 <mechanic41turk[m> OK. I added. And started the make command
10:16:26 <mechanic41turk[m> do you know how many hours is it going to take?
10:16:56 <sech1> no, I use gitian build on my PC
10:16:58 <sech1> much faster
10:17:16 <mechanic41turk[m> alright
10:17:50 <mechanic41turk[m> I will probably have to CTRL-c if it takes an hour or more, since I gotta take the laptop with me out and thus severe the ssh connection.
10:18:09 <mechanic41turk[m> anyways of shutting down the ssh connection without ctrl-c 'ing the build process?
10:27:09 <moneromooo> ^Z and bg might, though it the TTY dies, it might take it with it too.
10:27:19 <moneromooo> But you can ^C and run again with nohup maybe ?
10:27:27 <moneromooo> Or use screen.
10:29:41 <mechanic41turk[m> I ctrl-c 'ed
10:29:52 <mechanic41turk[m> I will try compiling it overnight
10:32:26 <pauliouk> hmm, I've done a p2pool and xmrig build on an android tv box running manjaro linux - didn't take too long, and worked without error
10:32:51 <pauliouk> I then did a benchmark on xmrig, looks like I'm the only person to have benchmarked a Cortex A55
10:35:32 <sech1> monerod build takes much longer
10:39:15 <pauliouk> 8gb mem card, wouldn't have risked building it :P
10:40:36 <pauliouk> https://xmrig.com/benchmark?cpu=ARM+Cortex-A55
10:40:44 <pauliouk> poor thing :)
10:45:02 <pauliouk> I will say one thing for it though, its been running nearly 24hrs, and its only just 'warm' - not worth the power its running on, but might come in handy as some under desk heating in my office when winter sets in
11:06:22 <pauliouk> Hmm I'll say another thing for it... it might only be working at 100H/s but the android box found a share last night.
12:29:24 <hyc> mechanic41turk[m: you should have built with PR#8001 / #8005
12:29:37 <hyc> then you wouldn't need to mess with cmake flags at all
12:31:24 <hyc> pauliouk: I think most of the cheap tvbox makers are still only using cortex-A53
12:31:57 <hyc> but now you see the point - even a lowly tvbox can productively mine.
12:32:29 <pauliouk> yup the bulk seem to be A53, I went with this one as it was cheap enough for me not to worry if I bricked it, and has an A55. However, seems the skipped the optional L3 cache :/
12:32:49 <hyc> all of the cheap builds do
12:33:47 <hyc> I think most of them also only have 32bit wide memory bus
12:34:13 <hyc> but despite the lame memory interface, still usable for mining
12:34:22 <pauliouk> the worst thing with this manufacture is they seem to alternate other components in their builds. wifi chips, memory etc
12:34:37 <hyc> which box do you have?
12:34:54 <pauliouk> X96 Air
12:35:46 <hyc> I've been looking for some newer ones to replace my current crop
12:35:57 <pauliouk> still surprised I'm the only person to bench mark an A55 on xmrig :P
12:36:09 <hyc> I have 10 here that I used to use for cryptonight mining
12:36:35 <hyc> but they only have 2GB RAM and I don't want to go thru the trouble of adapting randomx to fit in that
12:37:21 <hyc> the nice thing is they all ran on 5V/1A power supplies so I just bought a 10 port USB power supply to power them all
12:37:42 <hyc> with USB-to-barrel connector cables
12:38:54 <pauliouk> my plan if it had a decent enough hash rate, was to strip it out of the case, solder on the power connectors, and hook them up to a 5v 20A power supply, in a case with a 120MM fan front and back
12:39:22 <hyc> I just have them stacked on each other with bottle caps for spacers
12:39:27 <hyc> totally silent/passive
12:39:57 <pauliouk> thats one way :)
12:40:42 <hyc> so, 100H/s is not bad for 5W. or is it 10W?
12:41:55 <pauliouk> I imagine somewhere between the two, closer to 5W most of the time
12:42:44 <hyc> time for me to shop for new boxes... 
12:42:56 <pauliouk> if I had a batch of 20 or so of them, then it'd be worth the setup, but since I only have 2 (ones 32bit only) then, yeah not worth priming them from Amazon :)
12:43:32 <hyc> as I recall, I bought 10 from either GlobalSources or Aliexpress
12:43:42 <hyc> that was 5 years ago, don't really remember
12:44:08 <hyc> can't get a lower price, but shipping is slow
12:44:47 <hyc> direct from manufacturer. they will even custom print the case logos for you
12:45:01 <hyc> I got them plain, but this time around I might go for Monero logos
12:46:25 <pauliouk> hmm, the X96Air cost £32.99, so like $40 
12:46:57 <pauliouk> for the price of 10 of them, I'm closer to buying another Ryzen 3600 rig 
12:47:40 <pauliouk> 7KH/s vrs 1KH/s :P
12:48:00 <nioc> 5W for 100H/s ?  my 3900X system gets 102H/s per W
12:48:03 <pauliouk> but, if I can pickup some used ones for like £10 a piece, then a nice stack of them would be worth it
12:50:04 <hyc> yeah, Ryzen prob still wins perf/watt
12:51:06 <pauliouk> but at the end of it, you only have 1 Ryzen PC... at the end of using android TV boxes, you can have 20 nice little mini-pc's hooked up to every TV
12:52:09 <hyc> and you can get that box for less than $2 each https://www.aliexpress.com/item/1005003293371783.html
12:53:20 <hyc> rk3318 is only quad cortex-a53
12:53:55 <hyc> but it would work well enough. can't beat the price.
12:54:17 <pauliouk> was just about to mention that, although not the A55, it'd probably hash the same rate with memory and no l3 cache
12:57:59 <pauliouk> for people who buy Raspberry Pi's or other SBC's who don't use the GPIO etc, a cheap android box would probably be much better for them
12:58:10 <hyc> definitely
12:58:22 <hyc> most people buying pis never make good use of them 
12:58:36 <hyc> just like most people buying pi3 or pi4 don't evevn know they're only running 32bit OS
12:59:38 <pauliouk> I bought the 5 or so pi zero's I have in my draw because they were cheaper than buying the PIC microcontrollers and support components for some of the projects I was working on
13:01:40 <pauliouk> now a PCB that I could hook up say... 50 Pi compute modules to, and distribute mining to them, that _might_ be worth the setup
13:02:10 <hyc> p2pool is pretty greaat for that
13:02:47 <hyc> there's a 49 port USB2.0 hub out there
13:02:55 <pauliouk> but would have to go very large scale to make it worth while. Then we're talking another Ryzen 3600 build money :P
13:02:56 <hyc> would provide power & comms
13:03:22 <hyc> I had an idea to commission a mining-dedicated "tv stick"
13:03:30 <hyc> just plug them in to usb port
13:04:57 <pauliouk> I've got a box of those little bitcoin asic's in the draw next to the pi's
13:05:11 <hyc> the little usb ones?
13:05:17 <pauliouk> they were ideal, just give them .55A and away they go
13:06:00 <pauliouk> hmm damn, I bought a lot of them :o
13:06:06 <pauliouk> ANT USB1
13:06:27 <hyc> ah yep
13:07:32 <pauliouk> I have 14 of them for some reason
13:08:11 <hyc> you must have a pair of 7port hubs somewhere
13:09:42 <pauliouk> uww worse
13:11:53 <pauliouk> so I've hooked up 4 powered 5 port usb hubs to a 10A 5V power supply
13:12:24 <pauliouk> caseless - nice and safe :)
13:15:37 <hyc> lol
13:16:34 <pauliouk> also remembered that these things get toasty
13:17:15 <hyc> yeah. I suppose tvboxes running flat out on randomx will heat up too
13:17:41 <gingeropolous> that would be cool tho to bring back the usb stick miner for monero
13:17:49 <pauliouk> wonder what'd happen if I hooked them all up to a decent power supply, shared with an android box, 3d printed a nice case with a fan, and sold it on ebay as a bitcoin mining farm :D
13:18:03 <hyc> lol
13:18:31 <hyc> onero gingeropolous: you want to build m
13:18:43 <hyc> doh
13:18:54 <hyc> you want to build monero ASICs? (cat on keyboard)
13:19:46 <hyc> mine would still have been usable as HDMI tvsticks afterward
13:19:56 <gingeropolous> of course i do
13:19:59 <pauliouk> 1.6GH each, so thats a good 20GH/s bitcoin mining rig :P
13:21:14 <hyc> I've still got email contacts for the companies I talked to in China back in 2016
13:21:25 <hyc> if you really want to do it
13:21:33 <gingeropolous> .hrmmmm
13:21:38 <pauliouk> RandomX is ASIC resistant. However, if a mobile manufacturer decided to build a mobile logic board with 10x snapdragons on it, with enough memory, then it could be considered a monero asic
13:21:48 <pauliouk> only, it wouldnt be an asic
13:21:54 <hyc> yes, exactly
13:21:57 <gingeropolous> yep
13:22:20 <gingeropolous> i mean, the btc usb sticks were such a great product, marketing wise
13:22:40 <hyc> might be difficult to get a project off the ground right now, with chip shortages
13:22:57 <gingeropolous> tho it'd be kinda hilarious to plug in a 100 h/s usb stick into a PC that can do kh/s or more
13:22:58 <pauliouk> could save a fortune on the board as it wouldn't need all the trash that gets used on a phone, although might want to include the 5G so you could mine anywhere with a signal
13:24:00 <hyc> yeah, without on-chip cache not sure there's any point to building it
13:24:13 <pauliouk> or if a CPU manufacturer decided to bring out a 128core CPU, with 4MB of L3 cache per core, then... yeah
13:24:24 <hyc> it has to have at least a hash/watt advantage
13:25:05 <hyc> too bad we can't get our hands on a phytium FT-2000+
13:25:25 <gingeropolous> i've always thought its better for monero community itself to launch a monero "asic", open source, such that it creates a scorched earth effect. i.e., we shrink the profit margins of any competition. but really a fools errand competing with intel and amd prolly
13:25:47 <hyc> probably
13:26:02 <hyc> their economy of scale alone will kill that
13:26:18 <hyc> still... I wonder if we shouldn't be looking at RISC-V already
13:26:31 <pauliouk> https://www.cnx-software.com/2019/10/18/arm-microatx-motherboard-phytium-ft2000-4-arm-desktop-soc/
13:27:55 <pauliouk> I guess this is why Monero will last so long though, it can only really be mined on a standard gaming PC
13:28:18 <pauliouk> or on data centre hardware in xvb's case :P
13:30:07 <hyc> I was thinking of their 64core server chip https://semiaccurate.com/2016/10/04/phytium-shows-off-64-core-arm-server-systems/
13:30:23 <pauliouk> but, if you could get your hands on the logic boards from say 50 small android phones, you could probably find a way to rig them up into an ant-miner case, and get a good 200KH/s out of it
13:30:36 <hyc> they're still the only company I know designing chips with enough cache per core
13:30:48 <mechanic41turk[m> <hyc> "mechanic41turk: you should..." <- how do I do that?
13:31:05 <mechanic41turk[m> I am quite the newbie, I would appreciate some more descriptive advice.
13:31:06 <pauliouk> if they're the only company doing it, then they've got no one to compete with on price :/
13:31:38 <hyc> mechanic41turk[m: did you ue git to grab the monero source tree originally?
13:31:50 <mechanic41turk[m> hyc: yes.
13:32:01 <hyc> pauliouk: also the server chip is not for sale outside of China
13:32:39 <mechanic41turk[m> I have been following the instructions here: https://github.com/monero-project/monero/#cloning-the-repository
13:32:51 <hyc> great
13:32:53 <pauliouk> bah, thats a scary thought. If they're not allowed to mine BTC any more, are they going to move onto XMR and just start throwing up warehouses full of those?
13:33:17 <gingeropolous> "its just a datacenter, see? they are CPUs"
13:33:35 <mechanic41turk[m> gingeropolous: "it's cloud computing"
13:33:40 <hyc> mechanic41turk[m: to get the pull request onto your machine, go into the monero source directory\
13:33:54 <DataHoarder> setup a generic task-based system gingeropolous, then just keep sending randomx tasks :)
13:33:58 <pauliouk> "just ignore the fact our server admins are wearing dirty vests and cargo shorts with holes in them"
13:34:08 <hyc> then type "git fetch origin pull/8001/head:pr8001"
13:34:21 <gingeropolous> well, thats the beauty of permissionlessness ,  yah can't stop the signal
13:34:38 <mechanic41turk[m> hyc: so I do this within ~/monero/ directory, right?
13:34:38 <hyc> then "git cherry-pick pr8001"
13:34:49 <hyc> in the monero source directory, yes
13:35:08 <mechanic41turk[m> okay.
13:35:22 <hyc> after the cherry-pick, just rerun make
13:35:38 <hyc> with no special flags
13:36:11 <mechanic41turk[m> hyc: OK. If I do these, I don't have to add `-D NO_AES=ON` inside the Makefile, right?
13:36:17 <hyc> right
13:36:24 <mechanic41turk[m> Alright.
13:36:32 <mechanic41turk[m> Thanks for that. I will report back in 6 hours.
13:36:50 <hyc> shouldn't take 6 hours to rebuild from only changing 1 file
13:37:11 <mechanic41turk[m> hyc: 6 hours later I can be at home and try those.
13:37:17 <mechanic41turk[m> Raspberry pi is sitting at home currently.
13:37:19 <hyc> lol ok
13:37:24 <mechanic41turk[m> haha
13:38:35 <hyc> pauliouk: phytium is the only *ARM* mfr with enough cache per core. obviously AMD is doing that with ryzen
13:38:56 <gingeropolous> tho, with china banning bitcoin mining, monero mining taking root there is a reasonable scenario. that could get interesting if the same market shenanigans carry over
13:39:33 <hyc> that sounds unlikely. China has banned all crypto
13:40:07 <hyc> tho most ryzen chips only have enough cache for half their cores. hmmm.
13:40:22 <mechanic41turk[m> hyc: chinese govt isn't as omnipotent as to ban some random cpu processes, right?
13:40:31 <DataHoarder> they are :)
13:40:43 <mechanic41turk[m> I prefer to believe otherwise lol
13:40:45 <mechanic41turk[m> what a dystopia
13:40:54 <DataHoarder> bitcoin mining is also just a few random hash processes
13:41:04 <hyc> they don't ban CPU processes, but they will go after excessive elctricity users
13:41:06 <mechanic41turk[m> DataHoarder: *with awfully specialized machines.
13:41:10 <DataHoarder> I would prefer to believe otherwise but reality is reality
13:41:18 <DataHoarder> also GPUs mechanic41turk[m, or anything cryptocurrency
13:41:25 <mechanic41turk[m> hyc: home computer CPU power usage shouldn't stick out that much.
13:41:26 <gingeropolous> hyc, f2pool still reporting 170 mh/s
13:42:14 <hyc> 170MH is trivial compared to what, petahashes?
13:42:21 <gingeropolous> naw, for monero
13:42:22 <hyc> that says majority have gone already
13:42:24 <hyc> oh
13:43:12 <hyc> I wonder how many miners that represents
13:43:18 <hyc> prob spread out quite a lot\
13:47:56 <gingeropolous> doesn't help that their info page is crap https://www.f2pool.com/coin/monero
13:48:41 <gingeropolous> they don't seem to list per miner hr like some other pools, so no idea about distribution
14:05:17 <nioc> gingeropolous: according to  https://miningpoolstats.stream/monero  f2pool's HR never varies and never has info about their node height or last found block 
14:05:57 <nioc> and yes # of miners not listed
14:06:31 <gingeropolous> .hrmm
14:06:55 <nioc> first exchange they list is hitbtc lolololol
14:08:10 <nioc> and this is the first miner they list https://www.f2pool.com/miner?id=120
14:08:56 <nioc> well at least the link doesn't list monero
14:21:37 <gingeropolous> man the vega64 was so good pre randomx
14:33:09 <pauliouk> hmmm... Ampere bought the rights to the applied micro x-gene3
14:33:40 <pauliouk> Ampere eMAG 8180. Nice little arm CPU. 64bit, 32mb L3, 8 channels of DDR4-2666
14:34:16 <pauliouk> 32 cores, upto 3ghz each.
14:34:59 <pauliouk> would probably purr along nicely using 16 cores, leaving the rest for running p2pool and monerod heh
14:44:01 <pauliouk> $4697 for a pre-built eMAG Arm workstation
14:45:53 <pauliouk> hmm if only I could get a week alone in their warehouse, stack about 100 of those machines up and do some load tests :P
14:53:25 <hyc> not very cost-effective, seems
14:54:55 <Inge> can now get the 128-core version https://www.anandtech.com/show/16979/the-ampere-altra-max-review-pushing-it-to-128-cores-per-socket
14:55:26 <pauliouk> nah, no where near enough. could get 5 standard mobos, cpus, memory and hdd's - and a storage self to mount them to for the same price
14:57:21 <Inge> or get like a 3970x :)
15:03:34 <pauliouk> Not sure I could bring myself to spend nearly £2K on a cpu :P
15:07:25 <pauliouk> could build 2x R9 5960X rigs for the a little less and get about 1.1x the hashrate
15:08:36 <hyc> yeah, ARm ecosystem doesn't have the economy of scale on high perf chips that AMD has
15:08:59 <hyc> maybe Apple is getting that scale, but they're not selling their chips to anybody else
15:09:17 <pauliouk> hmm, actually I might treat myself to a R9 5950x - the rig I've got could take it easily enough
15:11:08 <pauliouk> how do I explain to the gf that I've spent some of our future house deposit that I've been saving on a CPU?
15:13:10 <abberant[m]> does anything beat 3900x in terms of roi?
15:15:01 <pauliouk> hmm I've mined nearly 1/4 a monero with a R5 3600 in about 2 months here.
15:15:06 <abberant[m]> If you want the gaming speed it might be more worth your money to get. 5600x and build a whole separate 3900x rig to solely mine on, it would fit within the cost of one 5950x at msrp 
15:15:13 <pauliouk> roughly half the price of a 3900x
15:15:57 <abberant[m]> s/If you want the gaming speed it might be more worth your money to get. 5600x and build a whole separate 3900x rig to solely mine on, it would fit within the cost of one 5950x at msrp/If you want the gaming speed it might be more worth your money to get a 5600x and build a whole separate 3900x rig to solely mine on, it would fit within the cost of one 5950x at msrp /
15:19:13 <abberant[m]> I'm really tempted to go all in on 3900x rigs but 2 year roi on paper is just too long, Its better than that in reality since you can sell the parts, but it still doesnt sound like a good idea.
15:20:14 <abberant[m]> even gpus at scalper prices are half that
15:22:56 <hyc> buying any hardware just for mining is seldom a good idea
15:23:11 <hyc> buying gaming rigs because you enjoy gaming is what makes the purchase worth it
15:23:19 <pauliouk> I have that 3600 just for mining (and keeping my office warm)
15:23:33 <pauliouk> its nearly paid for itself, if XMR goes up another $100 it will have done :)
15:28:52 <abberant[m]> observer api has been down right?
15:29:23 <pauliouk> DataHoarder didn't pay his electric bill
15:33:15 <DataHoarder> it should be up abberant[m] 
15:33:43 <DataHoarder> it *was* down due to a wide power outage that ended up taking 6 hours
15:33:47 <DataHoarder> and took down network as well
15:34:35 <DataHoarder> https://p2pool.observer/api/block_by_id/bc827aeecb5a5254cec6c2fc9a07f49c1d3c03b92dea39c5d9b3fbb19affe0ef
15:34:57 <DataHoarder> site itself uses the api locally
15:36:07 <pauliouk> why does Chrome suggest translating that from Luxemburghish to English? :D
15:44:19 <pauliouk> didn't realise, we didn't celebrate the 18,000,000th block being mined
15:45:02 <pauliouk> well the 18,000,000th Monero being mined
15:46:50 <pauliouk> 1 DOGE wallet holds nearly 28% of all the DOGE
15:50:01 <abberant[m]> <DataHoarder> "it should be up abberant..." <- works, thanks 
15:54:35 <gingeropolous> p2pool doesn't like when i stop right after starting it
15:54:37 <gingeropolous> some event loop
15:55:22 <DataHoarder> press enter gingeropolous 
15:55:37 <gingeropolous> well it fixed itself. 
15:55:41 <gingeropolous> i mean the loop timed out  guess
17:31:18 <mechanic41turk[m> <hyc> "then "git cherry-pick pr8001"" <- interestingly, git complains when doing this command and asks for my email and username for the git.
17:31:41 * mechanic41turk[m sent a code block: https://libera.ems.host/_matrix/media/r0/download/libera.chat/2aa7c6b92b0b455f556f8020e92a5c6e76b478e4
17:31:55 <mechanic41turk[m> I will try putting bogus name and email address and try again.
17:33:17 <mechanic41turk[m> Now getting this error:
17:33:26 * mechanic41turk[m sent a code block: https://libera.ems.host/_matrix/media/r0/download/libera.chat/d9dbc4b2175ecc5b6ce3b56542fe20855456abe3
17:33:56 <mechanic41turk[m> How should I move forward?
17:35:46 <mechanic41turk[m> cc: hyc 
17:36:47 <mechanic41turk[m> `git status` command returns this:
17:36:53 * mechanic41turk[m sent a code block: https://libera.ems.host/_matrix/media/r0/download/libera.chat/76240f81597e98d4ac4e2f2ed4c8117b6b374ef5
17:39:02 <moneromooo> git cherry-pick --abort
17:39:13 <moneromooo> git stash  # to stash your local changes where you can get them back later
17:39:23 <moneromooo> git cherry-pick pr8001
17:39:52 <moneromooo> If you want to get your changes back, git stash pop (probably after reverting pr8001).
17:40:15 <moneromooo> Alternatively, if you don't care about keeping those changes:
17:40:22 <moneromooo> git cherry-pick --abort
17:40:31 <moneromooo> git reset --hard
17:40:36 <moneromooo> git cherry-pick pr8001
17:40:39 <mechanic41turk[m> Here is what I got after doing your first set of suggestions:
17:40:44 * mechanic41turk[m sent a code block: https://libera.ems.host/_matrix/media/r0/download/libera.chat/bbd6518ec4fcdfbfae60a99d14711be6e2b583a2
17:40:57 <mechanic41turk[m> Is it all good?
17:41:02 <mechanic41turk[m> Can I move on with `make` command?
17:41:28 <moneromooo> Yes. I suspect you nuked local changes in between ?
17:41:49 <mechanic41turk[m> I didn't issue any commands in between the outputs I posted above.
17:42:06 <mechanic41turk[m> I don't know if I nuked local changed
17:42:09 <mechanic41turk[m> *changes
17:42:36 <mechanic41turk[m> I just want to get the relevant pull-request to be included as hyc suggested.
17:42:49 <mechanic41turk[m> It is abote the NO_AES stuff.
17:42:52 <mechanic41turk[m> *about
20:25:48 <abberant[m]> is there a disadvantage of having low custom diff?
20:28:00 <abberant[m]> I'm gonna run a lot of friends rigs through the proxy for api stats, dont know how low I should make share time be for my best worker, cause stats will be less consistent short term for people with low hashrate, and they're normies so they'll want more accurate stats
20:32:13 <pauliouk> hmm, doesn't seem to make any difference. I've got a little android box running at 10K difficulty, and it landed a share
20:47:08 <hyc> the idea is to submit a share every 30 seconds on average
20:47:38 <hyc> so whatever your hashrate, set difficulty = hashrate * 30
20:54:58 <pauliouk> just realised why I expected more hashrate out of my android tv box hyc. termux on my Samsung S20FE shows the CPU as Cortex-A55
21:04:59 <hyc> core clocks could be different, memory clocks different, etc...
21:06:28 <pauliouk> hell of a lot more cores available than what should be
21:06:49 <abberant[m]> <hyc> "the idea is to submit a share..." <- with nicehash mode are the workers aggregate in relation to difficulty?
21:07:06 <hyc> I know nothing about nicehash
21:07:15 <pauliouk> probably why xmrig won't let me submit a hash benchmark
21:10:51 <abberant[m]> hyc: its like a mining mode, I asked cause you said to calculate diff from hashrate but I have many workers with many hashrates
21:11:45 <abberant[m]> you reminded me though, theres a nicehash mining mode thats enabled by default that I think distributes jobs based on the custom diff depending on how strong the workers are
21:13:50 <hyc> the only point is to set the diff so that you'll submit shares frequently enough for your overall hashrate to be estimated 
21:14:12 <hyc> if you're within +/- 50% it'll probably be fine
21:15:01 <hyc> as long as the diff isn't so high that it takes hours or days to submit a share, you'll be reflected in the displayed hashrate estimate
21:16:18 <abberant[m]> think ill run into problems with like 5s shares on my faster rigs?
21:22:44 <hyc> nah
21:23:00 <hyc> a bit of wasted network bandwidth. no big deal
21:27:46 <pauliouk> my stratum server only see's the hash rate from the android box