Hey all,... (full message at <libera.ems.host/_matrix/media/v3/do…6e2603bb533ef789a4ec6537c0ed78d0a44>)

It would definitely be a great contribution

Might want to ask the people in #monero-outreach:monero.social for some tips/material/logistical support before creating pages and stuff, but this kind of outreach is very important

Ok, thank you! I will head over to Outreach.

(It's not a very active channel, but people are still in there so I'm hoping they will still come back to you)

Hej Fredrik! That would definitely be helpful. I keep seeing Monero stickers all over Stockholm, but people have no clue here

Hej! Cool. Yeah I've seen in Gothenburg also. But exactly, people do not seem to have a clue. But can't really blame people either, this is 'pretty far out there' at the moment. I'm so sick and tired of the old system I feel i have to do something.

jtgrassie: spamming the code?

You set the block template refresh to 120s. Please correct me in my assumption that was confirmed when I asked in my opening comment

The pool grabs the block template when a new block is mined, and then again in 120s

> it's NOT a terrible option. the pool has been set to 2 mins since launch and works absolutely fine as 2 mins. xmrig is a miner not a pool. You clearly have minimal understanding of the effect of this option. Please stop spamming the code.

Context:

here is what 2mins looks like: block 2821003

* here is what 2mins looks like: block 2821003

xmrchain.net/block/d1d9445c698c43ea…ef9718fbd39ac54806a9b90a65903b21069)

> xmrig is a miner not a pool. You clearly have minimal understanding of the effect of this option.

Did you forget xmrig and xmrig proxy connect to the daemon when solominjng? Its the same functionality. Man i hate liars

Meanwhile nanopool is updating every10s while jtgrassie makes false claims about load to excuse putting out an update that is roughly the same bullshit behavior as prior

(But user configurable).. lol, why in the world would some ship asshole defaults? Probably because theirs an asshole

They're*

ofrnxmr[m]: the profit difference between 2 mins and 20 secs is negligle. The load difference is significant.

so you might think you know what you're talking about but you clearly have never ran a high load pool and ran the numbers

you being bitchy with me for adding this as an option is bizarre, as before it was fixed at 2 minutes and now I've added the option for a pool op to lower via conf

unrelated: for matrix users (or if your app has this feature) - please disable thumbnail / image previews. this can be done via all settings -> preferences and finding the relevant options and disabling them. now back to mining discussion thank you

jtgrassie: Bitching because while everyone else runs 30 or better, youre releasing bullshit software

And its not about the fees, your defaults bloat the tx pool

I repeat: Terrible defaults

ofrnxmr[m]: "youre releasing bullshit software" <- I've made it a USER CONFIG OPTION

You are free to use or not, change or not.

I dont pool mine, im bust contacting people using tour software

don't like? don't use. Fuck off

Wondering why they are mining 0 tx

I dont use it lol

Good.

Good what? That your software leads to ddosing the tx pool?

Lol you have no fucking idea what you are on about. It doesn't ddos anything.

You're an idiot

If there are 100tx in the pool and you mine 0, you leave those to continue to pile up above the 300kb limit, dumbass

Dont act so smart when you dont know what youre talking about

and someone else get's them then

you talk a lot, but don't contribute

🥳

you don't have to use the pool, you don't have to use the default settings

you can set to whatever you want

trolling people that actually contribute something is not a good look

I dont have a mirror

Not concerned with how I look to liars

Anyway, thanks for the user config option

No thanks for leaving the default retarded

Sincerely - the person wasting Hus time cleaning up your mess

p2pool has a nice feature of including high fee transactions immediately in the block if seen in mempool. it cares about UX because it updates the template often also

yep, and p2pool works completely differently to a centralized pool

So kill your pool software

It really shouldn't be used with those settings, at all, by anyone

says the person who doesn't run a pool...

neither does jtgrassie unless you want to prove it

jtgrassie: Lol says the person who's pool is configs to mine 0 tx by default ^^

ofrnxmr[m]: except it doesn't work how you insinuate

plowsof11: monerop.com the public ref pool

plowsof11: also run 2 private pools

<jtgrassie> "trolling people that actually..." <- I think he means well but just wasn't properly socialised as a child and now has trouble expressing his concerns in a helpful way. Is there any possibility he may be misunderstanding you and the reasons for your defaults?

So you are talking shit here?

It works exactly the way I insinuate

thanks jtgrassie

Are you running the defaults on those pools?

Actually, I can probably just check

dsmlegend[m]: "I think he means well" <- I don't.

Answer DSM question then

dsmlegend[m]: "Is there any possibility he may be misunderstanding you the reasons for your defaults?" possibly, but it's not my job to educate

Thanks everyone for attending

guess I should make future updates privately and not release??

Sure

If thats how you feel

Luckily there are very few entitled idiots like you.

Personally, 👍 from me on that

Who's entitled? LMAO youre batty

You ask a q, I answer "do what you want" and you claim im entitled

Anyone can fork your repo and adjust the line properly. I commented directly because I figured you made a mistake

you have been trolling me all morning. calling me a liar etc etc. all for developing someing and releasing a FEATURE that one can use (or not) and change (or not).

I asked a question and you gave a false answer

"I figured you made a mistake" no mistake.

Claiming 2mins is perfectly acceptable lol.

jtgrassie: Yeah, I noticed

"you gave a false answer" in your opinion as you don't know the answer

Anyway, its your software and I made my comments

it is perfectly acceptable!

No it isnt 🤣

Ill agree to disagree

the *choice* is lower load vs small increase in reward

Wen taking repo offline?

So I can tell people to use the fork that has proper defaults

and it's a *choice*, and it causes no harm

how nanopool is able to handle 10-second template updates, and the "high load" pool defaults to 120 seconds?

is nanopool not high-load?

maybe their pool isn't running on a rasp pi

My android node has no serving templates every 5 secs

NP*

ofrnxmr[m]: so you're running a high load pool on an android phone? BS

You're runnign a high load pool on a rasp? Bs

Or stupid

I can't know what every user of monero-pool is running. Idiot

What kind of idiot assumes nanopool is going to run your shit on a pi?

Sub Nanopool for high load pool

I'm not assuming anything, that's the point!

Why not set the default to 30s and if someone is having load issues they can still increase it? I assume a Raspberry isn't the usual choice for a pool.

that's why I made it a user configtion option

I suggested 45 (1-2 times per block on avg)

"120s is perfectly acceptable" and im trolling 🐒

I know that it's user configurable but setting a sane default would be nice. Running a pool on a Raspberry is clearly an edge case.

perfectly acceptable to miners who only care about the coinbase reward* but the ecosystem suffer

but your choice in the end so whatever

it's a USER choice now, that's why I made the change (which nobody else bothered to do)

having you all moan about a choice I added is super weird

no one claims that it's not a user choice lol the discussion is about the default value which is your choice obviously

boycott pools that update less than every 45 seconds

Nobody is upset about there being a choice. People are upset that the default harms UX, and I think that's a very fair concern.

selsta: every option in the config file could be questioned

BusyBoredom[m]: "the default harms UX" every default has the ability to harm UX

Yep of course, and that's why we should be careful when decide on defaults.

the default was always 120s (and not changable without recompiling), I added a change to make it user configurable

Nobody's upset with your change to make it configurable. That was a great change, thank you for making it.

yet I get trolled and called a liar for contributing

Trolled lol

I only called you a liar after you tried to justify what is either a lie or a lack of knowledge. But then you claimed I didnt know what I was talking about

Thats a liar right there LOL

Not about about my approach though, its about the setting and understanding the justification for it - not that youre obligated to justify anything

But its nice when good people do good things, instead of deflecting while doing questionable things

ofrnxmr[m]: you call this a lie "the *choice* is lower load vs small increase in reward"

because that is factual

you might not like the fact (or agree with it), but it's still a fact

🍿 still claiming that load like are we

Line*

And talking about tx fees, ayye. Lol

pools auto-adjust difficulty to get shares from miners every 30 seconds, so why the default can't be set to 30 seconds? Is generating new miner job harder than checking a share? I don't think so

set it to whatever you like

(fwiw I have it set to 45s on the reference pool)

rookie numbers

all major pools are below 30 now

again, set it to whatever you like

I like 30 or less

45sec in theory averages out with 30 secs, with the last update before an expected block being 30 secs before

But 120secs is asshole numbers

linking rucknium.me/posts/monero-pool-transaction-delay if anyone has not seen the data

monero fees are so low, miners could probably get away with "mining empty blocks" but it would hurt UX

Ive seen 0.7 blocks

Usually because someone left 100tx on the table

ofrnxmr[m]: again, set it to whatever you like

Do us a favor and set to 45sec, pretty please?

you can run a pool and set to whatever you like

updating blocks every 120s allowed "in September 2021 [..] an estimated 72 XMR additional revenue for P2Pool miners."

so the p2pool miners shld thank you

that advantage is now gone

^^

But the people using monero shouldn't

i would also like to note that the monero pool on jtgrassies github has no built in donation fee

Its not gone

xmrchain.net/block/d1d9445c698c43ea…ef9718fbd39ac54806a9b90a65903b21069)

Heres s block from today where c3pool mined 0 tx while there were 30+ people waiting for the bus

Which had to wait 5+ more mins

plowsof11: thank you for reminding them

jtgrassie: why template-timeout = 60?

wouldnt something like 10 or 20 be better?

xfedex[m]: because that's what the reference pool is running

the cost of getting a new template is very low, much lower than the transaction fees you'd earn, at least on a coin like XMR

"i would also like to note that the monero pool on jtgrassies github has no built in donation fee" <- unlike node-js pool and xmrig miner

xfedex[m]: it's a user config option, change to you hearts content

P2pooler doesnt either (centralized pool on p2pool)

whats the reason for such a high default value tho?

p2pool has 10 seconds if i am correct

"whats the reason for such a high default value tho?" it used to be 120s and not changable

And xmrig is one of the lowest fee miners, one of the only open source, and advertised disabling it

Xmrig also GPU mines many different algos

xfedex[m]: Because someone might run a High load pool on a pi, or smth

Thats not an answer ^ lol

ofrnxmr[m]: I seriously hoe you don't talk like this to all Monero contributors

:( I do though

we've lost other developers because of being hounded by twats like you

if you want to put off peole from contributing, carry on

Tell me I hurt your feelings without telling me your feelings are hurt

Itd a quick change s/120/45

I'm pissed at even having this conv

Stop being a crybaby

i think it is now 45s and configurable*

You should be

I developed something, keep it updated, added a feature, and you bitch and moan

Its 60 now

Better than 120, but what a guy

i think it should be 20

or 30

ofrnxmr[m]: "You should be" < you should be ashamed of yourself

pls remember the extra 72 moneros p2poolers got and be thankfulfortoday

Why would I be? Im nit too prideful to make a change from 120-45

xfedex[m]: 20s would be lower than the vardif target of 30s, so pointless

You should be ashamed

jtgrassie, what's vardif?

ofrnxmr[m]: "You should be ashamed" < what for? developing open source code???

Variable difficulty adjustment when sending out jibs to miners

jtgrassie: For having too much pride to make a simple change

Man this is painful to watch. This whole conversation should have just been:

"Hey man, 120s is bad for UX, could you consider a lower default?"

"Oh damn you're right, no problem I'll check that out next weekend"

Alright. Though, I apologize for being mean lol. But I am how I am. Thanks for the change 👍

> <@busyboredom:monero.social> Man this is painful to watch. This whole conversation should have just been:... (full message at <libera.ems.host/_matrix/media/v3/do…0e6b8b91d2a9598217de33d18bb1ff0ee71>)

I commented on GitHub to confirm it was the update to the template time, then I suggested it should be set to lower, 45sec etc. Then was told no, its perfectly fineb

I do not *have* to make any changes, or engage with you.

And after some back and forth, told to stop spamming his git, so I brought the convo here.

Anyway.

All is ok now. 60 is better than 120 so ill shutup

I owe you nothing

Right, i know

I changed to 60s as that's what I run on the reference pool.

and should shut you up

Well, id still ask you to try 45 seconds for yourself and consider changing from 69-45

60-45. Not because I want you to, but because it works.

what I run is none of your busness or concern

you can set to whatever you like

now please, go away

So, how about the Superbowl?

Anyone watch or notice anything monero happen?

brb im setting my pool to 44 seconds (or 24.. maybe 25 .. none of you're business tbh)

Im going to try 24/25 seconds. Probably the best, right Sir?

sounds good to me 😡

Sir, it says error when enter 24/25

ofrnxmr[m]: Yes

breaking news: the Big Botnet has switched back to hashvault.pro

They are currently mining with ~1.0 GH/s to address 49ixgZU6cafjMkM6RjmR9UA3qhmRL5NrfQAyxR7BBcEjL2VQUSMbTs81scaTd6R7REaNFkbrebRruXQRP5sZHEjtRTKPyQC

they are using 9 different xmrig-proxies

so far the botnet has earned ~20k $

xfedex[m]: Actual proxies, or just bundling workers under the same names?

<jtgrassie> "plowsof: thank you for reminding..." <- Wait what lol, that piece of information implies that a 120s update time implies less revenue for miners on your pool

merope: oh sorry they have many workers with different names

s/time implies/time means/

out of context - his miner has no built in donation

They're using xmrig-proxy

A thoughtful botnet, how nice

<᷾s> Wait, matrix replies don't show up on IRC either!?!?

sech1 since you have xmrig fees, can't you get the hacked IP addresses and then call the hosting?

I dont know if donation mined to a different address or to a different pool but

not his problem

Why tf would he do that

GDPR complaint from the botnet owner incoming

w0w

drama was somehow unsatisfying

Slow Monday

We're selling MoneroKon tickets and exchanges are devaluing Monero price wtf

Trocador.app allows exchanging to localmonero

Hi everyone 😅 thanks for the debate!... (full message at <libera.ems.host/_matrix/media/v3/do…816870b5572b180d0b20bcb283d835c7f3e>)

Lovera[m]: "I think it’s proven that 30 seconds helps the network." has someone proven that 30s updates are meaningfully more helpful than 60s?

"Man this is painful to watch." +1

I’m not sure. It’s twice as long.

If I didn’t miss anything in the conversation, wasn’t the default value 120?

Lovera[m]: yes it used to *default* to 120s, I made a change so it could be easily changed by the pool op and then changed the default in the conf file to try and end this conv.

<jtgrassie> "I changed to 60s as that's..." <- Oh I see

the default is now 60s

there are a load of other settings that have to be changed in the pool conf which is why this whole conv is ridiculus

but instead, the "community" likes to harass people

it's a wonder anyone contributes to monero

Thanks for the change 👍🏼 it’s much better than 120 I guess.

It is also true that if you have the file. CONF should modify this parameter. The only drawback I see is that the user does not have much knowledge about that value and does not touch it.

Btw, A few years ago I used Monero pool to run my own pool in my neighborhood 😅 it helped me learn some things! I couldn’t find anything public, just monero-pool

glad it helped

"the user does not have much knowledge" <- yes, this is true of all the parameters

the reality is that even at 120s as the default setting, those that don't have the knowledge to change this won't be running a serious pool

same goes for all the other default settings!

Hence the impact to the "community" / network of leaving the default at 120s, is essentially zero

The #5 pool is still misconfigured. Until I contacted them, the #1 pool was as well. Size of pool =/= understanding of the settings

ofrnxmr[m]: "Size of pool =/= understanding of the settings" < that is fair

"The #5 pool is still misconfigured" it's not misconfigured though, that's their choice or ignorance

Its misconfigured. They will update

0 pools have decided to not update to as fast as reasonable. And I think the ones that went with 30s will likely change to 10s soon enough

it's configured as the set it

*they

C3pool uses monerooceans software. Moneroocean was configured badly, has been updated, c3 is running old version for now. Different language, so hard to communicate

All of them had defaults set (aka dont update or 2mins )

P2pool was the sole outlier

(In top 10)

Almost every pool I studied in rucknium.me/posts/monero-pool-transaction-delay never updated their block templates except when a block had been found. HashVault seemed to update every 120 seconds. xmrpool sometimes updated randomly. This set of histograms tells the story: rucknium.me/img/mining-pool-behavior-histogram.png

Thanks for making the update frequency configurable, jtgrassie

"configured badly" is not the same as "misconfigured"

Rucknium[m]: most welcome

Rucknium[m]: I’m curious what type of hardware do you need to run a pool like hashvault with 60 sec for example. I think they get enough rewards to update their hardware if it is the problem

Nano is 10s right now Lovera @btclovera:matrix.org:

And the majority of the hashrate is 10-30s. Solo mining with daemon is 1s, xmrig is 15s (configurable)

The cost of having infrequent block template updates is pretty clear at this point. The cost is measured in lost pool revenue and worse Monero user experience. The computational cost (hardware infrastructure cost) of frequent updates is less clear. ideally, we would know the tradeoff to pick good defaults.

Anyone want to run some benchmarks?

I can probably run on mobile hardware - what do you need me to do?

the difference in mining revenue is negligable ~0.01 XMR per block

<ofrnxmr[m]> "Nano is 10s right now Lovera @..." <- Good one

empty block 0.6 XMR, block with 57 txs 0.613 XMR

The difference in time to confirmation is not

...under current tx volume and exchange rate conditions.

"The difference in time to confirmation is not" agreed

<Rucknium[m]> "Anyone want to run some benchmar..." <- If it’s easy to do, I can try it 😅

ofrnxmr: I don't know what a reasonable benchmark test would look like.

if the big pools complain / switch back to 60 ~ 120 seconds , we'll know i guess

If the concern is UX (confirmation time), I'd prefer to see numbers of average confirmation time vs template referesh time

as I expect the difference between 60s updates and 10s updates is not that significant

pool revenue difference almost certainly insignificant

"empty block 0.6 XMR, block with 57 txs 0.613 XMR" <- that's 2% difference, a significant number if you compare to a typical pool fee

fair

10s updates mean any tx that show up within 10s before the block are mined.

pool that updates frequently and has 2% fee is better than a pool that mines empty blocks and has 1% fee (better for miners)

There is a 90+% difference between being mined by 120s vs 10s, in roughly 59% if cases

50%

I can try to run the numbers on what average time to first confirmation would look like for 10 sec update frequency vs 60 sec.

See if you can synthesize consistent incoming tx

spackle_xmr: might be able to do this

In my blog/Reddit post I did it for the actual observed confirmation times compared to a scenario when all pools had the average block update frequency as P2Pool

"that's 2% difference" <- not for every block, that's assuming the pool always mines empty blocks, which is not the case

I think I would just take the empirical arrival times of txs into the mempool and the empirical time of found blocks and then just "simulate" when txs would be confirmed if there was a 10 sec vs 60 sec lag.

Easy way to see the difference?

ofrnxmr: Can you rephrase?

Run 2 instances of xmrig-proxy

One with 10s, one with 60s.

Keep track of the number of tx in the last update for each before a block is mined

monte carlo simulation

would be quicker and easier

Yes. Monte Carlo would be easiest. Even easier than using the empirical data I collected over a month.

You would need to have the average tx volume to set the probability parameters on tx arrival. Besides that, no real external data needed.

Would Monte be as accurate as using real data in realtime

I mean, things are always changing.

Historical data must be from after nano did the switch

And tx volume and hashrate is all over the place

Depends what you mean by "accurate". You'd need a large enough sample size with realtime data. I have a large sample of 1 month already, but I didn't literally run xmrig-proxy

spamming testnet solves every problem

Oh

#no-wallet-left-behind:monero.social meeting

Preliminary results. I have not double-checked them. With the empirical data from December/January, if all pools were updating the block template every 10 seconds, then the average time between a tx entering the txpool and receiving 1 confirmation is 135.8 seconds (median is 98 seconds). If updating every 60 seconds, then the average confirmation delay is 208.5 seconds (median is 172 seconds).

What was the average blocktime during this period?

The Monte Carlo results are very similar. 10 second update frequency: 136.1 seconds average, 99.5 seconds median. 60 second update frequency: 208.2 seconds average, 172.6 seconds median.

ofrnxmr: 120.4 seconds

i wonder why average time to get 1 confirmations with pools updating template every 10 seconds isn't (120+10)/2 = 65 seconds

xfedex: It's because hashing is a memoryless process. The average time to the next block is _always_ 120 seconds regardless of how long we have already waited. (Assuming no major changes in hashrate)

The PoW protocol for finding blocks can be modeled as a Poisson process. The time interval between blocks is a random exponential distribution. The number of blocks found within any particular time interval will have a Poisson distribution.

This is a deep dive: arxiv.org/abs/1801.07447

"In the original Bitcoin paper, it was suggested that the blockchain arrivals occur according to a homogeneous Poisson process. Based on blockchain block arrival data and stochastic analysis of the block arrival process, we demonstrate that this is not the case."

When the paper says "this is not the case", they mean that it is not _exact_. But it's really pretty close.

To run the Monte Carlo simulation I calculated the average arrival rate of txs (this is just the transaction volume over a period of time) from the empirical data to get the rate parameter for an exponential distribution. Then I draw random exponential numbers to get the time between each simulated tx arriving.

I did something similar for blocks except I just set the mean arrival time to the target: 120 seconds.

ahh, statistics

If one minute has passed since the last block was found, does that mean that the next block will be found sooner? No. The same is true if I roll the dice and don't get a royal flush (mixing metaphors). The fact that I didn't roll a royal flush before does not affect whether I get a royal flush on next roll. That's the same idea as with hashing. The hashing attempts are independent.

It's counter-intuitive, but true.

Mean is not the same as median. Median time to next block with an exponential distribution is ln(2)/lambda. So ln(2)/(1/120) = 83 seconds

my brain keeps telling me that when you submit a transaction at a random moment you will be in average halfway from the next block

i might run a simulated blockchain to prove i'm wrong

Do it.

en.wikipedia.org/wiki/Gambler%27s_fallacy "The gambler's fallacy, also known as the Monte Carlo fallacy or the fallacy of the maturity of chances, is the incorrect belief that, if a particular event occurs more frequently than normal during the past, it is less likely to happen in the future (or vice versa), when it has otherwise been established that the probability of such events does not depend on what has happened in

the past."

^ tip of my tongue. I thought Monte sounded familiar

So, when you are looking at mempool.space and see that the most recent BTC block was mined 25 minutes ago, do not think "Any second now, my BTC tx will be confirmed."

Think "Damn, still have to wait 10 minutes on average."

You’re wrong Rucknium

rucknium, simulations prove something very important

you were right, after a tx is submitted it takes in average 120 seconds for it to be added in a block

That was a quick coding job. Nicely done.

source code: paste.debian.net/1270624

plain javascript

<Rucknium[m]> "So, when you are looking at..." <- This reminds me a little of Pieter Wuille’s pool about the average time of one block and the next picking a random point in time Everyone thought 10 minutes but the correct answer is 20. the expected time between the previous and the next block is 20 minutes