sgp: we specifically asked that this channel not be bridged

Sorry. Guess I should have asked here too before opping.

lkcl: sounds perfect

I don't think there's a relay bot I can kick though. Not sure how that works.

hv-bridge is here

don't kick it though :D

<\x> plz dont ban me

Where does that go ?

<\x> hashvault discord

you can kick investanto though, never liked that bot

I guess. It's useless anyway.

it only spammed with quits and joins :D

hyc: I can remove the bridge but what's the problem with it

We're really pushing to get people on Matrix

IRC simply isn't used by more and more community members

that's fine. enlarging this community would only dilute the conversation

granted, there's not much focused activity going on now

but when there is, we want active developers participation, not peanut gallery

<\x> sorry

<hyc> granted, there's not much focused activity going on now >>> the beauty of hacking a major industry so they do all the work for you

i still like to kick the can around about increasing the 2 gb thing to something greater, or have it algorithmically adjust somehow

anyone ever hear back from linzhi-sonia?

we could just build in an automatic RAM size increase every 3-4 years

might piss some people off, but ~3 yrs seems to be the average update cycle for desktop PCs

yeah, that presupposes that civilization maintains its steady march of progress. my main concern with a fixed growth like that is what happens in the case of some 10-30 year stall

difficulty adjustment kicks in?

would that offset the requirements for increased ram size?

maybe. if everybody's hashrate took a nosedive because the RAM requirements caused existing systems to start swapping

then lowering difficulty could ocunter the perf drop

counter

true, but it would further the divide between "professional", larger miners and smaller miners perhaps

<\x> hyc ddr5 will start with 16Gbit ICs

though that divide exists already. i mean the 2 gb thing is mostly botnet deterrence though, right?

remember the target was originally 4GB

<\x> so likely the normal sticks on ddr5 be like 2x 16GB

yeah

the aim was to make it too expensice to do an ASIC with all on-chip RAM

<\x> like on a home setup,

and id much rather be proactive about it instead of reactive

though HBM is getting pretty common these days

<\x> idk if this matters to this channel but

lest we go through "ermagerd another hf to kick asics thIs IsnT sUStAInAblE

<\x> files.catbox.moe/078zvy.pdf JESD79-5_1.pdf

<\x> Jedec DDR5 specification

backing down to 2GB somewhat compromised our goal, but it allowed a generation of older PCs to participate

but there's probably not a clean way to tie it to difficulty somehow... either thresholds or change in rate of change

hashrate acceleration factor

At least this time around, I don't see how we can avoid another HF after ~3 years of operation

at that point we'll need to look at both RAM scaling and CPU scaling, since the other factor was time to prove & verify.

Doubling the mining memory buffer also doubles the verification buffer IIRC.

and while RAM increases pretty much like clockwork, CPU perf doesn't

doubling RAM might leave us in a situation where verification becomes too slow

so the primary thread it still customized HW, and the general sense is that that would look like a CPU with massive RAM on chip

but if thats occuring as a specialized effort.... the tech is there for commodity CPU manufacturers to do it as well

yep

so a lot depends on what's the hot new commodity CPU in 3 years time

so i guess at that point it goes back to botnet deterrence?

with the assumption that botnets are on old computers because old computers run old software that doesn't get updated

but neither of the main OS providers even allow for internet-connected but not-updatable systems now afaik

I don't get the botnet deterrence thing. They're doing a public good instead of whatever nefarious thing htey'd otherwise to.

windows will eventually update no matter how many times you say "no i have way too many windows open"

hell, perhaps in 3 years time we could find ourselves lowering the 2 GB ram requirement in order to speed up verification

The more usage, the more of the verification time is taken by tx verification compared to PoW verification.

indeed. could also lower PoW verification with increase blocktime

if thats *really* a concern

moneromooo: the botnet deterrence thing is an issue of public image, I think. Instead of just letting them proliferate

and Monero being known as the botnet coin, we make botnets easier to detect, and become known as the good guys fighting them

well we're not the ones letting them proliferate :{

i meant :P

I suppose. but certainly they're encouraged to try, if it's so easy to mine and earn money

we're already in danger of becoming the new ransomware coin, taking that title away from bitcoin

Oh really :/

money's money. Not our fault society is broken

Guess it makes sense. Fourth amendment coin would be so much better, but I guess we live in a shit world.

but back to it... the easiest way to be proactive against any sort of custom HW development may be time-delimited RAM requirement increases .. ?

unless we could find a way for the protocol to be aware of a hashrate increase that deviates from the normal rate of increase

normal behind defined as some rollling window perhaps

not sure how practical or useful that is, esp in the face of Azure abuse

but time-delimited ram requirements expose us to the possibility of monero failing in some nuclear winter, civilization-pause event

anyway, right now I think our botnet deterrence story is good. slowdown is not ignorable, so botnets become easier to detect

<hyc> not sure how practical or useful that is, esp in the face of Azure abuse >>> it could be a long window

but yeah. its not clean

and it really is 2nd-guessing the difficulty adjustment algo.

if 50% of miners went offline in a global catastrophe...

roight roight roight

Moneros claim to fame in Norway is the highest profile kidnapping ransom in modern history (although police charged the husband with murder and now believe the crypto ransom was a wild herring chase or a red goose.