You don't need to use the proxy

p2pool can handle multiple connections

ah

i guessed that

like it seemed useless to have xmrig-proxy and p2pool on the same server

thank you sech1!

<c​rypt0_bear:matrix.org> Is the quibic-li client not open source? github.com/qubic-li/client

<c​rypt0_bear:matrix.org> I see binaries but I do not see any source code.

<c​rypt0_bear:matrix.org> (Trying to figure out if they are doing dynamic code execution or just shipping with xmrig)

it looks proprietary

you could do a strings on the exe file and see if they're using xmrig

i mean, if they are, it would be a violation of xmrig's license since they're not shipping the source code anywhere

<c​rypt0_bear:matrix.org> Yeah. I got the bins. The alienminer seems to be another proprietary one. They have a hiveos config as well that is just doing another random bin install. I can setup a vm and tear it apart but I am pretty sure that they are just shipping xmrig with it

<c​rypt0_bear:matrix.org> If they are not. Then the program is basically a RAT and just updates dynamic shellcode

<c​rypt0_bear:matrix.org> Which I doubt they implemented. They probably just had xmrig hardcoded on there and then the parent pool can basically send the command to run the program

if they are shipping a modded version of xmrig

it is grounds for GPL violation

unless they uploaded the source code

<c​rypt0_bear:matrix.org> There is no mention of it on the docs or anything like that

<c​rypt0_bear:matrix.org> I am trying to follow the logic of the ai training part. But the xmr stuff seems to be hardcoded since there is no reference for the ai training stuff yet

<e​longated:matrix.org> It’s a scam coin, they are mining xmr and paying in worthless tokens

just finding mentions of xmrig inside the binaries are suspicious

you should go dig deeper in the binaries (after extracting them) to see

elongated: exactly, the miner program being proprietary is already a red flag

<c​rypt0_bear:matrix.org> I was trying to figure out the mechanics of their xmr mining. They claim it is all done via their chain communication. which means you could bypass their greedy mining by snooping into their mempool basically.

and also tied to a centralized platform

<c​rypt0_bear:matrix.org> Also if they are going with full dynamic exeuction then that means that pools basically have full RCE over miners.

<c​rypt0_bear:matrix.org> Just trying to figure out the mechanics since they promise the sky and back but I don't see any source code.

<c​rypt0_bear:matrix.org> Looks like a marketing stunt betting on not everyone cashing out the new tokens. Basically funding via rehypothecation with xmr as the payout currency

the client is proprietary and tied to their site

so it's not decentralized

btw, "full dynamic execution" is vague

for example, xmrig does full dynamic execution, it has a JIT engine that allocates a big code buffer to recompile RandomX code into the native CPU instruction set

of course, if they wanted to do RCE, they could tweak xmrig's JIT to have a little.. something special to it

without arousing suspicion

<c​rypt0_bear:matrix.org> Yeah but xmrig is basically a vm. Their execution is basically done so as virtualization. This would be like they run xmrig in memory but I doubt it

xmrig is not a VM as i can see it

it's more closer to a video game console emulator

<c​rypt0_bear:matrix.org> Based on their docs, the full computation would mean you run actual executable bins via in memory.

but for some CISC set

i mean, it'll be recompiled to instructions for the native CPU and executed

(nobody would mine in interpreter mode)

and so i'd think that they could make a modification to xmrig's JIT engine, to have an opportunity to sneak in custom x86 code

but not actually include it directly in xmrig's code

<c​rypt0_bear:matrix.org> so virtual machine right? Not trying to split hairs but basically close to what something like the eth EVM would be. You just run itterations to then figure out the next inputs. But the actual binary that runs is native code. But the "virtual" code inside the calculations is running inside of that. It is not just executing into its own process. (Long day so language skills declin<clipped message>

<c​rypt0_bear:matrix.org> ing. if it does not make sense)

yes

but xmrig likes to recompile the "virtual" code into native code

for a ton of extra performance

(also i realized i was retarded, so yeah randomX is kinda like a virtual machine)

xmrig actually executes the native code within its own process

<c​rypt0_bear:matrix.org> I don't think they modified. They are just having their bins dump it as either a bundled embdedded exe or later on curl to get their version. But from their docs for them to have a full execution of ai training programs then that means their miners would be capable of executing new in memory shellcode right?

yes

they could allocate a big code buffer and download shit for their program straight into that

and execute off of the buffer

of that buffer*

<c​rypt0_bear:matrix.org> Yeah xmrig is self contained. It is a really simple instruction set too so not really something with a huge attack surface for busting out of the thread. (Especially since you are just hashing at the end of the day).

yes, but i'm referring to the case if they modified xmrig

since the client is proprietary

<c​rypt0_bear:matrix.org> Yeah so at best they are just including xmrig and at worse they are adding a shellcode loader into their miner

it's also important to note, xmrig is most commonly ran as root

because of its one-time init code (tbh xmrig should descalate after that)

<c​rypt0_bear:matrix.org> I downloaded the bins I could find. I'll spin up some vms next week and then reverse them. I was curious what they did for them. Was late to the 51% drama

<c​rypt0_bear:matrix.org> huge pages support too

all of those are done one-time

like setting up kernel parameters and MSR registers

xmrig can be ran as non-root if you do all of that yourself

they are not gonna reach 51% and push people to run their literal virus lol

to mine XMR

<c​rypt0_bear:matrix.org> Ok yeah I see they give instructions for huge pages an non root since their docs mention that the miner can download and exec runners which are apps

i like how cfb is saying their crappy shitcoin is decentralized

a decentralized shitcoin where you have to register for an account on their site

and mine with a proprietary executable

"decentralized"

<c​rypt0_bear:matrix.org> Yeah I'll dig through these bins next week

<s​hortwavesurfer2009:monero.social> matrix.monero.social/_matrix/media/…ero.social/suMbDNRevgRqVGDUoXlagfTV

<s​hortwavesurfer2009:monero.social> Yo, check out the new Monfluo wallet design.

<u​ser2570:unredacted.org> I like

<f​reddi99:matrix.org> anyone knows if this qubic shit is resolved or are they still trying to sabotage xmr?

resolved? why is there a problem?

<f​reddi99:matrix.org> okey well maybe "problem" isn't the right term here. but is this something serious? i see conflicting information everywhere (some lose their mind, some say it's just propaganda) and cannot make up my mind about whether this is legit or some made up bs

qubic is a paper tiger

it's mostly FUD

they're trying to pretend they'll get 51% but they won't get shit

Have an eye on this, as somebody said so spot-on recently "blocks don't lie": moneroconsensus.info

In a few minutes they will switch from mining Monero 50% of the time to mine 100%, for 24 hours.

Also check the red "piece of cacke" labelled "unknown" here, scroll a bit down beyond the list of pools: miningpoolstats.stream/monero

*piece of cake

"unknown" is, give or take a few percents, most probably the Qubic "miners"

but it probably includes some not-yet-listed pools or solo miners

rbrunner: on miningpoolstats are you referencing qubic dot org?

api error

No, the round graph just below that

Over the last 1000 blocks

Cindy: Well, yes, that's why it's called "unknown".

A year ago it was 2%: web.archive.org/web/20240607022708/…tps://miningpoolstats.stream/monero

<e​longated:matrix.org> Who is mining empty blocks ? What was % of empty blocks before cubic shit

for moneroconsensus , they added ~9 pools yesterday

The assumption is that's Qubic

with the empty blocks

unless there are 2 blocks very close together it is qubeic

<e​longated:matrix.org> So only 9 blocks mined by them 🤡

Over which time?

yesterday they had 3 or 4 in a row sometimes

<e​longated:matrix.org> Over say a month before cubic started mining empty blocks ?

Make of it what you like, but the Qubic fans, on their discord server, trust the "pool hashrate" display on the left side here: explorer.jetskipool.ai/xmr-tracker.html

<e​ddie:oblak.be> I was thinking, we should not publicly react to it anymore, just ignore those fools, the twitter beef is half the fun for them

tweeter? IRC is my social media lol

this will be a nothingburger

I don't see any problem in studying them like we would any other technical or operational problem regarding Monero, calmly and reasonably

yes

but the problem is interesting

<e​ddie:oblak.be> rbunner, I agree

<e​ddie:oblak.be> rbrunner, I agree

Just ignore them. They need hype, they need publicity. Don't give it to them.

With 2 GH/s they would be the biggest Monero pool right now.

They don't have enough hashrate, and don't worry, we are working behind the scenes. Just can't go public with anything right now.

qubic fans will throw the biggest fit when they realize that it was just a pyramid scheme

and cfb will run away with the mined XMR lol

What do we think of tevadors solution to make centralised pools (prohibitively?) expensive by increasing the bandwidth requirements from 1MB to 3GB a day per miner (pls confirm actual figures) monero-project/research-lab #136#issuecomment-3170565603

Essentially pushing people toward p2pool mining

3GB a day per miner would DDoS the biggest pools

so that sounds like a pretty good idea

<e​ddie:oblak.be> I do not have a problem with centralised mining pools per se, but it is worrying that most miners flock to the biggest pools

<e​ddie:oblak.be> you need to be able to attract miners that are technically unskilled, as to increase the potential pool of miners, imho

thx plowsof , I hadn't seen that b4

I just fell off my chair. tevador posted 2 hours ago. Wow, quite a relief.

i wonder how the 3GB will affect small mining pools (LAN-only with a bunch of computers)

<e​ddie:oblak.be> would it be possible to devise a mechanism where it gets more difficult for centralized pools only when they reach 5 or 10 % of the total hashrate?

p2pool is great except for the many small payouts you get which will then need to be consolidated

i wonder about modifying RandomX to include ed25519 signatures in the algorithm

of course this is "miners signing their own block" which i don't know works

but a local mining pool could just provide all of its own computers with the key

while a centralized online pool would have to devise a mechanism to sign data for miners

how would this affect p2pool?

btw, does xmrig support mTLS auth in its API service?

i mean if the service only supports tokens then i'm fine with locking it behind a HTTP forwarder that provides mTLS auth

Huh, 17 hours mining Nano on P2Pool at 9kH/s, and no Monero...

That sucks dude.

try mini

Maybe I should switch to Main?

i mined mini with my crappy laptop at 4KH/s

and i got atleast some payouts

28 shares accepted... whatever that means.

shares in p2pool or xmrig??

xmrig.

12 in p2pool.

p2pool likes to break challenges up into pieces for xmrig

so i shouldn't rely on xmrig's "shares accepted" thing only

I see.

into easier-to-solve pieces*

"Payouts 0.0000000/month"

Eh, 'thank you'.

p2pool-nano needs to find a block to pay you. It's all in the README

switch to p2pool-mini to get payouts more often

How about Main? I don't mind waiting for larger payouts.

24/7

main, you'll get crushed by other computers with 1000x more hashrate

mini is a equlibrum

Ah

9KH/s is more than enough for mini

i don't know why you're mining in nano

It's the default for Gupaxx.

... but I've stopped p2pool in Gupaxx, and the ability to choose which sidechain is still grayed-out.

Oh, I've cleared the Options and now can choose.

i thought the default was mini

I haven't messed with that 'till now.

I've hit Clear on Options, whereas before I used Advanced. Now I have settable Options.

... except p2pool and xvb are now staying orange, not going to green.

does xmrig have mTLS auth for its API?

In Gupaxx there's a TLS checkbox in xmrig.

that's for the pool server

i meant if xmrig is hosting a HTTP API server

Cindy, do you know whether anything writes to the blockchain, other than monerod?

IOW I want to run two VMs and have them share one disk with the bc.

i think its just monerod?

run monerod in a seperate server

I wasgoing to run monerod in one VM and not in the other, but they both need the ports it has.

why are you running two VMs

I could forward those ports as long as they're TCP and not UDP.

like why two VMs in the same PC

<o​frnxmr:xmr.mx> Cuz he's quantum computing

all you're doing is adding additional overhead

and killing your overall hashrate more :P

I'm pinning CPUs separately. I've found that when I run 55 threads in one VM hashrate crashes.

why are you running 55 threads

Throttle it back to 24 CPUs or so and I max around 9200.

why are you RUNNING 9200 THREADS?!?!

So I'm splitting those threads into two VMs.

Nah, 9200H/s

ah

... 2ith 24 CPUs

i doubt you're gonna get any performance with this

<o​frnxmr:xmr.mx> Are you usong cpus and threads onterchangeably

you're just adding more burden with the VM engine

<o​frnxmr:xmr.mx> Cindy, he's not looking for performance, he's looking for a long-term hobby

55 threads in 24 CPUs would obviously kill your hashrate

because like

I'm using cpus and threads onterchangeably

at some point, the kernel scheduler kicks in

ofrnxmr: oh really?

Well 9200H/s is the best I've ever done. In testing as I increase threads H/s drops off. It's the algo.

<o​frnxmr:xmr.mx> Seems like it. Lol. This seems like what happens when someone has too much time on their hands. Its not this (weeks) difficulty to mine monero

when you increase threads beyond the cores

your kernel has to juggle threads around the cores

which will be more overhead

Sure, this is why I'm matching threads to CPU threads.

... and testing shows that over a certain number of CPU threads performance drops dramatically.

also seriously, move the blockchain outside the VM

So I am splitting CPU threads between VMs. Problem is sharing the blockchai and monerod ports.

remember that even if the CPU is being passed through, the hard drive is obviously emulated

so youll gain a little by moving it outside the VMs

The bc drive is in my file server, and is pulled in to the VM. But I can't have two monerod's writing to one BC.

wait

you're remotely mounting the blockchain on some file server

and monerod is using it like that?

Yep

honestly that would be even worse performance than a physical HDD

KVM VMs.

It is a rotating drive.

oh wow, a combo :P

sorry for dunking on you a lot lol

but this setup seems way too overcomplicated

Yay. Is 9,200H/s bad for28 CPU threads?

and adding more burden to your miner

Best I've ever done.

28 threads in one VM and 28 in t'other.

I just have to share monerod somehow.

host monerod outside this mess?

This would double my performance over one VM.

And forward ports?

yes

also lol "double my performance"

bro discovered infinite CPU power glitch

just host more VMs

<e​ddie:oblak.be> 9 kh/s for 28 CPU is really bad

Well I'm out of CPU threads to pin now, with 2 VMs.

yes

eddie is right

<e​ddie:oblak.be> You should be able to do 25 or more

even my crappy laptop with 4 cores and 4 GB of RAM

O

can still get more than half of that

Huh...

the problem is the amount of overhead you're adding on this

<e​ddie:oblak.be> I do 4 on 8 cores

It hardly hits my disk.

<e​ddie:oblak.be> And that's just for giggles

but i discovered today that like

i can get unlimited hashrate if i host more VMs

(You're welcome)

yay

CPU cores are passed directly through.

Need to check my host load, before and after..

what about your blockchain

That disk is hardly hit while hashing.

I guess monerod updating.

Hm, if I run monerod in one VM, I don't need the disk in the other VM. Just monerod's ports. Hope they're all TCP.

... but on the host I can't differentiat between the qemu-system-86' between my herd of VMs...

is this all for security? :P

Yes

if i were you, i'd host monerod on a seperate machine

Ok monerod is indeed all TCP, 18080-3

flash linux and put xmrig on a flash drive

disconnect the hard drive off the computer and boot off of the flash drive

and i'd get 4x the hashrate with the same amount of security

This is my Big Bad Server wih copious cores.

I do have an ODroid I could host monerod on, but I don't see the point. The disk is hardly hit.

are you mining off of a remote node?

It's the raw horsepower I bought this server for, specifically monero.

you should check what node you're mining off of

like "hardly hit disk" is suspicious

Running Gupaxx. Should be p2pool.

iotop

check the p2pool tab

in gupaxx

Ok the VM's been down a bit so needs to sync up.

iotop shows the bc disk is hardly hit.

Had to copy the qcow2 for the second VM.

(and pin diff CPU cores)

I am doing New Science, you see.

Maybe 'burntcorpse' will pop in. (Hopefully someone will open a window if he does)

Ok p2pool is syncing now.

"<e​ddie:oblak.be> I do 4 on 8 cores" That's freakin' 500H/s per core. Unbelievable.

the most amount of security i'd be doing is doing all the configuration myself, running xmrig as non-root

maybe even running xmrig in a seperate linux namespace

You're suspicious that KVM is a ottleneck. It's now clear which qemu-system-x86 on the host is altcoin1 VM. 2,797% CPU for 28 cores. 99.8928571429% util.

I seem to be maxing my CPUs.

Idk where else a problem could be.

even if you're passing through the CPU, a system is a lot more than just the CPU

so KVM is emulating the other stuff

and it does that through software traps, stuff that cuts into the CPU performance

Those numbers on the host. On the guest 2,780% usage. Maybe the diff is overhead.

I don't believe 500H/s per CPU.

Just running top on host and guest cuts to 5,500H/s.

Well, stopping top on both I'm still at 5,600H/s. Must be the difficulty.

lol

I was at 9,200+

My daemon reorganized 5 times in the last 1.5 hours.

Just did No 6 :)

yep 6 here

last one was one of 6 consecutive empty blocks

I am sure that the "paper tiger" Cindy wrote about this morning UTC :)

oh wow!

6 empty blocks? :o

6 reorgs in 1.5 hours and 6 empty blocks happens every other week, right? :)

rbrunner you don't have to call me out like this :P

are they really mining empty blocks right now?

I looked at renting hashpower, only $500 for enough power to on avg find a block

all the blocks they mine for several days now are empty

what are they trying to achieve with that :P

slowing down transactions?

publicity

i see

made me turn on my rigs in the summer lol

i mean, honestly

weather has been ok so

CFB is just trying to do this for clout, but i wonder if the EU or other governments that are trying to ban monero will jump in on this too

like harness tons of computers for the purpose of attempting to sabotage monero

isn't that just a litterbox?

there are many types lol

including my car

<r​ucknium:monero.social> Your IoT litterbox can mine Monero, too :D

automated litterboxs are a bad idea

you need to observe the output

wasn't there that uhh

automated cat box from china

that literally killed cats

like crushed their necks with the (very powerful) window motors

hey guys

i doubt this is the first 51% attack attempt on monero btw

is there an amazon like shop to buy stuff for xmr?

hello Cind

hi Guess65

<r​ucknium:monero.social> Guess65: There have been multiple outfits setup for that, but they closed probably because not enough customers. Check out xmrbazaar.com

how come there arent enough customers?

wonders

Ok I have one monerod and blockchain present in two VMs.

SB #1 online

And she laughs because I am trying something new...

<c​rz:matrix.org> ofrnAI: Check my DM thx

shitbox #2 online

Are these reorgs actually happening? And if so, should I start thinking about how I can do a 0 conf scam on someone?

<r​ucknium:monero.social> midipoet: you can't do that because:

<r​ucknium:monero.social> 1) You would have to send a contradictory tx that spends to yourself to the Qubic pool, which doesn't accept direct tx submission AFAIK.

<r​ucknium:monero.social> 2) Qubic is mining blocks devoid of any txs anyway, so it would not mine any contradictory txs.

<r​ucknium:monero.social> But yes a few orphans are happening: moneroconsensus.info

<r​ucknium:monero.social> You would see this in the default log level of a Monero node, too.

<r​ucknium:monero.social> Your first tx to the victim would stay in the txpool, waiting to be mined by honest miners. You would still be sending them the money, just maybe a few minutes later.