<DataHoarder> "although if you restart it might..." <- yeah, this is beyond that tho

🤷

sech1: running the api branch off master, getting a bunch of err msgs. pasted in -dev

where does it come from? I don't see how it could be related to my changes

no idea

load_binary seems a strange call to be happening

what how people getting 11,3kh/s with 7 3800x meanwhile max I'm getting is 7,5k

is ram speed a really big factor?

yes, and channels

my ram at 2400mhz

probably bump that, "Enforce POR" option on server stuff for me is disabled :)

no need to OC specifically, but you can rise that

then the classic, hugepages, msr, etc.

and CL17-17-17-39

I never went towards optimizing that much

already running ECC here with registered/buffered RAM

my mobo supports ram tweaking but for that ram I dont seem to be able to change anything

there should be an option to remove the "baseline" DDR4 limit, on the original spec

even on servers

registered ECC is already going to be higher latency, nothing you can do about that

(yeah, I know, not trying to fix my setup)

DataHoarder, do you have deep knowledge in networking ?

Perhaps

What packets will receive DDoS target if there are two attackers with equal outgoing bandwidth and their cumulative bandwidth is much larger than incoming bandwidth of target ?

Would it receive uniformly chosen packets with 50%/50% probability from both attackers ?

or it would be complete disconnect without 0 incoming packets

s/without/with/

that is indeed deep, and I guess we are assuming paths taken by DDoS end up using same peerings on average towards final path, and previous paths don't die to it, let's say all reach

second is probably a no-no as "disconnect" doesn't just happen, packet loss due to lack of memory on what passes packets on to the final target

(that needs to be a programmed action)

can't say about 50%/50%, that gets too deep and from having written some firmware for 100G cards as hobby it still doesn't give me answers

don't know enough there

do you have an answer wfaressuissia, given it is DDoS I would say packets are random from both sides... so there is no single source to favor

I just know some of the code I saw preferred certain sources depending which packets had gotten to it slightly

anyhow, this is kind offtopic for this channel

18:21 <@hyc> sech1: running the api branch off master, getting a bunch of err msgs. pasted in -dev <-- it's unrelated to any PRs

it's just someone changing source code to spam nodes, it's harmless apart from spam

log spam*