RandomXv2 official pull request by sech1 tevador/RandomX #317 (shared in #monero-pow)

<user5864:matrix.org> plowsof: wait, +52% more operation mean 52% more performences ?

it's a different pow function

it's not directly equivalent but yeah, it does way more work (when measured in instructions, memory accesses) and at the same time higher hashrate on modern cpus compared to v1

look at github.com/SChernykh/RandomX/blob/v…c/design_v2.md#4-performance-impact benchmarks

look at VM+AES/s, that is absolute work

<user5864:matrix.org> yeah I stumble on that, thanks

<user5864:matrix.org> woah

or relative work/joule to compare

so while older gens might end up slower H/s they also do more work than v1

<user5864:matrix.org> how does more instruction outputed result in lower hashrate

<user5864:matrix.org> this look very good for zen5 but all other seem to receive more bad outlook

cause v2 does 384 program size instead of 256

for example

v2 hash is 1.5x more work than v1 hash

so don't look that v2 hashrate is lower

yeah, modern ones benefit more (we push their memory further)

multiply it by 1.5 first if you really want to compare hashrates

<user5864:matrix.org> hmm

<user5864:matrix.org> so I guess its not about hashrate, but higher share result ?

^ they are pow functions, you can't compare each other directly. compare VM+AES/s or relative work/joule

no

<user5864:matrix.org> hm

it's about work/joule (does more work per joule, aka more efficient)

<user5864:matrix.org> it just so counter intuitive. Hashrate lower but does output more operation

Well, mine Bitcoin if you want more hashrate

cause you are comparing say

bitcoin sha hashrate to randomx hashrate

they do different things

same on v1 to v2

they do "different" things

<user5864:matrix.org> just trying to warp my brain around, im non technical sorry

It's like dollar and euro

Similar things, but not 1:1 rate

imagine this. randomx 2x which is two randomx hashes

v1 hashes and v2 hashes

you have half hashrate

but 2x work

<user5864:matrix.org> ahh > <DataHoarder> they do different things

this is why you can't compare hashrate directly

<user5864:matrix.org> I see. Thanks for the explanation :))

in this case it's like randomx 1.5x :') but with more specific changes that CPUs do well

<user5864:matrix.org> you think it imght goes live next couple weeks ?

it needs a hardfork

<user5864:matrix.org> ohh. So I guess might came with fcmp one

<user5864:matrix.org> or around

<user5864:matrix.org> exciting times

<user5864:matrix.org> appreciate the hard work put behind everything of this

so modern CPUs are better but I can't build a new rig cause ram has 4x in price :(

If you undervolt and optimize for efficiency, a single RAM stick will do. 1 stick can handle up to 20-21 kh/s

you can use laptop memory :')

hmmm

<user5864:matrix.org> nioc: good news is, ddr5 didnt affected that much the hashrate, on V1 at least

<user5864:matrix.org> so now 1 kit of ram can become 2 rigs

<user5864:matrix.org> should be interesting thing to test it out also, how ram effect hashrate on Zen5 with V2

Zen5 is still capped on bandwidth

<user5864:matrix.org> so it shouldnt change ?

I have slow ram on a zen5 platform and it was also equivalent. The prefetch trick makes latency tuning less critical afaik (but you still benefit)

In my own implementation, it was like +2 KH/s over V2 without the memory prefetch

Without that there were no large differences across generations

<user5864:matrix.org> thanks for the insight

[CCS Proposals] Lee Clagett opened merge request #637: Mark 2025 Q4 Month 1 Completed repo.getmonero.org/monero-project/ccs-proposals/-/merge_requests/637