⚖️ RobertBunn.eth (BA, MA, JD)

Superb evening! #hyperscale test stressing out the latest changes. We hit a worthy milestone of 10k+ swaps/s with just 4 shards! Network was approx 80 validators from community, spread around the globe, variety of hardware inc laptops on battery over wifi😂 No machine consumed more than 40% of CPU & 8GB RAM during the 3 hours test of 35MILLION swaps! To put that into perspective, Solana swaps/s max 1200/s, in practice problems have occurred at around 700 just recently with TRUMP launch. A single Hyperscale shard is sustaining 3x that load on hardware specs a fraction of a Solana validator. Finality time remaining healthy at 5-6 seconds throughout the test, half of Solana's. While we were only drawing about 30% of the available performance in this network, 10k+ swaps/s is significant milestone. I'm not aware of any network demonstrating it either in main net, or a properly distributed test net.

Ok, go benchmark a ECDSA secp256k1 algo & see what results you get. Single core can easily handle 10000+/s. With aggressive optimizations 20000+/s. Grab the extreme, ultimate form, optimized Bitcoin secp lib and you're doing 40000+ Second "RADDIX" isn't Hyperscale Alpha. HA also uses BLS and before you exclaim that BLS is slow, it doesn't use it in the primitive way your thinking. Verifying a BLS signature IS slow, perhaps 500/s ... but you can aggregate signatures at approximately 150,000+/s PER CORE because its just a point field addition. So with an architecture which is designed for that property, you can actually validate 150+k/s BLS signatures & messages.

Any #radix projects that wanna pitch, DM me. First come first served. 2 minute pitch, 5 min Q&A If you're accepted to grants program gonna kindly say no, idea is for lesser known projects to get exposure 💪