johnwoods

Wow, this looks incredible! It looks like the release post has changed significantly since I saw and shared it.

Looking forward to the benchmarks and 1.0!

Congratulations to the Godot team! I have used Godot a lot in the past and can't wait to replace my Unity junk.

Alright looks like a proper release post was made. Ignore this one.

Benchmarks here: https://github.com/ohsayan/sky-benches

Apart from your buzzword marketing, what features are actually available for use today? Most are "on the roadmap." Reminds me of Spacedrive which literally launched with a "nice UI" and no actual features that they advertise for.

That's a hell lot of buzzwords (GPT-3 trained with HN titles, eh?). Does "serverless" and "cloud-native" database together even make any sense? Idk, but I feel like these folks are doing too much pointless marketing when the core product literally just "borrows" code from other databases without any credit to them.

Came across this rather interesting issue on their own repo: https://github.com/surrealdb/surrealdb/issues/103

I think it's time we say "Stable Diffusion Powered Hyperfast Scalable Ultimate Cloud-Native Kubernetes Native Document Graph Relational NewNoSQL Database"

Enough of the buzzwords on HN, honestly. Also, NOT FOSS.

if I'm doing my math right: redis is doing 99892 sets per sec, so 25X should be 2,497,300. the video however shows 1,023,030 which is more like 10X and NOT 25X

please note that I ran the bench with 30 threads with 1 connection per thread. you're not doing the same but doing 28 threads with 20 connections on each thread, effectively using 560 connections whereas I only used 30 connections. rerun the same with -t 30 and -c 1

agree with your article entirely. the dragonfly benchmark seems far too artificial to be true and used in practice (given the expensive hardware). my tests were on a relatively cheaper box, and Skytable topped all the other systems.

not if I can get equivalent (if not better) performance on my existing hardware

> My old friend Memcached just came by and want to say Hello. He is still alive and kicking. Netflix has been treating him well.

Forgot my dear buddy. Added it back.

updated the post with the overall result for now

updated the post. I ran the benches that were mentioned here[1] and here[2].

[1]: https://github.com/dragonflydb/dragonfly#benchmarks [2]: https://github.com/ohsayan/sky-benches

updated the post

reminds me about the need for AI/ML in random things. maybe it's to sell the "powered by AI/ML" tag. just another hypestorm

It's funny that so many startups are led to thinking that they "need" K8s

As noted by another HN user, it's clear that this doesn't do anything interesting on Linux. Also, higher codegen units will produce slower code at times. Overall, on Linux I do not notice any significant improvement in build times (we already use sccache)

Looks like it was released just a few hours back. After reading the post[1] it seems that this is intentional.

Anyways, I guess we could open a feature request.

[1]: https://ohsayan.github.io/2022/02/jotsy-a-self-hosted-notes-...