New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Chasing the highest GB5 scores is basically useless for most web apps
I have always looked for servers with the highest GB5 scores, thinking that it would make a significant difference.
But since I started to use a VPS with a GB5 score or just 585 I realized that for most webapps there is no practical difference in response times if you have a fast disk for databases and plenty of ram. On my SSDNodes VPS I have 22 apps of different types and I don't see any difference at all in performance even compared with other servers with scores up to 1200-1300 single core.
Do you see an actual difference with different GB5 scores with regular webapps?
Comments
wait till you get one with 2,000.
It depends as what you define as "regular webapps" I guess. I do see a big difference on the apps I self host.
I wont idle on anything less than a server that can pull 3000 gb5
And how much traffic you get in all of these web apps to conclude that CPU benchmark doesn't matter?
gb5 is outdated! gb6 is the deal!
Different types of apps. At the moment:
Good point. I was referring to self hosted apps for myself, so I haven't tried this kind of server with something that receives a lot of traffic. But comparing this server with more expensive ones with the same apps and usage I don't see any practical difference.
Wow. Which provider?
I always look at GBL when I'm trying to "score".
Why do you expect any difference between CPUs if it needs to generate one page per second maximum then?
If there would be any difference (maxing out single core per page generation) your server would be completly overloaded with like 10 visits per second.
Good shared hosting can push 200 visitors per second easily and youre wondering why you dont see difference on beefy VPSes for load that is 200x smaller.
You would be fine even on 10 year old 1vCPU with that kind of usage.
GB5 score is meaningful, its just that you need couple% of these CPUs, so even if CPU is 4x worse it wont matter, as resources are available.
Just for context: People are hosting 100 Wordpress sites on 2GB of ram & 1vCPU. And it works fine if there is not much visitors. You generate pages when visitors come, not generate just because it runs.
I think one place where GB5 single core performance is noticed, is on heavy WordPress/WooCommerce backoffice navigation.
There i noticed some diferentes in speed when using a server with high GB5.
Most web apps are IO limited. If you're developing on something lightweight like Node, you can handle thousands of concurrent connections on a single core without issue so you'll pretty much never need to worry about CPU. By the time CPU speeds actually matter, you'll probably be rich enough to afford good servers anyway.
There's the issue - no or little traffic. Start testing with traffic - in particular concurrent user traffic loads and see. Also, additionally start testing backend stuff in background or foreground tasks i.e. multi-threaded backup/compression and encryption/cryptographic tasks, including HTTP/2 HTTPS on the server combined with traffic usage, and you'll see a difference.
But yes Geekbench score isn't indicative for all usage test cases.