Quantcast
Channel: Raspberry Pi Forums
Viewing all articles
Browse latest Browse all 7605

General discussion • Re: Computer benchmarks can be useless

$
0
0
There are two other nice ones BUT one is full on
https://www.phoronix-test-suite.com/ which goes to
OpenBenchmarking.org
But you pick specific test or a suit of tests.
You want file access in database ,you have it. Etc.

Luckily they have benched a lot for you.
You pick specific ones to run.

Then also if you happen to be a creator with davinci, or Adobe products.https://www.pugetsystems.com/pugetbench/creators/
My impression is the Phoronix test suite indicates what happens when you compile open source without paying attention to performance tuning. While repeatable, the results may not indicate how hardware would work in production when set up by competent engineers.

As an example, simple tuning can affect the speed of some software by a factor of three.

https://sthbrx.github.io/blog/2018/08/1 ... on-power9/

I've seen similar performance problems when a Ryzen Pro 4650 was not identified by a certain software library which then switched to a slow compatibility mode. This made the 4650 seem significantly slower than an older 1700.

Statistics: Posted by ejolson — Mon May 12, 2025 7:48 pm



Viewing all articles
Browse latest Browse all 7605

Trending Articles