Re: Current ULE state and plans (was Re: HMP scheduling on FreeBSD)

> On Jan 22, 2026, at 12:59 PM, Minsoo Choo <minsoochoo0122@proton.me> wrote:
> 
> On Thursday, January 22nd, 2026 at 3:44 PM, Olivier Certner <olce@freebsd.org> wrote:
> 
>> 
>> 
>> Hi,
>> 
>>> Is there a set of tests that people can run to measure
>>> important characteristics of schedulers under various
>>> conditions (# of cores, multi-cpu, amoutn of memory,
>>> interactivity, network traffic, NFS etc.)? The same can also
>>> help with testing scheduler changes to see their effect on
>>> these characteristics. And can help with trying out new
>>> scheduling disciplines. And may be such testing can inspire
>>> peopel who want the 4BSD to improve it as well! No synthetic
>>> tests will sufficiently test real world conditions but at
>>> least they can provide a baseline. Better still may be they
>>> can be run against Linux as well!
>> 
>> 
>> Basically, any benchmarks with lots of threads, or very sensitive to latency can do. You can, e.g., use database or networking benchmarks for comparisons. And 'make buildworld', of course.
>> 
>>> Alternative question: what tests are currently used to test
>>> ULE vs 4BSD schedulers and what can be done to improve them,
>>> make them more comprehensive?
>> 
>> 
>> I'm using 'late' from Jeff, which I have changed significantly. At the moment, it's at https://github.com/OlCe2/late. The shell scripts allowing to run synthetic scenarios are quite rough. I've started documenting their purpose and use in some file that I haven't pushed to the repo. Only the main executable has been polished enough (self-documenting). I am supposed to import it into the base system at some point(tm).
>> 
>> Thanks and regards.
>> 
>> --
>> Olivier Certner
> 
> If we want real-world scenarios, phoronix test suite would be great.

Thanks, Olivier & Minsoo. I will look at the repo and look
for the Phoronix tests.

Eventually a comprehensive performance test suite would be
great.  The idea is to have specific repeatable tests for
measuring characteristics of specific features as the code
changes over time.  Over time one should be able to graph how
important characteristics are changing.

For scheduling one can think of measuring latency, fairness,
scalability, variability, realtime responsiveness etc. [Very
very similar to what one might want to measure in networking
scenarios. In fact there is a lot of similarity between OS &
Networking issues!]

Bakul