linux-kernel - 0.33 ms latency performance bottlenecks

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date:   Tue, 19 May 2020 15:45:27 +0200
From:   Ywe Cærlyn <Ywe.C4rlyn@...-Shed-Studio.eu>
To:     linux-kernel@...r.kernel.org
Subject: 0.33 ms latency performance bottlenecks

I tried Renoise, a good tracker app with 0.33 ms latency. It was really 
good. Tech Trance in the 90s was all about tracking.

( a little playlist here for interested 
https://www.youtube.com/playlist?list=PLpA7__w8yeJ0g6cF0m_y_h-HWjzCNTuWm )

Latency on the amiga at that time was very low. More like 0.33 ms 
latency, or even 0.2.

We want this kind of latency again. Also having in mind a few samples 
for clocking the D/A well if necessary.

D/A could be a 1bit converter with few analogue components on the 
output, maybe simply a 1 pole diode filter, if possible. This could work 
well on a bitstream shifted up in frequency.

It seems to use almost a whole cpu just at this latency. It works well 
though.

One might want to look into performance bottlenecks here, hardware aid, 
or cpu instruction set enhancements. Inner loop optimization beyond HZ 
timer etc, which is at 91 here.

Serene Greetings,
Ywe Cærlyn
I-T Shed Studio.