[Performance regression] BCM4359/9 on S905X2
Jerome Brunet
jbrunet at baylibre.com
Tue Mar 21 03:46:57 PDT 2023
On Tue 21 Mar 2023 at 11:40, Marc Gonzalez <marc.w.gonzalez at free.fr> wrote:
> Hello everyone,
>
> I've been benchmarking an Amlogic S905X2 board.
> It provides a BCM4359/9 WiFi chip connected through SDIO.
>
> There's a large performance gap between vendor kernel and mainline.
> (Downloading a 1GB file to /dev/null from a device inches away)
The title is misleading. You are comparing different sources. This is
not a regression. This is merely a difference.
If we are talking about mainline, then which board is it ? What is the
corresponding DT ? What is the MMC configuration in both case ? Have
you checked you are running with the same clock configuration to begin
with ?
>
> # curl -o /dev/null http://192.168.1.254:8095/fixed/1G
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 100 1024M 100 1024M 0 0 27.5M 0 0:00:37 0:00:37 --:--:-- 28.6M
> vs
> 100 1024M 100 1024M 0 0 11.0M 0 0:01:32 0:01:32 --:--:-- 11.0M
>
> Line 1 = vendor kernel (4.9.180 amlogic android)
This cannot help indentify a downstream kernel, and the infinite number
of forks and patches associated with it.
> Line 2 = mainline kernel (6.2.0-rc8)
>
> Why is the vendor kernel 2.5 times faster?
>
> (I'm using the same firmware files, but it seems the vendor kernel reads
> an additional configuration file that the mainline vendor seems to ignore.)
>
> Regards
More information about the linux-amlogic
mailing list