From: Marko Ristola <marko.ristola@kolumbus.fi>
To: Emmanuel <eallaud@gmail.com>
Cc: VDR User <user.vdr@gmail.com>, linux-media@vger.kernel.org
Subject: Re: [Q]: any DVB-S2 card which is 45MS/s capable?
Date: Tue, 03 Aug 2010 19:03:12 +0300 [thread overview]
Message-ID: <4C583DC0.9090405@kolumbus.fi> (raw)
In-Reply-To: <4C578AFF.2030804@gmail.com>
03.08.2010 06:20, Emmanuel wrote:
> Marko Ristola a écrit :
>>
>> Hi.
>>
>> I have written some Mantis bandwidth related
>> DMA transfer optimizations on June/July this year.
>> They are now waiting for approval by Manu Abraham.
>>
>> Those reduce CPU pressure, increasing the bandwidth
>> that can be received from the DVB card. 45MS/s bandwidth
>> support will surely benefit from those patches.
>>
>> Main features:
>> 1. Do one CPU interrupt per 16KB data instead per 4KB data.
>> My implementation benefits only Mantis cards.
>> https://patchwork.kernel.org/patch/107036/
>>
>> 2. Remove unnecessary big CPU overhead, when data is delivered
>> from the DVB card's DMA buffer into Linux kernel's DVB subsystem.
>> Number 2 reduces the CPU pressure by almost 50%.
>> This implementation benefits many other Linux supported DVB cards too.
>> http://patchwork.kernel.org/patch/108274/
>>
>> Those helped with my older single CPU Core computer with 256-QAM,
>> delivering HDTV channel into the network and watching the
>> HDTV channel with a faster computer.
>>
>> The performance bottlenecks could be seen on the
>> command line with "perf top".
>>
>> I had to increase PCI bus latency setting into 64 too from the BIOS.
>> Moving DVB device into separate IRQ line with Ethernet card helped too,
>> because Ethernet card did an interrupt per ethernet packet.
>>
>> So if the hardware can deliver 45MS/s data fast enough, software side
>> can be optimized up
>> to some point. My patches contain the most easy major optimizations
>> that I found.
>> If you can test some of those patches, it might help to get them into
>> Linux kernel
>> faster.
>>
>> Best regards,
>> Marko Ristola
>>
> OK these optimizations look like a step into the good direction. I
> guess what is also missing is a tuner which can handle that in DVB-S2,
> which does not seem obvious. The mantis card can do that?
According to Internet shops it can.
Manu Abraham has implemented the DVB-S2 support earlier than this summer,
so with my small improvements it could perform better.
Here is a page about DVB-S2 cards that work somehow with Linux:
http://www.linuxtv.org/wiki/index.php/DVB-S2_PCI_Cards
I myself have been testing only with 9600 MS/s with terrestial DVB-C,
so I have no experience with DVB-S2.
With 9600 MS/s and HDTV, the whole system must work, including VDR with
it's components,
not only the PCI DVB card.
With SDTV, any hardware will do.
Regards,
Marko Ristola
> Thx
> Bye
> Manu
>
prev parent reply other threads:[~2010-08-03 16:03 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-07-25 14:17 [Q]: any DVB-S2 card which is 45MS/s capable? Emmanuel
2010-07-26 4:14 ` Goga777
2010-07-26 19:01 ` Emmanuel
2010-07-26 21:03 ` VDR User
2010-07-27 12:52 ` Emmanuel
2010-07-27 15:11 ` VDR User
2010-08-02 13:55 ` Marko Ristola
2010-08-03 3:20 ` Emmanuel
2010-08-03 16:03 ` Marko Ristola [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4C583DC0.9090405@kolumbus.fi \
--to=marko.ristola@kolumbus.fi \
--cc=eallaud@gmail.com \
--cc=linux-media@vger.kernel.org \
--cc=user.vdr@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.