[MEncoder-users] Mencoder and PVM support

Wed Sep 21 23:47:29 CEST 2011

Just an uneducated question: Wouldn't the network overhead be prohibitively
expensive given the amount of work that is assigned to a process before it
has to sync up with other processes again?

-----Original Message-----
From: mencoder-users-bounces at mplayerhq.hu
[mailto:mencoder-users-bounces at mplayerhq.hu] On Behalf Of Markus Mohr
Sent: Wednesday, September 21, 2011 4:09 PM
To: mencoder-users at mplayerhq.hu
Subject: Re: [MEncoder-users] Mencoder and PVM support

Just i worry about the availability of PVM within new Software.
Within clustering i can read sometimes about PVM, other times about MPI
(similar project but more actual). 
On the other hand i found statements complaining about better writing an own
system for communicating since PVM/MPI had too much overhead.

But PVM is now rather old and literature seems to be old as well (latest is
2005, official PVM-Book), so i worry heavily about current developments.
Actually i asked at the "handbrake"- forums as it was my favorite standalone
encoder, but they said their multithreading system was really fit to their
needs, pvm changes would be hard to make. Especially since i dont know about
it yet, i cannot say I am deep into programming...
Also there was a "transcode" project including pvm. But it - and its version
was also from 2003.
so i worry if there could be problems by including pvm to actual software.

Am 21.09.2011 12:00, schrieb mencoder-users-request at mplayerhq.hu:
> Send MEncoder-users mailing list submissions to
> 	mencoder-users at mplayerhq.hu
>
> To subscribe or unsubscribe via the World Wide Web, visit
> 	https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
> or, via email, send a message with subject or body 'help' to
> 	mencoder-users-request at mplayerhq.hu
>
> You can reach the person managing the list at
> 	mencoder-users-owner at mplayerhq.hu
>
> When replying, please edit your Subject line so it is more specific 
> than "Re: Contents of MEncoder-users digest..."
>
>
> Today's Topics:
>
>     1. Re: Mencoder and PVM support (The Wanderer) (Markus Mohr)
>     2. Re: Mencoder and PVM support (Carl Eugen Hoyos)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Tue, 20 Sep 2011 22:26:00 +0200
> From: Markus Mohr<m.mohr at gmx.de>
> To: mencoder-users at mplayerhq.hu
> Subject: Re: [MEncoder-users] Mencoder and PVM support (The Wanderer) 
> Message-ID:<4E78F6D8.9000305 at gmx.de>
> Content-Type: text/plain; charset=ISO-8859-15; format=flowed
>
> That is correct. PVM would be the ability to encode one task one multiple
machines.
> Problem is (for me) that pvm-project is rather old and probably 
> difficult to implement. so i want to check if that is even interesting to
invest time in. if not then reasons would be interesting.
>
> Thank you!
>
>
> Am 20.09.2011 12:00, schrieb mencoder-users-request at mplayerhq.hu:
>> Send MEncoder-users mailing list submissions to
>> 	mencoder-users at mplayerhq.hu
>>
>> To subscribe or unsubscribe via the World Wide Web, visit
>> 	https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
>> or, via email, send a message with subject or body 'help' to
>> 	mencoder-users-request at mplayerhq.hu
>>
>> You can reach the person managing the list at
>> 	mencoder-users-owner at mplayerhq.hu
>>
>> When replying, please edit your Subject line so it is more specific 
>> than "Re: Contents of MEncoder-users digest..."
>>
>>
>> Today's Topics:
>>
>>      1. Re: Mencoder and PVM support (Carl Eugen Hoyos)
>>      2. Re: Mencoder and PVM support (The Wanderer)
>>      3. Re: Audio Sync issues (Jeff Cranmer)
>>      4. Re: Audio Sync issues (Rolf Ernst)
>>
>>
>> ---------------------------------------------------------------------
>> -
>>
>> Message: 1
>> Date: Mon, 19 Sep 2011 10:36:30 +0000 (UTC)
>> From: Carl Eugen Hoyos<cehoyos at ag.or.at>
>> To: mencoder-users at mplayerhq.hu
>> Subject: Re: [MEncoder-users] Mencoder and PVM support 
>> Message-ID:<loom.20110919T123543-243 at post.gmane.org>
>> Content-Type: text/plain; charset=us-ascii
>>
>> Markus Mohr<m.mohr<at>   gmx.de>   writes:
>>
>>> I have a project to confirm. I want to add PVM-support to Mencoder
>> You should probably add what PVM is (and why you don't want to 
>> implement it in FFmpeg if it is a media format).
>>
>> Carl Eugen
>>
>>
>>
>> ------------------------------
>>
>> Message: 2
>> Date: Mon, 19 Sep 2011 10:30:26 -0400
>> From: The Wanderer<wanderer at fastmail.fm>
>> To: MEncoder usage discussions<mencoder-users at mplayerhq.hu>
>> Subject: Re: [MEncoder-users] Mencoder and PVM support 
>> Message-ID:<4E775202.6080801 at fastmail.fm>
>> Content-Type: text/plain; charset=UTF-8; format=flowed
>>
>> On 09/19/2011 06:36 AM, Carl Eugen Hoyos wrote:
>>
>>> Markus Mohr<m.mohr<at>   gmx.de>   writes:
>>>
>>>> I have a project to confirm. I want to add PVM-support to Mencoder
>>> You should probably add what PVM is (and why you don't want to 
>>> implement it in FFmpeg if it is a media format).
>> According to VERA, PVM is "Parallel Virtual Machine". If that's what 
>> he's talking about, then he probably means the ability to process a 
>> single task - such as an encode - on multiple machines, treating them 
>> all as one big, heterogenous, high-communication-latency multi-core
processor.
>>
>
> ------------------------------
>
> Message: 2
> Date: Tue, 20 Sep 2011 23:22:58 +0000 (UTC)
> From: Carl Eugen Hoyos<cehoyos at ag.or.at>
> To: mencoder-users at mplayerhq.hu
> Subject: Re: [MEncoder-users] Mencoder and PVM support 
> Message-ID:<loom.20110921T011609-482 at post.gmane.org>
> Content-Type: text/plain; charset=us-ascii
>
> The Wanderer<wanderer<at>  fastmail.fm>  writes:
>
>> On 09/19/2011 06:36 AM, Carl Eugen Hoyos wrote:
>>
>>> Markus Mohr<m.mohr<at>  gmx.de>  writes:
>>>
>>>> I have a project to confirm. I want to add PVM-support to Mencoder
>>> You should probably add what PVM is (and why you don't want to 
>>> implement it in FFmpeg if it is a media format).
>> According to VERA, PVM is "Parallel Virtual Machine".
> And I was absolutely sure he wants to re-encode HP Photo Album files...
>
>> If that's what he's
>> talking about, then he probably means the ability to process a single 
>> task - such as an encode - on multiple machines, treating them all as 
>> one big, heterogenous, high-communication-latency multi-core processor.
> I implemented this twice in the past for a rendering system, once 
> using "Simple Sockets Library", once the remote object facilities of 
> GNUstep. At the time, hardware was so slow that transferring sources 
> and a rendered frame via network had little impact on the overall 
> performance. I do not believe this is true now with current hardware 
> and the requirements of real-world encoding (B-frames come to mind).
>
> If this is done, it should certainly be implemented in FFmpeg, and 
> please test actual multithreaded performance and the inter-process 
> communication requirements first.
>
> Carl Eugen
>
>
>
> ------------------------------
>
> _______________________________________________
> MEncoder-users mailing list
> MEncoder-users at mplayerhq.hu
> https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
>
> End of MEncoder-users Digest, Vol 85, Issue 15
> **********************************************
>
_______________________________________________
MEncoder-users mailing list
MEncoder-users at mplayerhq.hu
https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users