[CinCV TNG] Cinelerra CVE - disabled multithreading in codecs?

Einar Rünkaru einarrunkaru at gmail.com
Mon Jul 17 23:13:43 CEST 2017



On 07/17/2017 11:24 PM, Andrew Randrianasulu wrote:
> В сообщении от Monday 17 July 2017 22:55:26 Einar Rc3bcnkaru написал(а):
>> On 07/17/2017 08:21 PM, Andrew Randrianasulu wrote:
>>> В сообщении от Monday 17 July 2017 14:53:59 Einar Rc3bcnkaru написал(а):
>>>> On 07/17/2017 03:24 AM, Andrew Randrianasulu wrote:
>>>>> Hello again!
>>>>>
>>>>>
>>>>> I just compiled and installed Cinelerra-CVE
>>>>>
>>>>> Cinelerra-CVE 2.3.5
>>>>> (C) 2006 Heroine Virtual Ltd.
>>>>> (C) 2006-2017 The CinelerraCV Community
>>>>> Internal ffmpeg library versions:
>>>>>        libavutil      54. 31.100 / 54. 31.100
>>>>>        libavcodec     56. 60.100 / 56. 60.100
>>>>>        libavformat    56. 40.101 / 56. 40.101
>>>>>        libswscale      3.  1.101 /  3.  1.101
>>>>>        libswresample   1.  2.101 /  1.  2.101
>>>>> Compiled on Пн июл 17 01:43:25 MSK 2017
>>>>>
>>>>>
>>>>> and loaded very same mjpeg files I used to test Cinelerra-CV.
>>>>> They can be played - but even decoding in CVE is slow, with video
>>>>> output device set to x11_xv I only get 7-8 fps. (1.4 Ghz * 4 core, but
>>>>> cores NOT maxed out during decoding)
>>>>
>>>> The GHz is quite pointless number during last 15 years. I had several
>>>> computers during this time all about 2GHz. Always the next had 1.5 - 2
>>>> times better speed.
>>>
>>> sorry, here is cat /proc/cpuinfo
>>
>> According to
>> https://www.cpubenchmark.net/cpu.php?cpu=AMD+FX-4300+Quad-Core it is a
>> 3.8 GHz CPU. cpuinfo shows the current cpu MHz
> 
> And this about to be 'fixed' in 4.13+:
> 
> http://lkml.iu.edu/hypermail/linux/kernel/1707.0/01366.html
> 
> - Stop trying to export the current CPU frequency via /proc/cpuinfo
>   on x86 as that is inaccurate and confusing (Len Brown).
>   
> To be honest, I think new behavior will be more confusing for me, but .....
> 

Modern cpus can change their MHz very rapidly. You run something and try 
to get MHz, but it is already changed.. When you try to compute 
something based on this - you get very confusing results.

Bencmarking of computers is very complex, benchmarking of modern 
computers is even more complex.

Einar


More information about the Cinelerra mailing list