[FFmpeg-user] FFmpeg & hashing algorithm speed

[Peter B.]

Thanks for the infos!

And "yes": My PHP/Python examples were not related to FFmpeg, but this 
discovery led me to become curious how FFmpeg does it - and therefore 
ask this question :)

Will read the Giorgio-Thread and maybe return with reproducible 
(commandline + uncut console output) question scenarios.



Thank you very much!
Pb


On 07/11/2019 00:21, Carl Eugen Hoyos wrote:
> Am Do., 7. Nov. 2019 um 00:09 Uhr schrieb Carl Eugen Hoyos <ceffmpeg at gmail.com>:
>> Am Mi., 6. Nov. 2019 um 23:40 Uhr schrieb Peter B. <pb at das-werkstatt.com>:
>>
>>> Hashing for fixity and transfer-checking seems to be quite a bottleneck
>>> for archiving big video/film files for archiving.
>>>
>>> Therefore I was considering if using "simpler" algorithms (like CRC
>>> instead of SHA256) could speed up the process.
>>> I was surprised that e.g. MD5 vs CRC in Python as well as PHP
>>> seemed to be equally "fast".
>> But this is not related to FFmpeg, no?
>>
>>> So I was wondering:
>>> Does FFmpeg have/do any speed optimization for hashing functionality
>>> (framemd5, hash muxer, etc) in its code?
>> Some of Giorgio's commits were meant to increase hash function
>> speed:
>> http://git.videolan.org/?p=ffmpeg.git;a=search;s=vazzana;st=author
>> There is an email with a script to compare the implementation's speeds,
>> I don't know when it was sent to the development mailing list though.
> This mail leads to a long thread:
> http://mplayerhq.hu/pipermail/ffmpeg-devel/2013-May/143854.html
> In the end a file tools/ffhash.c was created to compare hash
> implementation speeds.
>
> Carl Eugen
> _______________________________________________
> ffmpeg-user mailing list
> ffmpeg-user at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-user
>
> To unsubscribe, visit link above, or email
> ffmpeg-user-request at ffmpeg.org with subject "unsubscribe".