More cores mean faster video editing?

Rafael Mestdag

Reputable
Mar 25, 2014
1,442
1
5,460
I mean, I have a very fast 4 core i5 3570K but I get the impression that my older and a lot slower in general FX 6300 with its 6 weaker cores was faster for video decoding/editing.

Is this true? Why?
 
Solution


You cannot go by core count alone.
That is like saying...my truck engine has more cylinders than my wifes car engine, so it is faster.


The overall architecture is what matters.

USAFRet

Titan
Moderator


You cannot go by core count alone.
That is like saying...my truck engine has more cylinders than my wifes car engine, so it is faster.


The overall architecture is what matters.
 
Solution

Rafael Mestdag

Reputable
Mar 25, 2014
1,442
1
5,460


And how can I tell which one of these processors is faster at video editing?
 

USAFRet

Titan
Moderator


You test. Just like every other "benchmark" or performance site does.

Last year, when I changed from an i5-3570k/16GB RAM to an i7-4790k/32GB RAM, I wanted to know how much faster it was, doing things that I normally do.
So I set up a test, of a repeatable operation.
100 random RAW images out of my camera, importing into Adobe Lightroom 5. ~32MB each, so ~3.3GB total.
How long does it take, with the exact same set of files.

2 min 33 sec(old i5) vs 1 min 55 sec(new i7).
About 35% faster. Quantifiable numbers.

Claims of "It's faster!!" or random benchmarks from dodgy sites mean absolutely nothing.
Show me, in your particular workflow...which is faster, and by how much.
 

Rafael Mestdag

Reputable
Mar 25, 2014
1,442
1
5,460


Unfortunately I haven't got both machines with me, I part exchanged the fx 6300 for the i5 3570K so that's all I have now.
 

USAFRet

Titan
Moderator


So then all you currently have is "I get the impression".
 

4745454b

Titan
Moderator
And the impression doesn't match what tests show. I can't seem to find Toms 6300 review (?), so I picked another sites review at random.

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/57615-amd-vishera-fx-6300-fx-4300-review-9.html

You mentioned converting video, so I'm pointing out their two pics. X264 and mediacoder 64. In this first link, the more frames you can process per second, the faster the video will encode. Here, even the older 2500 can match the 6300, while the newer 3570K passes it. Heck, even the way old "Stars" arch found in the X6 1100T is faster than the 6300.

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/57615-amd-vishera-fx-6300-fx-4300-review-7.html

Moving on to the Mediacoder 64 program, it's more of the same. Here it's a timed test to convert a 600MB file into an Mpeg4 file. The 2500K is mildly faster, while the 1100T and 3570K are much faster. Rather 37 seconds faster. They might be "4 core" CPUs, but the 3570K still is much more powerful than the 6300 at converting videos. Probably due to the backwards step in IPC FX has compared to Stars.
 

Rafael Mestdag

Reputable
Mar 25, 2014
1,442
1
5,460


It's interesting to see on that last 1080p encoding chart that the FX 8350 with its 8 weak cores is a lot faster than the i5 3570K with its way stronger 4 cores. If the 3570K wasn't so much better at gaming than the FX's I would rather have the 8350.
 

USAFRet

Titan
Moderator
That's the thing about PC's...people rarely use them for only one thing.

Gaming, video rendering, streaming, watching videos, writing a resume, CAD, database, image manipulation, animation, Excel, etc, etc, etc.

You want a balanced system. Anything beyond that is nitpicking and fanboyism.
 

Rafael Mestdag

Reputable
Mar 25, 2014
1,442
1
5,460


That's right, and I would have an FX 8350 just for video editing alongside my 3570K if I could. But I'm primarily
a gamer so...