SMP: need help to understand the results

[Lawrence74]

Hello.

Just read the test on Tom's Hardware

http://www6.tomshardware.com/cpu/02q1/0203131/dual-06.html

that compares Intel and AMD bi-processor
benchmarks results.

Could anyone help me to understand the graphics published on the site ?

I know the SMP computers obtain better results, but I don't understand why the speak about frames per second ?
What's the real difference between single and bi-processor computers, if the results are based on time used for encoding (eg:
100 mn instead of 115 mn) and not on frames ?

Thanks for your help because I'm a bit lost.

[khp]

OK lets do the math for a 100 minute movie

100 minutes equals

100*60*25 = 150000 frames, assuming that you are converting a pal movie.

so if you can convert 33.08 frames per second with a AthlonXP 2000, it would take 150000/33,08 = 4534.46 seconds to complete one pass.

[Lawrence74]

Thanks for the previous answer.

Just another question: would I have to use softwares optimized for working with two processors ?

eg: I use Virtual Dub to encode, using Divx 4.12. Would it work as fast as the benchmarks say, or slower ?

[Batman]

Dual processing, should speed up the process. I know, dual processors help in TMPGEnc.

[khp]

Originally posted by Lawrence74
eg: I use Virtual Dub to encode, using Divx 4.12. Would it work as fast as the benchmarks say, or slower ?

Well the divx3.11/4.12/5.0.1 codec itself can only use 1 processor, but because you also need to decode the source material and do resize and filtering, a dual processor system usually gives a very nice speed boost.

When frameserving a DVD with avisynth into vdub, you should be able to achive the same or better results than the THG test show.