mpeg encoding speed with athlon 64 X2

[frogman54]

Can anyone tell me approximately how long it would take an athlon 64 X2 processor to encode a 20 minute video file to mpeg 2 at say 6000 kbps CBR using tmpgenc? I realize this is kind of a specific request but I am trying to decide if it is worth it to buy a new processor and the above is mainly what I use my computer to do. So if anybody has a similar processor and can test this out or already knows I would REALLY appreciate it.

[redwudz]

Depends on the speed of the CPU. If all other factors remain the same compared to a single processor CPU, probably the same speed. The encoder program would have to take advantage of both processors and I don't think TMPGEnc can. But you should be able to run another CPU intensive program at the same time with less problems with the X2, so your multitasking abilities would improve. Much the same as a 64 bit processor VS a 32 bit processor. The programs used have to be written to make the most use of the CPU architecture.

[jagabo]

TMPGEnc Plus is multithreaded and will use both cores. It can encode nearly twice as fast using both processors (depending on the settings used) of my Athlon 64 X2 3800+.

I ran a test with TMPGEnc Plus 2.52.34.129 converting a 60 second, 720x480, 29.97 fps, progressive DV AVI file as the source (handheld camcorder footage). I encoded as 720x480, 29.97 fps, progressive MPEG2 as the output at CBR 6000 with Motion Search Precision set to Normal. Audio was 384 kbps, 48 KHz, MP2.

I used TMPGEnc's multithread option to switch between one thread and two (Option -> Envoronmental Settings -> CPU -> Use Multithread. This isn't exactly the same as comparing a single core CPU to a dual core CPU but it's close.

It took 109 seconds to encode with one core, 49 seconds with both. I was surprised that the 2-core result was more than twice as fast so I ran the test several times to be sure. Same thing every time. For another comparison I ran the same test on a non-hyperthreaded 2.8 GHz Pentium 4. It took 125 seconds.