For my needs please

[bbadalucco]

Just to follow-up on this. Stopped by my neighbors the other night as he has a 2009 2.26 Octo. I installed the latest build of handbrake and did a test run of 2 blu ray (rips) transcodes. He was saying that on average (running one at a time) they are taking around 5:15. Average FPS with the blu ray were around 9 and 77 for DVD.

I know I'm my Macbook Pro it was taking well over 10 hours so I'm not sure how I feel about it.

He did say that CPU usage was only around 40%

Settings in Handbrake:
h264
24 FPS
60% Constant quality
Output 1920x1080

Sounds like the 3.2 would be faster,anyone have an idea how long this would take on a 3.2 octo? I think I'll pull the trigger Monday.

General thoughts?

 

[gugucom]

My x5365 based 2006 octad does that in under 3h and loads the cores near 100%. It has no hyper threading but uses an SSD RAID0 boot array. The 2008 3,2 octad with two x5482 CPUs has the same Core CPU architecture just with 45nm manufacturing width instead of the 65 nm of the Clovertown . So it will simply be faster and behave the same way.

The 2009 machines have the new Nehalem CPU architecture which re-introduced hyperthreading that was not applied to the previous Core architekture. So the 2009 octads all have 16 virtual threads which handbrake does not recognize in the current version. It only supports 8 cores. This is the explanation of the poor performance of the Nehalem machines.

For facts on Xeon CPUs http://en.wikipedia.org/wiki/Xeon

 

[bbadalucco]

3.2 octad....ordered

thanks

 

[bbadalucco]

In Octad mode the memory becomes dominant though. The 2009 2,66 Octad machine beats the 2008 3,2 Octad machine because the higher core frequency gets beaten by the memory bandwidth issue.

Just curious, are you saying the 2.66 Octad would be faster than the 3.2 Octo (2008) when using handbrake? Or just in general?

 

[Mackilroy]

Wirelessly posted (Mozilla/5.0 (iPod; U; CPU iPhone OS 3_0 like Mac OS X; en-us) AppleWebKit/528.18 (KHTML, like Gecko) Version/4.0 Mobile/7A341 Safari/528.16)

I think the 2009
machine would generally be faster, but not necessarily on HandBrake.

 

[gugucom]

Just curious, are you saying the 2.66 Octad would be faster than the 3.2 Octo (2008) when using handbrake? Or just in general?

I was going by the cinebench 10 benchmark without considering some special aspects of handbrake in hyperthreading at that time. If handbrake did not have the hyperthreading bug - yes - the 2009 2.66 Octad would beat the 2008 3.2 Octad. Considering the hyperthreading bug I would not stay with that opinion. So it really is an issue of the handbreak programming if they release a version that supports 16 virtual cores.