i5 vs i7 - Final Cut Pro X

[derohan]

Hello fellow Apple people

I'm sitting here behind my wonderful machine (which is in the process of transcoding & rendering) wondering how my 1.8Ghz i5 would fair vs, let's say the 2.0 Ghz i7, and even perhaps a rMBP if anyone is aware?



361% CPU Usage EY EY EY





*THREAD STARTER VIRGINITY GONE*

 

[simsaladimbamba]

And i7 would not be much faster if you talk about dual-core i7s. If it would be a quad core, then that would be different.

 

[Mrbobb]

I have this idea, and if anybody end up patenting it, am gonna kill myself.

Why limit yourselves to the cpu in front of you.

I bet, right now, there are million of idling computer somewhere, and they are all hooked up to the Internet. Talk about MASSIVELY PARALLEL processing. Kinda like the SETI plug-in, give pieces of your work to remote cpus.

I want my laptop able to talk to those millions of idling cpus out there and, in a cooperative fashion, speed up my work, seamlessly (the keyword).

Doable?

 

[simsaladimbamba]

Doable?

If everyone has fiberoptic network connections, probably, but it also depends on the kind of data you want to have computed. SETI is mostly "text", thus it is relatively small, but video data is large, even the highly compressed data.
And the OP has shown use the CPU activity during the transcoding process of the source files and/or the rendering of timeline effects, which would mean, that your internet connection and that of the parallel clients has to be fast, as even one second of 1080p video can mean 4 to 90 MB/s of data.

 

[blesscheese]

And i7 would not be much faster if you talk about dual-core i7s. If it would be a quad core, then that would be different.

^ What he said. Dual-core i7's aren't much better at heavy-duty hefting than i5's. A quad core is what you would want, but I don't think Apple sells them in anything less than a 15" model.

Don't feel bad...your MBA is still doing a great job!

 

[derohan]

I have this idea, and if anybody end up patenting it, am gonna kill myself.

I remember reading (or dreaming) about OS X Server being able to share one computer's workload throughout it's network

So let's say an office was configured with a variety of iMacs, Mac Minis and so on, and I were to be working on one, it would distribute the tasks at hand to the other idling computers in that network (if they had been configured correctly)

Gigabit ethernet Yum Yum

A quad core is what you would want

Oh, I'm missing my i5 Mac Mini already...

 

[simsaladimbamba]

Oh, I'm missing my i5 Mac Mini already...

The i5 Mac mini was also dual core, only the i7 Mac mini is a quad core model, and not even every i7 model.

 

[ColdCase]

I have this idea, and if anybody end up patenting it, am gonna kill myself.

Why limit yourselves to the cpu in front of you.

I bet, right now, there are million of idling computer somewhere, and they are all hooked up to the Internet. Talk about MASSIVELY PARALLEL processing. Kinda like the SETI plug-in, give pieces of your work to remote cpus.

I want my laptop able to talk to those millions of idling cpus out there and, in a cooperative fashion, speed up my work, seamlessly (the keyword).

Doable?

There are several scientific data crunching projects doing something like this now, mostly in a university environment.

The compressor app can use several computers on your LAN to share video encoding work (it hands out small jobs to idle computers).

There is at least one quantum computer research project researching the concept, but if you want to share cores across different machines in a similar manner to what CPUs do within one machine, the data latency and rate will kill performance. With the current state of technology, best you can do is the app dividing the job up into small chunks and handing the chunks off to different computers.

 

[derohan]

The i5 Mac mini was also dual core, only the i7 Mac mini is a quad core model, and not even every i7 model.

Guess I didn't pay attention, but it was comfortably more powerful for editing, rendering and transcoding 1080p 50hz files at a high bandwidth/data rate

The 4 threads thing has always confused me - I had eight on my 1st generation i7 alienware m15x ^^

There are several scientific data crunching projects doing something like this now, mostly in a university environment.

A friend of mine mentioned they allocate data through a closed network to more powerful servers that are out of sight where he works - So plenty of space and not much maintenance required in the workspace - I believe they were doing physics simulations so mainly GPU's concerned I would assume.

I'll look up their website, I'm a few kilometers away from them