iTunes speed on 2.0GHz Intel iMac

[ecosse011172]

Hi,

I'm seriously considering a 20" intel iMac so that I canget rid of my PC. ( I wish thy'd bring out a 23" iMac)
When ripping (average) a CD to AAC VBR 256Kbps with my iBook I get 7-8x, with my 2-3 year old PC (P4 2.8c) about 18x.

What kind of speed have you encountered on the new iMacs?
I understand the limitation of optical drives, HD etc. but it would be interesting to see performance at the moment with an important task for me.

Thanks in advance,
Denis

 

[bigfib]

Rip speed

Hiya,

I just checked that for you.
On my intellimac 17" 512 I get from 9,5 to 16,8x which is about 50% faster than my g4 ibook.

Hope this helps.

 

[jacobj]

Hi,

I'm seriously considering a 20" intel iMac so that I canget rid of my PC. ( I wish thy'd bring out a 23" iMac)
When ripping (average) a CD to AAC VBR 256Kbps with my iBook I get 7-8x, with my 2-3 year old PC (P4 2.8c) about 18x.

What kind of speed have you encountered on the new iMacs?
I understand the limitation of optical drives, HD etc. but it would be interesting to see performance at the moment with an important task for me.

Thanks in advance,
Denis

According to the article on the macworld website the intel imac is 2.0 GHz is 1.12 x faster than the imac G5 2.1 GHz... that must blow your ibook out the water.

The Craig Wood tests show the 1.67 G4 PB encoding at 15x and the MacBook Pro encoding at 20x...

Your iMac would be a little faster than the MBP and your iBook would probably fall behind the PB...

Edit: I would really like to know if iTunes uses both cores if it can... all these stats indicate that it is running on a single core... can Apple make iTunes multi core aware (or whatever the technical term is?)

 

[ecosse011172]

According to the article on the macworld website the intel imac is 2.0 GHz is 1.12 x faster than the imac G5 2.1 GHz... that must blow your ibook out the water.

The Craig Wood tests show the 1.67 G4 PB encoding at 15x and the MacBook Pro encoding at 20x...

Your iMac would be a little faster than the MBP and your iBook would probably fall behind the PB...

Edit: I would really like to know if iTunes uses both cores if it can... all these stats indicate that it is running on a single core... can Apple make iTunes multi core aware (or whatever the technical term is?)

Thanks a lot, a bit dissapointing that the new dual core ships can't do much better than my 3 year old PC......

 

[BakedBeans]

Thanks a lot, a bit dissapointing that the new dual core ships can't do much better than my 3 year old PC......

they cant do much better in iTunes maybe.... they can do a lot better in a lot of other things though - if your job is using itunes, still with a fast PC. If you want a really fast user experience then get the intel imac

does it even use both cores?

 

[robbieduncan]

they cant do much better in iTunes maybe.... they can do a lot better in a lot of other things though - if your job is using itunes, still with a fast PC. If you want a really fast user experience then get the intel imac

does it even use both cores?

There has to be a point where the read speed of the optical drive is a limitting factor.

 

[DeSnousa]

Its not properly optimised for the intel chips. While on PPC they are. Give some more time, maybe a Quicktime upgrade soon will address the situation

 

[BakedBeans]

There has to be a point where the read speed of the optical drive is a limitting factor.

What does a Quad do it in?

We'll soon know

 

[DeSnousa]

What does a Quad do it in?

After a Google search the best I could find was a result of a conversion of a 43 min ACC file to MP3, which took 43 seconds. This makes me assume that they don't test imports because the drives are slow compared to the computing power. My guess is maybe near the full speed of the drive. Whats that in a Powermac?

 

[jacobj]

Thanks a lot, a bit dissapointing that the new dual core ships can't do much better than my 3 year old PC......

You are not really comparing like for like here. The only way to know is for 2 users to rip the same cd to the same settings and to have both machines with the same RAM.

If you are using iTunes as your measure as to how fast the CoreDuo is then you are making a mistake.

My favourite benchmarks so far have been the Handbrake tests...

Just look at this image from another thread! It is doing an 1.5 video clip in 37.5 minutes with 512MB RAM... that is unreal IMHO.

Edit: the 37.5 minutes is first pass only.

 

[ecosse011172]

You are not really comparing like for like here. The only way to know is for 2 users to rip the same cd to the same settings and to have both machines with the same RAM.

If you are using iTunes as your measure as to how fast the CoreDuo is then you are making a mistake.

My favourite benchmarks so far have been the Handbrake tests...

Just look at this image from another thread! It is doing an 1.5 video clip in 37.5 minutes with 512MB RAM... that is unreal IMHO.

Edit: the 37.5 minutes is first pass only.

Why is using iTunes a mistake? Because the results aren't as good as with Handbrake?

iTunes had been out on x86 machines for a long time and is one of Apple's most used and important apps so it should have been reasonably optimised by now. Maybe the optical drive speed is a factor but shouldn't they read at 48x?

 

[gnasher729]

Why is using iTunes a mistake? Because the results aren't as good as with Handbrake?

iTunes had been out on x86 machines for a long time and is one of Apple's most used and important apps so it should have been reasonably optimised by now. Maybe the optical drive speed is a factor but shouldn't they read at 48x?

The G5 beats the Yonah processor in two areas: Highly optimised floating-point code, and highly optimised Altivec code. (Why only highly optimised floating-point code? The G5 has TWO FPUs, but every instruction has six cycles latency, that is six cycles until the result is available. To take full advantage if the two FPUs, a program needs to execute TWELVE independent floating-point instructions before the result of the first instruction is used! In other words, I can write some code in two hours that runs about equal speed on G5 and Intel. If I spend another 10 hours, I can make it twice as fast on G5, but not on Intel).

The time critical part of iTunes is the encoding, and that is highly optimised Altivec code. That kind of code is extremely rare, and it is one of the few bits of code that is faster on G5. Handbrake on the other side is completely unoptimised C code, as is 99 percent of all code out there in the world.

 

[gnasher729]

Just look at this image from another thread! It is doing an 1.5 video clip in 37.5 minutes with 512MB RAM... that is unreal IMHO.

Edit: the 37.5 minutes is first pass only.

Please note that Handbrake is one application that uses very little memory. It works quite happily in 30 Megabyte or so, and it is one of the few applications that won't get faster if you add more RAM. Other applications are extremely wasteful with memory and will be unhappy with only 512Megabyte of RAM.

 

[jacobj]

Why is using iTunes a mistake? Because the results aren't as good as with Handbrake?

iTunes had been out on x86 machines for a long time and is one of Apple's most used and important apps so it should have been reasonably optimised by now. Maybe the optical drive speed is a factor but shouldn't they read at 48x?

iTunes is a mistake if it is your only benchmark, not that it is not a benchmark, only that in isolation it is misleading.

 

[SiliconAddict]

Thanks a lot, a bit dissapointing that the new dual core ships can't do much better than my 3 year old PC......

Step 1. . .Port the software
Step 2 . . .optimize the software
Step 3 . . .Take over the world.


At some point here in the next 12 months ALL of Apple's hardware will be sporting dual core goodness. Once that is assured you can bet your yearly salary that Apple will be tweaking their apps to use both cores. I'd say if you pick up a dual core whatever this year you should REALLY plan on upgrading to iLife 07 and Leopard next year. It will be worth it.

 

[mzd]

Maybe the optical drive speed is a factor but shouldn't they read at 48x?

the advertised 48x read speed on a drive is the peak speed when copying data off the drive. the speed fluctuates as the head moves across the disc, getting anywhere from probably 8x to a max of 48x. but, the key is that is data copy only. audio extraction is going to be a different number and it won't be advertised on the box or on the drive. you need to test that. i'm sure there are apps out for mac that do this, but i don't have a mac. on my pc, my 24x drive has 8x audio extraction. again, that is the max, so at some point on the disc, it will reach 8x, but the average is around 6x-7x for a given song. iTunes can encode at 4x on my pc, so that is an indication of the processor. 8x would be the fastest i could get given my drive.

 

[rhashem]

The time critical part of iTunes is the encoding, and that is highly optimised Altivec code. That kind of code is extremely rare, and it is one of the few bits of code that is faster on G5. Handbrake on the other side is completely unoptimised C code, as is 99 percent of all code out there in the world.

+1 insightful!