8 core won`t give me more CPU ? But are slower !

[nossnevs]

Hi

I used to run pro tools 10 native on my very old MacPro (1.1) 2 x 2.66 GHz dual core with 16 gb ram (667 Mhz) ( I know PT 10 only used 4 GB or 3 GB)

Now I upgrade to another old MacPro (1.5) 2 x 2.26 GHz. Quad core with 24 GB ram (1066 Mhz.) So double up core and faster ram, but a little slower Ghz.
I´ve installed PT ver. 12.7, (HD native, upgrade from my dv toolkit) running 64 bit and AXX plug ins.

I thought that AXX plugins use less CPU than RTAS ?

I did a test on my old MacPro, PT 10 vs the "new" MacPro PT 12, use the same PT session from a external ssd disk.I use the same plugins, only different is that the plugins are RTAS on PT 10 (the old MacPro) and AAX on PT 12 (the new MacPro)

10 tracks EQ, - 15 tracks D-Verb. - 5 tracks Izotope spectral d-noise and 5 tracks Izotope Voice D-noise.

On my old MacPro the CPU shows 65 %
On my "new" MacPro the CPU shows 75 %
This I find very strange because I thought that PT 12 used less CPU ( better written software) and I have double my core from 4 to 8. and I believe PT12 can use all the ram.
I know the Ghz is 2,26 vs 2.66 but I find I hard to believe that this could be the reason. After all I double the core and have faster ram.
I have a offer to replace my Processor to 2.93 Ghz for a very fair price. Will this give me a big increase and better performance ?
All inputs are more than welcome, I do appreciate all information I can have from your guys.



Thanks Michael

[Darryl Ramm]

You are wasting your time looking at CPU meters.


Sent from my iPad using Tapatalk

[albee1952]

From my perspective, each major update of Pro Toos(10>11>12) actually uses more cpu(all those new features require more horsepower to work). Then, I thought PT12 was not approved on older Mac Pro(pre 5,1)? Of course, "unsupported" doesn't mean it won't work, but performance might be less than great

[musicman691]

Originally Posted by nossnevs

Hi

I used to run pro tools 10 native on my very old MacPro (1.1) 2 x 2.66 GHz dual core with 16 gb ram (667 Mhz) ( I know PT 10 only used 4 GB or 3 GB)

Now I upgrade to another old MacPro (1.5) 2 x 2.26 GHz. Quad core with 24 GB ram (1066 Mhz.) So double up core and faster ram, but a little slower Ghz.
I´ve installed PT ver. 12.7, (HD native, upgrade from my dv toolkit) running 64 bit and AXX plug ins.

I thought that AXX plugins use less CPU than RTAS ?

I did a test on my old MacPro, PT 10 vs the "new" MacPro PT 12, use the same PT session from a external ssd disk.I use the same plugins, only different is that the plugins are RTAS on PT 10 (the old MacPro) and AAX on PT 12 (the new MacPro)

10 tracks EQ, - 15 tracks D-Verb. - 5 tracks Izotope spectral d-noise and 5 tracks Izotope Voice D-noise.

On my old MacPro the CPU shows 65 %
On my "new" MacPro the CPU shows 75 %
This I find very strange because I thought that PT 12 used less CPU ( better written software) and I have double my core from 4 to 8. and I believe PT12 can use all the ram.
I know the Ghz is 2,26 vs 2.66 but I find I hard to believe that this could be the reason. After all I double the core and have faster ram.
I have a offer to replace my Processor to 2.93 Ghz for a very fair price. Will this give me a big increase and better performance ?
All inputs are more than welcome, I do appreciate all information I can have from your guys.

Thanks Michael

Are you sure the PT10 plugs are just RTAS and not aax? aax came in on PT10. aax plugins don't necessarily use less cpu than RTAS.

'10 tracks EQ, - 15 tracks D-Verb. - 5 tracks Izotope spectral d-noise and 5 tracks Izotope Voice D-noise.' This is your problem right here. Do allt he tracks absolutely need eq? 15 tracks each having their own D-Verb? If so you're doing that allw rong. Use sends from each of the tracks to an aux track that has the D-Verb plugin on it and use the send levels in each track to control the amount of reverb. This will not only cut down your cpu usage but will also sound better as everything will sound as if it's in one place. Lastly all those iZotope d-noise plugins - do you really need them? Furthermore those were never meant to run in real time as it's well known they are major cpu hogs.

[DetroitT]

PT10 plugs are ether RTAS and/or AAX (10 transition)

[Darryl Ramm]

...and stop looking at CPU meters. Starting with I suspect you are comparing hyperthreading and non-hyperthreading CPUs. That comparison is pointless.

So what actual problem are you trying to solve that you with the new setup? e.g. what AAE errors do yuu get? (What the meters show is not an actual problem....).

[nossnevs]

Originally Posted by musicman691

Are you sure the PT10 plugs are just RTAS and not aax? aax came in on PT10. aax plugins don't necessarily use less cpu than RTAS.

'10 tracks EQ, - 15 tracks D-Verb. - 5 tracks Izotope spectral d-noise and 5 tracks Izotope Voice D-noise.' This is your problem right here. Do allt he tracks absolutely need eq? 15 tracks each having their own D-Verb? If so you're doing that allw rong. Use sends from each of the tracks to an aux track that has the D-Verb plugin on it and use the send levels in each track to control the amount of reverb. This will not only cut down your cpu usage but will also sound better as everything will sound as if it's in one place. Lastly all those iZotope d-noise plugins - do you really need them? Furthermore those were never meant to run in real time as it's well known they are major cpu hogs.

Hi Thank you for answer. I no this is not the right way to set up my PT and plugins. I only did this to push the system and compare the two Mac`s cpu.

[musicman691]

Originally Posted by nossnevs

Hi Thank you for answer. I no this is not the right way to set up my PT and plugins. I only did this to push the system and compare the two Mac`s cpu.

That is not a fair comparison by any means. There are better ways to compare cpu throughput and I'd suggest Googling for how to do that. You're comparing two different programs on two different computers.

[Bob Olhsson]

Didn't dverb change?

[musicman691]

Originally Posted by Bob Olhsson

Didn't dverb change?

I believe so; I know it's more efficient now than it was which makes the Dverb test kind of moot.