Is it worth upgrading the CPU in a 2year old PC?

[fese]

OK, let's hold up here. First, I5 vs i7 will not see a very drastic improvement.

2nd, we don't know what you are doing in your DAW! I would guess you are recording midi, not bouncing anything down to audio, and thus hitting a wall with many plugs running.

3rd, you have a strong CPU and a top tier audio card...but you are running horrificly high buffer settings! That setup should never exceed 512, if not < 256!

I mean thjs with respect, but it may be time to examine your recording processes and system settings rather than hope a CPU will change everything.

Schmidi

1. Yeah, I've come to that conclusion, too. Probably isn't with worth it.
2. Depending on the song I usually have 30/70% to 50/50% audio/MIDI, usually one to three synths, kontakt, drum sampler. That is not the problem.
3. What is the problem with high buffer settings? Asio load meter is much calmer than with 256 or lower and I don't get spikes and dropouts. For recording I use lower latency of course.

I had a UAD card not too long ago, and people may say what they will about DSP cards being obsolete and expensive but I had less problems then. Sold it for funding the RME, though.
I think the problem are all those new analog emulation plugins, and also guitar sims. Well, gotta live with it until I buy my next computer which won't be before 2016.

[4damind]

I don't think that the improvement will be so big. It's not an old Core 2 Duo and not to mention that a lot of plug-ins needing more and more CPU power so also with a i7 you can relatively quick push the ASIO meter to the limits.
I would wait for the next CPUs from Intel because you also need a new board and so it makes sense to wait some month and make a bigger step forward (more RAM, maybe a SSD, new motherboard, new CPU) so it will be good enough for the next 2-4 years.
If you run Windows it could be also a good idea to wait for the next release of Windows 9 (Spring 2015?) and make a complete clean install of all things on the new hardware.

[fese]

I don't think that the improvement will be so big. It's not an old Core 2 Duo and not to mention that a lot of plug-ins needing more and more CPU power so also with a i7 you can relatively quick push the ASIO meter to the limits.
I would wait for the next CPUs from Intel because you also need a new board and so it makes sense to wait some month and make a bigger step forward (more RAM, maybe a SSD, new motherboard, new CPU) so it will be good enough for the next 2-4 years.
If you run Windows it could be also a good idea to wait for the next release of Windows 9 (Spring 2015?) and make a complete clean install of all things on the new hardware.

This is exactly what I wanted to avoid. I did that two years ago when switching to win7 and it took me a week to get it all up and running, mostly thanks to Native Instruments. A real pain in the behind.
Until then, maybe some self constraint could help for a change...

[4damind]

I don't think that the improvement will be so big. It's not an old Core 2 Duo and not to mention that a lot of plug-ins needing more and more CPU power so also with a i7 you can relatively quick push the ASIO meter to the limits.
I would wait for the next CPUs from Intel because you also need a new board and so it makes sense to wait some month and make a bigger step forward (more RAM, maybe a SSD, new motherboard, new CPU) so it will be good enough for the next 2-4 years.
If you run Windows it could be also a good idea to wait for the next release of Windows 9 (Spring 2015?) and make a complete clean install of all things on the new hardware.

This is exactly what I wanted to avoid. I did that two years ago when switching to win7 and it took me a week to get it all up and running, mostly thanks to Native Instruments. A real pain in the behind.
Until then, maybe some self constraint could help for a change...

I use new hardware always to cleanup things. It needs some hours for me to have the main stuff running and the rest comes if it's needed the next days. This avoids also problems with wrong drivers and boot problems. But for a switch to a new SSD a image works of course fine and doesn't need a fresh install of all stuff.

[maarkr]

i upgraded abt 1 yr ago to an i5, 3570 w Z87 mobo & SSD and i decided not to change anything until later next year when Win 9 is stable and see where new procs and mobos are... not worth spending another $300 on an i7 for me.

[jancivil]

don't buy a CPU, buy Vienna Ensemble Pro and host the plugins in it in its process rather than Cubase's, competing for priority.

[fese]

That seems like a very powerful piece of software, except I don't need surround and network capabilities and I totally don't understand how it integrates with Cubase, their website doesn't really explain it. Is it a separate process I have to start, with extra sessions to load and save or does it save state data with Cubase? Does it route the audio streams from or to Cubase? Do I insert it on a Cubase track?
Support for 32/64 bit seems nice, but only for that there's jbridge of course...

[jancivil]

I don't need surround either.

You connect to it as a plugin in Cubase, a multitimbral instrument. You connect to its ports/channels as instruments, there's nothing exotic.
Cubase can save its data in 'coupled' mode but for me decoupled and its own sessions is superior; as coupled, the loading of it in Cubase is far slower, saves take far longer, etc and a crash means both crash. Decoupled and preserved means you can have a template stay up and loaded while you go to a different Cubase project and there is all this flexibility now. Say you want to load a whole different template, Cubase is not bothered with it, it's faster and overall this is more stable. It remains connected as you do.

Same as a Kontakt multi, you assign the outputs you have made available in the F11 rack. One can make 768 audio outputs, ie., 384 stereo. Also up to 48 MIDI ports (x a port's 16 channels).
I can't even feature what computer would handle that much I/O in an instance however. I mix with its automation mapping to the extent I don't use but a few stereo outs and I've never needed more than 16 ports.
That's another thing, all of the plugins you automate are in one map and it's all clearer than with most things, certainly Cubase.
Also you can send audio to it as if it's an outboard effect, however it's not an effect in the usual sense, it returns to Cubase via instrument channels, period. In terms of printing that audio, it is external, ie., a realtime render.

It works as 'local host', you don't need to network it past that. That's how I'm using it. All of these issues people go on and on about with teh CPU are not a part of my life in the same way. I was about to move away from Cubase because my virtual instrument use can be demanding but this sorted me. The performance of instruments particularly is just vastly more robust. I really believe that if you want to spend your way out of this issue this is wiser as I know it will take all of this load off of Cubase per se, rather than what I could possibly guess about replacing a CPU. You assign a number of 'threads per instance' to the metaframe or the particular viframes that populate it so let's say you have certain things you know are eating more CPU than normal - heavy effects or a bigger automation map, or synths that are known for it - you isolate these and give them more cores.

[aciddose]

One can make 768 audio outputs, ie., 384 stereo. Also up to 48 MIDI ports (x a port's 16 channels).
I can't even feature what computer would handle that much I/O in an instance however.

Assuming 48khz rate:
768 * 4 * 48000 = 147456000

Assuming 60fps at 1080:
1920*1080*4*60 = 497664000

So, one that is slightly more than a quarter as powerful as your GPU in typical use regarding moving data alone. (Not processing.)

Now as for what purpose 768 outputs would ever serve... Say you want to create a jungle ambiance consisting of the swaying of fronds in the breeze right down to each individual gnat fart.

[fese]

I am not sure I ever want to handle 768 channels in Cubase, though I am very sure that someone will eventually do that, just because he can...
But as far as I understand VEP is mostly for hosting virtual instruments which don't have a problem with, I mostly use DIVA for basses or monophonic parts and my other synths and few kontakt libs are usually not that demanding. Those damn guitar sims and analog emulation plugins seem to be the culprit more often than not...
But anyway, i think I keep my system as it is and live with the "limitations". I mean, twenty years ago I used a portastudio and a nanoverb, I am happy I don't have to work with those restrictions anymore. Besides, it is just a hobby for me. All is good

[jancivil]

Well, I send to the amp sims and multiple convolution reverbs. As I say, some things eat a lot more CPU than say a basic Kontakt or VSL instance so I distro to cores accordingly.
I have a huge guitar sound that I would_not be able to do just plugging them into Cubase {on my 8 core 24GB machine}. Hey, I'm not freezing teh tracks.
But it is VE PRO.

I didn't even know until I looked, that many outs. I do not think a lot of audio outputs is a great idea, this will bring the CPU to its knees as fast as anything, going nuts with that. You see these idiots at the VSL Forum, hundreds of audio outs after dozens or hundreds with the audio input and they wonder why their system isn't doing it. Typically it's going over a CAT6 cable both directions as well.
I only enable 9 stereo in any one instance, 1-18. I used to want to mix from audio and I printed a lot of audio, but that's history. Big clutter in Cubase, too.