upgrading pc- what platform too choose.

Hi there,
 
At the moment I am lucky enough to be expecting some extra cash coming in and I've decided to use it to upgrade my computer for use with Sonar X2.
At the moment I'm running on a 32bit Vista machine with 3GB of ddr2 667 a Q8200 core2quad proc and 1 7200rpm 500GB HDD.
This system is however not as stable as it used to be and as I'm getting more aquainted qith producing, I'm really starting to feel the bottlenecks here.
Now I've been looking around a lot to see what system I wanty to go for now, but can't seem to find the right info to make the decision on what system would suit my needs best.
At the moment I'm basically considering two setups. One would be Intel based with a I5 3450 proc. (quad core) and one AMD based with a FX-8120 proc. (octo core).
What I am mainly wondering is what will benefit me more, a faster core design from intel with 4 cores, or a somewhat slower cores but 8 of them. I certainly don't run very large projects, maybe 4-8 audio tracks and the same amount of synth tracks on a usual base. I do however use busses and want to be able to load at least 5 effects per track and bus.
I just can't find out what Sonar will benefit more from, more cores (more multitasking) or faster cores (faster processing).
Is there anyone who knows a bit more about this ?
 
Besides that I am also wondering what is the better choice for HDD. I'm going for SSD drives with a 500+ mb/s write speed, but am a bit puzzled as to what would be the best configuration. I see many times that people advice to use separate drives for system, libraries and tracking. Is this still applicable with the fast SSD drives ? Also, if I create a striped raid of two ssd's , will that be better than spreading it over two separate drives, or do I benefit more from the overall improvement of read/write due to the raid config.
If raid works best, which 1155 or am3+ mobo's could you advice that have good raid options (at least 2xsata600 raid 0)
So in recap, are separate drives even needed on fast ssd's and, if it is still indeed desired to spread load, is it better to spread it using separate drives, or those drives configured in RAID 0.
 
The last notion I want to make is that I'm aware memory is a very important factor. I'll be using ddr3 1600, as I'm not overclocking for stabilities sake. It will be at least 16GB, but probably 32GB. Now the 1155 as well as the AM3+ setup both support dual channel memory. The 2011 platform (i7 at 2011mhz fsb) will offer quad channel memory. Basically I'm not considering a 2011 setup as it is very heavy on the budget. However, if it really does make a significant difference in performance and memory really is the most important to consider, I might consider this setup afterall and then just start out with one SSD rather then more (separate or in RAID).
Does anyone have thoughts about this, would that be the better option, or is the performance difference marginal or just plain overkill and can I better choose for a more complete system with several ssd's as I won't be needing 2011 type performance.
 
Well, that's about it, as you can see there's a lot going on in my head and I really hope there's some friendly people here that might be able to shed a bit more light on these issues.
Thanks in advance for that and grtz,
 
Paul 

Tnx jcschild,

That's the kind of feedback I was hoping for.
I was myself also leaning towards Intel and you confirmed what I was kind of figuring myself.
The i7 will still impose a significant budget rise, however if what you say is indeed so, certainly possible by cutting on the ssd costs.
Here though, one thing is still a bit unclear. The way you say it, there is no gain from tracking on a ssd. That I can follow.
Looking at the size of the files created and the time they represent I figure there's still lots of tracks headroom when it comes to tracking.
My concern though is mainly the reading part.
These days I'm trying to create a more loop based workflow and the things I track will be cut up, the vocals many times comped from several takes etc.
Seeing that a sampler really benefits from faster reads and faster acces times led me to think this will also be the case when playing back audio in a project. The tracking is normaly fairly quickly done, but then starts the editing and so constantly jumping through the timeline, looping parts while adjusting etc. Also I'm really experimenting a lot with the matrix view these days, so I'm still wondering if the ssd really wouldn't improve overall performance and experience.
Is it really that the libraries in general are that much bigger, or the amount of files needing to be accessed per given period that much greater that the audio data in a (<10tracks) project will never come close to that same amount of workload for a hard drive?
If that is really so, it would be great as ssd's,though quite affordable these days, are still relatively small and expensive per GB.
Last, what did you mean by 'worth it' regarding libraries?
Is that concerning physical criteria (amount/size of samples), or criteria of a more personal nature?

Anyways, as you see still looking to learn more, but really thanks a bunch for your replie and the info ..:)

SSD has crazy read speeds and very low seek time.. so for heavier sample libraries they make sense.
things like LASS, HW Strings and a few others.

loop based stuff is generally synths not sample and is pointless to put on SSD
if its not orchestral its generally not heavy hitting.
however there are some exceptions, particularly in the drums area.. (ocean wave)

tnx again Scott,
 
really appreciate your time and just learned from fireberd that you're a
more specialised and experienced member on the matter than I had hoped finding here.
For Os I'm considering now, as I hear that ssd is not necessarily more reliable.
I'm not calculating the highest priced ssd's at the moment, but the cheapest (OZC Agility 3 120GB/240GB)
and I do want reliability, so I have to see what's best at what price.
For the libraries I still want one though.
Next to all the audio and DimPro/SD3 libraries coming with Sonar 8-X2, I also have NI Alicia's Keys (a 7gb piano),
received the HD strings package for DimPro and want to expand on drums (Battery) and other sampled piano's.
So that now seems to make sense me, right?
Then about the RAM (sorry if I'm firing of too many questions), what is best to avoid or look for. I see so many differences and wonder what is best to buy, as I understand that stability can be an issue.
I'm not looking to OC , want ddr3 @ 1600mhz and the best timing/price ratio.
Can I just get the cheapest : 4096MB Kingston HyperX Genesis DDR3 1600MHz CL9  ,
or do I really benefit from paying more for another brand or more expensive version of the same brand ?
(apart from any specific mobo/hardware compatibility issues that can't be foreseen)
Is this really an issue to consider, with reall differences between modules of the same specs.,
or will most common brand memory (when compatible) do just fine?
and again of course tnx so much  :)
 
 

The problem there being that an SSD large enough to accomodate an extensive collection of sample libraries is cost prohibitive...

Depends on what your definition of "cost prohibitive" is...   
I've got hard-core composer clients that have 6+ SSDs in a single DAW.
For these folks, being able to run 1500+ simultaneous notes of disk-streaming polyphony from a single machine is actually cost effective.
Note: You're better with multiple drives than a single large (512+GB) SSD.


Prices on SSD are starting to come down.
I just got a 256GB SATA-III unit (sustains 520MB/Sec on reads) for $200.

For larger scoring projects I am often streaming 32 tracks from the sampler comp on 2 standard 1TB drives with zero problems...so not in any rush to switch to SSD until the price comes down more...also not enamored with the idea of splitting my large orchestral libraries up on several drives. Will continue to watch SSD development...

In your situation, you don't need the speed of SSD... so it's not a cost effective solution.
I get that...   


If you needed 500+ simultaneous notes of disk-streaming polyphony (LASS, Hollywood Strings, etc), it would be a different matter.  For these folks (and they're the minority), SSD delivers disk-streaming performance that's just not possible with conventional HDs.

Curious why they would be running that much (1500+) simultaneous polyphony?  In a roughly 100 person symphony orchestra, with most of the instruments being monophonic or at best 2 note polyphonic (excepting harp, piano, timpani, etc.) at a 1 to 1 person/track ratio that would only be in the few hundreds of simultaneous polyphony...can you eloborate?  Seems like 500+ would even be high?  Are they doubling each chair position for some reason? Multiple mic positions perhaps?

Multiple mic positions perhaps?

Bingo...   



With LA Scoring Strings (LASS) or Hollywood Strings, polyphony piles up very quickly.