both matter.

These things influence your performance

• CPU itself (cores,ghz,cache,uarc implementation)
  
• mobo/RAM/OC/BIOS/EFI/HT
• Your DAW (how effetively it handles mle cores)
  
• Your FX
  
• How you arrange your proj.

If you have many trax, mle cores will be better.
If you have few tracks, but heavy FX on each, more ghz will matter.

This assumes your DAW is efficiently optimized for Mle cores. But all DAWS are not made equal.

The max. speed is the speed of the slowest path. C also: https://www.google.co.nz/search?q=critical+path&source=lnms&tbm=isch

If ur comparing CPUs to buy, use cpuboss.com

 

I'm thinking about switching to another DAW. Never easy but sometimes it's needed.
I'm gonna make some tests with CPU hungry plugins and tracks in each one...and check it out

 

That's the problem... I have 48 tracks projects (usually) so multi core should give me an edge. It does not. On the contrary.
may be it's windows 10... At this point i'm clueless

 

can confirm, Daw choice makes a huge impact.

When I use Logic pro X I can almost double the tracks and plugins compared to Studio One v3. Same cpu i7-4790k same OS and plugins. Some daws are better than others when it comes to handling plugins.

 

Because of how threading works with respect to real time audio processing, a balance of both is recommended. I always say 4+ cores each 3.5ghz or higher.

Each channel within a DAW and its inserts are processed on the same thread - so lets say a synth like Diva, an eq and a compressor all would be on the same core because the computations happen one after another.

Now you have two channels like the one above, they would be processed on two separate threads, but if you buss them and start putting plugins then they move to one core. This is because the calculations for the synths eqs must be done before the ones for the buss plugins.

Your CPU craps out when the core with the highest usage maxes out.

 

https://www.cakewalk.com/Documentation?product=SONAR&language=3&help=NewFeatures.002.html

 

I have a haswell 6 core @4300 as main with logic x with ssd, download and 2x raid5
As slave 1 more 6 core slave @4400 and more 3x 4 core, all with Windows 10 and VEP pro 5.
The main advange isnt The processing power (VEP slave is very light), but The ability of using Windows vsts with logic.
The 4 core slaves are mostly spares... Shit happens really its my 4th RMAof motherboard this year.. Im hating Asus so much... Probably will by a spare motherboard while one Goes To rma... Its crazy.
Logic X with Windows slaves is a dream, the only thing better is master with protools win, Windows vst slave and OSX au slave... Simply all.
Also x99 is The way.. Waiting for cheap 20 core at 4000 GHz..
People think a lot of power but an old computer as slave usually doesThe job

 

As simple as the question might seem, it is difficult to answer. For example, the term "multiprocessing" is not describing a specified action. It just says that the use of multiple cores is possible. It's up to both, the system and the DAW, how the cores are used. Which means differences, which means different results.
Higher clock speed definitely helps calculating faster, but there are many situations where a faster calculation is less important than the parallel processing of multiple tasks. VSTis make use of SSE (2 or higher), which allows for a parallel processing of instructions, independent from the number of cores (SSE exists since 2000 or so). For one VST instrument clock speed is more important than core count, since they already process instructions in parallel, which means on a per-cycle-basis.

Another aspect is that terms may be confused. A thread is a different thing than a task is a different thing than a core. This is due to Intel using the term "Hyper-threading", but a system thread is something different. A task can make use of how many threads it sees fitting, each thread it then worked on independent from the others. That's not multi-core-processing. But as you might already guess, the latter can be implemented on an instruction, thread or task level, where more and more DAWs tend towards the instruction level, while older ones where on the thread or even task level. One DAW that was built from the beginning with multi-core environments in mind was Reaper. Developed on Xeons, it might even help you @yabiss

 

Seeing as the dude has plenty of cores, yeah, it's the clock rate. Your i7 is much better than what? You didn't list a comparison. My 4ghz handles everything fine but a faster clock rate would obviously improve it.

 

Oh and this brings up a seldom discussed aspect: project arrangements. Use fx busses, etc. A lot can be done just with the way you work with the projects. A lot of people spend so much time on optimization, they forget about this part.

 

My I7 at 2.4ghz is much better at handling the exact same project than my bi cpu 16 cores at 2.6ghz. The more complex the project is the more i feel my multi core system is loosing time handling and synchronizing.

So it has to be more than just clock speed too.

My guess after reading all your comments guys is that how the DAW was conceived is definitely the most important aspect of all. I read about the Sonar Load Balancing function available in the platinum version only. It says a lot about this feature being a plus not included in many DAWs.

I'll give Reaper a try and compare

 

oh my, this thread has so much misleading info
let me clarify some things (english isn't my native language, so I apologize for eventual misinterpretations):
1) higher clock speed is ALWAYS better, but also means more power consumption and therefore more heat, workstations and laptops are therefore featuring lower-clocked cpus, just because of more reasonable cooling and power requirements (this is also why there are practically no cpus being clocked higher than 4.4GHz out of the box)
2) more physical cores are handy IF the software can make use of splitting the tasks (notice that AMD FX series although featuring (8000 series) for ex. 8 cores, have those actually "packed" into modules consisting of 2 cores each, sharing the bandwidth, so that's why often FX-8350 performs barely like decent desktop Core i5
3) more threads (hyperthreading, Core i7, Xeons and Core i3 cpus for ex.) are useful IF the software can make use of splitting the tasks onto them (this is why sometimes Core i5 having just 4 cores performs almost same as Core i7 with 4 cores but 8 threads)
4) bigger CPU cache is actually what matters A LOT, because every tasks is loaded from HDD/SSD into RAM and then into the CPU cache to be processed, so that's why multi-core cpus have advantage, not because having more cores at first place, but because having bigger cache to work with (just for ex., i7-6700K has 8MB L3 cache, i7-5820K has 15MB L3 cache, Xeon E5-2697v3 has 35MB cache, i7-3630QM has 6MB cache, i3-6320 has 4MB L3 cache)
5) having more cores isn't essentially better at all, if the task cannot effectively use them "at once", for ex. Cakewalk Sonar recently introduced Plug-in load balancing which actually doesn't perform better IF there aren't many heavy plugins to handle - it tries to evenly split load across all available cores/threads, which often means some of the cores are just "waiting" for the others without processing anything useful, only being loaded consuming power and producing heat (I tested some of the projects myself on my rig with OCed i7-5820K so this is not a theory)
6) in terms of reliability and performance, it's usually always the audio interface which is being responsible for dropouts and latency, something like RME HDSPe AIO can handle multiple running audio softwares of various bit+sample rates, zero latency monitoring and loopbacks and this all with minimum CPU load
7) DAWs are treating things differently and it's always recommended to read the technical manuals, this will help understanding the processing chains and optimal arrangements- for ex. having 10 tracks each with same effect stresses PC more than routing all those tracks into one bus and applying the effect just once over there
-EDIT:
8) Xeons used to be top tier cpus in terms of reliability and 24/7 operation, which as far as I know no longer matters, nowadays lower-end Xeons are basically crippled Core i7 (meaning, disabled intergrated graphics or limited clock speeds, locked multiplier etc..), only advantage of Xeons nowadays is support of ECC RAM (error correction, which appears to be valued among professionals)
9) LGA 2011-V3, X99 chipset eventually is what "home professionals" (especially "all-round" multimedia creators) should definitely check out (more PCIex lanes, 8 RAM slots, bigger cpu socket with better mounting - screws not holes in mobo)

 

Last edited: Nov 2, 2016

Really? Well, to be honest, you are repeating everything been said by others. Basically.

 

What this guy said.

 

?! Its like comparing ram speed and ram quantity in benchmarks. It's a non sense. This is different part of the unit doing different things, you cant compare how they perform in VS in a given domain. In the Audio domain both are important though. But if you want to truly compare CPU processing power watch for the biggest cache not cores or frequency.

Edit : What Tzzsmk and part of others posts just said answer everything. [SOLVED]

 

Last edited: Nov 2, 2016

And to people glorifying Xeons, just dont. Would be great to see a complex 96 tracks project with tons of FX chains with a i7 4790K VS Xeon 26XXX. For tons of reasons you'd be suprised.

 

Before the i7-4790k i had a 2 core i3-550 3.20 ghz. This little guy made wonder up to 150 track EDM project full of reverbs and delays in Logic Pro X, no hyper-threading, Frequency stepping is another big part for cpu for audio purpose, less steps is better IMO.

 

Reaper has ReaMote for that. Very easy to utilise. You just connect two PCs with a crossover ethernet cable, install Reaper and the plugins on both, start Reaper on the first and ReeMote on the second. Done.

Regarding the topic, how plugins use the CPU is up to the DAW. Reaper likes more cores and runs plugins on different cores in a balanced way nicely, so the more cores you have the better. But a plugin shouldn't need more CPU power than one core can provide or we have a little problem there...

Cheers!

 

At least this point is worth to take into consideration.

Starting with 48 tracks seems to me a bit of an overkill, but of course cpu load depends on what every track is holding. With having samples only to be fired, that would be no problem, but with 48 softsynths plus sidechain compressors plus other fx, the idle status alone would have a big impact on the cpu meter. Thinking of that, I once found that sidechain compressing can be a realtime killer for some reasons.

I wonder how we made tracks happily with a single 200 MHz CPU and 32MB RAM