Yes, Windows in general gives much better performance for audio than OSX, especially at lower latencies. A good comparison can be found here:

http://www.dawbench.com/win7-v-osx-3.htm

GUIs are also much snappier on Windows, this is also true for VSL products, especially VEPro.

It is true that the later versions of VIPro have vastly improved memory allocation routines - something where OSX is known to be slow. This is usually experienced by software becoming slower and slower the more RAM it has allocated. And in case of samplers with huge orchestral templates, it gets very apparent.

But this is not the only issue. I have actually never experienced a software being snappier or better performing on OSX than on Windows (where both platforms are supported). As a developer there are definitely more hoops to jump through when developing for OSX, than for Windows.

I have seen that benchmark, it's full of holes.  If cubase is used, it's invalid. Cubase does not port to Core Audio directly.

It's disturbing to see you follow that. I'm confident it's true that VE Pro works better on windows and this makes me wonder if some of it's because you are biased towards an OS. Logic is extremely efficient under OSX, as efficient as I think can be.

The operating systems must be very different to work with, but an absolutist sort of assertion is quite suspect to me.

Cool. I use Digital Performer and I couldn't imagine VEP getting too much faster. I can do almost any complexity of arrangement with a low buffer. But VIPro could use some improvements. I use VIPro instruments that have many matrices. VIPro is quite slow switiching between them.

sorry for the typos ... but with my ipad that forum is only partly useable. 

best

Now, if you can show this memory is slower, and I know that technically the machine I have has slower RAM by specification than for windows boxes potentially, using VI Pro/VE Pro, that is acceptable. This benchmark certainly is not. I have studied it and stand by my assertion. It shows confirmation bias in the extreme IME.

Not that you've shown any benchmarks to back up your claim, but as I pointed out, the tests also use Reaper, Pro Tools and Studio One.

DG

MAKE A BENCH MARK ! 

and even so all depends of the programer !

I have been programming for 20 years in Hex, Cobol, Basic, Algol, Postscript, depending on how you write a program, it will go fast or slow

The company I was working for, bought a Postscript program from a Canadian company. It was very slow. I complettly re-wrote it ; It was running 10 time faster, no wait for the user !

I have post on the Core Audio develloper conference :

On 07/10/2011, at 10:30 PM, Cyril Blanc wrote:

Hello   xxxxxx

Do you have a way to do a bench: ASIO v.s. Core audio on the same computer ?

Answer :




Short answer, not that I know of without using specialised gear like logic analysers to look at the exact timing.

Long answer: You're measuring latency, i.e the time it takes for the source samples coming from memory buffers, via hard drives, Audio Units etc, to be copied to the render buffer, then to be DMA'ed in the kernel driver to the hardware DAC and ultimately your ears. Audio is real-time. If you put 512 samples into the render buffer at 44.1khz it will take 11ms for those samples to be processed plus the hardware delay of the DAC. It can't be any quicker and it can't be any slower. Unless you have an OS that cannot do real-time audio or a badly written driver that adds latency. Anything else is app specific hence why well written DAWs have things like PDC and other complex scheduling solutions.

regards
peter

I'm not saying that Logic doesn't perform better, but to quote figures without actually stating where you got them from doesn't really back up that claim. However, 4x better? I don't believe that for one moment. Are you sure that when you did that test you had all of the tracks in your Logic project in Live mode (to eliminate the process buffer)?

Of course it is possible that all cross-platform applications are coded better for Windows. It's also possible that Windows is just a better operating system. No way of knowing, unless you see all of the source code for all existing cross-platform applications.

There is no fight. Nothing to win. If you were interested solely in performance you wouldn't be using OSX at all. The fact that you are, means that there are other things that are more important to you, such as the way the OS handles for you and the sequencer you choose to use. Nothing wrong with that. For me I don't care what OS or machine that I use, as long as I can get the performance I need. OSX won't do that for me.

DG

You are shooting the messenger. I am a long time Mac user and the author of 2 Logic Pro books but I am building a PC slave because East West, for whom I am aa part  time employee, has  told me exactly what VSL is telling you here, that it is very difficult to make the OSX 64 bit version of Play as efficient as the Windows 64 bit version. That is just the reality.