Hello fellow members,
I’m trying to conclude and understand the situation here so bear with me.
I recently started using a Mac Pro 12 Core top of the line, 64GB system.
This system is new, formatted flash HD, new OS, under cubase 8.5 (latest version).
Audio interface of the studio is an Apogee Symphony I/O.
A truly top of the line system by all means.
My initial expectation was that I would improve the system performance by a multiple of times than that of my previous system which was a Macbook Pro 2.6 ghz I7 16GB (under the same audio interface and same configuration, including same sample rate/buffer).
I know obviously that we are also talking about threads of operation, which means that the system is not measured as one whole thread but of distribution to multiple threads (in the Mac Pro - 24 threads, in the previous system 8 threads. Those systems also differ in clock speeds).
I was testing the system with a sample rate of 48 khz and a buffer size of 192.
Cubase V8.5.20 (169) :
As an initial test, just to understand the performance in a Cubase “prospective” I loaded up 3 instances of Spectrasonics Omnisphere 2 (running a patch which a bit more on the heavy side of things).
Cubase could not keep up. This utterly surprised me.
I could see the average peak level rise and the real-time peak hitting max point. (ASIO guard is set to default factory settings which is in the the middle).
From this point on I already started to understand there is a deep problem with cubase performance wise and multi core setups like this real heavy weight Mac Pro.
I also tried to load only one instance of HZP (Hans Zimmer Piano) with Full and Bright preset running all 4 mics. Again, the system could not keep up and did not give me the performance I was expecting.
Logic Pro X (latest version):
As a direct comparison to Cubase and with the exact same sample rate and buffer settings, I loaded the same 3 instances of Omnisphere 2 in Logic Pro X - Surprisingly (or not at all!) I was finally getting performance that made more sense to me. I could see proper distribution to the cores (something which Cubase’s performance meter truly lacks), and I could steadily and easily load up those patches.
Now the question to Sternberg and also to the members of the group who also have a similar Mac Pro setup:
What is going on here?
Thank you,
David Ben Dayan.