Performance of new 10900k VS 10940X (Cascade Lake) ?


Once again if you send me small Cubase Song I can run this on all Systems we have open right now.
That is 10940X, 10980XE, 10900K, 3970X and we also have standard 2666 Value RAM VS 3200 with very tight timings and 3600 with low timings.
Currently I think about another 9900K to have a comparrsion. Right now I am not into Benckmarks but will do after summer break.

I cant spend hours on this task, but tell we what you want to know and I will make a short test. Untill now I have no NI Products installed but I have a full NI Komplete 12 NFR, which I can use. So if its easy to install NI Kontakt (or even maybe a demo), I am in for a halve hour or so.

Who are these “real champs”? I’ll soon be embarking on upgrading our PC’s and having already looked around, the figure is likely going to be around 6-8k each going by estimated specs excl network, NAS drives etc. Any hints to great Audio PC builders would be greatly appreciated.


That’s awesome, thanks :smiley: :smiley:

Thanks a lot for this offer, that would be great!! I will be preparing a test session. Do you have uheDiva or Repro? These are the hungriest synths I use, so I would implement one of them into the test. Kontakt installation is easy and goes fast, but the libraries are quite big (like VSL). However, I found this free orchestra library that should do fine for the test (small size, no install needed): https://alpineproject.wixsite.com/main/libraries

I’ll send you the session tonight. Thanks again! :slight_smile:

I have Diva on all Systems, yes thats a good one to test with, I like it alot, a few instances and your ready :slight_smile:

Thanks for the link. I can start mid of next week.

Hey Daniel, sorry for the delay. Here is the project file.

One important note. I realized that this free Kontakt library is very light on CPU, so unfortunately it’s not a fair comparison with professional tools like OT BerlinStrings or CS Strings.

Those 8 insert effects on every channel are a bit exaggerated, but as external plugins are much hungrier than Cubase internal plugins I thought this might be useful for creating a similar scenario. The Diva Patches do modulate a lot so I think it can barely get even hungrier than that for a soft synth track. However, I haven’t put them on Divine mode because I wasn’t sure if that would be overkill. If your machines handle to run this Project, it would be very interesting to see how they’d perform with Divine mode on though.

I’m very excited to see the results. Take your time and thank you again for this opportunity! :slight_smile:

And before I forget, there is something that makes me curious. Also on the website of DA-X, I’ve come across this chart, which states that the 9940X performs better than the 10940X. Would be interesting to verify this, as the 10940X should be the (slightly) better model, right?

hello

i have plan to upgrade my pc i like the threadripper 3960x …?
but i heard there is some issue crackling sound ? Daniel can u make a benchmarks between threadripper and i9 .?
and told us what ur experience

i have cubase 10.5 pro
knotakt , omnisphere
plugins (real time)
40 tracks live sessions

In case that some might be interested, I could successfully build my Hckintosh on an 10940X and after running a few A/B comparisons I came to the conclusion that it depends a lot on the used plugins, which OS does better real time performing.

From what I’ve tested, Kontakt on Windows 10 can deal 25% better than on OSX Mojave. Divadoes even 100% better on Windows, however, OSX can do 100% better with Serum and 20% better with Ableton stock plugins. I was paying attention to configure a solid hardware setup and to provide Latency Monitor with good results, but I’m no expert and cannot say how precise my testing was.

In the end for me as a film composer, Windows might be the better choice due to it’s performance with Kontakt. However if I weren’t relying on these two, there wouldn’t be a significant difference between the operation systems, as the more plugins loaded the more it would balance out and come to a similar performance result.

Just my impression from a bit of testing. :slight_smile:

Could you list the specs of your new Build?

Which motherboard did you choose?

Asus Prime X299-A II. Both of the onboard m.2 slots operate on their own dedicated PCI lanes, and the first two PCI slots can be set to 16+16 at the same time. I added an ASUS Hyper M.2 card for additional m.2 drives and that’s working well. Would definitely recommend that motherboard.

[/quote]

Asus Prime X299-A II. Both of the onboard m.2 slots operate on their own dedicated PCI lanes, and the first two PCI slots can be set to 16+16 at the same time. I added an ASUS Hyper M.2 card for additional m.2 drives and that’s working well. Would definitely recommend that motherboard.
[/quote]
Thanks. I notice there’s a third m.2 slot on that mb, did you utilize it (looks like if installs vertically)?

Not using that one. It uses the chipset PCIE interface, so it’s sharing bandwidth with USB and basically everything else on the motherboard. Might not make a huge differenece in practice, but I didn’t want to take any chances. The add on card w/ 4 m.2 slots was only like $50 and each of those slots is a dedicated connection.