Intel Core i9-12900K vs. AMDs Ryzen 9 5950X on Cubase Pro 11?

Hello Community and Developers!

Has Cubase Pro 11 already been tested on a Intel Core i9-12900K?

And are there reference values to the performance on a Ryzen 9 5950X?

Thank you for any helpful information!


yes id also like to know this!

I should have my new 12900KF and Windows 11 next Wednesday and I’ll have a good idea soon after how well it works. I have old projects that are so packed with tracks and plugins I can’t even open them anymore, I hope with this new rig they will open right back up and I can finally remix the drums on one of them.


You can use the benchmark proposed by Dom Sigalas to have a basis for comparison.

Here you will find a compilation of the results proposed by users according to their configurations:


Thank you very much, Thalys!

Thanks for the second video you posted. I admit I didn’t know that one and hadn’t even found it on this forum here (I may have missed it). In the second video, Dom’s method is validated, so to speak, or cross-checked. The last thing I wanted to do, though, is cast doubt on Dom’s test. On the contrary, his show is amusing and interesting, and the comments under his video are also very diverse and interesting. But neither Dom’s Test nor the comments are conclusive, they are often even contradictory. So anyway, there remain at least two problems aka open questions that I like to formulate with reference to Dom’s video:

  1. Dom’s test is not fair. He compares a two-year-old i9 from Intel with the latest flagship from AMD. That says nothing about an i7 or even i9 that is comparably up to date compared to the big AMD. Dom’s test doesn’t say anything about the i9 that my post refers to, because the latest Alder Lake wasn’t on the market yet.

  2. There is no clear picture of whether, and if so when, a higher clock rate is preferable to more cores or vice versa. There seems to be a tendency (in the comments) that the advice is to use many cores for many tracks with many plugins (so maybe rather AMD?), otherwise (well: in which cases?) rather higher clock speed (so rather Intel?).

My original question, with which I opened this thread, may not have immediately shown that I am (also) concerned with this complex of topics, which obviously occupies many Cubase users.

Dom’s video and the second video here don’t really help either, because of the already mentioned points and also, because at least I personally am not able to set up two or more fat test machines myself and make comparative tests with the demonstrated test method - or of course even better: with my own projects.

I had hoped that the statements from Steinberg or the developers of Cubase would be more helpful than they are, however. On the website (, the information seems relatively vague to me and almost as if they themselves don’t really know what they could really recommend. And with the components specified on the page with the system requirements page (System Requirements for Steinberg Products | Steinberg) one comes in any case quickly to unbearable limits!

I would find it very helpful if some typical project types are named and clear recommendations are given (e.g. “for an orchestral project with symphonic and other sounds, over sixty tracks, numerous subgroups and multiple plugins on each track this and that, for a recording project with a metal band or something similar with standard instrumentation that and such etc.”). This here How does Cubase utilize multiple cores? gives many interessting points, but isn’t really conclusive either.

One way or another, I would be very happy if perhaps one or the other experienced user can help even further, gladly also the developer team of Cubase!

Anyway, thank you very much in advance!

Thank you very much, Bernard_J!

I’m very excited to hear about your experience!

From here: - View Single Post - DAWBench DSP / VI Universal - Cross Platform DAW Benchmarks :

The first thing to note about that set of test results is that whomever made the chart decided to include 12900K results using DDR5, not DDR4. If you look at the first chart for VI Polyphony you see the processor reach 1820 voices at 64 a sample buffer. If you look at the next chart comparing DDR4 to DDR5 you see that at the same 64 samples the number appears in green which is marked as DDR5.

So, we can learn some really interesting things from this set of tests.

  1. At the lowest tested buffer size the 5950X is better than the 12900K with DDR5. Not by much, but better.

  2. As the buffer size increases the 12900K with DDR5 gets clearly better.

  3. The improvement going from DDR4 to DDR5 as seen in the second chart I posted is pretty remarkable. 50%-70% improvements! (using the VI polyphony test, the DSP test showed little difference)

  4. Using the numbers we find in the DDR4 vs DDR5 chart and looking at the first chart again it’s clear that the 12900K is way behind the 5950X. 1180 vs 1860 at 64 samples, 1920 vs 2920, and then stays behind all the way up to 256 samples (less difference there).

  5. It’s only at the largest 512 buffer that the 12900K with DDR4 beats the 5900X 12-core by AMD!

  6. In the DSP test the 5950X wins at all buffer sizes.

So if we wanted to boil it down based on these charts there seems to be three basic conclusions:

A: If you are going to run mainly DSP and not virtual instruments the best CPU is the 5950X.

B: If you are considering the 12900K with DDR4 and are using a lot of VIs at low buffer sizes the 5950X is still better, and so is the 5900X.

C: If you are using a lot of VIs at low buffer sizes the 12900K with DDR5 is better.

If I got any of that wrong please correct me.


Your are drawing your conclusions from the VI Benchmark mostly. For me that type of load is not interesting. Even pure DSP load is not interesting for me I figured.
Analyzing my larger projects yields that none of them ever made use of the many cores as the DSP test does. Most cores are “idling” around, whereas a single core typically is fully/overloaded (i.e. more complex chains or the Output bus/chain) - and thus completely limit multicore usage.
So I do not need that much parallel processing (my projects do not make use of it).
What I need is fastest core-performance and here the Alder lakes seem to reign.

Now have a 12900k system running and it runs great with Win11/C11 - flawlessly with a bunch of plugins. On those single chains mentioned above I do observe more or less the same core-performance improvements as regular single core benchmarks suggest. However, coming from a 7900x and had no chance to compare to Ryzens.

1 Like

The fundamental problem we have is that most people don’t sit on both Intel and AMD systems with multiple CPUs on either with the ability to then compare them all. All we have in the way of a ‘fair’ test is what Pete and Vin have done. Also SoundOnSound did a test not too long ago. But that’s all we have to go on. Anything else isn’t really apples-to-apples but anecdotal.

Except it doesn’t reign there.

If you say that neither test is “interesting” then you have no data really. And if you do want to look at the tests for guidance then the 5950X is better. It’s ahead on the DSP test on all buffer sizes, and on the VI test it’s ahead of the Alder Lake CPU at 64 samples and only marginally behind at 128.

The question then is whether or not people who use Virtual Instruments typically work with buffers larger than 128. If they do then I agree that the difference is significant between the 12900K and the 5950X on the VI test, but even if that’s the case if you wanted to sort of ‘sum it up’ the latter would still beat the former on most test results (5 out of 8).

1 Like

Synthetic benchmarks do say so for single core (P).
As well you can assume from dsp tests above: 5950x is approx 10% faster in multicore dsp. If the e-cores were same speed as p-cores, each core would be ~0.7% slower than the ryzen cores. Since we know e-cores are less performant than p-cores by a wider margin, the p-cores are thus faster than ryzen cores.


You should lobby Vin (TAFKAT) and Pete to run the tests without locking all cores to the same clockspeed then. Virtually nobody will isolate a single core so most users likely will run the CPUs at stock configuration with them boosting cores “natively” in the most appropriate way.

1 Like

I was able to go to 95 tracks before it stopped using the exact same files Dom was using. It crackles at about 25 though, but I have a two cheap sound cards - that may be why. One drops out at 10 the other at 25. The cheaper one drops out earlier, but has really low latency. (I’m ordering an RME card soon, so we’ll see if that helps keep the sound from dropping out).

That was also while I was still setting up the new machine downloading a large library at 25 Mbps and watching this video of DOM. The CPU never was above 85%. It must be my sound card or the P and E core thing was causing it to hit a wall earlier than I’d like. Quite surprising still. I’m also hitting new lows for latency which I didn’t think was possible:

I kind of think the top of the line 5950x AMD is still better for heavy workloads. Their next gen flagship with DDR5 will likely destroy the i9 12900KF - if you can wait for it.

Though I have to say the switch from i7 8700 is still amazing and satisfying. I’m most impressed by the M2 sample library load times -they are quite literally instantaneous. Not sure if that’s from being able to use the full band width of the drives, or what, but it is very very noticeable.

Look forward to getting started on real projects, higher sample rates, and seeing how much I can really push it in pursuit of some new sounds. One thing is for sure, new machines are really good at doing many things at once. You don’t really need to optimize every last scrap of power anymore… they just work really well.

1 Like

It’s like Captain Charles Pizer said: “We are waiting!” :rofl:

I went for the 5950x with Windows 11, until now its pretty stable and fast.
Not sure if 5950x vs 12900K will matter that much, but it was much cheaper (1000 euro in parts).

Next stop is a good interface…you can have a big performance gain, going from most USB interfaces to, for example, RME PCIe cards…So if you have an older USB interface, you can get the 5950x, and with the saved money, get a new RME internal card. This will be faster than 12900K + old USB interface, and probably more stable, since RME drivers are one of the best in class.

1 Like

look at this intel i9 12900KS

I think your analysis matches my experience using the 12900KF working with large projects in Cubase. I noticed that the lower/faster settings on the buffer quickly need adjusting as I add more Virtual Instruments - but there seems to be no limit to how many I can add as I increase the buffer. A project I remixed that would crash my old i7 8600 now runs smoothly and only uses 24% of my CPU, and I added many complex plugins, like East West convolution reverb. It is a dream machine for virtual instruments, which is perfect for me since I only play the bed tracks in simple piano and then get to work from there.

I believe the new gen of intel and any future hybrid cpus pay the price of having different cores and an AI to interpret the work load with some additional latency. The 5950x and the previous gen of intel CPUs the 11900 are the last of the simple high performance low latency chips. These new gens with increasingly hybrid designs and chiplets all pay a price with latency but you gain big workload returns for huge projects. I was able to unfreeze all 120 tracks without any issue. It’s just astonishing.

Well, at least AMD’s next gen, Zen 4, will still be one core architecture throughout the CPU. Their engineer said that they’re very confident that the production node and the core architecture is so energy efficient that future power management will enable them to run the CPU at very low power, and then it’ll respond appropriately when high performance is needed. I think in terms of pure engineering Intel has for a long time been behind in terms of energy efficiency. It didn’t matter back when AMD wasn’t competing, but now that it does it matters.

So the Intel CPUs are great but consume a lot more power. For AMD to match I think all they’ll need to do is drop performance on cores, and at that point really the only question is if that lower power consumption will let the cores match Intel’s efficiency cores in power/performance.

Either way, we’ve got great options these days. I’m kind’a loving my machine right now.

Soon this debate will be extended to Gen 15 from Intel.

I haven’t noticed any huge power usage outside of benchmarks. It’s very efficient. I do agree, the AMD has a slight edge in performance but the premium is not worth it. The amount you pay extra for a 5950X does not equate in to real savings on my power bill.

1 Like

That could be true.

I can’t recall if I ever did any calculation on actual cost of extra watts used but I’d imagine it wouldn’t be that much money. Plus, my apartment uses a thermostat so it really becomes a question of who pays for heat, right? If you’re in a house that’s using a thermostat and the cost of heating is about the same as the cost of electricity then whatever you lose in power efficiency is what you end up spending less in heating, quite literally. Of course that only applies when heating is working. If it’s not on then that’s another matter.

Generally though I find the question of power efficiency to be mostly interesting from the perspective of future CPU designs and maybe what cooling solutions are or will be required.

1 Like