I’m not asserting that at all.
@nettydrown
To Answer your question as to the performance difference. We were not saying there would be a drop in latency. The older R7 3900 series of CPU’s has worse internal latency than the equivalent intel’s. The 5900 series is now on par with regards to latency but the capasity to load more into the project without having issues is greater. As I have previously stated in this tread Cubase tracks are single threaded but more tracks more threads can be used. I don’t know how VSTi’s are handled though. So what this means is that with the 5900 series with AMD the latency is about the same as equivalent Intel but the size of project it will handle will be greater due to more threads. This also depends on track count plugins per track etc.
I think there is something said for buying what you know to work and be reliable and what you are comfortable with.
1 Like
The general idea behind DAWbench is that it stresses any DAW’s CPU in a couple of different types of workloads. The idea is that the DAW isn’t coded specifically for a CPU and instead the actual test looks at how the CPUs perform as the tasks change.
If you wanted to see if this is correct you could check the data over time (years) and see what the delta in performance is between a higher clockspeed for the same architecture and brand versus the delta comparing an increased amount of cores. If the delta as a percentage remains roughly the same when comparing Intel to AMD then that implies that Cubase/Nuendo is more or less agnostic to the brand of CPU. So in other words I think there’s a good case for being able to find out if there is a discrepancy or not by just looking at the data (which I’m sure people have not done).
Another thing is why the assumption should be that there is a difference specifically due to Cubase/Nuendo. Why wouldn’t the assumption be the exact opposite - that there is no difference? And if we assume there is a difference, why would that difference be different from other DAW’s differences? Why would we assume that Cubase/Nuendo would favor a different brand than Reaper?
As for that one post on Gearslutz? Why would you post that one? It’s curious. That poster said they saw no improvement on either Cubase OR Ableton. The poster also never returned to the forum to report any changes and has posted only once since. It’s entirely possible that the problem was a bad driver (Windows related, since Windows was also updated in the process) or a problem with the graphics card.
Here’s a couple of posts of Intel users who get drop outs:
Gearslutz - View Single Post - The “today we build our studio pc” thread
Gearslutz - View Single Post - The “today we build our studio pc” thread
The above two posts do NOT prove there’s a general problem with Intel CPUs running Cubase, right?
Can we really talk about CPU problems/drop outs without also inspecting the mobo as well?
“Can we really talk about CPU problems/drop outs without also inspecting the mobo as well?”
You tell me.
depends if the problems are across multiple mobos and to what extent
“depends if the problems are across multiple mobos and to what extent”
Right.
I Started a new thread.
That article on DAW bench, they said they actually dropped their testing of Cubase on the second round and did not report on Cubase, why, must be because there was something “inconsistent” in the results or even contradictory, else why would they do that? Hum, I wonder, not very scientific. At least they did mention there was “something” they were not telling us. It was of interest to me, but they did say the results were, “an irregularity” - uh what? Ok well gee wonder what that was? Whatever it was it might not have supported the thesis that AMD is out performing Intel within DAW’s… No – that couldn’t be… When I was still learning back in my day, some really smart folks told me about a concept of “missing information”. The idea, that, sometimes it is informative to study what is called “missing information”. That when there are topics that are not reported it can be meaningful, that there may be a “sensitivity” to a segment of information or data. Well, what I notice is that there are no tests to be found of Cubase’s multicore capabilities, and I just wonder why that is. But I know what I have seen in MY DAW with a very multicore system and it has me wanting a much faster single core performance – I will know that answer for me, in a moon phase or so. My main point was that I can’t find tests of multicore with Cubase. I wish there were. Why haven’t anyone done that and reported?
Well using said analogy… Are we saying that Ford or GM don’t leave the same flaws in their product designs when they update? Well they sure do… And no, it’s only cutting edge when you dip your toes into waters where few have tread, AMD has had more incentive to risk their necks trying stuff that was more or less a gamble when it was undertaken where Intel has felt comfortable milking what they felt were their strengths and failing to advance… Kinda just like the legacy auto makers.
Had AMD has more marketshare and been a worthy target for attackers I’m pretty sure the additional money and effort would have been invested in infecting them too, but they don’t even hold 10% in the datarooms so why would anyone waste their time… Their is little safety in obscurity
Alittle snooty aren’t we? Just stating some counterpoints to consider for one who may trust their paycheck to their equipment, so while you don’t mind waiting for proper drivers or compatibility or whatever other speedbumps arise from AMD’s present obscurity in the market, some may not be able to afford to… Maybe plugins are your thing and you need all those virtual instruments to make music for you so all that horsepower is necessary, some are only worried about that final audio that gets laid to disk so keeping it on something of which there are forum after forum after forum to help you correct any errors which may arise is of utmost importance… This is why I try to offload all my labor intensive activities onto machines that if they glitch it’s a minor setback and not complete frickin mess… That’s been my entire input here, if you’re zooming around in a hotrod you’re bound to hit the wall at some point, will it be while your kids are in the car? Too much to risk drivin the kids around in the hotrod is all I’m saying
Yeah, it all seems to be a big conspiracy.
Look, if you actually care about this (I don’t think you do) then you can contact Vin directly and ask why he stopped testing on Cubase. As far as I recall he stated why that was publicly, and if I remember correctly it had to do with Cubase not being predictable or it didn’t handle loading the cores in a good manner - along with Vin’s distaste for the people at Steinberg. I have absolutely zero recollection of it having anything at all to do with either AMD or Intel, and Vin has absolutely been a big outspoken proponent of Intel and a vendor of Intel systems, so he would definitely not have shied away from saying that the problem was as you implied if that’s what it was…
Vin has gone through the various ways in which Cubase assigns work to different threads/cores. He did that on at least three different versions of Cubase where there were three officially different ways of handling multicore loading.
Ask in a new thread dedicated to it maybe?
No, it’s not like that.
That’s utterly beside the point. We weren’t discussing the cause for being susceptible to an attack, someone was saying that Intel was the safe bet. It wasn’t apparently for data centers, regardless of the cause of it not being so.
If you don’t like my attitude then put me on ignore. At least I didn’t call you a name, right?
How many years will my Ryzen 7 1700 system have to be up and running and producing content for major networks in order for it to satisfy your demands for “proper drivers or compatibility or whatever”? How many years for Zen 3? I’m genuinely curious, because I have the feeling that it’ll never ever be enough.
“Some people” said “wait” when my generation came out.
Zen+ came out and people said “wait”.
Zen 2 came out and people said “wait”.
2+ came out and people said “wait”.
Zen 3 is out and people are saying “wait”.
What determines when the wait is over? Benchmarks aren’t enough. User reviews aren’t enough. What’s the determining factor?
So don’t get the latest Intel CPUs then is what you’re saying? When their next gen comes out people should wait how long before buying? Can they buy the gen that just came out?
Cool, good answers! I don’t know who Vin is.
Much appreciate the discourse.
Such test would really have been helpful to pick out a new system with like has been discussed many many variables.
I based my choices in good part on the spec by Chris Selim - Mixdown Online (I like that guy) and then with updates to the newer generation of cpu’s.
“TAFKAT” on Gearspace.com, formerly known as Gearslutz.com. He’s the one one who created the first DAWbench test, and the subsequent ones, and has been in pretty close contact with Pete at Scanproaudio and other builders. He also has a podcast on Spotify. I’ve only heard one episode so far.
Not sure that it’s still useful for any, but as it’s seems that there is a heated AMD vs Intel debate, here is my position, based from experience, concerning it…
I made a new DAW build 16 months ago to replace my aging i7-870 Lynnfield / Windows 7 based one I was using since more than 9 years (resons : aging HDDs : even if they were WD Black Caviar ones, I was beginning to get nervous and no more possible upgrades in the narrow future, as more and more stuff were already requiring W10 - Cubase, among others…).
My first motion was to select an Intel/Asus/WD combo, as my previous one was still working perfectly and I’ve nearly always been (excepting very early at the beginning of the century…) an Intel aficionado… until I looked at this further to see the i7-9700K price at nearly 400 € and saw here and there a lot of praises for the new AMD Zen 2 architecture. After one month scratching my head, I finally choosed the AMD way : Ryzen7 3700X / Asus Prime X470-Pro, 2 SSDs (one of them being a M2 WD one) and W10 Pro. What decided me against the Intel i7-9700K was :
- 7nm architecture (i7-9700K : 14 nm)
- 65 W TDP (i7-9700K : 95 W). It’s important for me, and to be sure of the whole thing, I put on it a Noctua NH-U12S cooler.
- The benchmarks already available which showed in average that the 3700X was 15-20% more powerful than the i7-9700K, which was something like 100 € more costly.
Result : after 16 months of use, I just remember the whole thing crash once and it was a bad move from the user : I simply forgot that Cubase was running when I powered down the DAW, all its windows having been reduced for a while.
With an empty template project including 7 different VSTis (2 multiouts, one being a sampler), 8 audio tracks, 3 external instruments, 4 different buses, 3 reverb sends, something like 7-8 different inserts and the control room active, the audio processing load is more or less 15% (it was something like 35 with my previous setup, using Cubase 10.0.x).
And I guess that the new Zen 3 units are even better. Not that sure about it, concerning the Intel i7 10XXX ones… To anyone worrying about any instability using AMD based setup : from my experience, you shouldn’t.
3 Likes
https://www.youtube.com/watch?v=_jVAfk4AG3A
https://youtu.be/mxiuvQPL_qs
Anyone is entitled to their own preferences, but I find some of the way Intel is defended in this thread (and other threads) quite amusing.
As soon as AMD can deliver (which is the only real problem right now), I will not waste one single thought any longer on even remotely considering Intel for my next Cubase desktop PC in 2021. It will OF COURSE be a significantly stronger AMD alternative. I regard this choice as something being beyond any reasonable doubt at this point in time for myself, according to all information I could gather.
In former years, I decided the same alternative for Intel, as long as it made sense to me. At this point though, my understanding for Intel defenders and some of their funny arguments is rapidly vanishing. We owe nothing to either company, and just want to have the best Cubase experience, like Dom Sigalas.
That’s like Apple users saying arguments for using PCs is rapidly vanishing… yet Apple only has %5-7 of the market. That is a funny argument isn’t it?
I might build an intermediate AMD machine just to run Windows 10, before I build my $10k machine, but, Intel Alder Lake is looking pretty darn good… So I’ll probably wait for their release instead of jumping on the AMD hype train pretty much blind - I mean, the train is moving and the track looks good for as far as I can see which is… the length of the train…
There we go again with what I call funny arguments. Market shares change over time, like they have time and again in this kind of business. They are as irrelevant for buying decisions of a new Cubase PC, as irrelevant can get. Actual delivery, in sharp contrast, is really relevant to me.
If you regard yourself as blind for what AMD delivers right now, I can’t cure your blindness. 
If you want to regard Intel’s promises of future greatness as more real, than powerful, existing, tested alternatives, that’s your business, not mine.
I think it’s pretty clear who will have more fun using Cubase for quite a while to come. I for one don’t care one bit, if you wait for miles long Intel future dream trains, while I simply use Cubase on a powerful Computer, like Dom Sigalas does.
Not blind. just not dumb to change brands because AMD for a - sliver - of time has slightly higher performance. Too me, that is very funny argument and justification. I could buy a BWM right now which goes faster than my current car - but BMWs need a lot of maintenance, and that maintenance/parts replacement costs more than my current car which is a stable workhorse. Should I get another stable workhorse, or should I get a BMW so I can post its benchmark tests on internet forums and claim I’m the best despite maybe that BWM will be garbage in a year.
In a few months time, performance differences will be negated, and Intel will still be the more stable, widely tested, supported, and integrated brand that is made in the USA. And then in a few more months time it will be faster and have support for DDR5 and PCIe5 before AMD does. That’s why it’s silly to jump on the AMD benchmark hypetrain.
whops, more “funny” arguments from me… or maybe just something called… patience!
Or maybe just something called “how about you start your own pro-Intel thread already and leave the evangelism for it”?
I mean, you don’t think you’ve made your point already in this Ryzen thread?
Ffs…