The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Ryzen Threadripper Virtual Instrument Plugins
Old 4th December 2018
  #451
Gear Nut
 

Quote:
Originally Posted by thedberg View Post
I don’t think Dawbench claims to prove anything except the difference in how processor’s react to those specific tests.

Dawbench can’t tell us how many actual instances of our favorite plugin that will run on our new studio computer. But it can give us a decent hint of the *difference* between CPU’s in a few types of workflow.

Two benchmarks, one of only one VI and one of only one effect is not even close to being statically valid yet
people are making claims that Ryzen has trouble based purely off those results! (which are hardly very detailed either, e.g what are the RAM timings and are they being tested under cubase or reaper????).

Anyone who has even a passing interest in CPUs benchmarking knows how different the results can be from one program to the next! (the fact that the old DSP test and the SGA test show different CPUs being faster proves my point!).

It would hardly be a challenge to make a better benchmark (e.g one that uses more than one plugin!)
Old 4th December 2018
  #452
Quote:
Originally Posted by NamesAreAPain View Post

Two benchmarks, one of only one VI and one of only one effect is not even close to being statically valid yet
people are making claims that Ryzen has trouble based purely off those results! (which are hardly very detailed either, e.g what are the RAM timings and are they being tested under cubase or reaper????).

Anyone who has even a passing interest in CPUs benchmarking knows how different the results can be from one program to the next! (the fact that the old DSP test and the SGA test show different CPUs being faster proves my point!).

It would hardly be a challenge to make a better benchmark (e.g one that uses more than one plugin!)
The people here working with Dawbench have repeatedly explaimed RAM timings and the DAW used. And more: they have repeatedly said that they would welcome participation in improving the benchmark. So if you are in a situation that makes it “hardly a challenge” then go do it
Old 4th December 2018
  #453
Lives for gear
 
Pictus's Avatar
 

Quote:
Originally Posted by wr41th View Post
Hello everyone.
Would someone be so kind to clarify the number differences in Petes DAWBench-VI tests to me?
In the Intel test he gets pretty low values, while in the AMD test he has much higher ones, if I'm reading this correctly [I'm visually impaired so maybe I'm missing something]
F.e. one result for the 7820x @64 is 400, the other 1020?

http://www.scanproaudio.info/wp-cont...Test-Chart.jpg

http://www.scanproaudio.info/wp-cont...ch-vi-TR-2.jpg

This is a pretty huge difference, 280 for the TR would basically mean it could theoretically have trouble handling a piano being added to a small project, depending on how many voices the library uses for one key.
AMD's Ryzen Threadripper 2920X CPU reviewed - The Tech Report - Page 7


The tests clear shows that AMD CPUs have power to calculate, so can deal well with physical
modelled instruments
, but its architecture do have problems when dealing with sampled audio.
It will not run lots of KONTAKT pianos like it probably can run lots of Pianoteq ones.

The future AMD Ryzen/Threadripper 3000 (ZEN 2) may change this.
AMD X570 Chipset For Ryzen 3000 Series Zen 2, 7nm CPUs Leaks Out

Old 4th December 2018
  #454
Here for the gear
 

Ah Pictus that is interesting, I wasnt aware of the new line coming next year already. I plan on getting a new machine in ~mid 2019, so who knows I might look into that then, too.
The question still remains though how ScanProAudio can have a 1950x result of 64/280 in one test and 64/880 in the other - six-hundred voices difference?
Curiously the TechReport test also shows hugely different result where it has a 1920x result of 128(!)/240, which is almost 300 below the lowest ScanPro.
I'm not interested in validity discussion but as a consumer trying to orient himself these results are - not as helpful as they could be. Or as I said, I'm overlookin something here, the difference seems to be around the 600 mark in SPA's case, maybe he deactivated cores and I overread it or something?
Old 4th December 2018
  #455
Lives for gear
 
Pictus's Avatar
 

Too much variables to account for...
BIOS settings, Windows version, different audio interfaces, driver versions and etc...
Old 4th December 2018
  #456
Lives for gear
 

Quote:
Originally Posted by wr41th View Post
The question still remains though how ScanProAudio can have a 1950x result of 64/280 in one test and 64/880 in the other - six-hundred voices difference?
You're not the only one, but we went through that discussion at one point. It might have been a different thread. Pete from Scan basically said that because the setups weren't 100% identical that would possibly account for the difference. The BIOS/firmware wasn't necessarily the same, nor the OS version, etc.

Quote:
Originally Posted by wr41th View Post
Curiously the TechReport test also shows hugely different result where it has a 1920x result of 128(!)/240, which is almost 300 below the lowest ScanPro.
One thing you have to remember is that the interface/driver makes a difference when it comes to playing back audio. A 'lesser' interface/driver would give you crackling earlier at a lower buffer size. At one point TechReport had results where several CPUs with different architectures scored exactly the same in some cases, which to me seems completely unlikely. More likely is that there was a bottleneck in the system. And so the test setup actually makes a big difference it appears.

So really it seems that when you look at these results you have to look at an individual tester and see what the differences are between the CPUs for that one tester. You can then do the same for another individual tester. Then you can average the differences perhaps (i.e. for one tester CPU X is 10% faster than CPU Y, for another tester that difference is 15%, calculate the average 12.5% or whatever) . But it's definitely difficult to figure something out by comparing different testers' results to each other "horizontally", for lack of a better word.

Quote:
Originally Posted by wr41th View Post
I'm not interested in validity discussion but as a consumer trying to orient himself these results are - not as helpful as they could be. Or as I said, I'm overlookin something here, the difference seems to be around the 600 mark in SPA's case, maybe he deactivated cores and I overread it or something?
My opinion has always been to set a budget, define what the absolute minimum needs are, define what one wants but doesn't need, and then find the best performance/$. I think that's still possible.

(In the case of Scan's testing I do find it unfortunately that he gave up on testing the 2xxx series Threadripper CPUs since TechReport apparently managed to test them just fine. And supposedly the TR CPUs were looking really, really good until he stopped testing. I think that according to my way of thinking it would have been possible for a user to find that a TR CPU would have been the best purchase despite of the things that made Pete stop testing.)

I agree that using a wider range of plugins during the test would probably be better, but the problem is who is going to decide which those plugins are going to be and how to decide which ones. When we (me and Pete) discussed that briefly my point was that we (I) don't know what it is about some plugins that make them tax the CPU a certain way and not another, and he didn't know either. What I suggested was to either run more than just two tests where different types of plugins were used in each, or to balance the load more using more different types of plugins.

I say this because there have been swings over these past couple of years where the value proposition changed greatly between Intel and AMD, back and forth as it were, and it is hard to decipher where the best value is. A more balanced test would possibly help with this.

But then again, I think it's well worth pointing out: Pete and Vin (TAFKAT) and TechReport are among the very few that make these comparisons and make them possible. If it were so easy to spend the time doing them wouldn't we all do them?
Old 5th December 2018
  #457
Lives for gear
 

Quote:
Originally Posted by wr41th View Post
Hello everyone.
Would someone be so kind to clarify the number differences in Petes DAWBench-VI tests to me?
Different test builds, different OS installs, different Cubase versions, full driver updates and patches installed and a couple of service packs in the middle (2 years worth).

We also received a multi-core load balance optimization tweak between those two dates (that is now implemented in C10), so that would have helped the lowest buffer results on any chip with more than 14 threads.

I freeze the install whenever I start a chart to stop variances over the course of that time frame, but two different charts are not comparable.

Quote:
Originally Posted by wr41th View Post
This is a pretty huge difference, 280 for the TR would basically mean it could theoretically have trouble handling a piano being added to a small project, depending on how many voices the library uses for one key.
At the lowest buffer setting, yeah. That was pretty much my take home from all of the TR testing up until now.

Quote:
Originally Posted by NamesAreAPain View Post
Two benchmarks, one of only one VI and one of only one effect is not even close to being statically valid yet
people are making claims that Ryzen has trouble based purely off those results!
I making those claims off having to support the setups. I get queries all day long, if they were solid platforms then I'd be selling them, it's poor business to turn away sales like that.

As it is, the ones people insist on custom building often come back as not being suitable for the task in hand, so until one starts working properly through updates or architecture changes I'm going to maintain my stance.

Quote:
Originally Posted by NamesAreAPain View Post
(which are hardly very detailed either, e.g what are the RAM timings and are they being tested under cubase or reaper????).
RAM timings make little to no difference for audio work, something that has been tested by ourselves and other users repeated. I retested the first AMD generation with 3200MHz after being pulled on it at the time, absolutely no difference.

I've also tried the Infinity Fabric update from last month that was supposed to optimize all this, once again, zero difference.

DAWBench Vi is under Cubase.
DB DSP is under Reaper.


Quote:
Originally Posted by NamesAreAPain View Post
Anyone who has even a passing interest in CPUs benchmarking knows how different the results can be from one program to the next! (the fact that the old DSP test and the SGA test show different CPUs being faster proves my point!).
Something I've always maintained too, so no argument. What's the answer though, build a test rig with every plugin in existence?

Go back to the original Zen testing, I was so surprised that I went and tested under Sonar and a few other platforms that I don't normally use, simply because I wanted to rule out specific platforms and it's all replicable over all of them.

Yes, you're right that different software, drivers, bug fixes, service patches etc... can skew it. It's why I standardize my benches over the course of a generation and it's why I say that new and old charts are not remotely comparable.

Quote:
Originally Posted by NamesAreAPain View Post
It would hardly be a challenge to make a better benchmark (e.g one that uses more than one plugin!)
Great, and as I say to everyone else, go ahead and I'll use it.

The reason for using a single plug in's is so you can get a proper metric. All plugins have different loads, so how do you compare accurately once you have more than one type of plugin being added/removed?

I don't have time to sit and develop tests, I barely have time to keep up with what's coming out, so I welcome more validation options that are useful in gathering data.

Ultimately I do this testing to establish what I can and can't sell reliably and offer guidance to my own sales desk. It's not done me wrong so far and little of the real world feedback has contradicted me so far. If you've benched your TR machine and are seeing 90%+ CPU loads on a 64 buffer, then please do share the kit list you've used that manages it as I would love to have more solutions to offer clients.

Quote:
Originally Posted by mattiasnyc View Post
One thing you have to remember is that the interface/driver makes a difference when it comes to playing back audio. A 'lesser' interface/driver would give you crackling earlier at a lower buffer size. At one point TechReport had results where several CPUs with different architectures scored exactly the same in some cases, which to me seems completely unlikely. More likely is that there was a bottleneck in the system. And so the test setup actually makes a big difference it appears.
Yes indeed, and to note, Techreport is also using an RME (as do most other people doing this testing) which I could do as I've got units on my desk to do so.

I choose to use a £200 interface as A. we've already seen those other results B. most users don't have a £1000+ best in class interface.


Quote:
Originally Posted by mattiasnyc View Post
(In the case of Scan's testing I do find it unfortunate that he gave up on testing the 2xxx series Threadripper CPUs since TechReport apparently managed to test them just fine. And supposedly the TR CPUs were looking really, really good until he stopped testing. I think that according to my way of thinking it would have been possible for a user to find that a TR CPU would have been the best purchase despite of the things that made Pete stop testing.)
It was simply the fact that it would drop from a stable 70% load to a 45% glitchy mess often unannounced, with no way to recover. Yes, it's fine getting a benchmark result out of it, but I couldn't provide a solution to a studio where the performance ceiling is actively changing in such a fashion. The first time it happens mid-session and I'd have someone drive up and kneecap me!
Old 5th December 2018
  #458
Lives for gear
 

Quote:
Originally Posted by Pete Kaine View Post
Yes indeed, and to note, Techreport is also using an RME (as do most other people doing this testing) which I could do as I've got units on my desk to do so.

I choose to use a £200 interface as A. we've already seen those other results B. most users don't have a £1000+ best in class interface.
I might be wrong of course but I'd think that the more expensive the CPU the more expensive other parts of the system are as well, statistically. In other words; those considering a >8 core CPU are likely to be the ones with the more expensive interfaces.

I'd also think that if pricier converters perform better at lower latencies then that is what you should use in order to test CPU performance. Otherwise the interface will be a de facto bottleneck. This would simply be a matter of a consistency of approach since your approach is to try to isolate CPU performance.

Obviously we who comment on your testing have for the most part not bought your systems and the only support we've given you is a virtual high-five and maybe steered a customer in your direction, so since it's your money I'm certainly not complaining about you not spending more money on an interface....

Quote:
Originally Posted by Pete Kaine View Post
It was simply the fact that it would drop from a stable 70% load to a 45% glitchy mess often unannounced, with no way to recover. Yes, it's fine getting a benchmark result out of it, but I couldn't provide a solution to a studio where the performance ceiling is actively changing in such a fashion. The first time it happens mid-session and I'd have someone drive up and kneecap me!
Knees are quite useful. They make stuff like walking easier, not to mention kneecapping people!

I actually didn't mean that you should have kept on testing at that time considering you had the experience you did. I just meant I'd have liked to have seen you getting a system up and running during re-testing. If TechReport can do it (or have problems with both brands) then it's clearly possible.

But it might be worth reminding people (if it's correct) that you're sharing 'cause you're cool and your main business is actually building DAWs you sell. In that context of course it's absolutely reasonable to do what you did.

So, I wasn't 'criticizing' you.

:-)
Old 6th December 2018
  #460
Lives for gear
 
b0se's Avatar
16 cores @ 5Ghz for $499. I'll believe it when I see it :¬)
Old 6th December 2018
  #461
Here for the gear
 

1920x vs 9900k

Hi all,

I am sorry for jumping into the conversation. I am new in this forum and I saw that you are experienced based on your conversations.

I am planing to build up new Production Machine. Mainly using Cubase and NI Kontakt, Reaktor, Massive, Spark Drum Mashine, Omnisphere, Isotope etc..

So my question is which is does it worth buying AMD TR 1920x over Intel's 9900k?

Still 1920x is 350 euros right now in Germany, and 9900k is 600 euros.

As i can read from many forums including Abletons official forum, they are advising that both things are essential core numbers and clock speed as well, but still 9900k seems to be really good but expensive, but does it worth it spending so much money?

Some people are saying that 8086k would be great choice based on the fact that can be OC up to 5ghz with not so hot temp. Or 8700k would be enough?

What are your advises?

Thank you in advance.

Cheers,
Nikola
Old 6th December 2018
  #462
Gear Addict
 
Lesha's Avatar
Some interesting insights after a BIOS update
Does the AGESA 1.1.0.2 really make the Threadripper 2990WX fly? : Amd
Old 6th December 2018
  #463
Lives for gear
 

Quote:
Originally Posted by Nikola Petrov View Post
Hi all,

I am sorry for jumping into the conversation. I am new in this forum and I saw that you are experienced based on your conversations.

I am planing to build up new Production Machine. Mainly using Cubase and NI Kontakt, Reaktor, Massive, Spark Drum Mashine, Omnisphere, Isotope etc..

So my question is which is does it worth buying AMD TR 1920x over Intel's 9900k?

Still 1920x is 350 euros right now in Germany, and 9900k is 600 euros.

As i can read from many forums including Abletons official forum, they are advising that both things are essential core numbers and clock speed as well, but still 9900k seems to be really good but expensive, but does it worth it spending so much money?

Some people are saying that 8086k would be great choice based on the fact that can be OC up to 5ghz with not so hot temp. Or 8700k would be enough?

What are your advises?

Thank you in advance.

Cheers,
Nikola

I think the 1920x would be a better fit for your workload, as the more threads are better for mixing, especially for this price. Just get lotsa ram!
Old 6th December 2018
  #464
Lives for gear
 
Pictus's Avatar
 

Quote:
Originally Posted by Nikola Petrov View Post
Hi all,

I am sorry for jumping into the conversation. I am new in this forum and I saw that you are experienced based on your conversations.

I am planing to build up new Production Machine. Mainly using Cubase and NI Kontakt, Reaktor, Massive, Spark Drum Mashine, Omnisphere, Isotope etc..

So my question is which is does it worth buying AMD TR 1920x over Intel's 9900k?
No
Quote:
Still 1920x is 350 euros right now in Germany, and 9900k is 600 euros.
As you can see here, the 9900k has more than double performance(DAW Bench VI) to deal with sampled audio.

Quote:
As i can read from many forums including Abletons official forum, they are advising that both things are essential core numbers and clock speed as well, but still 9900k seems to be really good but expensive, but does it worth it spending so much money?

Some people are saying that 8086k would be great choice based on the fact that can be OC up to 5ghz with not so hot temp. Or 8700k would be enough?
May help Intel’s i9 9900K and the Coffee Lake refresh. | Scan Pro Audio
Old 6th December 2018
  #465
Here for the gear
 

Hi Pete and others,

well quite frankly new drivers, BIOS, and Windows updates giving about 300 percent more CPU power is - impossibru.
That being said, in fact the expensive RME and the updates seem to lead to a hefty decrease not an increase in performance.
Anyhoo, after looking at it again, I think one of the reasons is actually not all the updates, but the tests have been made with different samplerates. I think thats the main culprit and also a thing to consider I guess, if one dosnt necessarily work with 96KHz.
Thanks for all the infos though, this seems to be one of the more complicated PC buys I have looked into so far.
Old 6th December 2018
  #466
Lives for gear
 

Quote:
Originally Posted by mattiasnyc View Post
I might be wrong of course but I'd think that the more expensive the CPU the more expensive other parts of the system are as well, statistically. In other words; those considering a >8 core CPU are likely to be the ones with the more expensive interfaces.
It doesn't really break it down like that. I know users doing sound design with £3000 worth of multi-core systems and 2in/2out interfaces. Although, I'm also aware of guys running i5's as hard disk recorders and feeding them with 32 channels over ADAT. Whatever gets the job done really.

Quote:
Originally Posted by mattiasnyc View Post
I'd also think that if pricier converters perform better at lower latencies then that is what you should use in order to test CPU performance. Otherwise, the interface will be a de facto bottleneck. This would simply be a matter of a consistency of approach since your approach is to try to isolate CPU performance.
I get that and its a fair enough comment if we're starting to see a 64 buffer bottleneck become apparent.

I need to add an additional test in the new year and whilst I didn't really want to restart the main charts so soon after setting up the current ones, if I'm going to do it with a different test then it seems that I might as well look at a new interface at the same time.

Quote:
Originally Posted by mattiasnyc View Post
Obviously, we who comment on your testing have for the most part not bought your systems and the only support we've given you is a virtual high-five and maybe steered a customer in your direction, so since it's your money I'm certainly not complaining about you not spending more money on an interface....
I didn't see an issue before, it's only really now that we have multiple results coming in, I can see the weakness with the current setup. It's not that I've got a problem with using another interface, it's just having the need to justify the time to do so hasn't really existed before this point.

Quote:
Originally Posted by mattiasnyc View Post
I actually didn't mean that you should have kept on testing at that time considering you had the experience you did. I just meant I'd have liked to have seen you getting a system up and running during re-testing. If TechReport can do it (or have problems with both brands) then it's clearly possible.
I was going to do a full write up on the retest, but we hit autumn and my time went out the window. The fact that the benches didn't make any real difference after that last BIOS rework just meant that it wasn't interesting enough for me to devote time to writing it up even though I could have done with doing so. I'll consider if the new one makes any significant impact of course.

Quote:
Originally Posted by mattiasnyc View Post
But it might be worth reminding people (if it's correct) that you're sharing 'cause you're cool and your main business is actually building DAWs you sell. In that context, of course, it's absolutely reasonable to do what you did.
Well, yeah. I do it so I know what's worth supporting and so that my sales desk has comparison charts to work from. Given that half of them don't have an audio focus it's also more or less a cheat sheet. Besides that, I make it public facing because my boss says I don't publish enough to our social accounts!

But I'm taking the point. I kept it cheap to reflect the average user and it seemed to do so well enough. Your right though, if it's a bottleneck that hampers testing then it needs to be reconsidered.

Quote:
Originally Posted by mattiasnyc View Post
So, I wasn't 'criticizing' you.
And don't worry about that, I wasn't suggesting that you were. You were more than fair with any previous response, I was just responding to your comment and using it to try and explain my stance to the wider forum. I know that you know, as we've had these conversations before, but for anyone else reading, it may help clarify it a bit.
Old 13th December 2018
  #467
Lives for gear
 
Avgatzeblouz's Avatar
Quote:
Originally Posted by Pictus View Post
the 9900k has more than double performance(DAW Bench VI) to deal with sampled audio.
tracks with heavy
Hi there, do you know if the same applies for heavy mixing tasks ? Lots of tracks with heavy plugins (acustica, Equilibirum, etc...).
Old 13th December 2018
  #468
Lives for gear
 

Quote:
Originally Posted by Avgatzeblouz View Post
tracks with heavy
Hi there, do you know if the same applies for heavy mixing tasks ? Lots of tracks with heavy plugins (acustica, Equilibirum, etc...).
No. Just follow the thread to the images and you can see for yourself the non-VI score ("DSP" score).

Also remember that for mixing a larger buffer is more ok and that makes a huge difference in performance.

On top of that there's for example Steinberg's ASIO-guard which again makes a huge difference.
Old 13th December 2018
  #469
Lives for gear
 
Avgatzeblouz's Avatar
Quote:
Originally Posted by mattiasnyc View Post
No. Just follow the thread to the images and you can see for yourself the non-VI score ("DSP" score).

Also remember that for mixing a larger buffer is more ok and that makes a huge difference in performance.

On top of that there's for example Steinberg's ASIO-guard which again makes a huge difference.
Thanks, found them !
Old 2 weeks ago
  #470
Lives for gear
 
Pictus's Avatar
 

Coreprio – Bitsum
Coreprio was initially Bitsum’s custom implementation of AMD’s Dynamic Local
Mode; more configurable and robust than AMD’s implementation, allowing the
user to set the prioritized affinity, thread count and refresh rate.
Dynamic Local Mode applies to the Threadripper 2990wx and 2970wx CPUs,
addressing their asymmetric die performance.

Now Coreprio also offer an experimental function labelled ‘NUMA Dissociater’.
This feature applies to EPYC and TR systems with 4 NUMA nodes.

Old 2 weeks ago
  #471
Gear Guru
 
UnderTow's Avatar
Quote:
Originally Posted by Pictus View Post
Coreprio – Bitsum
Coreprio was initially Bitsum’s custom implementation of AMD’s Dynamic Local
Mode; more configurable and robust than AMD’s implementation, allowing the
user to set the prioritized affinity, thread count and refresh rate.
Dynamic Local Mode applies to the Threadripper 2990wx and 2970wx CPUs,
addressing their asymmetric die performance.

Now Coreprio also offer an experimental function labelled ‘NUMA Dissociater’.
This feature applies to EPYC and TR systems with 4 NUMA nodes.

Ah! I was just reading about this and wanted to post an article link when I found out I had accidentally closed the relevant browser tab.

I would be curious to see how audio benchmarks fare with this.

Alistair
Old 2 weeks ago
  #472
Lives for gear
 
b0se's Avatar
Indeed, looking forward to seeing how this fares in the audio world. Cheers for sharing.
Old 2 weeks ago
  #473
Gear Addict
 
Lesha's Avatar
If the rumors turn out to be true, new mainstream 7nm Ryzen CPUs with 12 or 16 cores and higher clocks would be great to have. I guess we will find out soon, at CES.
Old 2 weeks ago
  #474
Lives for gear
 
b0se's Avatar
FYI @monkeyman2019 has a 1950x build and has been posting in the El Rey thread. He/she is getting some great numbers - 120 instances of El Rey running at once. Crazy for an AA plugin. Would be interested to see that combined with VSTi's.

Quote:
Amd Threadripper 1950x- Stock

32gb ram at 3000mhz Xmp 2.0 profile

Asrock Taichi x399m

Samsung 970 nvme x2 in bootable raid 0.

Windows Server 2016.

Reaper X64
44K @ 512

Link to the post and thread: El Rey is finally out - Greg Wells' Signature Tube compressor plugin
Old 1 week ago
  #475
Here for the gear
 

Quote:
Originally Posted by b0se View Post
FYI @monkeyman2019 has a 1950x build and has been posting in the El Rey thread. He/she is getting some great numbers - 120 instances of El Rey running at once. Crazy for an AA plugin. Would be interested to see that combined with VSTi's.



44K @ 512

Link to the post and thread: El Rey is finally out - Greg Wells' Signature Tube compressor plugin




What vsti are you thinking about. If there are demo avalable. Would be willing to try time permitted
Old 1 week ago
  #476
Lives for gear
 
b0se's Avatar
Quote:
Originally Posted by monkeyman2019 View Post
What vsti are you thinking about. If there are demo avalable. Would be willing to try time permitted
Hey MM. Omnisphere perhaps, or if you really want to test the machine, U-He Diva!

Diva: The spirit of analogue | u-he

Create a track with Diva first with some programmed midi, and whack on some AA's (El Rey being fine too). How many of those you could duplicate would be a serious test.

Thanks!
Old 1 week ago
  #477
Here for the gear
 

Quote:
Originally Posted by b0se View Post
Hey MM. Omnisphere perhaps, or if you really want to test the machine, U-He Diva!

Diva: The spirit of analogue | u-he

Create a track with Diva first with some programmed midi, and whack on some AA's (El Rey being fine too). How many of those you could duplicate would be a serious test.

Thanks!

If there is a demo available no problem at all. I think i will post all my future reporting in this threadripper thread.

Regrading omnisphere I have a client who i swilling to let me test in the next month or so.
Old 1 week ago
  #478
Gear Maniac
 

Interesting to see threadripper experiences, thanks for sharing. Diva will definitely run in demo mode, either with dropouts or noise as demo limitations. CPU usage depends greatly on the patch, maybe it's easiest to play chords with the init patch in divine mode so everyone can easily replicate.
Old 1 week ago
  #479
Lives for gear
 
b0se's Avatar
Quote:
Originally Posted by monkeyman2019 View Post
If there is a demo available no problem at all. I think i will post all my future reporting in this threadripper thread.

Regrading omnisphere I have a client who i swilling to let me test in the next month or so.
Yup, there's a demo - download link on the page I posted previously. Would love eo see how your machine handles it.

BTW, when you get to the 50% CPU load and adding another El Rey crashes your machine - have you tried loading a non AA plugin? Or does it happen with any plugin, as soon as you hit 50% CPU?
Old 1 week ago
  #480
Here for the gear
 

Quote:
Originally Posted by b0se View Post
Yup, there's a demo - download link on the page I posted previously. Would love eo see how your machine handles it.

BTW, when you get to the 50% CPU load and adding another El Rey crashes your machine - have you tried loading a non AA plugin? Or does it happen with any plugin, as soon as you hit 50% CPU?
Another one on the list to try.

Thanks for the recommendation
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump