Reason 9.5 & above CPU Stress Tests (2017, two different songfiles included)!

This forum is for discussing Reason. Questions, answers, ideas, and opinions... all apply.
User avatar
EnochLight
Moderator
Posts: 8405
Joined: 17 Jan 2015
Location: Imladris

17 Feb 2020

Grumbleweed wrote:
16 Feb 2020
What a difference a couple of years make! My Ryzen 9 3900X just played through the complex track with no problem and played through (with crackles) the simple track to the end.
Sweet :D

Grum.
tronam wrote:
16 Feb 2020
It's true. I clearly remember not being able to run the Complex Benchmark song at all a few years ago; The audio engine would just die the moment I hit play. Now I can play through the entire song with power to spare. A pretty wild reality check for just how far we've come.

Mac (Retina 5K, 27-inch, 2019) | Intel 8-Core i9 9900K | Radeon Pro Vega 48 | 48GB RAM
Indeed - and those cutting edge monster power machines you guys got are pretty sweet, too. :o My aging 3770K still whimpers at the thought of attempting these song files.
Win 10 | Ableton Live 11 Suite |  Reason 12 | i7 3770k @ 3.5 Ghz | 16 GB RAM | RME Babyface Pro | Akai MPC Live 2 & Akai Force | Roland System 8, MX1, TB3 | Dreadbox Typhon | Korg Minilogue XD

User avatar
tronam
Posts: 486
Joined: 04 Mar 2015

17 Feb 2020

EnochLight wrote:
17 Feb 2020
Grumbleweed wrote:
16 Feb 2020
What a difference a couple of years make! My Ryzen 9 3900X just played through the complex track with no problem and played through (with crackles) the simple track to the end.
Sweet :D

Grum.
tronam wrote:
16 Feb 2020
It's true. I clearly remember not being able to run the Complex Benchmark song at all a few years ago; The audio engine would just die the moment I hit play. Now I can play through the entire song with power to spare. A pretty wild reality check for just how far we've come.

Mac (Retina 5K, 27-inch, 2019) | Intel 8-Core i9 9900K | Radeon Pro Vega 48 | 48GB RAM
Indeed - and those cutting edge monster power machines you guys got are pretty sweet, too. :o My aging 3770K still whimpers at the thought of attempting these song files.
Well, thankfully the prices keep coming down and the core counts keep going up. Once the time comes to upgrade it should be a pretty significant boost in performance. That's the beauty of waiting it out for a bit; very satisfying upgrades.
Music is nothing else but wild sounds civilized into time and tune.

User avatar
Grumbleweed
Competition Winner
Posts: 214
Joined: 16 Jan 2015

18 Feb 2020

The only thing that bothers me with the new set up is the boot time. It used to be a fast PC booted quickly and a slow one was, well slow. The new one is slow due to the motherboard peeing about doing self checks or something rather than just getting on with things.
Once it is up though it is a beasty boy. Moving the Reason content (that you can't choose where to install) from the C to E drive was over in a flash - I was most impressed :D .

Grum.

User avatar
EnochLight
Moderator
Posts: 8405
Joined: 17 Jan 2015
Location: Imladris

18 Feb 2020

Grumbleweed wrote:
18 Feb 2020
The only thing that bothers me with the new set up is the boot time. It used to be a fast PC booted quickly and a slow one was, well slow. The new one is slow due to the motherboard peeing about doing self checks or something rather than just getting on with things.
Once it is up though it is a beasty boy. Moving the Reason content (that you can't choose where to install) from the C to E drive was over in a flash - I was most impressed :D .

Grum.
You should be able to turn most of that self-check stuff of in your motherboard UEFI settings and speed things up during start-up.
Win 10 | Ableton Live 11 Suite |  Reason 12 | i7 3770k @ 3.5 Ghz | 16 GB RAM | RME Babyface Pro | Akai MPC Live 2 & Akai Force | Roland System 8, MX1, TB3 | Dreadbox Typhon | Korg Minilogue XD

User avatar
TheDragonborg
Posts: 285
Joined: 10 Feb 2020

18 Feb 2020

lol my computer played the entire complex song and not stopped once and it topped out at around 75% usage...

And I didn't even bother changing any options... I use the aisio4all driver cause it works better with my PreSonus Audiobox USB than it's official driver does... also there is no CPU limit option in Reason 10.

It also played the entire fx chain track yet started to get choppy halfway through...

My PC has an Core i7 8700k (hex core) overclocked at 4.8Ghz with 32GB of RAM.

Anyway... any project of mine never exceeds 2% CPU usage. And they are full of VSTs and REs.

stekme
Posts: 29
Joined: 19 Sep 2017

02 Dec 2020

Any Apple M1 tests please =)

stekme
Posts: 29
Joined: 19 Sep 2017

06 Dec 2020

Reason 9.5 2017 COMPLEX RE benchmark song V2 -----> http://s.go.ro/0tpcy7fd

Apple M1 MacMini 2020, 8Gb RAM:

SampleRate / Time
Снимок экрана 2020-12-06 в 12.44.20.png
Снимок экрана 2020-12-06 в 12.44.20.png (137.78 KiB) Viewed 2518 times
vs MacBook Pro 2018 i7 6-core, 2.2Ghz, 32gb RAM:
128: 32 sec
256: 57 sec
512: 1:06 min
1024: 1:17 min

Mac mini 2018 i7 6-core 3.2Ghz 32RAM:
128: 1:00 min
256: no limits!

:puf_unhappy:

ReasonStudios... We waiting M1 optimisation! :exclamation:

User avatar
EnochLight
Moderator
Posts: 8405
Joined: 17 Jan 2015
Location: Imladris

06 Dec 2020

stekme wrote:
06 Dec 2020
Reason 9.5 2017 COMPLEX RE benchmark song V2 -----> http://s.go.ro/0tpcy7fd

Apple M1 MacMini 2020, 8Gb RAM:

SampleRate / TimeСнимок экрана 2020-12-06 в 12.44.20.png

vs MacBook Pro 2018 i7 6-core, 2.2Ghz, 32gb RAM:
128: 32 sec
256: 57 sec
512: 1:06 min
1024: 1:17 min

Mac mini 2018 i7 6-core 3.2Ghz 32RAM:
128: 1:00 min
256: no limits!

:puf_unhappy:

ReasonStudios... We waiting M1 optimisation! :exclamation:
Which version of Reason did you test? (I realize you used the old 9.5 test file).
Win 10 | Ableton Live 11 Suite |  Reason 12 | i7 3770k @ 3.5 Ghz | 16 GB RAM | RME Babyface Pro | Akai MPC Live 2 & Akai Force | Roland System 8, MX1, TB3 | Dreadbox Typhon | Korg Minilogue XD

stekme
Posts: 29
Joined: 19 Sep 2017

06 Dec 2020

R11.3.6
Not find newer test files, but this file already help to understand that currently, M1 under rosetta2 work very slow.
Not listen anybody, who talk that M1 working today better than oldest intel-macs. Because Rosetta working as is as...

I can say more.
Optimised Reaper Native with internal plugins for M1 version (!!!) in tests also show that this Intel-macs models working much much better than M1.
And Reaper M1 is native.. working slower than Reaper on Intel machine...
But.. in tests M1 is 4 core, i7 is 6-core.. If look at test as per-core results, M1 core near to the i7 8generation.

Don't know, M1 Reason version is help or not to have better results than Reaper native M1...
So, for music better to wait 12-core M1X in 2021 Q1..

User avatar
EnochLight
Moderator
Posts: 8405
Joined: 17 Jan 2015
Location: Imladris

06 Dec 2020

I’m sure Reason will eventually be properly recompiled for the Apple M1 chips. Will probably take some time though, if history is any indication.
Win 10 | Ableton Live 11 Suite |  Reason 12 | i7 3770k @ 3.5 Ghz | 16 GB RAM | RME Babyface Pro | Akai MPC Live 2 & Akai Force | Roland System 8, MX1, TB3 | Dreadbox Typhon | Korg Minilogue XD

stekme
Posts: 29
Joined: 19 Sep 2017

06 Dec 2020

Yes, of course Reason for M1 must been published to work properly without Rosetta under MacOs with M1 as native app.
I just talk about current M1 chip. M1 not created for music production at all.

If we look to the Reaper created for M1, we can understand that M1 very near to the previous generation (8th) of the Intel.
And users who bought right now new M1 mac, must understand, that he bought Mac, that work slower than Intel MacBook Pro 15' 2018 2.2Ghz 6-core.

Reaper M1 app test show that Intel Macs 2018 (mini and mbp) is faster than current M1.
In this situation, I recommend not run to change Intel-macs to the M1 machines right now.

Better to wait announced by insiders next generation in 2012Q M1X (12-core mbp 16").. Just wait few month and you can buy much more powerful Mac.
Current M1 machines is like a office machines at near line of Arm-macs.

User avatar
fieldframe
RE Developer
Posts: 1037
Joined: 19 Apr 2016

06 Dec 2020

I don't think the M1X is going to fix this, actually, as it appears to be something in the way Reason is interacting with Rosetta, probably with multi-core support.

YouTuber Jon Sine had a few videos last month where he originally concluded the M1 wasn't ready for music production, but then followed a tip from a commenter and changed Logic's multi-core settings. Originally, the M1 MacBook Air could only run 9 instances of Serum, versus 21 on an 8-core Intel i9 MacBook Pro. However, after changing the setting, the M1 could run 18 instances:


So the M1 + Rosetta is clearly capable of rivaling a 2019 Intel 8-core in DSP tasks, and should be able to do considerably more than a 2018 6-core, but something is holding it back. Has anyone tried with and without hyperthreading enabled?

User avatar
tronam
Posts: 486
Joined: 04 Mar 2015

06 Dec 2020

fieldframe wrote:
06 Dec 2020
I don't think the M1X is going to fix this, actually, as it appears to be something in the way Reason is interacting with Rosetta, probably with multi-core support.

YouTuber Jon Sine had a few videos last month where he originally concluded the M1 wasn't ready for music production, but then followed a tip from a commenter and changed Logic's multi-core settings. Originally, the M1 MacBook Air could only run 9 instances of Serum, versus 21 on an 8-core Intel i9 MacBook Pro. However, after changing the setting, the M1 could run 18 instances:

So the M1 + Rosetta is clearly capable of rivaling a 2019 Intel 8-core in DSP tasks, and should be able to do considerably more than a 2018 6-core, but something is holding it back. Has anyone tried with and without hyperthreading enabled?
The main issues were that Logic Pro defaults to setting the Core Audio processing threads to "Automatic" and on the M1 Macs this will only end up using 4 out of 8 CPU cores, so setting it manually to 8 greatly boosts the performance potential. Also, when Logic itself is running as a native app, it loads Intel AU plugins under a separate emulation process that doesn't tend to be very performant or reliable at high plugin counts. The trick for much better performance is to instead run Logic as an Intel app, then all 3rd party plugins can just natively run within it without having to be individually translated. It's an interesting workaround that works surprisingly well. I suspect Reason AU/VST runs a bit differently than most plugins, but I'd be curious to know if it works better in Logic like this as opposed to standalone under Rosetta 2.
Music is nothing else but wild sounds civilized into time and tune.

User avatar
Boombastix
Competition Winner
Posts: 1929
Joined: 18 May 2018
Location: Bay Area, CA

07 Dec 2020

R9.5 Complex file
Windows 10
Buffer 1024
Rate 44.1kHz
Reason 10.4
Ryzen 9, no OC

Bar 44: Load average 31%. What can I say, this shit is awesome :lol:

CPU Core shows that one core gets more, so the song file is perhaps a bit unbalanced and slower computers may perhaps underperform due to one core hitting 100%?
.
R9.5 Complex.gif
R9.5 Complex.gif (18.13 KiB) Viewed 2423 times
10% off at Waves with link: https://www.waves.com/r/6gh2b0
Disclaimer - I get 10% as well.

stekme
Posts: 29
Joined: 19 Sep 2017

07 Dec 2020

fieldframe wrote:
06 Dec 2020
So the M1 + Rosetta is clearly capable of rivaling a 2019 Intel 8-core in DSP tasks, and should be able to do considerably more than a 2018 6-core, but something is holding it back. Has anyone tried with and without hyperthreading enabled?
Youtubers already saiyng that M1 is revolution. But in music productio - better not look at not deeper tests by youtubers, but make own deeper tests to understand how really M1 work today. Running the same projects on various machines.
And looking not only to the quantity of one-plugin instances... Youtubers make tests of instances quantity at all.. and I think its bad tests to understand what M1 can do really..
And ofcourse - you can listen before - about DAWBench?
Tests used - DAWBench, its very famous real test of DAWs. And it been used in tests. (Youtubers not going this way and just run instances to max, its absolutely synthetic tests, out of the real work).
Does anbybody from youtubers try to use DAW bench? No)))) But they talk M1 is super powerful.. No.. Real tets show anothe for native M1 apps (without rosetta!)

I already write there, that we can test anything that work without Rossetta2 on arm-mac.
My example is a REAPER M1 app. With internal M1-optimised plugins. Its all. Optimised for M1.

What we can do with Reaper?

Two mac mini: 2018 and M1.
Image
Intel.png
Intel.png (200.33 KiB) Viewed 2409 times
Image
M1.png
M1.png (221.45 KiB) Viewed 2409 times
Checking in Reaper for this two namchines Longest RT block : 2.77/2.67 ms (hold)

Final results:
INTEL: 344 PLUGIINS
M1 : 200 PLUGINS

Finaly, Intel-Mac Mini 2018 x1.72 more powerful than M1.
Ok, Intel Mac 6-core, M1 - 4core. Anyway, per-core, its nearest results. But at the end, intel-Mac in native M1 app tests, going after Intel-Mac Mini 2018

No needs to wait Reason M1 version, to understand that M1 not better in DSP levels than Intel Mac Mini 2018 6-core.
Last edited by stekme on 07 Dec 2020, edited 2 times in total.

stekme
Posts: 29
Joined: 19 Sep 2017

07 Dec 2020

fieldframe wrote:
06 Dec 2020
Has anyone tried with and without hyperthreading enabled?
Logic Pro for M1. With enabled HT on М1 (8-cores M1 and 12-Intel):
Intel mac mini 2018 - 272 ChromaVerb
М1 mac mini - 224 ChromaVerb

stekme
Posts: 29
Joined: 19 Sep 2017

07 Dec 2020


Its true in this video, about midi tracks... Logic on buffer 32

Intel - 89 tracks
М1 - 103 tracks (8-cores) / 67 tracks (4-cores)
For midi tracks yes. M1 good.
For audio tracks - no.

User avatar
miyaru
Posts: 624
Joined: 28 Oct 2019
Location: Zaanstad, The Netherlands

07 Dec 2020

It is interesting to read all the test stuff, I'm not a Mac guy, and proberly will not be so in the near future, but it is fun to read all about it.

In the meanwhile, I downloaded the test file - the complex one - and ran it on my 3 yr old Intel i7-7700 4 core rig with 16 GB ram and two NVMe's. One for the OS and software the second for data.

And to my surprise it runs well with the file until the last few bars, were my cpu max's out to 100%. Not bad for such a machine! Was fun....
Greetings from Miyaru.
Prodaw i7-7700, 16Gb Ram, Focusrite Scarlett 18i20 3rd gen, ESI M4U eX, Reason12, Live Suit 10, Push2, Presonus Eris E8 and Monitor Station V2, Lexicon MPX1,
Korg N1, Yamaha RM1x :thumbup:

User avatar
miyaru
Posts: 624
Joined: 28 Oct 2019
Location: Zaanstad, The Netherlands

07 Dec 2020

Here you can see it stresses my system to it's paces......
Attachments
CPU @100%.png
CPU @100%.png (284.09 KiB) Viewed 2367 times
Greetings from Miyaru.
Prodaw i7-7700, 16Gb Ram, Focusrite Scarlett 18i20 3rd gen, ESI M4U eX, Reason12, Live Suit 10, Push2, Presonus Eris E8 and Monitor Station V2, Lexicon MPX1,
Korg N1, Yamaha RM1x :thumbup:

User avatar
Kategra
Posts: 327
Joined: 18 Jan 2015

07 Dec 2020

Sample rate = 44,100 Hz
Buffer length = 256 samples

Hardware:
  • CPU Ryzen 5950X PBO active, SMT BIOS deactivated, Air cooled.
    RAM 32 GB Quad Kit 3000 Mhz, CL15
    SSD Samsung 250 GB (Windows 10 Creators edition 64 bit + Reason installed on SSD)
    RME ASIO Fireface 400 audio interface
Software:
  • Windows 10 Home (version 2004)
    Reason 10 (4d4)

Results:
  • Complex RE = plays the whole project without any crackles or pops ,(SME deactivate, only real cores active) CPU load in task manager=52%

    Simple FX Chain = varies wildly:
    first and also the best play: 2 min and 5 seconds perfect then crackles and pops start to appear.
    (restarted Reason) second and third play: 1min 15 seconds
    (restarted Reason) forth play: 1min 35 seconds

COMPLEX
44100 Hz buffer 256 52% CPU max Ryzen 5950X.jpg
44100 Hz buffer 256 52% CPU max Ryzen 5950X.jpg (777.42 KiB) Viewed 2359 times

SIMPLE FX: 1st play:
Simple 44100 Hz buffer 256 79% CPU max Ryzen 5950X 2m5s .jpg
Simple 44100 Hz buffer 256 79% CPU max Ryzen 5950X 2m5s .jpg (985.93 KiB) Viewed 2359 times


Simple FX 4th play:
Simple 44100 Hz buffer 256 67% CPU max Ryzen 5950X 1m40s .jpg
Simple 44100 Hz buffer 256 67% CPU max Ryzen 5950X 1m40s .jpg (947.72 KiB) Viewed 2355 times

User avatar
Kategra
Posts: 327
Joined: 18 Jan 2015

07 Dec 2020

I did a Windows restart and I've changed the buffer length to 1024 samples (max for my audio interface) and the results are both amazing and weird for the SIMPLE FX "benchmark".

What happens is that between ~2 min 40 seconds and 3 min I think I hear one or two small pops at the end of the loop, BUT after that (~ 3 min mark) the audio is perfectly clean and plays the rest of the project with no clicks or pops while the CPU goes to 100% utilization.

I was expecting to hear clicks and pops at some time because the CPU is hitting 100% utilization minutes before the end of the project... but no - audio stream is crystal clear. I guess the "Simple FX benchmark" is broken somehow or Reason has some really great processing tricks implemented :D
Simple 100% CPU 1024 buffer.jpg
Simple 100% CPU 1024 buffer.jpg (892.84 KiB) Viewed 2351 times

Fusion
Posts: 128
Joined: 24 Mar 2017
Location: UK

29 Aug 2022

256 buffer Crackles after about 10 secs.Max 4096 gets to about 30 secs with same crackles. Only using about 30% cpu.

MacBook Pro 2015 i7 2.8Ghz Intel Graphics 16gb ram 512gb SSD, Traktor Z2 Soundcard.

Just tried installing windows via bootcamp and performance was similar although the GUI was a lot more glitchy so will stick with Mac.

Thinking the Mac mini i7 2018 would be a good upgrade choice? or anyone know of a cheaper low power option? I see the Hp EliteDesk with same 8700 i7 cpu as the mini , but not sure if performance will be the same.

Post Reply
  • Information
  • Who is online

    Users browsing this forum: No registered users and 25 guests