Do the dropouts only happen occasionally ? Or constantly ?

Do you have the same issues with the same project, but with 44.1 sample frequency ?

In regards to LatencyMon measuring:

I wrote it already many times in similar threads. You can find some other threads where this has been described as well. Here again:

For Windows 7 I would take LatencyMon v4.02, which still measures kernel latency timer.
https://www.dropbox.com/s/imcbu6lcac5vg … 2.exe?dl=1
https://www.dropbox.com/s/9h6uoqgcaqqvd … e.md5?dl=1

LatencyMon needs to be run on an IDLE machine, where no application is running! Even avoid mouse movements, which generates additional interrupts for the CPU.

LatencyMon needs to run more than one minute if you want to catch problems maybe caused by automatically started jobs from Microsoft taskmanager, background jobs that wake up for some activity and alike, which all can also happen during a longer recording session. For a 1st good start I would let it run for 10 Minutes and once you got it fine tuned for better values, then for validation maybe 30-90 minutes.

Before measuring, freshly reboot / login and wait for approx 3 minutes, until all startup/background jobs have run.

YOU> I ran LatencyMon for about a minute, it tells me that everything is OK (green)

Which doesnt mean anything. LatencyMon shows green with kernel timer latencies up to 1000 microseconds
if I remember right. Such high values would not be good even if the tool shows green.
And as you record in 96kHz sample frequency your system is even more under stress, as the CPU has to process the double amount of data compared to 44.1/48 and you generate more i/o to/from disk.

While the tool is running I recommend that you make some notes on a paper about the shown values of kernel timer latencies in microseconds. Because the tool shows you only in the 2nd bar the maximum value. But this maximum value is so to say the highest (worst) peak of kernel timer latency. But also interesting is to know
- what was the lowest value your system was able to reach ? On my system 1,75 microseconds
- in what range are often recurring low values ... on my system 5..11 microseconds
- in what range are often recurring peaks values .. in my system there are no often recurring peaks, sometimes 40
- my highest measured kernel timer latency is 123 microseconds

Then please post a screenshot of the measurement in this thread and report the values which I suggested to observe while the tool was running. Yes its boaring to watch 10 min a screen but its required. Sadly the tool doesnt present a graph for these values which would be extremely useful.

Please post also a screenshot of the TAB "Drivers" in LatencyMon.
Sort the values 1st by Highest Execution in ms and make the colum size a little bit narrower so that all is better visiable.

And post the content of the TAB "Stats".

Thanks

Did you try already all USB3 ports on the system for the UFX+ ?
Were they all from chipset or do some come from an add-on chip, pls check mainboard description / manual for that.

> Since my UFX+ is not anywhere near then I probably do something terribly wrong.
> I even closed all VMs and apps except for TotalMix and Sonar but that did not help at all.

VMs ? Virtual machines ? Do you use the system for virtualization ? If yes which vendor/solution: VMWare, .. ?

Just to confirm that the UFX+ is capable of handling a severe audio load... Driver Version 0.9612 ...

So it must be something on your system, therefore the many questions, where I hope you answer them completely.

As some of you know I made already a performance test with 400 audio tracks with 44.1kHz,
each track having 2 VST inserts loaded. Even with ASIO buffersize of 32 samples no dropouts.
http://www.tonstudio-forum.de/blog/inde … cks-de-en/

I changed this project now to 96 kHz .. the audio files become now twice as large and the CPU has more data to process.

But even with 400 Audio tracks @96 kHz I can playback the project without audio loss via UFX+ -> AES -> ADI-2 Pro without any loss at lowest ASIO buffersize of 64 samples, which is the lowest for projects @96kHz.

Each of the 400 audio tracks have these 2 inserts loaded:

From taskmanager you can see my systems memory is full and some data has been moved to pagefile.
And you see that the CPU load of all 6 cores / 12 threads of the E5-1650v3 is around 39-55% with average about 45%.

In the RME driver settingsdiaload you can also see 96kHz and ASIO buffer size.

Simple thing to try. Change the graphics card driver to standard VGA or the included in windows (choose update driver from device manager and choose 2 times choose yourself)

Agreed!

Thank you Vinark for a hint, I will try it at some point. Before that I indeed would like to collect all initial info as Ramses suggested.

LatencyMon, measure on almost idle machine, just Skype and this browser was running:

_________________________________________________________________________________________________________
CONCLUSION
_________________________________________________________________________________________________________
Your system appears to be suitable for handling real-time audio and other tasks without dropouts.
LatencyMon has been analyzing your system for  0:10:02  (h:mm:ss) on all processors in the system.

_________________________________________________________________________________________________________
SYSTEM INFORMATION
_________________________________________________________________________________________________________
Computer name:                                        HELIOS
OS version:                                           Windows 7 Service Pack 1, 6.1, build: 7601 (x64)
CPU:                                                  GenuineIntel Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
Logical processors:                                   8
Processor groups:                                     1
RAM:                                                  32710 MB total

_________________________________________________________________________________________________________
CPU SPEED
_________________________________________________________________________________________________________
Reported CPU speed:                                   3997.0 MHz
Measured CPU speed:                                   832.0 MHz (approx.)

Note: reported execution times may be calculated based on a fixed reported CPU speed. Disable variable speed settings like Intel Speed Step and AMD Cool N Quiet in the BIOS setup for more accurate results.

_________________________________________________________________________________________________________
MEASURED KERNEL TIMER LATENCIES
_________________________________________________________________________________________________________
Highest measured kernel timer latency (µs):           377.045124

_________________________________________________________________________________________________________
MEASURED SMIs AND CPU STALLS
_________________________________________________________________________________________________________
Highest measured SMI, IPI or CPU stall (µs)           0.0

_________________________________________________________________________________________________________
REPORTED DPCs
_________________________________________________________________________________________________________
Highest DPC routine execution time (µs):              477.057793
Driver with highest DPC routine execution time:       ndis.sys - NDIS 6.20 driver, Microsoft Corporation

Highest reported total DPC routine time (%):          0.325621
Driver with highest DPC total execution time:         iusb3xhc.sys - Intel(R) USB 3.0 eXtensible Host Controller Driver, Intel Corporation

Total time spent in DPCs (%)                          0.472110

DPC count (execution time <250 µs):                   1435578
DPC count (execution time 250-500 µs):                0
DPC count (execution time 500-999 µs):                3
DPC count (execution time 1000-1999 µs):              0
DPC count (execution time 2000-3999 µs):              0
DPC count (execution time >=4000 µs):                 0

_________________________________________________________________________________________________________
REPORTED ISRs
_________________________________________________________________________________________________________
Highest ISR routine execution time (µs):              161.030773
Driver with highest ISR routine execution time:       dxgkrnl.sys - DirectX Graphics Kernel, Microsoft Corporation

Highest reported total ISR routine time (%):          0.096773
Driver with highest ISR total time:                   dxgkrnl.sys - DirectX Graphics Kernel, Microsoft Corporation

Total time spent in ISRs (%)                          0.161721

ISR count (execution time <250 µs):                   864871
ISR count (execution time 250-500 µs):                0
ISR count (execution time 500-999 µs):                0
ISR count (execution time 1000-1999 µs):              0
ISR count (execution time 2000-3999 µs):              0
ISR count (execution time >=4000 µs):                 0

_________________________________________________________________________________________________________
REPORTED HARD PAGEFAULTS
_________________________________________________________________________________________________________
Process with highest pagefault count:                 none

Total number of hard pagefaults                       72
Hard pagefault count of hardest hit process:          70
Highest hard pagefault resolution time (µs):          2950.217663
Total time spent in hard pagefaults (%):              0.000488
Number of processes hit:                              0

_________________________________________________________________________________________________________
PER CPU DATA
_________________________________________________________________________________________________________
CPU 0 Interrupt cycle time (s):                       18.432363
CPU 0 ISR highest execution time (µs):                149.311984
CPU 0 ISR total execution time (s):                   4.524244
CPU 0 ISR count:                                      703945
CPU 0 DPC highest execution time (µs):                477.057793
CPU 0 DPC total execution time (s):                   10.317062
CPU 0 DPC count:                                      678224
_________________________________________________________________________________________________________
CPU 1 Interrupt cycle time (s):                       4.997062
CPU 1 ISR highest execution time (µs):                152.834626
CPU 1 ISR total execution time (s):                   0.959452
CPU 1 ISR count:                                      45080
CPU 1 DPC highest execution time (µs):                245.854391
CPU 1 DPC total execution time (s):                   3.530980
CPU 1 DPC count:                                      169461
_________________________________________________________________________________________________________
CPU 2 Interrupt cycle time (s):                       1.556279
CPU 2 ISR highest execution time (µs):                135.726795
CPU 2 ISR total execution time (s):                   0.053923
CPU 2 ISR count:                                      3068
CPU 2 DPC highest execution time (µs):                211.363523
CPU 2 DPC total execution time (s):                   0.180284
CPU 2 DPC count:                                      21531
_________________________________________________________________________________________________________
CPU 3 Interrupt cycle time (s):                       4.465268
CPU 3 ISR highest execution time (µs):                161.030773
CPU 3 ISR total execution time (s):                   0.70860
CPU 3 ISR count:                                      32526
CPU 3 DPC highest execution time (µs):                238.859144
CPU 3 DPC total execution time (s):                   3.194049
CPU 3 DPC count:                                      178526
_________________________________________________________________________________________________________
CPU 4 Interrupt cycle time (s):                       1.79570
CPU 4 ISR highest execution time (µs):                130.317738
CPU 4 ISR total execution time (s):                   0.049236
CPU 4 ISR count:                                      3074
CPU 4 DPC highest execution time (µs):                171.184388
CPU 4 DPC total execution time (s):                   0.220853
CPU 4 DPC count:                                      24196
_________________________________________________________________________________________________________
CPU 5 Interrupt cycle time (s):                       2.858066
CPU 5 ISR highest execution time (µs):                152.654491
CPU 5 ISR total execution time (s):                   0.623343
CPU 5 ISR count:                                      29383
CPU 5 DPC highest execution time (µs):                234.165624
CPU 5 DPC total execution time (s):                   1.912515
CPU 5 DPC count:                                      100083
_________________________________________________________________________________________________________
CPU 6 Interrupt cycle time (s):                       2.622846
CPU 6 ISR highest execution time (µs):                137.893420
CPU 6 ISR total execution time (s):                   0.235451
CPU 6 ISR count:                                      17705
CPU 6 DPC highest execution time (µs):                210.477858
CPU 6 DPC total execution time (s):                   0.613526
CPU 6 DPC count:                                      87992
_________________________________________________________________________________________________________
CPU 7 Interrupt cycle time (s):                       3.874437
CPU 7 ISR highest execution time (µs):                151.903928
CPU 7 ISR total execution time (s):                   0.639181
CPU 7 ISR count:                                      30090
CPU 7 DPC highest execution time (µs):                243.167376
CPU 7 DPC total execution time (s):                   2.781961
CPU 7 DPC count:                                      175568
_________________________________________________________________________________________________________

25us

35..90us

110-150us

DRIVERS TAB
How to insert pictures? here is the link to dropbox
https://www.dropbox.com/s/jikbum94x1fcu … s.jpg?dl=0

You should best close all applications when running the tests so that its really an IDLE system.
Especially Webbrowser and Chat applications might keep tcp sessions / network communication open...

And when I see that you have there a high number of DPCs with the ndis driver.

Highest DPC routine execution time (µs):              477.057793
Driver with highest DPC routine execution time:       ndis.sys - NDIS 6.20 driver, Microsoft Corporation

But also with the USB3 driver ... Do you have the DAW application open or why is there so much communication via USB3, this I do not understand.

Highest reported total DPC routine time (%):          0.325621
Driver with highest DPC total execution time:         iusb3xhc.sys - Intel(R) USB 3.0 eXtensible Host Controller Driver, Intel Corporation

But if your lowest kernel latencies are not below 25us and usually at around
35..90us and 110-150us

Then I assume, that you most likely have not disabled power saving in your BIOS.

How are the settings for C-States (eventually also P-States, T-States), enhanced Halt (C1E) ?
You should disable them all .. sleep / halt states of a CPU cause latencies.

In my blog article there is a table where you can see which sleep state causes which latency.
This can be quite huge: http://www.tonstudio-forum.de/blog/inde … -X10SRi-F/

EDIT:
You need to completely disable energy saving in the BIOS, C-States, T-States, P-States.
Also C1E needs to be disabled.

The presentation of these values differs a little bit between products. On some systms you can disable them, on some systems you need to limit the energy saving state to C0 (no saving) or at least C0/C1. C1E has to be disabled.

Best is you make a picture with camera and post screenshots of the BIOS page(s) where the settings are displayed to see what the current setting ist.

And please answer the rest of the question, especially what you mean by "VM".

I think your machine has simply too much stress for recording.
Doing virtualization with potentially some load on the VMs will not work on your system.

I also use my system for many things, BUT, only one thing at a time, either office, or gaming, or recording, see what I mean ?

And .. you cannot compare and argue with your old interfaces with a handful of channels with the UFX+, which is a 188 channel interface. No mattter how many channels you actually use in your DAW project in the USB3 communication always all 188 channels are active.

SSD alignment is ok, you see it by the performance values themselves, but explicitely by the driver in use printed in green letters.

If your blower start blowing and making noise this is an indication CPU load, heat. Either this is really required because your CPU becomes too hot based on load rr the blower control need adjustment or your blowers are simply (sorry) crap and you need silent blowers for your case and the CPU.

As your system runs now with a high multiplier instantly everything in terms of energy saving issues should be solved. To disable CPU core parking with a desktop PC is fully valid. Steinbergs Cubase does this as well if you activate an options in the settings to use Steinbergs optimized energy profile. Then it overrides the systems energy profile during runtime.
If you do this change with the normal system profile (Highest Performance), then you have the benefits out of it also for other application and when you are recording not using cubase.

I think you should either take care of shutting all the VM stuff down, so that recording can make full use of the systems resources or you should use 2 different PCs. One as Home server for all the stuff that you are doing and one system which is simply optimized for recording.

If you have a PCIe x4- x16 Slot free you could try to get the  Sonnet Allegro and see whether this makes a difference.
https://www.heise.de/preisvergleich/son … 79334.html
This card uses MSIs (message signaled interrupts)
https://docs.microsoft.com/en-us/window … interrupts

But what you are doing there, recording on a device which seems to be more something like a server with the 4 VMs, that is not a valid basis for a stable recording system.

I added some stuff to my previous posting #18, please re-read (https://www.forum.rme-audio.de/viewtopi … 48#p125648).

I strongly suggest to split it into 2 machines. Or do a parallel installation of Windows for only recording purposes.
The way you use your system now is counterproductive for the purpose of recording without any issues.

You can give it a last try with the Allegro Pro as I wrote as the driver with the MSI might be superior to normal chipset / USB3 driver on your system.

In terms of noise reduction, you can view articles about my own system.
http://www.tonstudio-forum.de/blog/inde … mponenten/
http://www.tonstudio-forum.de/blog/inde … -X10SRi-F/
http://www.tonstudio-forum.de/blog/inde … arrow-ILM/
http://www.tonstudio-forum.de/blog/inde … al-Design/

If you can turn C-States completely off with one switch, then do this.
With my BIOS I cant do this and needed to tweak several options, you have it simpler with this one switch.

You can eventually get some more robustness out if you disable EIST and maybe let TURBO activated.
In a system where I had much problems I had a lower but 100% static CPU clock.
That was for that system (i7 quad core) perfect.

In my current system I let EIST enabled, so that I have under Win7 the control over CPU core parking and system clock by using park control to tweak each of the Energy Profile to my demands....

EDIT: btw many thanks alex for the many screenshots and information.

Thank you, Ramses, for all your answers.

Yes, I agree that VMs and other apps do not affect recording in a good way Of course. And when I was testing UFX+ I had not run VMs. I do not run anything extra and yet UFX+ still cannot reach Konnekt. All I say is that Konnekt 8 did not seem too be affected by VMs. But again - I have not been running VMs with UFX+, it's not capable to run on my idle machine, leave alone VMs.

Having another PC for recording is not an option at the moment.

Truly speaking, I think that my system in general -- i7-4790K + ASUS Z97A + 32GB RAM + SSD drives -- is not bad for audio recordings. I am fine to shut down the unused apps. I have no servers/services running. I am not recording 50 channels simultaneously, I have 5-10 tracks, recording 1 per time so I expected that the flagship RME interface would be capable to do better than being behind 12-years old TC Konnekt 8.

TC Konnekt 8 doesn't require all those CPU tweaks - it just works. While RME interface apparently requires a lot of tuning, and even new hardware - to the extent of buying a new PC! I am not sure if that's what one expects from a flagship.

May be I made a mistake and I should have bought UCX instead... but it has 2 preamps only.

I am going to try Thunderbolt as recommended by RME. That should help me to issue the final verdict.

Thank you for your answers and suggestions.

So you do not recommend to proceed with Thunderbolt? You think that USB2/USB3 on my machine is not an issue - but something else?

Actually, an installation of another instance of Windows could be a good idea. That would reduce bad factors down to hardware. I just don't have a free SATA port for a new SSD drive... in fact I do not have an extra SSD too Let me see, I might be able to re-use one of the existent drives.

And to your remark about 188 channels - I thought some of them on UFX+ are digital so they are not being transmitted? What's the point to transmit no-signal? Also, even for analog inputs, can't UFX+ detect the inputs are not being used and exclude them from transmission? This way there could be no load on USB2/USB3 at all, just like for Konnekt 8.

Thank you!

For €50 you can get ~128GB SSD which allows installation of Win7 and recording applications.
The already existing SSD you can use then to store data/recording projects.
Then you have even split the OS and Recording related I/O load across 2 SATA ports ...

> I will format it before installation

Formatting is completely useless on a SSD and only causes wear on the SSD without any need.
Better repartition the SSD to make some more room at the end by using tools like easeus partition manager free.

Also to think of .. your SSD might require approx 10% free unpartitioned space at the end of the SSD
("overprovisioning") to have enough free blocks when heavy I/O occurrs.

> Actually, would you recommend Win7

Win7 but before doing any Windows update pls change the parameters NOT to install the recommended updates along with important updates. If you remove this option, then only the important updates will be installed (this is whats not possible with Win10 anymore, so its an instantly changing system).

Yes, I have that done.
regarding SSD formatting - I meant a simple format, I think it just destroys the address table and doesn't do anything to the actual content, i.e. no wear occurs.

1) Isn't Iso Streaming disabled by default?
2) There is no need to detect that in real-time. Say if I'd like to disable MADI permanently (or, in general, disable any I/O) then I (un)check the corresponding input/output checkbox in settings and UFX just ignores that input / doesn't send anything to that output. This way we'd save bandwidth and processing time, wouldn't we?
In fact, you guys already do that when UFX+ switches to USB2. You limit the number of channels anyways. It would be good to give user that sort of control, correct?

Thank you for your reply, Matthias.

Yes, I did, on my machine USB2 is identical to USB3.

When you say "broken system" - what's your criteria for "being broken"?
Sorry, I do not consider my system to be broken to the extent of not being able to play only 3-4 guitar tracks (Guitar Rig 5 and BiasFX) without constant dropouts on buffer=256 samples with 100% disabled parking.

Frankly, I still do not understand what is broken in my configuration.Whether it's :
- USB3 (and apparently USB2).
- CPU-related issues.
- Windows 7 / BIOS misconfiguration
- misconfigured peripherals
- bad software
- bad hardware

or a combination of those.

Believe me, I like UFX+ and I would like to keep it, hence I am trying to find the flaws in my system that hard.

I will try the same setup on clean Windows 7 and clean Windows Server (if find all drivers).

LATENCY ON FRESH WIN7

I have installed a fresh Win7 on a separate SSD partition.

First, I measured latency on it:
Minimal kernel timer latency: 1.02us
average was really fluctuating around 5-20us, quite unstable, peaks at 20..100us, about 10-20% of time (each 5th..10th)

Then I uninstalled nvidea video driver:
Minimal kernel timer latency: 1.02us
average had become stable around 3-7us (mostly around 3-4), with rare peaks at 8..12us

_________________________________________________________________________________________________________
CONCLUSION
_________________________________________________________________________________________________________
Your system appears to be suitable for handling real-time audio and other tasks without dropouts.
LatencyMon has been analyzing your system for  0:10:02  (h:mm:ss) on all processors in the system.

_________________________________________________________________________________________________________
SYSTEM INFORMATION
_________________________________________________________________________________________________________
Computer name:                                        UFX-PC
OS version:                                           Windows 7 Service Pack 1, 6.1, build: 7601 (x64)
CPU:                                                  GenuineIntel Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
Logical processors:                                   8
Processor groups:                                     1
RAM:                                                  32710 MB total

_________________________________________________________________________________________________________
CPU SPEED
_________________________________________________________________________________________________________
Reported CPU speed:                                   4000.0 MHz
Measured CPU speed:                                   4592.0 MHz (approx.)

Note: reported execution times may be calculated based on a fixed reported CPU speed. Disable variable speed settings like Intel Speed Step and AMD Cool N Quiet in the BIOS setup for more accurate results.

_________________________________________________________________________________________________________
MEASURED KERNEL TIMER LATENCIES
_________________________________________________________________________________________________________
Highest measured kernel timer latency (µs):           41.215494

_________________________________________________________________________________________________________
MEASURED SMIs AND CPU STALLS
_________________________________________________________________________________________________________
Highest measured SMI, IPI or CPU stall (µs)           0.511994

_________________________________________________________________________________________________________
REPORTED DPCs
_________________________________________________________________________________________________________
Highest DPC routine execution time (µs):              184.799250
Driver with highest DPC routine execution time:       ataport.SYS - ATAPI Driver Extension, Microsoft Corporation

Highest reported total DPC routine time (%):          0.005205
Driver with highest DPC total execution time:         rspLLL64.sys - Resplendence Latency Monitoring and Auxiliary Kernel Library, Resplendence Software Projects Sp.

Total time spent in DPCs (%)                          0.012255

DPC count (execution time <250 µs):                   560957
DPC count (execution time 250-500 µs):                0
DPC count (execution time 500-999 µs):                0
DPC count (execution time 1000-1999 µs):              0
DPC count (execution time 2000-3999 µs):              0
DPC count (execution time >=4000 µs):                 0

_________________________________________________________________________________________________________
REPORTED ISRs
_________________________________________________________________________________________________________
Highest ISR routine execution time (µs):              11.466250
Driver with highest ISR routine execution time:       ataport.SYS - ATAPI Driver Extension, Microsoft Corporation

Highest reported total ISR routine time (%):          0.028395
Driver with highest ISR total time:                   ataport.SYS - ATAPI Driver Extension, Microsoft Corporation

Total time spent in ISRs (%)                          0.030570

ISR count (execution time <250 µs):                   1407652
ISR count (execution time 250-500 µs):                0
ISR count (execution time 500-999 µs):                0
ISR count (execution time 1000-1999 µs):              0
ISR count (execution time 2000-3999 µs):              0
ISR count (execution time >=4000 µs):                 0

_________________________________________________________________________________________________________
REPORTED HARD PAGEFAULTS
_________________________________________________________________________________________________________
Process with highest pagefault count:                 none

Total number of hard pagefaults                       2694
Hard pagefault count of hardest hit process:          1321
Highest hard pagefault resolution time (µs):          124912.272250
Total time spent in hard pagefaults (%):              0.183291
Number of processes hit:                              0

_________________________________________________________________________________________________________
PER CPU DATA
_________________________________________________________________________________________________________
CPU 0 Interrupt cycle time (s):                       2.581707
CPU 0 ISR highest execution time (µs):                11.466250
CPU 0 ISR total execution time (s):                   1.473107
CPU 0 ISR count:                                      1407652
CPU 0 DPC highest execution time (µs):                184.799250
CPU 0 DPC total execution time (s):                   0.569659
CPU 0 DPC count:                                      532881
_________________________________________________________________________________________________________
CPU 1 Interrupt cycle time (s):                       0.047058
CPU 1 ISR highest execution time (µs):                0.0
CPU 1 ISR total execution time (s):                   0.0
CPU 1 ISR count:                                      0
CPU 1 DPC highest execution time (µs):                11.83450
CPU 1 DPC total execution time (s):                   0.001453
CPU 1 DPC count:                                      2379
_________________________________________________________________________________________________________
CPU 2 Interrupt cycle time (s):                       0.068399
CPU 2 ISR highest execution time (µs):                0.0
CPU 2 ISR total execution time (s):                   0.0
CPU 2 ISR count:                                      0
CPU 2 DPC highest execution time (µs):                10.510
CPU 2 DPC total execution time (s):                   0.000315
CPU 2 DPC count:                                      337
_________________________________________________________________________________________________________
CPU 3 Interrupt cycle time (s):                       0.141082
CPU 3 ISR highest execution time (µs):                0.0
CPU 3 ISR total execution time (s):                   0.0
CPU 3 ISR count:                                      0
CPU 3 DPC highest execution time (µs):                12.142250
CPU 3 DPC total execution time (s):                   0.007122
CPU 3 DPC count:                                      7767
_________________________________________________________________________________________________________
CPU 4 Interrupt cycle time (s):                       0.076579
CPU 4 ISR highest execution time (µs):                0.0
CPU 4 ISR total execution time (s):                   0.0
CPU 4 ISR count:                                      0
CPU 4 DPC highest execution time (µs):                12.760
CPU 4 DPC total execution time (s):                   0.001060
CPU 4 DPC count:                                      789
_________________________________________________________________________________________________________
CPU 5 Interrupt cycle time (s):                       0.084682
CPU 5 ISR highest execution time (µs):                0.0
CPU 5 ISR total execution time (s):                   0.0
CPU 5 ISR count:                                      0
CPU 5 DPC highest execution time (µs):                7.15450
CPU 5 DPC total execution time (s):                   0.001136
CPU 5 DPC count:                                      1015
_________________________________________________________________________________________________________
CPU 6 Interrupt cycle time (s):                       0.064325
CPU 6 ISR highest execution time (µs):                0.0
CPU 6 ISR total execution time (s):                   0.0
CPU 6 ISR count:                                      0
CPU 6 DPC highest execution time (µs):                9.42850
CPU 6 DPC total execution time (s):                   0.000350
CPU 6 DPC count:                                      351
_________________________________________________________________________________________________________
CPU 7 Interrupt cycle time (s):                       0.156121
CPU 7 ISR highest execution time (µs):                0.0
CPU 7 ISR total execution time (s):                   0.0
CPU 7 ISR count:                                      0
CPU 7 DPC highest execution time (µs):                4.1170
CPU 7 DPC total execution time (s):                   0.009450
CPU 7 DPC count:                                      15438
_________________________________________________________________________________________________________

Drivers

Processes

CONCLUSION
On buffer=256 RME UFX+ became better but still in some places crackles were audible. Even if I played a single track, dropouts were audible at constant time spots. Konnekt 8 sounded approximately like that but on buffer size=128.

Original Win 7
Konnekt 8 Firewire === 512 samples === OK
UFX+ === 512 samples === dropouts

CPU/BIOS optimization, no CPU parking
Konnekt 8 Firewire === 256 samples === some dropouts
UFX+ === 512 samples === OK now

Fresh minimal Win 7, no nvidea driver
Konnekt 8 Firewire === 256 samples === OK
Konnekt 8 Firewire === 128 samples === minor dropouts exist even when playing back a single guitar track
UFX+ === 256 samples === minor dropouts exist even when playing back a single guitar track
UFX+ === 128 samples === massive dropouts

Fresh minimal Win Server 2012R2, no nvidea driver, no LAN driver
Sonar did not produce proper sound for UFX+ no matter what buffer size was set. Most probably due to system misconfiguration.

As we can see, the latency on this fresh Win7 install is tremendously better than on my regular Win7 and yet that did not make a too significant difference for UFX+ playback. Though a difference exists, it's pronounced but still doesn't allow to run a single track in Sonar without dropouts when buffer size=256. Konnekt 8 improvement due to latency had the same magnitude as UFX+ improvement.

So I guess my next step would be purchasing Thunderbolt card + TB3-TB2 adapter (which is a bit expensive for an experiment). The problem is that I am not sure if that is going to improve the performance. Matthias said it should but my concern is whether buffer size = 256 could be reachable on my system with TB.

Yeh, I will need to prepare the test environment first, it will take time.

I am using USB cable shipped with my UFX+

I connect it directly to m/b. There are 4 USB3 ports and only one is used -- by UFX+.

With Lindy cables I had no issues so far
http://www.lindy.de/CROMO-USB-3-0-Kabel … p;pi=41612

3m is the maximum for USB3 I tend to say, so pls try with 2m to have some "headroom".

Another thing .. if you keep the RME driver settings dialog open, then you see there an USB Diagnosis.
Keep this window open for some time, also when you work with DAW and are getting these crackles.

Do you see any USB transport errors there ?

So I have conducted a series of new tests yesterday. Below are the results with my comments

An existent, small 96/24 project in Reaper, on my original Win7
- 7 Guitar Rig tracks
- 1 track Dimension-based pad, (soft synth)
- basic XLN Addictive Drums 2.

Both UFX+ and Konnekt 8 played back audio in Reaper at any buffer size, including 64 (WTH?)
Initially I thought something is wrong because there was absolutely no difference when switching buffer sizes, somehow Reaper manages to handle that. Not sure how. However, once I enabled echo for 1 guitar track then the dropouts appear right away - on 64 samples. Nevertheless, 256 samples made it stable in Reaper with one track with echo, i.e. it's possible to record and hear what you play.

That was on my original "broken" Windows machine. Then I moved to the fresh Win7:

UFX+/USB2, Win7-fresh installation, small NEW project in Sonar
- 5 guitar tracks
- 8 voice tracks (duplicated)

Sonar was able to play back and echo 1 guitar with no issue on buffer size = 128. What?

How come my reference Sonar project cannot do that even on 256 although the number of unmuted tracks is about the same? OK, I began to delete tracks and after deletion of 6 muted tracks from the reference project I was able to play Sonar on buffer=128 with no dropouts! I do not understand what's so special about those tracks. Furthermore, if I add NEW guitar tracks (different VSTs though - but not anything small - full-scaled GR5 presets) - that doesn't bother UFX+ and Sonar still plays on buffer=128 with no dropouts! What's going on? The old presets are also pretty standard, the only difference was that they were applied on a bus, i.e. old 6 tracks routed signals to each own bus while new tracks have Guitar Rig directly in track's FX bin. It feels almost like if UFX+ somehow detects the tracks were created with a rival Konnekt 8 and doesn't want them to play properly ))

The same case was observed with another reference Sonar project, now with Bias FX guitar tracks (Bias is more demanding than Guitar Pro). And again I deleted a lot of MUTED tracks and suddenly UFX+ could playback with no dropouts on buffer size = 128.

Then I switched back to USB3 and the results were the same as for USB2.

How come muted tracks still influence the playback? Unless user unmutes them they should not participate in anything?

Anyways, it appears that the initial issue is quite resolved. I can have a decent buffer size = 256 with UFX+ (and probably even
128, which is excellent for guitar on 96KHz) if I keep the number tracks small (10-15) in the project OR if I find a method to somehow mute the tracks properly so they do not affect the playback. I believe any DAW should have such option, too bad I am not aware about such in Sonar.

P.S. I have ordered Asus Thunderbolt EX II card for my ASUS Z97A m/b -- apparently the last item in thousands km around -- it should arrive in a week or so. Once I try TB2 with UFX+ and post the results.

Thank you Ramses for your great help and support, thanks to RME for having a patience (I heard that is not common in German customer service? Although Ramses' example tells the opposite) and thank you all for your attention.

Maybe also another thing to consider.
I personally think its not required to record at 96 kHz.
You put your system only under much more stress.
Audio Files only become longer, it wastes more space etc.
I have doubts that in a double blind test you would be able to differentiate between 44.1 and 96 kHz.
More important is to strictly use at least 24 bit depth for higher dynamic and SNR.
Eventually you do yourself a favour to go back to 44.1 especially as you seem to work with a lot of VSTs and also VST instruments. Could definitively give you more headroom in terms of system load / bigger projects.
Maybe try it out.
Also maybe cubase in terms of the ability to freeze tracks of VST instruments but still having the capability of changing the volume and doing panning of this instrument.

Summary for my own, where we stand.

At the end it turns out to be an application problem, as with Reaper you didnt have these severe problems, is this correct ?

So you / we hope that Thunderbolt give you a little bit more headroom so that it runs eventually without issues with Sonar.

Which version of Sonar do you have exactly ?

I would think yes absolutely. If the problems are USB related I would expect the UFX+ to outperform firewire on TB by a margin. Here at least PCI outperforms firewire.

Yes, it appears that your theory might be correct. Though I do not expect a magic from MSI, otherwise all other USB controller vendors would have employed this technique. Perhaps, it's not just MSI.

My PC has no Firewire on m/b, i.e I have a PCI card with TI chip on it so essentially my Firewire is PCI. Or you mean that Firewire itself (TI chip) is the bottleneck? If so then yes, TB PCI card should be at least on par with Firewire PCI card, since they share the same PCI but TB itself is much faster than Firewire.

Let's wait for the delivery of TB Card. Once it gets here I'll perform tests and post the results.