Thank you so much for the detailed post! My external HDD is "WD My Passport" USB 3.0 - guessing its a 5400 RPM. I'll try moving these projects to the internal HDD and see if that helps. Though when HDD is the problem Ableton shows DISK usage blinking, not just audio crackles. Audio crackles seem to suggest a CPU/latency issue.

I'll get back to you with the details on the rest - thanks a ton for the info. Also on the RayDAT are you suggesting RayDAT (ADAT)->Babyface (ADAT) and Raydat (ADAT2) -> ADA8000?? i'd like to use the BF for main outs for the remote.

Well, the good news is I tried the BIOS thing and it seems to have done the thing. Actually the options you mentioned weren't shown in my BIOS but there was a High Performance setting and I turned that on. Additionally I disabled a bunch of unwanted startup programs also. My windows is already configured to be DAW spec, so I didn't have to fool with that.

Now its running fine (barring an occasional click/pop) with 128 samples latency. CPU usage seems reduced but that could be placebo. In either case it is a HUGE improvement over the previous scenario.

Also latencyMon shows no problems, but as you described it is far from perfect. If I can tweak this to the level you describe, it'll be a blessing.

The External USB vs. internal HDD had no measurable impact - when I copied my project to internal SATA HDD the performance/CPU/etc. was the same. So the external could be a 7200. I can afford to run my whole project in RAM, but I doubt if that would have a significant impact.

Anyways, I've ordered a 9652 card while this was going on, so with that it should bring the CPU even lower.

I have a Asus P8Z68-V LX Intel Z68 Chipset with UEFI BIOS (EZ mode). I'm going to check out that core parking thing tomorrow, but I'd rather not mess with it tbh - it's not clear whether having them run all the time is a good idea for the CPU and power consumption. Disabling C-states is out of the question for me, considering the unknowns.

I have 20 GB but looking at the bios it looks like it's slotted like 2 - 8 - 2 - 8 GB, so I'm not sure if that plays a role (the 2 GB sticks are different from the 8 gb).

Much of the stuff you've described I've already disabled, but I'll take a closer look. From your setup, would you say that the disabling cpu core parking gave you the biggest benefit? or what is it that made the biggest change?

Also I had avg antivirus but I've uninstalled it fearing a cpu hit.

Ah - I just tried it - i.e. set the option in Cubase, left the cubase app open (otherwise it deletes the profile), and ran the project in Live - the crackles were noticeably worse than before.

The crackles seem better without core parking (i.e. without the cubase setting). Or should I only use Cubase to get the benefit?

Sorry, I meant WITH core parking. i.e. with the checkbox enabled. I'll have to experiment with this more but for the moment I'm happy about getting the system back to a usable state, thanks to your tip on the BIOS settings, amongst others.

I'm pretty cautious about making any kinds of changes in fear of breaking things. For instance, I don't know what change I made these last 3 days now the sleep mode is not waking up. i.e. after I set the computer to sleep I can't wake it up with keyboard/mouse. It used to work perfectly before. Things like this that I make one small tweak and it breaks elsewhere is a complete nightmare for me. :-/

You say on one day its good on the other its bad .. is that related to changes that you introduce every day ?
If its good, does it correlate to better LatencyMon values ? In what range are they then (kernel latency timer) ?
If its bad, do you see worse LatencyMon values ?
But please always measure without applications running !! Its all about to see on an IDLE (non busy) system, how much latency there is already at that "quiet" level.

You should have maybe grouped your work into pieces that belong togther.
1. Document BIOS, make pictures of old status, then make changes until LatencyMon kernel latency values on an IDLE system come close to 2 µs and values doesnt show improvements anymore (there are only a few settings which influence this)
2. Backup OS (Disk Image, Data Backup) make changes from checklist, then remeasure with LatencyMon
The values at this point can't be worse as they achieve that you remove load from the system, silly tasks that simply keep the system busy at times where you dont wont it. Then you should have reached baseline for your system, better should not be possible then
3. Then Hardware changes one by one and re-measure on each change
4. If HW introduces Latency issues, try other PCI/PCIe slots. Sometimes depending on the board and PCIe lane distribution etc they influence each other a little. But if you simply remove all cards, then you ought to be back at normal, because without cards there can be no impact on the system. Against this baseline you can measure later when you put the HW back in piece by piece.

Nothing is lost now in my opineon. I would remove all cards in the system, as also the placement of equipment in a PCI/PCIe can have side effects depending on the design of the board. If there are no cards in the PCI/PCIe slots (besides Graphic card), then nothing bad can happen on that side. Then I would consequently on Points 1-4.

One caveat: if I tell backup of system and data I assume you have working backup / restore procedures.
Not that you necessarily need it, its IMHO usual procedure.
For BIOS changes you only need to document the changes, you can go back at any time.
For Windows changes you have the checklist and best practice is to document every step whether you needed to make a change or whether it was already set (i.e. in the area of automatic windows taks, elsewhere).
The only part where you make changes to registry by special tools is in regards to cpu core parking.
But also there make pictures with digital camera how the settings were before, then after.
With that you should at any time be able to revert all changes without requiring any kind of backup.
But I see no point why you shall go back in any way, as these changes are to reduce the kernel latency, which is a basement of having the CPU cores mostly free of doing any tasks that disturb and the CPU frequencies not to change too often up to completely fix at base frequency of the CPU.

After proper tuning of the system (you measured kernel latency of IDLE system during the whole process) you should reach min. values down to 1-2 µs at times, usually around 10 up to 50, but not very high SPIKES. This is where you can't so more on BIOS / system level.

Whats left then are influences of the drivers in your system if you add HW ... maybe unfortune Slot positions for HW influencing each other .. A good strategy is to keep the graphic card in the 1st PCIe slot if you look from CPU. Dont place anything in another PCIe-x16 Slot, as then your graphic card might automatically change to work only in "x8" mode. Not that it has much impact on normal office performance, but I want to simply tell you that.

Then look to place each of the required add-on cards first in the PCIe x1 sockets or eventually into x4 sockets.
If this is a PCI card, then you need to use the PCI sockets, dont remember exactly.

That is then a structured approach of work.

I can't exclude finally that certain drivers have negative impact, but if a driver should introduce latency of the system then you also can use LatencyMon to find out, which drivers cause the highest DPC latency.

Congrats so far. In regards to your measurements.

Important are the *lowest* and *usual* measured kernel timer values, how low the lowest values can get and what the usual values are. This you can only observe while looking at the values within lets say 3 minutes.

The peaks that you tell are also nice to know, but .. especially valuable is how low you can get to rate the results of the tuning. Along with that tuning the max values / the "spikes" of course also will go down usually.

When I look at my current values within 3 minutes:
the lowest one I reach is: 1,75 µs
The usual range is: 5-9 µs
The higher values that sometimes arise are in the range: 30-50µs
The peak value is: 79µs

You need to close all applications (Firefox and what not), run the tool, do not move the mouse.

You enabled the Windows power profile 'HighPerformance'?

And I have used the DPC Latency Checker with no "obvious program" running in the background.
I have about 1000µs constantly.

Alright, so now I have used the Latency Monitor.
And the figures are far from fine.
The first minute or so it showed all green, except the bottom one. But after 2 minutes the numbers went up...

http://imgur.com/kwGpYjE

And here is the report in text form:

_________________________________________________________________________________________________________
CONCLUSION
_________________________________________________________________________________________________________
Your system seems to be having difficulty handling real-time audio and other tasks. You may experience drop outs, clicks or pops due to buffer underruns. One or more DPC routines that belong to a driver running in your system appear to be executing for too long. One problem may be related to power management, disable CPU throttling settings in Control Panel and BIOS setup. Check for BIOS updates.
LatencyMon has been analyzing your system for  0:08:00  (h:mm:ss) on all processors.

_________________________________________________________________________________________________________
SYSTEM INFORMATION
_________________________________________________________________________________________________________
Computer name:                                        BEAST
OS version:                                           Windows 8 , 6.2, build: 9200 (x64)
Hardware:                                             W150ER                          , CLEVO CO.                       
CPU:                                                  GenuineIntel Intel(R) Core(TM) i5-3210M CPU @ 2.50GHz
Logical processors:                                   4
Processor groups:                                     1
RAM:                                                  16274 MB total

_________________________________________________________________________________________________________
CPU SPEED
_________________________________________________________________________________________________________
Reported CPU speed:                                   2494 MHz

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 INTERRUPT TO USER PROCESS LATENCIES
_________________________________________________________________________________________________________
The interrupt to process latency reflects the measured interval that a usermode process needed to respond to a hardware request from the moment the interrupt service routine started execution. This includes the scheduling and execution of a DPC routine, the signaling of an event and the waking up of a usermode thread from an idle wait state in response to that event.

Highest measured interrupt to process latency (µs):   339,918502
Average measured interrupt to process latency (µs):   5,353798

Highest measured interrupt to DPC latency (µs):       337,455324
Average measured interrupt to DPC latency (µs):       1,100948

_________________________________________________________________________________________________________
REPORTED ISRs
_________________________________________________________________________________________________________
Interrupt service routines are routines installed by the OS and device drivers that execute in response to a hardware interrupt signal.

Highest ISR routine execution time (µs):              81,187650
Driver with highest ISR routine execution time:       ACPI.sys - ACPI-drivrutin för NT, Microsoft Corporation

Highest reported total ISR routine time (%):          0,000743
Driver with highest ISR total time:                   Wdf01000.sys - Kernellägesdrivrutin för Framework Runtime, Microsoft Corporation

Total time spent in ISRs (%)                          0,00170

ISR count (execution time <250 µs):                   6542
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 DPCs
_________________________________________________________________________________________________________
DPC routines are part of the interrupt servicing dispatch mechanism and disable the possibility for a process to utilize the CPU while it is interrupted until the DPC has finished execution.

Highest DPC routine execution time (µs):              1442,667602
Driver with highest DPC routine execution time:       ACPI.sys - ACPI-drivrutin för NT, Microsoft Corporation

Highest reported total DPC routine time (%):          0,032259
Driver with highest DPC total execution time:         ntoskrnl.exe - NT Kernel & System, Microsoft Corporation

Total time spent in DPCs (%)                          0,123969

DPC count (execution time <250 µs):                   427372
DPC count (execution time 250-500 µs):                0
DPC count (execution time 500-999 µs):                2
DPC count (execution time 1000-1999 µs):              3
DPC count (execution time 2000-3999 µs):              0
DPC count (execution time >=4000 µs):                 0

_________________________________________________________________________________________________________
REPORTED HARD PAGEFAULTS
_________________________________________________________________________________________________________
Hard pagefaults are events that get triggered by making use of virtual memory that is not resident in RAM but backed by a memory mapped file on disk. The process of resolving the hard pagefault requires reading in the memory from disk while the process is interrupted and blocked from execution.

NOTE: some processes were hit by hard pagefaults. If these were programs producing audio, they are likely to interrupt the audio stream resulting in dropouts, clicks and pops. Check the Processes tab to see which programs were hit.

Process with highest pagefault count:                 svchost.exe

Total number of hard pagefaults                       37936
Hard pagefault count of hardest hit process:          28722
Highest hard pagefault resolution time (µs):          1941652,180834
Total time spent in hard pagefaults (%):              2,471991
Number of processes hit:                              20

_________________________________________________________________________________________________________
PER CPU DATA
_________________________________________________________________________________________________________
CPU 0 Interrupt cycle time (s):                       2,277267
CPU 0 ISR highest execution time (µs):                81,187650
CPU 0 ISR total execution time (s):                   0,025770
CPU 0 ISR count:                                      4281
CPU 0 DPC highest execution time (µs):                1442,667602
CPU 0 DPC total execution time (s):                   0,879226
CPU 0 DPC count:                                      237457
_________________________________________________________________________________________________________
CPU 1 Interrupt cycle time (s):                       1,606514
CPU 1 ISR highest execution time (µs):                9,559743
CPU 1 ISR total execution time (s):                   0,006929
CPU 1 ISR count:                                      2259
CPU 1 DPC highest execution time (µs):                99,782277
CPU 1 DPC total execution time (s):                   0,202004
CPU 1 DPC count:                                      24063
_________________________________________________________________________________________________________
CPU 2 Interrupt cycle time (s):                       1,977440
CPU 2 ISR highest execution time (µs):                3,899358
CPU 2 ISR total execution time (s):                   0,000007
CPU 2 ISR count:                                      2
CPU 2 DPC highest execution time (µs):                158,431836
CPU 2 DPC total execution time (s):                   0,697484
CPU 2 DPC count:                                      107687
_________________________________________________________________________________________________________
CPU 3 Interrupt cycle time (s):                       2,020703
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):                104,799118
CPU 3 DPC total execution time (s):                   0,605674
CPU 3 DPC count:                                      58170
_________________________________________________________________________________________________________