I run Cubase 5.3 64 bit latest update on dual core 3.76 CPU and 8 GB RAM with Tascam MKII 122 soundcard. Every 17 sec I have a CPU spike up till more than 90% capacity. What’s wrong? Can anyone tell what to do?
Regards
Mr Maarloew 
Probably some poorly programmed driver.
Easiest way to resolve this is to download DPC Latency Checker:
http://www.thesycon.de/deu/latency_check.shtml
and look if it “sees” your spikes. If it does, remove/disable every non-essential device (network interfaces, wireless devices, etc, etc …) on our PC until DPC Latency Checker shows no spikes anymore. Now you have found the source of problem. Then it’s up to you if you want to live without the problematic hardware or not.
Note that this does not properly indicate DPC latency under Win 8.1. It tends to register a lot of ~1ms latencies, even though they may be a lot less, but even 1ms is OK for streaming audio.
However, it will indicate with red bars those spikes that will produce problems.
For a more intensive diagnosis, see How to Diagnose and Fix High DPC Latency Issues with WPA (Windows Vista/7/8).
Thanks for this information. I didn’t know. I’m in Win7 … and probably will be until it gets unsupported (just like changed to W7 from XP only few weeks ago).
Nice tool! Tried it, but couldn’t find a way to display highest single latency for a chosen module. This would be the most usable value when searching for source of CPU spikes. Maybe it’s there somewhere, but just didn’t find it.
With that tool you can drill right down and see graphically which routines/processes have conspired to create a problem, if only it were easy to relate that to a specific audible hiccup. The problem with using it is due to the substantial time before one can see the data, and that when you get it, it is tedious to drill down into it. It gives credence to the term ‘data mining’, as in being stuck down underground, well away from the world above, where you really want to be!
LatencyMon is another tool that gives a current reading, as well as some deep stats, but it is not graphical, so they both lack the visual immediacy of DCPLat.
However, DPCLat lacks the ability to select a spike that you just saw happen at the problem audio event and drill down into it to the depth that the other tools give.
Sort of wish that there was some evolutionary leapfrog going on here rather than independent divergent development biases!