Does anyone have any further experiences or light to shed on this.  I would have thought a modern Thinkpad with Intel USB 3.0 chipset would work ok with the XT.  If they are not compatible this would totally influence my purchasing (I'm on a T430).
ExpressCard ports are missing from next gen notebooks.

Have you closed down unnecessary services ie. wifi, lan and so on , tweaked your visual performance. BIOS setting no S.M.A.R.T and so on. Laptops is generally optimized for anything else but Realtime Audio:-).

Hendrup.

Hello together,

my Lenovo Thinkpad Edge E530c (i5) Win7 with Samplitude ProX shows exactly the same behaviour.
Unfortunately it does not have any expresscard slot. So it must work with usb. Which should be the future in recording multitrack with notebooks, because there is no other interface. And thats obviously the reason for the strong investment and research of RME in usb.

The behaviour is as follows:
without starting Samplitude, DPC realy low, less than 250.
When I start the DAW, DPC rises up to 1800. Sometimes Samplitude has cpu-overloads while recording the 2 analog inputs.
There are 3 usb 3.0 Ports which are not burdoned with other devices.
I checked that with the device manager. The single usb 2.0 Port shares with some other devices but behaves similar to the usb 3.0 ports.
I already switched off the dispensable devices in the device manager under the sight of DPC latency checker.
By the way my Madiface XT works perfect with an old Core2Duo Desktop PC even connected to the USB2 port.
I hope there is one "golden" Switch, which has to be toggled, to solve this problem.

Best Stefan

Bingo! that´s been it. "Greenhorn Error". Essentially it has been the change of the energy saving status of win 7 which now is in "Ballanced Mode".
Even choosing "Redundancy Mode" in the driver dialog of Madiface XT helped improving the performance and brought down the dpc to an average load of around 700 µs.
"Ballanced mode" seems to be a bit better than "Maximum Performance Mode" because the cpu-allocation looks a bit more ballanced in the Task Manager.
Besides this, I followed some of the instructions of http://www.timos-welt.de/win7/ which I found in the link collection in this forum.
This fairly new notebook was equipped with a lot of software-supplement-security-and attention-rubbish which I also tidied up some way.

so, thank you and best regards
Stefan

Here is some more background information.

First of all: < the effect is ... caused outside our drivers.> I have to take that back.

While this is usually true, it wasn't with the MADIface USB driver. Its record part uses kernel functions and therefore causes a visible DPC latency. This latency scales with the current buffer size setting, and is therefore only noticeable at 256 and 512 samples.

As this DPC latency is caused within our own driver it does NOT affect recording or playback. DPC values like in the above posts of more than 1 ms do not mean that audio will be impossible at low buffer sizes. It just means that the current computer is most likely suffering from 'normal' dropouts because it is not optimized, so the buffer transfer to the software or the audio processing within the audio software will fail - the driver will not, and those failures are not caused by the DPC latencies of the driver (note that an optimized and powerful computer does not show values over 1 ms, as stated above).

Nevertheless - once we got aware of this we knew we had to improve it for several reasons. And that is what we did in the just released driver 0.9380. The DPC values are down more than 50%, making fully clear that if you have audio problems they are not caused by the driver's DPC latencies.