Well, here's one to chuckle over

Asus M2V AMD 3800 X 2

Sonar 3

Soundcard: er, the builtin 'crap' realtek sound device on the motherboard, driven through ASIO4ALL 2.7

Only took one data point: at 5.8ms latency, CPU was 30%.

Not bad I thought for the crappiest possible output device you could pick.

Interestingly, on my laptop (Pentium M 1.5 IBM T40) with UA25, I cannot get this test to run even at max latency setting (48mS reported). What actually happens is that every time I try loading the test, Sonar crashes catastrophically with an addressing exception. This is worrying - but checking, it looks like there's a new driver for the UA25 - will grab that and retry in a moment.

Update: No, version 2.0.1 of the UA25 driver does not fix this problem. Sonar will crash every time the project is loaded, even with latency set to maximum. The crash is a 0xc000005 addressing exception. This is interesting because the laptop has been quite stable for a very long time indeed. (WinXP SP2, S3P 3.1.1)

If I disable the audio driver i.e deselect the UA25, then the project will load. This looks like we may have a driver issue for Edirol to look at. Can anyone else run this test against a UA25 - would be interesting to see if the problem can be reproduced.

In some cases, the speedup came out to more
than 5%, presumably because the measurement isn't really exact.

Microsoft Office 2010