On 4/9/12 2:06 AM, Denis Shelomovskij wrote:
Why will recording the average produce so much noise?
As I explained, the average takes noise and outliers (some very large, e.g. milliseconds in a benchmark that takes microseconds) into account. The minimum is shielded from this issue. In the limit, the minimum for infinitely many measurements is the sought-after result.
As I see, floating point arithmetic is now used without a strong reason so it looks like a time of this part isn't valuable. Or is it just a temporary solution?
I don't understand "time of this part".
Anyway it should be configurable using a CT parameter, so it will not abuse one who doesn't need it.
We'd like the framework to do the right thing, and the average does not seem to be it.
IMHO, a histogram gives lots of information and will be a good addition.
I disagree. Andrei
