On Jan 26, 2007, at 4:32 PM, Tim Ellison wrote:
Geir Magnusson Jr. wrote:
One issue will be repeatability. We should prejudice our benchmark
tracking based on configurations using project facilities, like
build-test.
Even then we have not required that a build-test machine is
dedicated to
the task of compile/function testing. We may need such a requirement
for meaningful performance benchmarking.
Agreed - but we need the first step to make the second meaningful.
At least w/ the first step, a continuously running b-c on any
platform can raise an alarm when we have a performance regression
geir
Tim
