One of my earliest posts notes how I managed to tweak the performance so that Manager loads quickly. So much so, we used the function in EISNet v1.6.1 which was being installed over the Summer holidays.
The function I reviewed again today is the same function.
The origins if this function is a mystery to me but I do know it's one that has existed in EISNet since the beginning of time and it possibly the most fundamental procedure in the entire product - yet it's also the weakest. This procedure is referenced constantly throughout the entire source code at all levels. It's even shared among the other applications in the EISNet tools set. It's slow, clunky and quite frankly poorly written.
Although my previous change to the function had improved the performance, I still was not happy with it. It's not smart code - and I like all code to be 'smart'.
So I took the opportunity to have a quick go rewriting it just to see what would happen. I wasn't disappointed.
The end result is another 500ms off the Manager start up time. Manager now loads in 1.5 seconds on a typical Primary School system. Not bad when you consider Manager v1.5 would take 5 or 6 seconds.
I don't think we will be able to squeeze anything else, performance-wise out of Manager's start up time.
No comments:
Post a Comment