Even the FB5's editor's FB modern made project takes a lot of time to open comparatively (15-19 seconds), so it's not likely my code nor my project. It is Fn AddItem and whatever it is doing, here on PPC. It does a lot, and it does it 450+ times or more for each file on the big project. Could be... XRef, or getting the proj Handle, or SizeOf, or FN SetHandleSize_Wrapper...or Fn GetHandleSize_Wrapper Or does Fn FileExist or...Blockzero...or (I think my logging pointed to this as being one of the biggies) fn GetFileIconRef Any of those ring a bell for being slower on PPC for anyone? Robert On Apr 12, 2011, at 9:03 PM, Brian S wrote: > > On Apr 12, 2011, at 5:27 PM, Robert Covington wrote: > >> Repeat of same tale of woe from October , 2010 (I just gave up >> using FB back then). Is something repeatedly resizing a handle in >> a non-optimal way? >> >>> I added a log tracker to Fn OpenProject and associated calls. >>> >>> The loop in Fn ReadProjectIntoDataBase is the culprit, 2500 ticks. >>> >>> For each item added during that loop (just below For to bottom), >>> it takes 5 to 9 ticks. >>> For each call to Fn AddItem within that loop, that part takes 2 >>> to 6 ticks. >>> >>> "Item Added" is total time per item from loop start to loop >>> end...that's ticks reset just under the For, ticks counted just >>> above Next. >>> >>> Fn AddItem" is the call in the middle of the loop. It is likely >>> the bulk of the time, but not all. >>> >> > > > On Oct 23, 2010, at 9:35 PM, Robert Purves wrote: > >> I made a project with 1000 files. On my Mac, the native Intel >> editor opens the project in a small fraction of a second. With the >> editor set for PPC/Rosetta there is a noticeable delay but it's >> only about 0.5 s. >> >> Robert P. >> > > Robert, the obvious difference is the PowerPC CPU vs Intel CPU. RP > ran in Rosetta mode and still got good results. The only thing I > can imagine is there is possibly something in your project that is > different. Maybe if you will consider loaning the team ( or one of > us ) your source the problem can be identified. > > Brian S > > > > > > > > -- > To unsubscribe, send ANY message to: futurebasic- > unsubscribe@... > To access the list archives, go to: http://freegroups.net/groups/ > futurebasic/ >