Re: Testing wwCache with different browsers

Just a guess but for cases like this I'd suspect memory usage. VFP maybe needing to grab another 'chunk' of memory to use. There's no real pattern for this easily uncovered because of what Rick suggested - the current process load. And this approach is so deeply embedded inside of VFP other than some magic Calvin may have done in place I'm sure this is still the approach.

I remember WAY back a friend of mine had a routine written in FoxBase that was faster from a serial port than the ASM equivilent. I asked him about his thoughts on why that was the case and he thought is was that Fox was faster because it controlled the memory rather than the O/S. I've seen this pattern continue through the years but I also am now starting to think that this approach has reached the point where the new hardware setups (particularly VMs, solid state drives, etc) make the former GREAT advantage not as great as it once was.

Best,

DD

VFP performance just varies depending on load on the machine. I see this a lot and I'm scratching my head at some of this frequently as well. But I doubt any of it has anything to do with browsers at all.

Are you seeing this on a desktop box in testing or on a live box? Live machines tend to give much more consistent behavior because there's no desktop related stuff sucking on the CPU usually.

+++ Rick ---

I've been playing around with wwCache and I'm seeing some weird results that have made me curious. wwCache seems to be working exactly as it is supposed to, no problems there. However, I'm seeing different results in the time to process the page depending upon which browser I use. I've got a couple of test pages that I'm caching for 10 seconds. Using Google Chrome or Safari, the first hit takes 0.004-0.006 and each subsequent cache hit takes 0.004. For IE, Opera, and Firefox, they all report 0.004-0.006 for the first hit and each subsequent cache hit takes 0.000.

Also, with Chrome, it takes a half second to a second for the Reload button to come back for me, whereas all of the other browsers it's immediate or a very slight delay.

Like I said, just made me curious and wondered if anyone knew an explanation.

BTW, wwCache is a very handle all-purpose cache. We have used some fine tuned (read sloppily hard coded) custom caching in the past but I'm finding a lot of it is easily ported to wwCache without any noticeable loss in performance. FWIW, we are opting for FixedFilename (i.e. shared access) and we roll it daily like we do our log file to keep bloat and such to a minimum.

~Brett

EDIT: BTW, I realize there's nothing IN wwCache that's intrinsically "browser dependent".

---