Interesting, As a matter of fact I use stored procedures in my database for a lot of the big looping of 10s of thousands of records. But that is not portable and doesn't solve the issue of looping speed within Witango.
I see now that dynamics within loops with compiler based languages have come a long way since I last tried. I just don't want to loose the capability of dynamically creating object references. There is a plus side, the continual speed increase on processors. When I run it on a Pentium 2 400Mhz server it takes a minute and a half :-) Ben Johansen - http://www.pcforge.com Authorized Witango Reseller http://www.pcforge.com/WitangoGoodies.htm Authorized MDaemon Mail Server Reseller http://www.pcforge.com/AltN.htm -----Original Message----- From: Robert Garcia [mailto:[EMAIL PROTECTED] Sent: Wednesday, July 02, 2003 10:16 PM To: [EMAIL PROTECTED] Subject: Re: Witango-Talk: Boosting Server Performance, and TCFs I thought of that, so I made the instantiation a domain scope, then removed it for the subsequent times. I was surprised, it didn't not seem to have any effect. It must have only added a few milliseconds. Either way the TCF method was slightly slower or the same. I agree that no one is going to look at that many years, but it was a good example to use as a comparison. The reason I am so hot on this, is that I have several reports that I need to write, that require alot of looping and processing, and it just chokes the witango server, and I need to right them in RB. I think if we had the choice quick and easy compiled code, then this would be less of a problem. Robert. On Wednesday, July 2, 2003, at 02:55 PM, Ben Johansen wrote: > No doubt the instantiation of the object is the add second or 2. > > If I can find some time I might try server side javascript. > > But realistically, I did it for 1 year which is about the max period > someone would do, the TCF version on my 1.6gig P4 Laptop took a second. > > Start 14:54:31 > > from 07/02/2003 to 07/02/2004 > Total Nbr Days 366 > Total Work Days 262 > > End 14:54:32 > > Ben Johansen - http://www.pcforge.com > Authorized Witango Reseller http://www.pcforge.com/WitangoGoodies.htm > Authorized MDaemon Mail Server Reseller > http://www.pcforge.com/AltN.htm > > > -----Original Message----- > From: Robert Garcia [mailto:[EMAIL PROTECTED] > Sent: Wednesday, July 02, 2003 12:49 PM > To: [EMAIL PROTECTED] > Subject: Re: Witango-Talk: Boosting Server Performance, and TCFs > > I have tested your new TAFs. The TAFs that don't call the TCF method, > both run 14 to 15 seconds consistently. The TAF that calls the TCF > takes 15 to 16 seconds consistently. > > Does anyone have any updated docs on developing beans for use in > Witango? > > Robert. > > > > On Wednesday, July 2, 2003, at 12:56 AM, Ben Johansen wrote: > >> Ok, so you are cleaning my clock :-) >> >> I have changed the calcs into a TCF and also matched your logic. >> >> Robert, Could you run the test against your server with the new >> "workdates3.taf" which calls "daycalc.tcf"? >> >> This will be a more accurate time run because it is using the seconds > >> model >> you used in RB. >> >> also included is the seconds logic in just a taf "workdates2.taf" (no > >> tcf >> call) >> could you run this time test on your server >> >> >> Ben Johansen - http://www.pcforge.com >> -Authorized WiTango Reseller >> http://www.pcforge.com/WitangoGoodies.htm >> -Authorized Alt-N Reseller >> http://www.pcforge.com/AltN.htm >> >> -----Original Message----- >> From: Robert Garcia [mailto:[EMAIL PROTECTED] >> Sent: Tuesday, July 01, 2003 7:03 PM >> To: [EMAIL PROTECTED] >> Subject: Re: Witango-Talk: Boosting Server Performance, and TCFs >> >> >> OK, I created an application in RealBasic, and compiled for all >> platforms. It can be downloaded here: >> >> http://public.bighead.net/workdaytest/ >> >> I also set up the workday taf on my test server and it can be run > here: >> >> http://poh.bighead.net/workday/ >> >> With the default values, spanning 46020 days my witango v5.1.058 > server >> running on Windows 2000 server, AMD proc. running at 1.5ghz, with >> 512megs ram, and a 266FSB, it should take 15 seconds. If it takes >> longer, it is because a couple of you are hitting it at once. >> >> Now If I run my RB app on the same machine, it takes about 1/3 of a >> second if I do not give a progress indication, and about a second if I >> do. That is even more of a performance increase than I expected. On my >> mac, the results are about the same, except the progress calc takes >> just over 2 seconds. >> >> Also, I am running my calc within its own thread. If I were to not >> thread the process, and disable background tasks, I could probably >> squeeze out another 10% performance. >> >> Here is my code from the RB method, you can download the rb project >> from the link above: >> >> Sub calcWDays() >> dim sDate,eDate,iDate as date >> dim nDays,wDays,sTicks,eTicks as double >> dim i as integer >> dim s as string >> >> sTicks = ticks >> >> sDate = new date >> eDate = new date >> >> sDate.year = sDateDC.year >> sDate.month = sDateDC.month >> sDate.day = sDateDC.day >> >> eDate.year = eDateDC.year >> eDate.month = eDateDC.month >> eDate.day = eDateDC.day >> >> nDays = (eDate.totalSeconds - sDate.totalSeconds)/86400 >> >> wDays = 0 >> iDate = new date >> iDate = sDate >> for i = 1 to nDays >> s = str(iDate.DayOfWeek) >> if iDate.dayofweek > 1 and iDate.dayofweek < 7 then >> wDays = wDays + 1 >> end if >> iDate.TotalSeconds = iDate.TotalSeconds + 86400 >> next >> eTicks = ticks >> s = "The total number of dates between "+sDate.abbreviatedDate+" >> and >> "+eDate.abbreviatedDate+" is "+str(nDays)+". " >> s = s + "The number of workdays in between is "+str(wDays)+" and >> the >> time taken to calculate was "+str((eTicks-sTicks)/60)+" seconds." >> msgBox s >> >> End Sub >> >> I tried to make the RB method as Identical to the one in Ben's taf as >> possible as you can see. sDateDC and eDateDC are the date control >> objects that the user interacts with. >> >> Robert. >> >> On Tuesday, July 1, 2003, at 06:09 PM, Ben Johansen wrote: >> >>> Cool >>> and while you at it make the week day taf into a tcf and create a new >>> custom >>> tag called <@DATEDIFFWK> >>> LOL!. >>> >>> >>> Ben Johansen - http://www.pcforge.com >>> -Authorized WiTango Reseller >>> http://www.pcforge.com/WitangoGoodies.htm >>> -Authorized Alt-N Reseller >>> http://www.pcforge.com/AltN.htm >>> >>> -----Original Message----- >>> From: Robert Garcia [mailto:[EMAIL PROTECTED] >>> Sent: Tuesday, July 01, 2003 5:05 PM >>> To: [EMAIL PROTECTED] >>> Subject: Re: Witango-Talk: Boosting Server Performance, and TCFs >>> >>> >>> Actually, you could instantiate one field, then build an array of >>> fields to the nth, building an array of field objects, and then loop >>> through them referencing like so: >>> >>> for x = 1 to 10000000 >>> field(x) = "some value" >>> next x >>> >>> I just downloaded your weekday taf, I will do a test with it. >>> >>> Robert. >>> >>> On Tuesday, July 1, 2003, at 04:51 PM, Ben Johansen wrote: >>> >>>> Yes, when all factors are know the compiled code is quick. >>>> >>>> In your example you have 10 and it looks nice an simple. >>>> Now multiply it by 10 and then maintain it. :-P >>>> >>>> Take the example I just sent with workdates.taf >>>> write a case statement for all possible dates. Aaaaaah. >>>> >>>> The number of times that I had a loop that I know all the factors I >>>> could probably count on my hands and ok I will take off my shoes > also >>>> ;-) >>>> >>>> The nature of a loop is dynamics, placing the case within the loop > is >>>> to >>>> me not intuitive but is required for a compiler. >>>> >>>> I am just going by experience and IMHO I will take dynamics over >>>> compiled. >>>> >>>> Ben Johansen - http://www.pcforge.com >>>> Authorized Witango Reseller > http://www.pcforge.com/WitangoGoodies.htm >>>> Authorized MDaemon Mail Server Reseller >>>> http://www.pcforge.com/AltN.htm >>>> >>>> >>>> -----Original Message----- >>>> From: Robert Garcia [mailto:[EMAIL PROTECTED] >>>> Sent: Tuesday, July 01, 2003 4:37 PM >>>> To: [EMAIL PROTECTED] >>>> Subject: Re: Witango-Talk: Boosting Server Performance, and TCFs >>>> >>>> There really is no increase in code. I do this all the time in a >>>> single >>>> >>>> conditional statement like: >>>> >>>> for x = 1 to 10 >>>> >>>> select case x >>>> >>>> case 1 >>>> field_1 = "some value" >>>> case 2 >>>> field_2 = "some value" >>>> case 3 >>>> field_3 = "some value" >>>> case 4 >>>> field_4 = "some value" >>>> case 5 >>>> field_5 = "some value" >>>> case 6 >>>> field_6 = "some value" >>>> case 7 >>>> field_7 = "some value" >>>> case 8 >>>> field_8 = "some value" >>>> case 9 >>>> field_9 = "some value" >>>> case 10 >>>> field_10 = "some value" >>>> next x >>>> >>>> There are other ways also, this is the simplest. Another way is to >>>> build an array of objects, and loop through them. >>>> >>>> I think you underestimate the level of performance gain you get when >>>> you compile versus interpret on the fly on complex tasks. Especially >>>> processing loops. I will do a test and get back to this with some >>>> results. >>>> >>>> Robert. >>>> >>>> On Tuesday, July 1, 2003, at 02:32 PM, Ben Johansen wrote: >>>> >>>>> You are correct, and this is where I question the speed boost. >>>>> >>>>> On one hand you have the quicker run speed but on the other hand > you >>>>> have the increased amount of code due to lose of this dynamic. >>>>> >>>>> There is no easy conversion of code and mindset between them. >>>>> >>>>> Ben Johansen - http://www.pcforge.com >>>>> Authorized Witango Reseller >>>>> http://www.pcforge.com/WitangoGoodies.htm >>>>> Authorized MDaemon Mail Server Reseller >>>>> http://www.pcforge.com/AltN.htm >>>>> >>>>> >>>>> -----Original Message----- >>>>> From: Robert Garcia [mailto:[EMAIL PROTECTED] >>>>> Sent: Tuesday, July 01, 2003 2:20 PM >>>>> To: [EMAIL PROTECTED] >>>>> Subject: Re: Witango-Talk: Boosting Server Performance, and TCFs >>>>> >>>>> dynamic referencing a table is one thing, that should be ok. What I >>>>> mean is referencing an object dynamically like you can do with >>>>> javascript/flash with the eval() function. >>>>> >>>>> For instance, in javascript/flash, if you have a set of fields, >>>>> named >>>>> field_1, and field_2, and so on to field_10, they could be >>>>> referenced >>>>> dynamically like so: >>>>> >>>>> for x = 1 to 10 >>>>> eval("field_"+str(x)) = "some value" >>>>> next x >>>>> >>>>> The same could be done in witango with a bunch of arguments, <@arg >>>>> field_1>, and so on. >>>>> >>>>> <@for start=1 stop=10 step=1> >>>>> <@addrows local$newArray <@arg field_<@currow>>> >>>>> </@for> >>>>> >>>>> I don't believe this could be done in a compiled language. You > would >>>>> have to use a select case method or build an array of objects or >>>>> something. >>>>> >>>>> Robert. >>>>> >>>>> On Tuesday, July 1, 2003, at 01:47 PM, Atrix Wolfe wrote: >>>>> >>>>>> you could still do dynamic referencing if they made allowances for >>>> it, >>>>> >>>>>> like >>>>>> if they stored a table w/ the names of the vars and pointers to >>>>>> them. >>>>>> >>>>>> If you can make java bytecodes though and then use gcc to make > that >>>>>> into an >>>>>> executable or dynamic library, would you really need With to make > a >>>>>> witango >>>>>> compiler? That is assuming you dont lose anything like dynamic >>>>>> referencing >>>>>> in either of those 2 steps. >>>>>> >>>>>> I've been kind of wondering about this since With said you could >>>>>> make >>>>> >>>>>> tango >>>>>> into java bytecodes though...if you made your code into java >>>>>> bytecodes, does >>>>>> it support dynamic referencing TCF's, includes etc? >>>>>> >>>>>> Just wondering if could expect any and all code to work the same >>>>>> whether it >>>>>> was tango or java bytecodes. >>>>>> >>>>>> >>>>>> ----- Original Message ----- >>>>>> From: "Robert Garcia" <[EMAIL PROTECTED]> >>>>>> To: <[EMAIL PROTECTED]> >>>>>> Sent: Tuesday, July 01, 2003 1:25 PM >>>>>> Subject: Re: Witango-Talk: Boosting Server Performance, and TCFs >>>>>> >>>>>> >>>>>>> Just because our tafs would be compiled in some native format, >>>>> doesn't >>>>>>> mean we would have to start worrying about things garbage >>>> collection. >>>>>>> The compiled tcf would would have in/out points and process data >>>>>>> within. It would run within the memory space of the witango >>>>>>> server, >>>>>>> and >>>>>>> the server would handle the garbage collection and the nasty >>>>>>> stuff. >>>>>>> The >>>>>>> main benefit would be that complex methods in a tcf would not be >>>>>>> scripted, but compiled so that there would be no tag processing >>>>>>> and >>>>>>> the >>>>>>> other performance drags that are inherant with an interpreted >>>>>>> language. >>>>>>> It would mean, though, that certain things you can do in a tml or >>>>> taf, >>>>>>> you could not do, like dynamic referencing. >>>>>>> >>>>>>> Robert. >>>>>>> >>>>>>> On Tuesday, July 1, 2003, at 09:49 AM, Scott Cadillac wrote: >>>>>>> >>>>>>>> But it does add several layers of complexity too. I'm not an >>>>>>>> expert >>>>> >>>>>>>> at >>>>>>>> this >>>>>>>> level by any means, but it seams to me that Witango isn't ready > >>>>>>>> to >>>>>>>> support >>>>>>>> this kind of low-level environment (nor any interpreted language >>>> for >>>>>>>> that >>>>>>>> matter, e.g., ColdFusion). >>>>>>>> >>>>>>>> For example, garbage collection and other memory allocation >>>>>>>> issues >>>> - >>>>>>>> what >>>>>>>> sort of Metatags would we need for this? >>>>>>> >>>>>>> -- >>>>>>> >>>>>>> Robert Garcia >>>>>>> President - BigHead Technology >>>>>>> CTO - eventpix.com >>>>>>> 2781 N Carlmont Pl >>>>>>> Simi Valley, Ca 93065 >>>>>>> ph: 805.522.8577 - cell: 805.501.1390 >>>>>>> [EMAIL PROTECTED] - [EMAIL PROTECTED] >>>>>>> http://bighead.net/ - http://eventpix.com/ - > http://theradmac.com/ >>>>>>> >>>>>>> >>>>> > ____________________________________________________________________ >>>>> _ >>>>> _ >>>>>>> __ >>>>>>> TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf >>>>>> >>>>>> >>>>> >>>> > _____________________________________________________________________ >>>> _ >>>> _ >>>>>> _ >>>>>> TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf >>>>>> >>>>>> >>>>> >>>>> -- >>>>> >>>>> Robert Garcia >>>>> President - BigHead Technology >>>>> CTO - eventpix.com >>>>> 2781 N Carlmont Pl >>>>> Simi Valley, Ca 93065 >>>>> ph: 805.522.8577 - cell: 805.501.1390 >>>>> [EMAIL PROTECTED] - [EMAIL PROTECTED] >>>>> http://bighead.net/ - http://eventpix.com/ - http://theradmac.com/ >>>>> >>>>> >>>> > _____________________________________________________________________ >>>> _ >>>> _ >>>>> _ >>>>> TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf >>>>> >>>>> >>>> > _____________________________________________________________________ >>>> _ >>>> _ >>>>> _ >>>>> TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf >>>>> >>>>> >>>> >>>> -- >>>> >>>> Robert Garcia >>>> President - BigHead Technology >>>> CTO - eventpix.com >>>> 2781 N Carlmont Pl >>>> Simi Valley, Ca 93065 >>>> ph: 805.522.8577 - cell: 805.501.1390 >>>> [EMAIL PROTECTED] - [EMAIL PROTECTED] >>>> http://bighead.net/ - http://eventpix.com/ - http://theradmac.com/ >>>> >>>> > _____________________________________________________________________ >>>> _ >>>> _ >>>> _ >>>> TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf >>>> >>>> > _____________________________________________________________________ >>>> _ >>>> _ >>>> _ >>>> TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf >>>> >>>> >>> >>> -- >>> >>> Robert Garcia >>> President - BigHead Technology >>> CTO - eventpix.com >>> 2781 N Carlmont Pl >>> Simi Valley, Ca 93065 >>> ph: 805.522.8577 - cell: 805.501.1390 >>> [EMAIL PROTECTED] - [EMAIL PROTECTED] >>> http://bighead.net/ - http://eventpix.com/ - http://theradmac.com/ >>> >>> > ______________________________________________________________________ >>> _ >>> _ >>> TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf >>> >>> > ______________________________________________________________________ >>> _ >>> _ >>> TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf >>> >>> >> >> -- >> >> Robert Garcia >> President - BigHead Technology >> CTO - eventpix.com >> 2781 N Carlmont Pl >> Simi Valley, Ca 93065 >> ph: 805.522.8577 - cell: 805.501.1390 >> [EMAIL PROTECTED] - [EMAIL PROTECTED] >> http://bighead.net/ - http://eventpix.com/ - http://theradmac.com/ >> >> > _______________________________________________________________________ >> _ >> TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf >> >> > _______________________________________________________________________ >> _ >> TO UNSUBSCRIBE: Go to >> http://www.witango.com/maillist.taf<WorkDates2.zip> > > -- > > Robert Garcia > President - BigHead Technology > CTO - eventpix.com > 2781 N Carlmont Pl > Simi Valley, Ca 93065 > ph: 805.522.8577 - cell: 805.501.1390 > [EMAIL PROTECTED] - [EMAIL PROTECTED] > http://bighead.net/ - http://eventpix.com/ - http://theradmac.com/ > > _______________________________________________________________________ > _ > TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf > > _______________________________________________________________________ > _ > TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf > > -- Robert Garcia President - BigHead Technology CTO - eventpix.com 2781 N Carlmont Pl Simi Valley, Ca 93065 ph: 805.522.8577 - cell: 805.501.1390 [EMAIL PROTECTED] - [EMAIL PROTECTED] http://bighead.net/ - http://eventpix.com/ - http://theradmac.com/ ________________________________________________________________________ TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf ________________________________________________________________________ TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf
