| |
|
|
Konrad Flachs | Hi,
I have one Program for a Navigationssystem written, the mitunter different Formate konvertieren can. thereby is me noticed, the at reading of/ one csv-File (Strings through semicolon separated) The Ladezeit integral longer become is. The commands @substr$ needed now with Profan 8.0 about twice such a long time How with Profan 5.0.
is the one bow of Profan 8.0? ...or there a ploy thereby?
greeting Konrad |
|
|
| |
|
|
|
| now, XProfan operates in the Contrast To Prf5 with Large-Strings, means Strings The not only 256Chars, sonder even 32kB Chars on length having can. These Umstellung in the string-creature can unfortunately such Auswirkungen mitsich bring.
Also should itself well Roland to a little express.
I have the time short tested, and must admit substr is in XProfan unfortunately 5-time slower.
subesquent code have I for gepinselt: CompileMarkSeparationprint
declare t&,z&,s$,ss$
let t&=&gettickcount
let z&=0
let s$="SO KANNS WEITER| GEHGEN ODER OMIT | DLKFJS LKDFJ SO KANNS NICHT WEITER| GEHGEN ODER OMIT | DLKFJS LKDFJ SO KANNS WEITER| GEHGEN ODER OMIT | DLKFJS LKDFJ SO KANNS WEITER| GEHGEN ODER OMIT | DLKFJS LKDFJ "
while 1
if gt(&gettickcount,t&)
let t&=add(t&,1000)
locate 1,1
print z&
let z&=0
endif
let z&=add(Z&,1)
let ss$=substr$(s$,add(mod(z&,10),1),"|")
wend
end
@Frank, want we whom substr on asm-base beschläunigen? XPSE could from substr one _asm_substr make, a Procedure namens _asm_substr lay out filled with a ASM-Piece from you.
@Konrad, if you want schnuddle I you quick ne small Mini-DLL together, The the CSV-Parsen for you Done. would be naturally for a multiple zügiger. Müsstest me only whom Profan-code give, the To translate is. whom remainder erledige I already. an mini-example-cvs very To this Topic would to that testing necessary. Perhaps one small statement from the Originalen.
To denne, iF |
|
|
| |
|
|
|
Frank Abbing | Hi,
means I find not, the The function sooo slow operates, the one tappt im dunkeln immediate supplant should. Konrad, take off Strings verarbeitest thou ? One Beipspiel ? |
|
|
| |
|
|
|
Konrad Flachs | Hello If and Frank,
thanks for speedy response.
i have noticed, not only substr$ separate too mid$ slower become is.
to that comparison the reading of 20000 csv-Lines:
Profan 5.0: ca.27sek. Profan 8.0: ca.82sek. Profan 8.0 without if and mid$ : 47sek.
in the Program pieces I each row with z.B. Let rtecomment$ = @substr$(lineinp$,12,;) . then go the parts gefiltert with z.B. Let rtecomment$ = @mid$(rtecomment$,1,12) .... or of/ one if-request.
The Vergrösserung the maximalen Stinglänge ( of verse.5.0 To 8.0) shining The Verarbeitungszeit of some commands To verlängern.
ideal would naturally a Einschränkbarkeit the String-Length and one therefore resultierenden Zeitgewinn.
greeting Konrad |
|
|
| |
|
|
|
| As I said, or NEN mini-Dll, müsstest only say what The make should. Z.b. me your Profan-Source give the slow runs. spare geschätzt would The DLL not 27, sonder 3 sec for the same need.
To denne, iF |
|
|
| |
|
|
|
Frank Abbing | Hi,
The List view.dll z.B. using CSV-Files and bid some functions for on. i'd CSV-Files too always with reaches Edit, not whom slow Strings If you something Assembler learn want, can you you with the XPIA yourself functions write. such Maschinenfunktionen would not 27 or 3 sec need, separate only few Millisekunden. |
|
|
| |
|
|
|
Uwe ''Pascal'' Niemeier | Hello people!
only the Vollständigkeit halber: CSV-Files can too integrally well by ODBC/SQL To Leibe back. Obs faster is as The manuelle method or even a ASM-routine, faith I though not. For this has one Sortier- u. Suchmöglichkeiten and u.u. direct Access to single Lines/Split (according to Dateiaufbau).
SeeYou Pascal |
|
|
| |
|
|
|
Jens | Hello Frank,
self-evident have You yourself with the Geschwindigkeitsangaben not vertan, anyway I'm going of it from. Überrascht be I though utterly.
gives it (self-evident problemabhängige) factors, How quick z.B. Profan compiliert in the comparison to profane in c++ Translated, To c++ purely or. To Assembler-code is?
greetings
Jens |
|
|
| |
|
|
|
| now, in manchen affairs I had a factor of *65000 determined, in others z.B. only *10000. so meant is the Difference one (reinen) CPPs (not prf2cpp!) in the Ggs. to a PrfExe. the are naturally everything ca-indicated. in the Grunde are The Compilat of hochsprachen (How z.B. cpp) not slower, as ASM-code, there Yes The Compiler (under Berüchsichtigung vieler another factors adjustable with Compilerschaltern) too in ASM Compilieren. its but objectively so, that if one directly in ASM komponiert one some unwichtige-Operationen omit can. On well german, if one CPP-Prog in a second 1mio time add, would one reines asm the Perhaps 10-30% faster create. But this is too only then so, if one as CPPer not The right Compilierschalter in the Source respect. there remaining it the Programmer self give over, cost and benefit abzuschätzen, where C++ naturally with of/ one yummy Syntax and really einfachen Befehlsmöglichkeiten lockt. If one but z.B. only a tab 1Mio male add must, then is well naturally ASM faster. so could one insist, the so plainer The task, so More rewards ASM.
To denne, iF |
|
|
| |
|
|