BorlandTalk.com Borland discussion newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  Avoiding floating points Goto page 1, 2  Next          BorlandTalk.com Forum Index -> Delphi Non-Technical View previous topic :: View next topic   Author Message Jonathan Neve[Microtec] Guest Posted: Thu Apr 26, 2007 8:12 am    Post subject: Avoiding floating points Hi all, I hate floating points with a passion. Does anyone know of a simple way to get rid of them and somehow be able to do elemetary calculations without random imprecisions and unpredicitble and meaningless error messages popping up out of no-where? Thanks! -- Best regards, Jonathan Neve _______________ CopyCat - advanced Delphi components for database replication, supporting CodeGear Interbase, FirebirdSQL and MS SQL Server! _______________ CopyTiger - the ultimate database replication solution for CodeGear Interbase, FirebirdSQL and MS SQL Server! _______________ More information : http://www.microtec.fr/copycat Back to top Jonathan Neve[Microtec] Guest Posted: Fri Apr 27, 2007 8:12 am    Post subject: Re: Avoiding floating points Brad White wrote: Quote: David Farrell-Garcia wrote: Jonathan Neve[Microtec] wrote: If your application processes everything to integers or rounds to a fixed number of decimal places why do you care how the database stores it? 5 = 5.000000000 no matter how you slice it. But that only applies to a few numbers. For example, on the face of it, you would expect it to apply to 0.1, but it does not. Instead of 0.1 = 0.100000000 you get 0.1 = 0.100,000,000,000,000,000,001,355,252,715,606,880, 542,509,316,001,087,427,139,282,226,562,5 No matter how you store 0.1 in the database, you will get something else back out. Or did I miss your point? Brad. Exactly, and for most business applications, that's simply unacceptable. Anyway, the point of this thread was, I admit, primarily a rant -- I'm sick of floating point imprecisions, on the one hand, and their weird errors (perhaps compiler-related) on the other. In none (or hardly any) of my applications can I conceive a need for decimal points that float around, I *always* want to get back the *exact* decimal representation that I put into my database. So I'm not really trying to change the face of computer programming, I'm just frustrated with floats, and I need to make an effort to replace them with something more suitable for my needs. I suppose BCDs or simply Currency fields would be good, and I'll make sure I never make another application with floats... :) -- Best regards, Jonathan Neve _______________ CopyCat - advanced Delphi components for database replication, supporting CodeGear Interbase, FirebirdSQL and MS SQL Server! _______________ CopyTiger - the ultimate database replication solution for CodeGear Interbase, FirebirdSQL and MS SQL Server! _______________ More information : http://www.microtec.fr/copycat Back to top Paul Scott Guest Posted: Fri Apr 27, 2007 8:12 am    Post subject: Re: Avoiding floating points On Thu, 26 Apr 2007 20:13:30 +0100, John Herbster <herb-sci1_AT_sbcglobal.net> wrote: Quote: Having those worthless date-times scattered about the EXEs makes detecting other changes more of a problem that it need be. Ah, so *that's* the reason the exe's are different! (or one of the reasons?) I was supplied with two versions of an exe built only seconds apart - the only difference was a one character change to the MainForm's caption - one was "v5.40" the other "v5.41". (The changed caption was an easy visual check that when testing that our "new" DLL generated exactly the same results as the old version, that we were actually running the appropriate app+DLL combination) I expected a "fc /b" to show only one byte difference - but I was wrong - and by several orders of magnitude! -- Paul Scott Information Management Systems Macclesfield, UK. Back to top AlexB Guest Posted: Sat Apr 28, 2007 8:12 am    Post subject: Re: Avoiding floating points Mark Andrews (The Other One) wrote: Quote: I simply do not buy the idea that assigning zero is failing. We've all had situations where we start to doubt the compiler, the CPU and the arrangement of the planets. In every one of those cases I've eventually (often after a lot of sweat and tears) located the problem in my own code--sometimes a subtle memory overwrite problem. Mostly agreed. But (as I've described in this thread earlier) I really got problem with code generation in small test program (where I had no chance to make my own mistake). -- Alex Back to top AlexB Guest Posted: Sat Apr 28, 2007 8:12 am    Post subject: Re: Avoiding floating points John Herbster wrote: Quote: ... Wrong code generation. Does anyone know of a QualityCentral report on this? I think no. QC requires reproducing steps. This issue looks like ghost and (I think) can't be reproduced reliably. Some days ago I'd met with this issue again and again with the same ActiveX (new version). Steps was: - debug session, internal structured exception in ActiveX (0xc0000008- invalid handle) - some correction in program and rebuild - new debug session and wrong results in FP constant assignment (x=1.0; actual x value is trash) - CPU window shows wrong address for constant (1.0) in this line (but correct address for this constant in some other lines) - rebuild had no effect - new small test project in the same IDE session - the same result (wrong address) - reload IDE - test progect now compiles and works OK - rebuild of the first project - all seems OK. Quote: Has it be fixed in later versions or updates? I'd found this issue in BCB5. Jonathan, what version you use? Quote: Is or was there a workaround? Reload IDE, rebuild all or rebuild all with command line compiler. Then do 'Make' only and avoid recompiling of 'problematic' unit (i.e. use 'good old' obj for link) where compiler generates wrong addresses for FP constants ... -- Alex. Back to top Mark Andrews (The Other O Guest Posted: Sat Apr 28, 2007 8:12 am    Post subject: Re: Avoiding floating points John Herbster wrote: Quote: I have made two QC reports that are related to the subject: http://qc.borland.com/wc/qcmain.aspx?d=5928 http://qc.borland.com/wc/qcmain.aspx?d=8399 Great detective work John. I hadn't been aware of this problem before. Quote: In the first, Danny Thorpe's replies were interesting If by "interesting" you mean borderline dismissive, then yes. Quote: but he did not want to fix all parts of the problem. I'll be implementing your workaround. Thanks. Any idea why #5928 is marked closed? Cheers, Mark. Back to top AlexB Guest Posted: Sat Apr 28, 2007 8:12 am    Post subject: Re: Avoiding floating points Mark Andrews (The Other One) wrote: Quote: I simply do not buy the idea that assigning zero is failing. We've all had situations where we start to doubt the compiler, the CPU and the arrangement of the planets. In every one of those cases I've eventually (often after a lot of sweat and tears) located the problem in my own code--sometimes a subtle memory overwrite problem. AFAIK text (code) and constant data segments are read-only. How memory overwriting can be the reason of wrong FP constant address appearance? -- Alex. Back to top Mark Andrews (The Other O Guest Posted: Sat Apr 28, 2007 8:12 am    Post subject: Re: Avoiding floating points M.E. wrote: Quote: "John Herbster" <herb-sci1_AT_sbcglobal.net> wrote in message news:46309033 (AT) newsgroups (DOT) borland.com... Please, please, anyone: Give me code for an example of where assigning 0 fails! I can't. It does not fail "regularly". It does fail randomly, and it *does* fail. Like the OP stated. I call shenanigans. (I also can't believe some of the things I'm reading in this thread.) I simply do not buy the idea that assigning zero is failing. We've all had situations where we start to doubt the compiler, the CPU and the arrangement of the planets. In every one of those cases I've eventually (often after a lot of sweat and tears) located the problem in my own code--sometimes a subtle memory overwrite problem. I'm not saying that codegen bugs don't exist, but this particular claim needs some rather compelling evidence. Cheers, Mark. Back to top Loren Pechtel Guest Posted: Sun Apr 29, 2007 7:56 am    Post subject: Re: Avoiding floating points On 26 Apr 2007 14:45:56 -0700, "Nicholas Ring" <Nicholas.Ring (AT) smf (DOT) com.au> wrote: Quote: Just wondering if there was a Loren Pechtel wrote: On 26 Apr 2007 05:54:42 -0700, "AlexB" <b.a.v (AT) inbox (DOT) ru> wrote: Next semester the teacher asks me about it. I utterly astounded the whole department with the following: 10 X = 3000000 20 X = X + 1 30 If X < X + 1 then goto 10 40 Print "X = X + 1" Just wondering if line 30 has a typo, in that it should be going to line 20, not line 10? Of course--you don't really think I cut & pasted the old code, do you? Back to top Loren Pechtel Guest Posted: Sun Apr 29, 2007 7:56 am    Post subject: Re: Avoiding floating points On Thu, 26 Apr 2007 17:19:19 -0500, "John Herbster" <herb-sci1_AT_sbcglobal.net> wrote: Quote: 10 X = 3000000 20 X = X + 1 30 If X < X + 1 then goto 10 40 Print "X = X + 1" "Nicholas Ring" <Nicholas.Ring (AT) smf (DOT) com.au> wrote Just wondering if line 30 has a typo, in that it should be going to line 20, not line 10? Nicholas, I think that you are right. I looked at it, noted the concept that Loren was intending, and overlooked that error. Will this ever finish? var x: int64; begin x := 0; While (x < high(x)) do inc(x); ShowMessage('Finished with x='+IntToStr(x)); end; Rgds, JohnH Yup. It finishes when x = high(x). Back to top Loren Pechtel Guest Posted: Sun Apr 29, 2007 7:56 am    Post subject: Re: Avoiding floating points On 26 Apr 2007 09:32:35 -0700, "Jonathan Neve[Microtec]" <jonathan (AT) microtec (DOT) fr> wrote: Quote: I don't want integers, I want decimal-point numbers that "work" (from a user perspective), and so I'm not interested in the decimal point "floating"... Yes, yes, I know, I can hack it out by hand, I was primarily ranting at the utter uselessness of floats (for most of the practical purposes they're usually used for -- I don't doubt that they are useful for some people), and secondarily wondering what other people do to handle common every-day business operations (dealing with money, or any other sensitive data that has to be given back exactly the way it was put in). Floats are fine when you don't need perfect precision (note that in many cases perfect precision is unattainable anyway.) If you need it to be perfect (example: most financial calculations) then do it fixed point. My approach to handling this sort of problem is define a unit that all internal variables will be stored in that has no decimal points--it's only scaled on input or output. I have never used the Currency type because it's *NOT* an answer--having too many decimal points will cause trouble in some cases. When you're doing financial calculations you generally want the fractional pennies to vanish, not get carried. Back to top Loren Pechtel Guest Posted: Sun Apr 29, 2007 7:56 am    Post subject: Re: Avoiding floating points On Thu, 26 Apr 2007 12:08:14 -0500, "John Herbster" <herb-sci1_AT_sbcglobal.net> wrote: Quote: "Jonathan Neve[Microtec]" <jonathan (AT) microtec (DOT) fr> wrote ... the compiler produces different code from the same source depending on its mood does seem to me to be a compiler bug. ... It is not the compiler's mood; it is the date-time of compilation. I don't know about the later compilers in this regard, but D7 and earlier compiler encoded the date in two or three dozen places within the EXE. And yes, I think that it is a bug, too. I have suspected (but never proved) that there are cases where the compiler emits pure garbage. I have grown suspicious of garbage bytes due to alignment. If so, this is a totally harmless failure other than making the .EXE's not match so I've never dug into it. Back to top Loren Pechtel Guest Posted: Sun Apr 29, 2007 7:56 am    Post subject: Re: Avoiding floating points On Fri, 27 Apr 2007 10:31:42 +0200, "Anders Isaksson" <i.rather@not> wrote: Quote: John Herbster wrote: It is not the compiler's mood; it is the date-time of compilation. I don't know about the later compilers in this regard, but D7 and earlier compiler encoded the date in two or three dozen places within the EXE. It's also a question of the linker. Any disk sector of the EXE that isn't filled up completely by compiled/linked code (or data) will contain whatever bytes was there before being assigned to the current file. Code alignment will create such 'holes' in the disk image, all over the place, containing 'random' data from your previously erased files. At least up to D5, I haven't taken a closer look at later versions. But the whole sector is probably going to be written anyway, you won't pick up garbage bytes off the disk. I do think you pick up garbage bytes from RAM this way, though. Back to top John Herbster Guest Posted: Sun Apr 29, 2007 8:11 am    Post subject: Re: Avoiding floating points "Loren Pechtel" <lorenpechtel (AT) hotmail (DOT) invalid.com> wrote Quote: ... you won't pick up garbage bytes off the disk. I do think you pick up garbage bytes from RAM this way, though. Loren, et. al, I would like to see some proof or examples of this. Rgds, JohnH Back to top David Erbas-White Guest Posted: Mon Apr 30, 2007 8:11 am    Post subject: Re: Avoiding floating points Loren Pechtel wrote: Quote: On 28 Apr 2007 00:58:46 -0700, "AlexB" <b.a.v (AT) inbox (DOT) ru> wrote: Mark Andrews (The Other One) wrote: I simply do not buy the idea that assigning zero is failing. We've all had situations where we start to doubt the compiler, the CPU and the arrangement of the planets. In every one of those cases I've eventually (often after a lot of sweat and tears) located the problem in my own code--sometimes a subtle memory overwrite problem. AFAIK text (code) and constant data segments are read-only. How memory overwriting can be the reason of wrong FP constant address appearance? Segments? Segments are effectively not used anymore. Were they ever 'effectively' used??? <G> David Erbas-White Back to top Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First         BorlandTalk.com Forum Index -> Delphi Non-Technical All times are GMT Goto page 1, 2  Next Page 1 of 2   Jump to: Select a forum C++ Builder Newsgroups----------------C++ BuilderC++ Builder (ActiveX)C++ Builder (Announcements)C++ Builder (Command Line Tools)C++ Builder DatabasesC++ Builder Databases (ADO)C++ Builder Databases (Desktop)C++ Builder Databases (InterBase Express)C++ Builder Databases (Multi-Tier)C++ Builder Databases (SQL Servers)C++ Builder (Graphics)C++ Builder (IDE)C++ Builder (Internet)C++ Builder (Internet Socket)C++ Builder (Internet Web)C++ Builder (Jobs)C++ Builder (Language)C++ Builder (Language C++)C++ Builder (Compatibility with MFC)C++ Builder (Native API)C++ Builder (Non-Technical)C++ Builder (Students)C++ Builder (Third Party Tools)C++ Builder (Upgrade)C++ Builder (VCL)C++ Builder (VCL Components Usage)C++ Builder (VCL Components Development)C++ Builder (VCL Usage)C++ Builder (WinAPI)C++ Builder Installation Delphi Newsgroups----------------Delphi (General)Delphi ActiveX Controls UsageDelphi ActiveX Controls DevelopmentDelphi As400Delphi DatabasesDelphi Databases (ADO)Delphi Databases (dbExpress)Delphi Databases (Desktop)Delphi Databases (InterBase Express)Delphi Databases (Multi-Tier)Delphi Databases (SQL Servers)Delphi DeploymentDelphi GraphicsDelphi IDEDelphi InternationalizationDelphi InternetDelphi Internet Isapi-WebBrokerDelphi Internet WebSnapDelphi Internet WinsockDelphi JobsDelphi Language BASMDelphi Language ObjectPascalDelphi Model Driven Architecture (General)Delphi Model Driven Architecture (Third Party)Delphi MultimediaDelphi Native APIDelphi Non-TechnicalDelphi ObjectPascalDelphi OLE AutomationDelphi OO designDelphi OpenToolsAPIDelphi Reporting-ChartingDelphi RTLDelphi StudentsDelphi TeamSourceDelphi Thirdparty ToolsDelphi Thirdparty Tools (General)Delphi Thirdparty Tools (IntraWeb)Delphi Thirdparty Tools (RaveReports)Delphi UpgradeDelphi VCLDelphi VCL ComponentsDelphi VCL Components UsingDelphi VCL Components WritingDelphi WebServices SOAPDelphi WinAPIDelphi XMLDelphi Installationcomp.lang.pascal.delphi.components.misccomp.lang.pascal.delphi.components.usagecomp.lang.pascal.delphi.components.writingcomp.lang.pascal.delphi.databasescomp.lang.pascal.delphi.misccomp.sources.delphi JBuilder Newsgroups----------------JBuilderJBuilder AnnouncementsJBuilder Applet IssuesJBuilder As400JBuilder CompilerJBuilder CORBA/RMIJBuilder DatabasesJBuilder DB SwingJBuilder DebuggerJBuilder DeploymentJBuilder DocumentationJBuilder EnterpriseJBuilder HandheldJBuilder IDEJBuilder JavaJBuilder JavaAPIJBuilder Java LanguageJBuilder Java SwingJBuilder Java BeansJBuilder Java Beans UsageJBuilder Java Beans DevelopmentJBuilder JBCLJBuilder JDataStoreJBuilder JobsJBuilder LinuxJBuilder MacintoshJBuilder & MobileSetJBuilder MultiLingual AppsJBuilder Non-TechnicalJBuilder OpenToolsJBuilder Servlets (JSP)JBuilder StudentsJBuilder Team DevelopmentJBuilder Thirdparty ToolsJBuilder UpgradeJBuilder Installation InterBase Newsgroups----------------InterBaseInterBase Bug DiscussionInterBase GeneralInterBase IBConsoleInterBase OpenSourceInterBase ReplicationInterBase SQLInterBase ThirdpartyInterBase Installation Miscellaneous Borland Newsgroups----------------AppCenterAppServerEnteraVisiBroker  You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum Powered by phpBB © 2001, 2006 phpBB Group SEO toolkit © 2004-2006 webmedic.