- Pharo-dev - lists.pharo.org

Re: [Pharo-dev] Confused about Floats on 64-bit Pharo 6

by Henrik Sperre Johansen

For a solution that works without change in 32/64 migration, one could use self class maxVal digitLength ... but there's quite a bit of calculation in digitLength. Speaking of excessive calculation; SmallInteger class >> startUp: isImageStarting "The image is either being newly started (isImageStarting is true), or it's just been snapshotted. If this has just been a snapshot, skip all the startup stuff." | next val | isImageStarting ifFalse: [^self]. val := -32768. "Assume at least 16 bits" [next := val + val. next class == self] whileTrue: [val := next]. minVal := val. maxVal := -1 - val ... really? If this actually runs on 64bit image startup, that's quite impressive! -- View this message in context: http://forum.world.st/Confused-about-Floats-on-64-bit-Pharo-6-tp4948805p494… Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

June 1, 2017

Re: [Pharo-dev] Confused about Floats on 64-bit Pharo 6

by Henrik Sperre Johansen

SmallInteger >> digitAt: has an n > 4 ifTrue: [^0] check, in 64bit this needs to check against 8. (The value represents max number of 8-bit digits in a SmallInteger) "I thought BoxedFloat64 should not be used in 64-bit ?" It will be used for some large exponents, search for "immediate float" in https://clementbera.wordpress.com/2014/01/16/spurs-new-object-format/ for a good description. "Also, should #emin, #emax and friends not be different for SmallFloat64 ? " No, it's an implementation detail. Both float classes represent 64bit floating points with emin/emax as listed. When one is insufficient; it should be swapped to the other automatically, in a manner similar to Byte/WideStrings instances. Cheers, Henry -- View this message in context: http://forum.world.st/Confused-about-Floats-on-64-bit-Pharo-6-tp4948805p494… Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

June 1, 2017

Re: [Pharo-dev] Confused about Floats on 64-bit Pharo 6

by Nicolas Cellier

2017-06-01 15:37 GMT+02:00 Sven Van Caekenberghe <sven(a)stfx.eu>: > > > On 1 Jun 2017, at 15:29, Nicolas Cellier <nicolas.cellier.aka.nice@ > gmail.com> wrote: > > > > Hi Sven, > > SmallFloat64 consumes 3bits taken from exponent for tagging immediate > value. > > So very small values and very large values don't fit in SmallFloat64 and > require a BoxedFloat64. > > > > I don't think we should change emin and emax though. > > The reason is that emin and emax are public methods for accessing the > underlying floating point model. > > Small/Boxed Float are implementation details, they both implement the > same float model with different levels of optimization. > > If I am on 64-bit and SmallFloat64 is used (the default, right ?), then I > have a smaller exponent range (right ?), so I should be able to access meta > info about the currently applying limits, no ? > Yes and no. Yes if you care about implementation details. No if you delegate to abstract class Float, ignore the details and let the VM entirely deal with those. My opinion is that most applications should not care of those details and should avoid duplicating the VM job. Except maybe very optimized or low level libraries (Fuel?) > > I recently added that to NeoJSON to protect against overly large/small > exponents when reading floats, like 1e1231231231. > Good thing! This should be backported to number parsers, See https://codegolf.stackexchange.com/questions/7197/crash-your-favorite-compi… > > I understand about this being an implementation detail, but if the choice > of implementation changes the range, it seems important to be able to know > this, one way or another. > > It does not change the range of the underlying abstract Float model. If those meta-data are really needed, you must find another accessor than emin and emax or you will break the model. (2.0 timesTwoPower: Float emax // 8 + 1) is a BoxedFloat64. It's predecessor is Small. I let you find the ones near emin by yourself. > For the rest of the report, it's a bug we will have to track catch and > fix ASAP > > Thanks! > > > 2017-06-01 14:53 GMT+02:00 Sven Van Caekenberghe <sven(a)stfx.eu>: > > I am working in Pharo 6 RC, 64-bit on macOS 10.12.5 > > > > One of the tests of STON fails, #testFloat and I got confused about > Floats on 64-bit Pharo 6. > > > > Is there a write-up that explains the new/changed situation ? > > > > Because, what I see is the following: > > > > Float pi. > > 1.3. > > > > Both the above give me SmallFloat64 instances. > > > > (10 raisedTo: 100) asFloat. > > 1.0e100. > > 10.0 raisedTo: 100. > > > > All 3 above give me BoxedFloat64 instance. > > But the first 2 give wrong results (correct answer is 1.0e100) ! > > > > (10 raisedTo: 100) asFloat. > > "5.15323791002091e91" > > > > 1.0e100. > > "5.15323791002091e91" > > > > Why ? > > > > BoxedFloat64 allInstances. > > SmallFloat64 allInstances. > > > > The first is not-empty, the second is (logical since these are immediate > values). > > I thought BoxedFloat64 should not be used in 64-bit ? > > > > Also, should #emin, #emax and friends not be different for SmallFloat64 ? > > > > Sven > > > > >

June 1, 2017

Re: [Pharo-dev] Confused about Floats on 64-bit Pharo 6

by Sven Van Caekenberghe

> On 1 Jun 2017, at 15:29, Nicolas Cellier <nicolas.cellier.aka.nice(a)gmail.com> wrote: > > Hi Sven, > SmallFloat64 consumes 3bits taken from exponent for tagging immediate value. > So very small values and very large values don't fit in SmallFloat64 and require a BoxedFloat64. > > I don't think we should change emin and emax though. > The reason is that emin and emax are public methods for accessing the underlying floating point model. > Small/Boxed Float are implementation details, they both implement the same float model with different levels of optimization. If I am on 64-bit and SmallFloat64 is used (the default, right ?), then I have a smaller exponent range (right ?), so I should be able to access meta info about the currently applying limits, no ? I recently added that to NeoJSON to protect against overly large/small exponents when reading floats, like 1e1231231231. I understand about this being an implementation detail, but if the choice of implementation changes the range, it seems important to be able to know this, one way or another. > For the rest of the report, it's a bug we will have to track catch and fix ASAP Thanks! > 2017-06-01 14:53 GMT+02:00 Sven Van Caekenberghe <sven(a)stfx.eu>: > I am working in Pharo 6 RC, 64-bit on macOS 10.12.5 > > One of the tests of STON fails, #testFloat and I got confused about Floats on 64-bit Pharo 6. > > Is there a write-up that explains the new/changed situation ? > > Because, what I see is the following: > > Float pi. > 1.3. > > Both the above give me SmallFloat64 instances. > > (10 raisedTo: 100) asFloat. > 1.0e100. > 10.0 raisedTo: 100. > > All 3 above give me BoxedFloat64 instance. > But the first 2 give wrong results (correct answer is 1.0e100) ! > > (10 raisedTo: 100) asFloat. > "5.15323791002091e91" > > 1.0e100. > "5.15323791002091e91" > > Why ? > > BoxedFloat64 allInstances. > SmallFloat64 allInstances. > > The first is not-empty, the second is (logical since these are immediate values). > I thought BoxedFloat64 should not be used in 64-bit ? > > Also, should #emin, #emax and friends not be different for SmallFloat64 ? > > Sven >

June 1, 2017

Re: [Pharo-dev] Confused about Floats on 64-bit Pharo 6

by Nicolas Cellier

Can't reproduce in Squeak trunk 64bits with latest opensmalltalk-VM, So for me it should'nt be a VM problem, (or it's fixed already). 2017-06-01 15:31 GMT+02:00 Nicolas Cellier < nicolas.cellier.aka.nice(a)gmail.com>: > > > 2017-06-01 15:29 GMT+02:00 Nicolas Cellier <nicolas.cellier.aka.nice@ > gmail.com>: > >> Hi Sven, >> SmallFloat64 consumes 3bits taken from exponent for tagging immediate >> value. >> So very small values and very large values don't fit in SmallFloat64 and >> require a BoxedFloat64. >> >> I don't think we should change emin and emax though. >> The reason is that emin and emax are public methods for accessing the >> underlying floating point model. >> Small/Boxed Float are implementation details, they both implement the >> same float model with different levels of optimization. >> >> For the rest of the report, it's a bug we will have to track catch and >> fix ASAP >> >> (is looks like a shift of 1 bit) > > >> 2017-06-01 14:53 GMT+02:00 Sven Van Caekenberghe <sven(a)stfx.eu>: >> >>> I am working in Pharo 6 RC, 64-bit on macOS 10.12.5 >>> >>> One of the tests of STON fails, #testFloat and I got confused about >>> Floats on 64-bit Pharo 6. >>> >>> Is there a write-up that explains the new/changed situation ? >>> >>> Because, what I see is the following: >>> >>> Float pi. >>> 1.3. >>> >>> Both the above give me SmallFloat64 instances. >>> >>> (10 raisedTo: 100) asFloat. >>> 1.0e100. >>> 10.0 raisedTo: 100. >>> >>> All 3 above give me BoxedFloat64 instance. >>> But the first 2 give wrong results (correct answer is 1.0e100) ! >>> >>> (10 raisedTo: 100) asFloat. >>> "5.15323791002091e91" >>> >>> 1.0e100. >>> "5.15323791002091e91" >>> >>> Why ? >>> >>> BoxedFloat64 allInstances. >>> SmallFloat64 allInstances. >>> >>> The first is not-empty, the second is (logical since these are immediate >>> values). >>> I thought BoxedFloat64 should not be used in 64-bit ? >>> >>> Also, should #emin, #emax and friends not be different for SmallFloat64 ? >>> >>> Sven >>> >> >> >

June 1, 2017

Re: [Pharo-dev] Confused about Floats on 64-bit Pharo 6

by Nicolas Cellier

2017-06-01 15:29 GMT+02:00 Nicolas Cellier < nicolas.cellier.aka.nice(a)gmail.com>: > Hi Sven, > SmallFloat64 consumes 3bits taken from exponent for tagging immediate > value. > So very small values and very large values don't fit in SmallFloat64 and > require a BoxedFloat64. > > I don't think we should change emin and emax though. > The reason is that emin and emax are public methods for accessing the > underlying floating point model. > Small/Boxed Float are implementation details, they both implement the same > float model with different levels of optimization. > > For the rest of the report, it's a bug we will have to track catch and fix > ASAP > > (is looks like a shift of 1 bit) > 2017-06-01 14:53 GMT+02:00 Sven Van Caekenberghe <sven(a)stfx.eu>: > >> I am working in Pharo 6 RC, 64-bit on macOS 10.12.5 >> >> One of the tests of STON fails, #testFloat and I got confused about >> Floats on 64-bit Pharo 6. >> >> Is there a write-up that explains the new/changed situation ? >> >> Because, what I see is the following: >> >> Float pi. >> 1.3. >> >> Both the above give me SmallFloat64 instances. >> >> (10 raisedTo: 100) asFloat. >> 1.0e100. >> 10.0 raisedTo: 100. >> >> All 3 above give me BoxedFloat64 instance. >> But the first 2 give wrong results (correct answer is 1.0e100) ! >> >> (10 raisedTo: 100) asFloat. >> "5.15323791002091e91" >> >> 1.0e100. >> "5.15323791002091e91" >> >> Why ? >> >> BoxedFloat64 allInstances. >> SmallFloat64 allInstances. >> >> The first is not-empty, the second is (logical since these are immediate >> values). >> I thought BoxedFloat64 should not be used in 64-bit ? >> >> Also, should #emin, #emax and friends not be different for SmallFloat64 ? >> >> Sven >> > >

June 1, 2017

Re: [Pharo-dev] Confused about Floats on 64-bit Pharo 6

by Nicolas Cellier

Hi Sven, SmallFloat64 consumes 3bits taken from exponent for tagging immediate value. So very small values and very large values don't fit in SmallFloat64 and require a BoxedFloat64. I don't think we should change emin and emax though. The reason is that emin and emax are public methods for accessing the underlying floating point model. Small/Boxed Float are implementation details, they both implement the same float model with different levels of optimization. For the rest of the report, it's a bug we will have to track catch and fix ASAP 2017-06-01 14:53 GMT+02:00 Sven Van Caekenberghe <sven(a)stfx.eu>: > I am working in Pharo 6 RC, 64-bit on macOS 10.12.5 > > One of the tests of STON fails, #testFloat and I got confused about Floats > on 64-bit Pharo 6. > > Is there a write-up that explains the new/changed situation ? > > Because, what I see is the following: > > Float pi. > 1.3. > > Both the above give me SmallFloat64 instances. > > (10 raisedTo: 100) asFloat. > 1.0e100. > 10.0 raisedTo: 100. > > All 3 above give me BoxedFloat64 instance. > But the first 2 give wrong results (correct answer is 1.0e100) ! > > (10 raisedTo: 100) asFloat. > "5.15323791002091e91" > > 1.0e100. > "5.15323791002091e91" > > Why ? > > BoxedFloat64 allInstances. > SmallFloat64 allInstances. > > The first is not-empty, the second is (logical since these are immediate > values). > I thought BoxedFloat64 should not be used in 64-bit ? > > Also, should #emin, #emax and friends not be different for SmallFloat64 ? > > Sven >

June 1, 2017

Release deferred to next Tuesday :(

by Esteban Lorenzano

Hi list, yes, this release is being a pain. anyway, yesterday after asking for testI found two blocking issues and one ânice to haveâ and I needed to work on them. Blocking issues were: 1) Athens (with Fonts ) was not working on 64bits and because of that, Roassal was not either. 2) There is a problem in VM packaging which packages as files what should be symlinks.\ The nice to have: It seems there are english problems in the welcome browser (and in the STAGE pages). It would be nice if some native speaker takes a look :) I already fixed (1) and Iâm working on (2), but the release is deferred to Tuesday because: a) we want to test. b) you never release on friday c) monday is holiday here :) So well, that. Iâm sorry. Esteban

June 1, 2017

Confused about Floats on 64-bit Pharo 6

by Sven Van Caekenberghe

I am working in Pharo 6 RC, 64-bit on macOS 10.12.5 One of the tests of STON fails, #testFloat and I got confused about Floats on 64-bit Pharo 6. Is there a write-up that explains the new/changed situation ? Because, what I see is the following: Float pi. 1.3. Both the above give me SmallFloat64 instances. (10 raisedTo: 100) asFloat. 1.0e100. 10.0 raisedTo: 100. All 3 above give me BoxedFloat64 instance. But the first 2 give wrong results (correct answer is 1.0e100) ! (10 raisedTo: 100) asFloat. "5.15323791002091e91" 1.0e100. "5.15323791002091e91" Why ? BoxedFloat64 allInstances. SmallFloat64 allInstances. The first is not-empty, the second is (logical since these are immediate values). I thought BoxedFloat64 should not be used in 64-bit ? Also, should #emin, #emax and friends not be different for SmallFloat64 ? Sven

June 1, 2017

Minor glitch in source code diff tool

by Sven Van Caekenberghe

It seems there is a minor glitch in the source code diff tool: The first difference is marked (coloured) OK, each next one is one character position more off. This is in the latest 6.0 RC image. Should I create an issue ? Sven

June 1, 2017