gnu@sun.uucp (John Gilmore) (08/09/85)
I forwarded some of the discussion to a knowledgeable friend at IBM, and he gave me permission to post his response: Date: Wed, 24 Jul 85 11:29:58 pdt From: ucbvax!ibmpa!lmb (Larry Breed) Subject: 360 floating point Hi, John. Actually that report is a shade too kind. You have to expect to lose three, not one or two, bits from hex normalization. Yes, "on the average" you've got only one or two leading 0 bits in the fraction, but the overall accuracy of your results is controlled by the minimum accuracy rather than the average accuracy of intermediate results. This blanket statement isn't really accurate either -- leading zeros don't hurt on many adds and subtracts, and hurt only statistically on multiplies and divides -- but it is accurate in specifying worst-case behavior. The second overkindness was that IBM noted and corrected FP errors after first shipment. As I recall, it wasn't until late 1966 that the retrofitted guard digit made its appearance, and then only because W. Kahan and other numeric analysts outside IBM made an issue of it. Once you can point out to IBM that X*1.0 != X, you can usually get its attention. Just to put things in perspective, floating point has been done wrong in many ways, by many people. (1/3 == 9/27 on a 370, and not every manufacturer can make that claim!) This is why it's so valuable to have the IEEE 754 standard -- it doesn't have these anomalies, it DOES have valuable facilities lacking in other designs, and you can tell when you've implemented it right. There's a test suite (driver plus about 20,000 test cases) developed by the 754 designers and available from Berkeley.
ark@alice.UUCP (Andrew Koenig) (08/11/85)
> Just to put things in perspective, floating point has been done wrong in > many ways, by many people. (1/3 == 9/27 on a 370, and not every > manufacturer can make that claim!) This is why it's so valuable to have > the IEEE 754 standard -- it doesn't have these anomalies, it DOES have > valuable facilities lacking in other designs, and you can tell when you've > implemented it right. There's a test suite (driver plus about 20,000 test > cases) developed by the 754 designers and available from Berkeley. Can someone tell me how to get a copy?
meissner@rtp47.UUCP (Michael Meissner) (08/13/85)
John Gilmore <2588@sun.uucp> writes: > > Just to put things in perspective, floating point has been done wrong in > many ways, by many people. (1/3 == 9/27 on a 370, and not every > manufacturer can make that claim!) This is why it's so valuable to have > the IEEE 754 standard -- it doesn't have these anomalies, it DOES have > valuable facilities lacking in other designs, and you can tell when you've > implemented it right. There's a test suite (driver plus about 20,000 test > cases) developed by the 754 designers and available from Berkeley. > How does one get a hold of this test suite? Michael Meissner Data General ...{ ihnp4, decvax }!mcnc!rti-sel!rtp47!meissner
dick@ucsfcca.UUCP (Dick Karpinski) (08/15/85)
In article <4127@alice.UUCP> ark@alice.UUCP (Andrew Koenig) writes: >> manufacturer can make that claim!) This is why it's so valuable to have >> the IEEE 754 standard -- it doesn't have these anomalies, it DOES have >> valuable facilities lacking in other designs, and you can tell when you've >> implemented it right. There's a test suite (driver plus about 20,000 test > >Can someone tell me how to get a copy? Well, the right place to get IEEE standard documents is from The Secretary IEEE Standards Board 345 East 47 th Street New York, NY 10017 I have a few copies of the draft that was approved as well as copies of the binary and decimal format free draft (IEEE p854). There are test vectors available from UCB (address on request) for $100 and a program to show you how bad your non-IEEE arithmetic is. This last item is Paranoia, written up in the Feb 85 Byte. It is available in three languages (BASIC, FORTRAN, and Pascal) on floppy disk for $15. Call or write for more info, as if this weren't too much already. Dick -- Dick Karpinski Manager of Unix Services, UCSF Computer Center UUCP: ...!ucbvax!ucsfcgl!cca.ucsf!dick (415) 666-4529 (12-7) BITNET: dick@ucsfcca Compuserve: 70215,1277 Telemail: RKarpinski USPS: U-76 UCSF, San Francisco, CA 94143
ken@turtlevax.UUCP (Ken Turkowski) (08/19/85)
In article <400@ucsfcca.UUCP> dick@ucsfcca.UUCP (Dick Karpinski) writes: >... >program to show you how bad your non-IEEE arithmetic is. This last >item is Paranoia, written up in the Feb 85 Byte. It is available in >three languages (BASIC, FORTRAN, and Pascal) on floppy disk for $15. Has anyone converted Paranoia to C? Could you send me a copy or post to the net, please? -- Ken Turkowski @ CADLINC, Menlo Park, CA UUCP: {amd,decwrl,hplabs,seismo,spar}!turtlevax!ken ARPA: turtlevax!ken@DECWRL.ARPA