Pentium FDIV

Haven't seen too much discussion of this and how it would affect crypto in 
general.  Would the FDIV problem cause any problems when looking at large 
prime numbers?  Actually, let me rephrase that, would it affect determining if 
a large number is prime?  Floating point errors can be so much fun.

Also, does anyone have the "test" that was floating around the net and in 
various newspapers a few weeks to see if your machine has the FDIV problem.  I 
lost the little "post-it note" I had it written on.  Since then, I haven't 
been able to reproduce it.  (Either the bug went away, or I entered the wrong 
numbers.  **GRIN**)

--Brad

>>>>>>>>>>>>>>>>>>>>>INTERNETWORKING THE DESKTOP<<<<<<<<<<<<<<<<<<<<<<<
Brad Shantz                      bshantz@spry.com
Senior Software Engineer
SPRY Inc.                        Direct #:     (206)-442-8251
---------------------------------------------------------------------
PGP Public Key at:    http://www-swiss.ai.mit.edu/~bal/pks-toplev.html
Or email:             pgp-public-keys@pgp.ai.mit.edu Subj: GET bshantz
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

-----BEGIN PGP SIGNED MESSAGE-----

> Haven't seen too much discussion of this and how it would affect crypto in 
> general.  Would the FDIV problem cause any problems when looking at large 
> prime numbers?  Actually, let me rephrase that, would it affect determining 
> if a large number is prime?  Floating point errors can be so much fun.

I don't know how you missed it; it was certainly here. Let me summarize:

	- crypto typically uses integers
	- RSA crypto uses large integers
	- the Pentium FDIV bug only affects floats

- -Paul

- -- 
Paul Robichaux, KD4JZG       | Good software engineering doesn't reduce the 
perobich@ingr.com            | amount of work you put into a product; it just 
Not speaking for Intergraph. | redistributes it differently.
		  ### http://www.intergraph.com ###

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBLwLtj6fb4pLe9tolAQGNXAQAy61AjCjq10ZnsmozlpzanYLqtOib1Hfw
hW1M6uh6sYe7fAqhgILJ0OSKuOkCtXoHHX/rMBchrYdPr/LCUSZHoXQHvxQJpYWb
B+jBegxBO+vKa/yOV/JgzVSsg2jCraxnextAXSXuniZU149+MWhp0wQ1Pmh19BqD
2J7wRHVehMs=
=kebO
-----END PGP SIGNATURE-----

> paul@poboy.b17c.ingr.com (Paul Robichaux) writes:
> I don't know how you missed it; it was certainly here. Let me summarize:

> - crypto typically uses integers
> - RSA crypto uses large integers
> - the Pentium FDIV bug only affects floats

Yes, but Prof. Nicely found the problem because he was using FDIV when
doing something or other with large potential primes.  For most of our
crypto stuff integer is fine, but for some operations on 33- to 50-bit
integers it may be faster to use the floating point ops, since IEEE fp is
supposed to be exact for integers up to some number of bits in the 50's.

	Jim Gillogly
	Sterday, 7 Afteryule S.R. 1995, 18:07

Try	(4,195,835 x 3,145,727)
	-----------------------
		3,145,727

A faulty Pentium returns 4,195,579.

Regds

------------------------------------------------------------------------------
 jrt@AsiaOnline.Net   john@AsiaOnline.Net   PO Box 86141, Govt PO, Kln, HKG.

 Help protect the environment : This message is made from recycled electrons
------------------------------------------------------------------------------

On Thu, 29 Dec 1994 bshantz@spry.com wrote:

> Haven't seen too much discussion of this and how it would affect crypto in 
> general.  Would the FDIV problem cause any problems when looking at large 
> prime numbers?  Actually, let me rephrase that, would it affect determining if 
> a large number is prime?  Floating point errors can be so much fun.
> 
> Also, does anyone have the "test" that was floating around the net and in 
> various newspapers a few weeks to see if your machine has the FDIV problem.  I 
> lost the little "post-it note" I had it written on.  Since then, I haven't 
> been able to reproduce it.  (Either the bug went away, or I entered the wrong 
> numbers.  **GRIN**)
> 
> --Brad
> 
> >>>>>>>>>>>>>>>>>>>>>INTERNETWORKING THE DESKTOP<<<<<<<<<<<<<<<<<<<<<<<
> Brad Shantz                      bshantz@spry.com
> Senior Software Engineer
> SPRY Inc.                        Direct #:     (206)-442-8251
> ---------------------------------------------------------------------
> PGP Public Key at:    http://www-swiss.ai.mit.edu/~bal/pks-toplev.html
> Or email:             pgp-public-keys@pgp.ai.mit.edu Subj: GET bshantz
> >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
> 
>

> Also, does anyone have the "test" that was floating around the net and in 
> various newspapers a few weeks to see if your machine has the FDIV problem.  I 
> lost the little "post-it note" I had it written on.  Since then, I haven't 
> been able to reproduce it.  (Either the bug went away, or I entered the wrong 
> numbers.  **GRIN**)

This is from the djgpp (DOS port of gcc) mailinmg list, posted by the
author of said port himself. It's short enought to post and might be 
helpful to some. Flames will be forwarded to postmaster@intel.com.

=D.C. Williams	<dcwill@ee.unr.edu>

==============INCLUDED MESSAGE BELOW===========================

Subject: Intel recalls Pentium
Status: RO

CNN reported this weekend that Intel has announced that they will
replace ALL defective Pentium(tm) processors, no questions asked.  To
determine if your Pentium(tm) is devective, run the attached MS-DOS
program FDIVBUG1.COM (source included also).  It will tell you if you
have the bug and where to call for a replacement if needed.

DJ

--

begin 644 fdivbug1.zip
M4$L#!!0``@`(``B*F1U]4E@Z>@$``+<"```,````9F1I=F)U9S$N87-M=5)+
M3\,P##Z[4O^#A9!ZZ5#;O<JN3$CC`)-X7%&;I".C;4:2CO'O<=*"!HQ#9>?S
MYT<_FQFQ`?H:T5HL/ZS`75?6DN$94UR<A0$4QG2-`&86C'@Q[RT%E-Y`FB0O
M81`&QA;:+@C<-COPC\S!50K\#2;IY30?3R\2`C('C-/)=)[-':#WP-\A<>3&
M;)Z)7T(ZCC%-8HS6-)3L&KQ>/V+9;9`+*YA5.J9)<7F#2U$K+45$?4^F+5=/
M/N\\ZJNK5\>+I,';NP?<:6&(&5'6D/E=X2NA++C/6-W_8?<]HZNBKG'56E%C
M83$=Y4DRFF7Y*)_E,ZR41BUV=<&$UU>VA#2%E:K]-?./SH.<F=.S4230(495
M5498]!H-</$27Y(K6PM9ZK<`E6RE=;;F4#E>Q>4>JLQY35</GNE*/82EL:#W
MY#':F]['[K>V`J0AK7S%H?]1>])DV#-7SP2$@6<O3I+[*CUQ\?_<'C[$$^;/
MZ3CD;@U$RPW!9/K;"H-/4$L#!!0``@`(``F*F1UJ&TZHR````.H````,````
M9F1I=F)U9S$N8V]M>[F:`0@.U#$&.`+IAO__W8$T+U=`:EY)9FFN@EM`J$)2
M:;I"2FI):G))?I&.0E*E@HN7@DMJ3GY19BJ20A?/,+!*E<QB!3__$(6"HM1B
MH!PO%R^7BF<P@NN<F).CX)E7DIJCD%BB8*AK86"@:V9DH6MA9F&FD)9?I%"4
M6I"3F)R:"U2MD)D'%,E-+,G,SP,;M$N4<0OG6<79MQ_/OLO&S#C[CADWD.`#
M$7I`[GTQ8<9F.V%&AA*V7:F,KUDF[`H"J]_!X'-6$0!02P$"%``4``(`"``(
MBID=?5)8.GH!``"W`@``#``````````!`"``````````9F1I=F)U9S$N87-M
M4$L!`A0`%``"``@`"8J9'6H;3JC(````Z@````P````````````@````I`$`
C`&9D:79B=6<Q+F-O;5!+!08``````@`"`'0```"6`@``````
`
end