From mboxrd@z Thu Jan 1 00:00:00 1970 MIME-Version: 1.0 In-Reply-To: <1a3de3d2869ab078c9a4fdf9bce44982@brasstown.quanstro.net> References: <20111002163800.GA12773@polynum.com> <20111002175227.2D7F1B856@mail.bitblocks.com> <20111002184015.CD088B852@mail.bitblocks.com> <1a3de3d2869ab078c9a4fdf9bce44982@brasstown.quanstro.net> Date: Sun, 2 Oct 2011 12:59:36 -0600 Message-ID: From: andrew zerger To: Fans of the OS Plan 9 from Bell Labs <9fans@9fans.net> Content-Type: multipart/alternative; boundary=20cf303bf65ca9b13604ae5575c1 Subject: Re: [9fans] circular fonctions: precision? Topicbox-Message-UUID: 2f7c86f2-ead7-11e9-9d60-3106f5b1d025 --20cf303bf65ca9b13604ae5575c1 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable havent read any of these specs but im sitting here thinking that bits on a processor that can subtract =3D=3D more preicision .. On Sun, Oct 2, 2011 at 12:44 PM, erik quanstrom wrot= e: > > IEEE754-1985 didn't specify circular, hyperbolic or other > > advanced functions. You can have 754 compliant hardware and > > not implement these functions. In any case the standard can > > not dictate the accuracy of functions not specified in it. An > > iterative algorithm may lose more than 1 bit of accuracy since > > iterations won't be done in infinite precision. One can not > > assume accuracy to a bit even where these functions are > > imeplemented in h/w. For x86, accuracy may be specified in > > some Intel or AMD manual. > > that wasn't my reading of the spec. so you're saying that if > the iterative algorithm loops 53 times, it's free to return any > answer whatever and still be compliant? > > - erik > > --=20 =E2=8E=BC=E2=8E=BA=E2=8E=BA=E2=94=9C@=E2=94=BC=E2=90=8A=E2=94=9C=E2=94=9C= =E2=89=A4-=E2=90=8D=E2=8E=BC=E2=90=8A=E2=96=92=E2=90=8D:/=E2=90=A4=E2=8E=BA= =E2=94=94=E2=90=8A/=E2=8E=BC=E2=90=A4=E2=8E=BA# --20cf303bf65ca9b13604ae5575c1 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable =C2=A0havent read any of these specs but im sitting here thinking that bits= on a processor that can subtract =3D=3D more preicision .. =C2=A0

<= div class=3D"gmail_quote">On Sun, Oct 2, 2011 at 12:44 PM, erik quanstrom <= span dir=3D"ltr"><quanstro@quan= stro.net> wrote:
> IEEE754-1985 didn= 9;t specify circular, hyperbolic or other
> advanced functions. You can have 754 compliant hardware and
> not implement these functions. In any case the standard can
> not dictate the accuracy of functions not specified in it. An
> iterative algorithm may lose more than 1 bit of accuracy since
> iterations won't be done in infinite precision. One can not
> assume accuracy to a bit even where these functions are
> imeplemented in h/w. =C2=A0For x86, accuracy may be specified in
> some Intel or AMD manual.

that wasn't my reading of the spec. =C2=A0so you're saying th= at if
the iterative algorithm loops 53 times, it's free to return any
answer whatever and still be compliant?

- erik




--
=E2= =8E=BC=E2=8E=BA=E2=8E=BA=E2=94=9C@=E2=94=BC=E2=90=8A=E2=94=9C=E2=94=9C=E2= =89=A4-=E2=90=8D=E2=8E=BC=E2=90=8A=E2=96=92=E2=90=8D:/=E2=90=A4=E2=8E=BA=E2= =94=94=E2=90=8A/=E2=8E=BC=E2=90=A4=E2=8E=BA#


--20cf303bf65ca9b13604ae5575c1--