[Shootout-list] What people say about shootout
Dave
davejf@frontiernet.net
Fri, 29 Apr 2005 07:43:44 -0500
This is a multi-part message in MIME format.
------=_NextPart_000_0004_01C54C8F.2BC719B0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Paraphrased from the comments of more than one compiler developer: "99% of
developers don't care about blazing floating point performance, but the
other 1% bitches alot" <g> Obviously, these are not Fortran compiler
developers ;)
Seriously though, one of the things I like about the Shootout is that it is
not focused on floating point, as so many benchmarks seem to be. Over the
past year of so I've written alot of C/++ for 'high-performance' programs
('alot' and 'high-performance' are relative of course ;). I just grep'd that
repository and it appears that the above ratio holds there as well.
Another thing is that so often double precision floating point is the only
thing measured. And is perhaps over-used in production code as well. For one
recent floating point intensive app. that I tuned, all I did was basically
'refactor' it to use SP instead of DP and the performance soared and mem.
usage dropped. All w/o a lick of difference to the end user because the
extra precision was not needed in this case. My point being, why are all of
the floating point benchmarks I see out there (not just the Shootout) double
precision when in reality single precision is (or should be) used in more
day-to-day applications?
FP perf. is important, but I'd hate to see the Shootout focus a whole lot
more on it.
-----Original Message-----
From: shootout-list-admin@lists.alioth.debian.org
[mailto:shootout-list-admin@lists.alioth.debian.org]On Behalf Of Robert
Seeger
Sent: Friday, April 29, 2005 6:40 AM
To: shootout-list@lists.alioth.debian.org
Subject: Re: [Shootout-list] What people say about shootout
Heh. Those are the types of tests that I like. Since my real world
problems tend to have more to do with text processing than the math-oriented
ones most people here are interested in...
Rob Seeger
On 4/29/05, Isaac Gouy <igouy2@yahoo.com> wrote:
There's always the danger of insularity on a low-volume list like this
one, so it's a healthy exercise to look around the web and see what
people say about shootout.
comp.lang.ada is still talking about count-words and word-frequency -
not my choice for the most interesting benchmarks, but interesting is
in the eye of the beholder.
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
_______________________________________________
Shootout-list mailing list
Shootout-list@lists.alioth.debian.org
http://lists.alioth.debian.org/mailman/listinfo/shootout-list
------=_NextPart_000_0004_01C54C8F.2BC719B0
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Diso-8859-1">
<META content=3D"MSHTML 6.00.2900.2627" name=3DGENERATOR></HEAD>
<BODY>
<DIV><FONT face=3DArial color=3D#0000ff size=3D2>
<DIV><SPAN class=3D328145111-29042005><FONT face=3DArial color=3D#0000ff =
size=3D2>Paraphrased from the comments of more than one compiler =
developer:=20
"99% of developers don't care about blazing floating point performance, =
but the=20
other 1% bitches alot" <g> Obviously, these are not Fortran =
compiler=20
developers ;)</FONT></SPAN></DIV>
<DIV><SPAN class=3D328145111-29042005><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D328145111-29042005><FONT face=3DArial color=3D#0000ff =
size=3D2>Seriously though, one of the things I like about the Shootout =
is that it=20
is not focused on floating point, as so many benchmarks seem to be. Over =
the=20
past year of so I've written alot of C/++ for =
'high-performance'=20
programs ('alot' and 'high-performance' are relative of course ;). =
I just=20
grep'd that repository and it appears that the above ratio holds =
there as=20
well.</FONT></SPAN></DIV>
<DIV><SPAN class=3D328145111-29042005><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D328145111-29042005><FONT face=3DArial color=3D#0000ff =
size=3D2>Another thing is that so often double =
precision floating point=20
is the only thing measured. And is perhaps over-used in production code =
as well.=20
For one recent floating point intensive app. that I tuned, all I =
did=20
was basically 'refactor' it to use SP instead of DP and the =
performance=20
soared and mem. usage dropped. All w/o a lick of difference to the end =
user=20
because the extra precision was not needed in this case. My point =
being,=20
why are all of the floating point benchmarks I see out there (not just =
the=20
Shootout) double precision when in reality single precision is (or =
should be)=20
used in more day-to-day applications?</FONT></SPAN></DIV>
<DIV><SPAN class=3D328145111-29042005></SPAN> </DIV>
<DIV><SPAN class=3D328145111-29042005>FP perf. is important, but I'd =
hate to see=20
the Shootout focus a whole lot more on it.</SPAN></DIV>
<DIV><SPAN class=3D328145111-29042005></SPAN> </DIV></FONT></DIV>
<BLOCKQUOTE=20
style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff 2px =
solid">
<DIV class=3DOutlookMessageHeader dir=3Dltr align=3Dleft><FONT =
face=3DTahoma=20
size=3D2>-----Original Message-----<BR><B>From:</B>=20
shootout-list-admin@lists.alioth.debian.org=20
[mailto:shootout-list-admin@lists.alioth.debian.org]<B>On Behalf Of =
</B>Robert=20
Seeger<BR><B>Sent:</B> Friday, April 29, 2005 6:40 AM<BR><B>To:</B>=20
shootout-list@lists.alioth.debian.org<BR><B>Subject:</B> Re: =
[Shootout-list]=20
What people say about shootout<BR><BR></FONT></DIV>Heh. Those are the =
types of=20
tests that I like. Since my real world problems tend to have more to =
do with=20
text processing than the math-oriented ones most people here are =
interested=20
in...<BR><BR>Rob Seeger<BR><BR>
<DIV><SPAN class=3Dgmail_quote>On 4/29/05, <B =
class=3Dgmail_sendername>Isaac=20
Gouy</B> <<A =
href=3D"mailto:igouy2@yahoo.com">igouy2@yahoo.com</A>>=20
wrote:</SPAN>
<BLOCKQUOTE class=3Dgmail_quote=20
style=3D"PADDING-LEFT: 1ex; MARGIN: 0pt 0pt 0pt 0.8ex; BORDER-LEFT: =
rgb(204,204,204) 1px solid">There's=20
always the danger of insularity on a low-volume list like =
this<BR>one, so=20
it's a healthy exercise to look around the web and see =
what<BR>people say=20
about shootout.<BR><BR>comp.lang.ada is still talking about =
count-words and=20
word-frequency - <BR>not my choice for the most interesting =
benchmarks, but=20
interesting is<BR>in the eye of the=20
=
beholder.<BR><BR>__________________________________________________<BR>Do=
=20
You Yahoo!?<BR>Tired of spam? Yahoo! Mail has the best =
spam=20
protection around <BR><A=20
=
href=3D"http://mail.yahoo.com">http://mail.yahoo.com</A><BR><BR>_________=
______________________________________<BR>Shootout-list=20
mailing list<BR><A=20
=
href=3D"mailto:Shootout-list@lists.alioth.debian.org">Shootout-list@lists=
.alioth.debian.org=20
</A><BR><A=20
=
href=3D"http://lists.alioth.debian.org/mailman/listinfo/shootout-list">ht=
tp://lists.alioth.debian.org/mailman/listinfo/shootout-list</A><BR></BLOC=
KQUOTE></DIV><BR></BLOCKQUOTE></BODY></HTML>
------=_NextPart_000_0004_01C54C8F.2BC719B0--