From: Robin Vowels <robin.vowels@gmail.com>
Subject: Re: Numerical calculations: Why not use fixed point types for everything?
Date: Fri, 18 Jan 2013 15:06:57 -0800 (PST)
Date: 2013-01-18T15:06:57-08:00 [thread overview]
Message-ID: <23f3f31f-1f8b-4f04-98ce-025b7a81e3e0@th3g2000pbc.googlegroups.com> (raw)
In-Reply-To: 3l3jf85ae05qgsl2l0avomebmg2ogl17rq@invalid.netcom.com
On Jan 19, 5:15 am, Dennis Lee Bieber <wlfr...@ix.netcom.com> wrote:
> On Fri, 18 Jan 2013 01:17:37 -0800 (PST), Ada novice <shai.l...@gmx.com>
> declaimed the following in comp.lang.ada:
>
>
>
> > I have read (John Mc Cormick: Building parallel...with Ada's book) that fixed point arithmetic is faster than floating-point arithmetic since integer instructions are faster so I would ask about the second point put by Adam that mathematical functions are slow with fixed-point numbers. Can you please elaborate on that?
>
> Integer operations may be faster than floating point... But fixed
> point with a non-zero decimal place add in the overhead of having to
> track where the point is and shifting results to match the declared data
> type.
>
> 25 * 25 => 725
>
> 2.5 * 2.5 => 7.25 -- but if the data type is only one decimal place this
> now has to be rounded and shifted to become 7.3.
>
> All these adjustments and tracking are being done at the high-level
> instruction set, not in the processor hardware. In contrast, a floating
> point processor does all the mantissa and exponent adjustments in the
> hardware, which may be faster than having to do things like --
>
> ld i1, 25
> mult i1, 25
> add i1, 5
> div i1, 10
It may not be like that at all.
In the first place, 25 is a constant.
In the second place, adjusting the value (scaling) is achieved
with a shift, not an actual division by 10.
Thus, multiplication is simply
multiply a,b
shift.rounded a
> -- the sequence to treat integer registers as fixed 1-decimal place data
> compared to (the add/div would be added by the compiler based upon the
> number of decimal places to be kept in the result, based upon the number
> of places that were in the source
>
> ld f1, 2.5E0
> mult f1, 2.5E0
next prev parent reply other threads:[~2013-01-18 23:06 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-01-17 10:33 Numerical calculations: Why not use fixed point types for everything? Ada novice
2013-01-17 14:40 ` Nasser M. Abbasi
2013-01-17 16:16 ` Adam Beneschan
2013-01-17 17:00 ` Georg Bauhaus
2013-01-17 16:25 ` Adam Beneschan
2013-01-18 9:17 ` Ada novice
2013-01-18 17:24 ` J-P. Rosen
2013-01-18 17:52 ` Jeffrey Carter
2013-01-18 18:15 ` Dennis Lee Bieber
2013-01-18 18:59 ` Adam Beneschan
2013-01-19 4:41 ` Dennis Lee Bieber
2013-01-19 6:26 ` Jeffrey Carter
2013-01-19 14:14 ` Robert A Duff
2013-01-25 12:16 ` Paul Colin Gloster
2013-01-24 10:55 ` Ada novice
2013-01-24 11:47 ` Simon Wright
2013-01-24 14:21 ` Ada novice
2013-01-20 0:05 ` Robin Vowels
2013-01-18 23:06 ` Robin Vowels [this message]
2013-01-18 19:09 ` Adam Beneschan
2013-01-18 21:39 ` Randy Brukardt
2013-01-19 7:02 ` Ada novice
2013-01-25 12:09 ` Paul Colin Gloster
2013-01-25 12:23 ` Paul Colin Gloster
2013-01-28 9:09 ` Ada novice
2013-02-01 10:53 ` Ada novice
2013-02-01 15:01 ` Shark8
2013-02-02 18:55 ` Ada novice
2013-02-03 4:05 ` Shark8
2013-02-04 6:23 ` Ada novice
2013-02-04 6:43 ` Niklas Holsti
2013-02-04 7:27 ` Ada novice
2013-02-04 9:37 ` Niklas Holsti
2013-02-04 10:09 ` Ada novice
2013-02-04 14:24 ` Niklas Holsti
2013-02-04 16:44 ` Jeffrey Carter
2013-02-04 21:12 ` Niklas Holsti
2013-02-04 17:31 ` Robert A Duff
2013-02-04 21:20 ` Niklas Holsti
2013-02-02 21:08 ` Nasser M. Abbasi
2013-02-04 6:17 ` Ada novice
2013-02-05 2:27 ` Randy Brukardt
2013-02-06 7:11 ` Ada novice
2013-02-07 6:03 ` Randy Brukardt
2013-02-07 8:43 ` Shark8
2013-02-08 3:17 ` Randy Brukardt
2013-02-08 6:20 ` Ada novice
replies disabled
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox