i had searched on internet but i am unable to find...WHY FIXED POINT PROCESSOR HAS HIGHER CLOCK SPEED THAN FLOATING POINT PROCESSOR? regards abhishek This message was sent using the Comp.DSP web interface on www.DSPRelated.com
fixed vs floating
Started by ●April 20, 2005
Reply by ●April 20, 20052005-04-20
On Wed, 20 Apr 2005 08:53:17 -0500, abhi_15sep wrote:> i had searched on internet but i am unable to find...WHY FIXED POINT > PROCESSOR HAS HIGHER CLOCK SPEED THAN FLOATING POINT PROCESSOR? > regards"Why" is often a really difficult question to answer in general terms, but here's a stab at it. Roughly: floating point arithmetic takes more effort than fixed point: there are the extra steps of normalizing and so on. There are also a bunch of exceptional conditions to catch and do something about. Consequently, floating point units tend to be bigger and slower than their fixed point (i.e., integer) brethren. Clock speed follows more or less directly from that. Cheers, -- Andrew
Reply by ●April 20, 20052005-04-20
To simplify greatly: all other things being equal, the larger and more complex the logic, the slower it will operate. The logic required to implement floating point math is more complex than that required to implement fixed point math. This same fact is also the reason that floating point DSPs typically consume more power than their fixed point counterparts (again, assuming all other things are equal, which they never are!). -- Jon Harris SPAM blocked e-mail address in use. Replace the ANIMAL with 7 to reply. "abhi_15sep" <abhi_15sep@yahoo.com> wrote in message news:6YGdnQtpqfZQw_vfRVn-3A@giganews.com...> i had searched on internet but i am unable to find...WHY FIXED POINT > PROCESSOR HAS HIGHER CLOCK SPEED THAN FLOATING POINT PROCESSOR? > regards > abhishek
Reply by ●April 20, 20052005-04-20
I think in normal manual calculation multiplying 23.2345666 e+23 * .987324345 e-47 is much more complex than multiplying 234567890 * 9808756. The thing is that in fixed point you are limiting the size and the the processor does not take care of decimal point and the exponent. The desing of hardware is simple and hence when it runs the speed is more. Dspwhiz