home *** CD-ROM | disk | FTP | other *** search
- Path: nntp.teleport.com!sschaem
- From: sschaem@teleport.com (Stephan Schaem)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: TMapping again!
- Date: 1 Mar 1996 18:00:58 GMT
- Organization: Teleport - Portland's Public Access (503) 220-1016
- Distribution: world
- Message-ID: <4h7e0q$222@maureen.teleport.com>
- References: <38232526@kone.fipnet.fi> <4gn1r9$vm@maureen.teleport.com> <38232686@kone.fipnet.fi> <4h0agu$6n9@maureen.teleport.com> <4h463r$j0d@sunsystem5.informatik.tu-muenchen.de>
- NNTP-Posting-Host: julie.teleport.com
- X-Newsreader: TIN [version 1.2 PL2]
-
- Juergen "Rally" Fischer (fischerj@informatik.tu-muenchen.de) wrote:
-
- : |> : > on x86 doing a 32bit immediate add is as fast as a reg to register
- : |> : > add?
- : |>
- : |> : Should be, like on a 040.
- : |>
- : |> On the 040/030 , because of the code/data cache are separate the
-
- : aah wow on 030, too ? 2 cycles for a add.l #nn,Dn ? sure ?
- : i.e not the "2 cycles/cache_word" rule of 020 ? would be very nice.
-
- I wanted to mention that the 030 too as a separate inst/data cache
- , the x86 dont have this untill the 586...
- But ok the 040 take 1 cycle at a addi.l vs 6 for the 030 (1 vs 4 for
- for addi.w) Do you know how much a 486 and 386 take to add a 32bit
- or 16bit imediate value to a register VS register to register?
- If its like the 040 it should be as fast as register to register, right?
-
- Stephan
-