LOGIN   :::   RECOVER PASS   :::   GET ACCOUNT    
Browse
  • Projects
  • Code (CVS)
  • Forums
  • News
  • Articles
  • Polls
  •  
    OpenCores
  • FAQ
  • CVS HowTo
  • Mission
  • Media
  • Tools
  • Advertise
  • Mirrors
  • Logos
  • Contact us
  • Find Resources
  • Job Opportunity
  •  
    Tools
  • Search
      
  • Download Cores (CVSGet)
  •  
    More
  • Wishbone
  • Perlilog
  • EDA tools
  • OpenTech CD
  •  
    Navigation: All forums > Openrisc > Message List > Message Post

    Message

    Reply | Reply all
    Date Prev | Date Next | Thread Prev | Thread Next Date Index | Thread Index

    From: Matjaz Breskvar<phoenix@o...>
    Date: Wed Feb 16 01:45:21 CET 2005
    Subject: [openrisc] or1ksim stable_0_1_0
    Top
    * György 'nog' Jeney (nog@s...) wrote:
    > > > Regarding FAST_SIM i'd really be interested in some numbers. Does it
    > > > really speed up execution by some meaningful margin (As far as i know
    > > > faster simulation time is the only it exists) ? If it's
    > > > negligable i'm all for throwing that one out.
    > >
    > > If I can get a working simulator with --enable-fsim, I'll do some benchmarks.
    > > If we actually want to keep FAST_SIM, then we'll need to rethink the peripheral
    > > cleanups. FAST_SIM is based on the assumption that all the configuration is
    > > static and the compiler will remove always true or always false conditions.
    > > With the peripheral cleanups however, everything became dynamic...
    >
    > I've done some benchmarks. The results are surpising to such an extent that I'm
    > sure I'm doing something wrong. I am useing stock cvs. I have attached a copy
    > of fast_config.c which has been generated from the attached sim.cfg (sim
    > --output-cfg). I have also attached my `hack' to linux to prevent it from
    > recalibrating its delay loop (just to make it boot faster).
    >
    > +-----+-------------------------+----------------------------+
    > | Run | Real with --enable-fsim | Real without --enable-fsim |
    > +-----+-------------------------+----------------------------+
    > | 1 | 3m35.566s | 2m55.641s |
    > | 2 | 3m24.519s | 3m8.849s |
    > | 3 | 2m45.621s | 3m32.061s |
    > | 4 | 3m34.219s | 3m33.546s |
    > +-----+-------------------------+----------------------------+
    >
    > I have pasted the output of the time commands below.

    So you're saying that --enable-fsim is actually --enable-slowsim ;)...
    (for all but one misterious case)...

    Just to make sure, are you running this on otherwise idle machine ? Do you
    have any daemon that renices processes hard on cpu or similar ?

    Please give me a few days to confirm this (or anybody else if you want
    to help) and if it turns out our results are consistent it's bye-bye
    "FAST_SIM"... Agreed ?

    best regards,
    p.

    ReferenceAuthor
    [openrisc] or1ksim stable_0_1_0=?unknown-8bit?Q?Gy=F6rgy?= 'nog' Jeney

     
    Copyright (c) 1999 OPENCORES.ORG. All rights reserved.