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: Bill Cox<bill@v...>
    Date: Mon Jun 14 17:06:27 CEST 2004
    Subject: [openrisc] Area result in ViaMask array
    Top
    Hi.

    I got permission to post the results I got when mapping the OR1200 into
    our .13u ViaMask fabric.

    I used Amplify ASIC with an unattainable required clock, so
    timing-driven was full-on. We don't yet have characterized .13u timing,
    so I wont report those results. However, I do expect this design to run
    in the 300Mhz range when we're done.

    I used 4K instruction and data caches, and no QMEM (internal physical
    address space memory).

    After synthesis, ViaPath mapped the logic into 3208 logic cells. This
    easily fits into 13 logic tiles, which has a total area of 1.02mm^2.
    However, there is also distributed SRAM in this region. The various
    register files and the instruction and data caches use a total of 22 of
    the ViaMask fabric's 128 word by 32 bit dual-port SRAMs. Each logic
    tile has one of these, so the total size of logic plus SRAM was SRAM
    limited, and fits in a total of 1.72mm^2. Of this area, 9 logic tiles
    worth of logic remains free for use, or about 41%. With 2K byte caches,
    the entire design plus cache should fit in 14 tiles, or about 1.1mm^2.

    Bill

    On Mon, 2004-06-14 at 03:49, Bill Cox wrote:
    > Hi.
    >
    > I showed off the OR1200 at the Design Automation Conference last week.
    > I think many people there were quite interested in the OR1200 (which
    > made it a good demo for our stuff). I also used it in a Structured ASIC
    > tutorial I gave on Thursday, which had 16 people.
    >
    > I used a 4K instruction and data caches, and turned off both MMUs, and
    > QMEM. There was still a ton of SRAM on the 9x10 ViaMask base available,
    > but I wanted the whole CPU (with memory) to fit in one corner of the
    > array.
    >
    > I'll talk to my boss about giving out more detailed results (area,
    > timing, etc). A bunch of guys copied down the data they saw on the
    > screen, so I think the info is already in the hands of our competitors.
    >
    > While I'm not much of an HDL designer, I was impressed by the quality of
    > the code in the OR1200. I had very little difficulty using it. Thanks
    > to the authors for making it available.
    >
    > Bill
    >
    >
    > _______________________________________________
    > http://www.opencores.org/mailman/listinfo/openrisc

    ReferenceAuthor
    [openrisc] Showed OR1200 at DACBill Cox

    Follow upAuthor
    [openrisc] Area result in ViaMask arrayDamjan Lampret

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