Barcelona Article Online

Article: Inside Barcelona: AMD's Next Generation
By: David Kanter (dkanter.delete@this.realworldtech.com), May 16, 2007 12:38 pm
Room: Moderated Discussions
Dean M (dean.m8@gmail.com) on 5/16/07 wrote:
---------------------------
>Great read. Thanks.
>
>Just a few things:
>
>"...16 XMM registers, which actually occupy 2 slots, since >they are 128 bits wide."
>
>This is true for the K8, but not for Barcelona.
>From the AMD 10h optimization guide:
>
>"A floating-point XMM register is viewed as one 128-bit >register internal to the processor.
>Current generation processors cannot write a 64-bit half of >a 128-bit XMM register
>without having a merge dependency on the other 64-bit >half."

Interesting, thanks for pointing that out. I'll update my article as such.

>Also, the K8/Barcelona diagrams would seem to suggest that >data is stored from
>LSU1 and loaded to LSU2, but that's not the case as you >yourself have written in the article.
>Stores are written (after retirement) from LSU2 >exclusively. While loads can be
>initialized from either LSU1 or LSU2, the actual data is >passed directly to the
>IFFRF (and can be cough by a reservation station or ALU on >the way there).

So my diagrams did take a little liberty. I am extremely constrained by the available space that I have. My diagrams must be 750p wide, no more, and the Barcelona one was already pushing up against the limits to the point where at the end I had to clip out 16 pixels from the L2 and L3 cache basically to get it to fit.

The diagram is meant to indicate that the 64 bit addresses go from LSU1-->cache tags. 128 bit data comes back from the cache, and 64 bit data goes in. Unfortunately, I didn't have enough room to stick in address, read data and write data lines, without things getting ugly. I'd call it an unfortunate artistic compromise.

>From the same diagram: FPMISC should present its data >directly to LSU1, instead of going through a reservation >station.

Isn't FPMISC also used for integer<-->FP conversions? I was trying to make the point that the FPMISC pipeline is used to get data between the two functional unit clusters, rather than FPSTORE.

>And a typo: 44 Entry Integer Future File – should be 40.

You're right!

DK
< Previous Post in ThreadNext Post in Thread >
TopicPosted ByDate
Barcelona Article OnlineDavid Kanter2007/05/16 03:20 AM
  Barcelona Article OnlinePiedPiper2007/05/16 05:12 AM
    Yes, I left out a sentence there. Fixed (NT)David Kanter2007/05/16 12:07 PM
  Barcelona Article Onlineanonymous2007/05/16 06:01 AM
    Barcelona Article OnlineAnonymous2007/05/16 06:28 PM
      Barcelona Article Onlineanonymous2007/05/16 07:52 PM
  Barcelona Article OnlineAnonymous12007/05/16 07:08 AM
  Barcelona Article OnlineDean M2007/05/16 11:09 AM
    Barcelona Article OnlineDavid Kanter2007/05/16 12:38 PM
      Barcelona Article OnlineDean M2007/05/16 02:10 PM
  Barcelona Article OnlineIntelUser20002007/05/16 02:59 PM
    Barcelona Article OnlineLinus Torvalds2007/05/16 03:24 PM
    Barcelona Article OnlineDavid Kanter2007/05/16 04:57 PM
    Barcelona Article OnlineMichael S2007/05/17 05:07 AM
      Barcelona Article OnlineIntelUser20002007/05/18 08:58 PM
  8 socket serversDoug Siebert2007/05/16 04:58 PM
    8 socket serversMichael S2007/05/17 05:20 AM
      8 socket serversJoe Chang2007/05/17 07:38 AM
      8 socket serversAlex Jones2007/05/17 09:35 AM
      8 socket serversJose2007/05/23 08:23 AM
        8 socket serversMichael S2007/05/23 11:37 AM
          8 socket serversanonymous2007/05/26 03:49 PM
            8 socket serversJoe Chang2007/05/27 01:46 PM
        8 socket serversDoug Siebert2007/05/23 09:56 PM
          8 socket serversJoe Chang2007/05/24 04:33 AM
          8 socket serversAnonymous2007/05/24 11:18 AM
            8 socket serversDoug Siebert2007/05/24 10:47 PM
              8 socket serversLinus Torvalds2007/05/25 10:35 AM
          8 socket serversNick2007/05/25 02:29 AM
  Performance estimation seems oddHotar2007/05/17 01:54 AM
    Performance estimation seems oddDavid Kanter2007/05/17 08:38 AM
  microops vs macroops on page 4Peter Lund2007/05/17 12:04 PM
    microops vs macroops on page 4David Kanter2007/05/21 04:51 PM
      microops vs macroops on page 4EduardoS2007/05/21 05:42 PM
        microops vs macroops on page 4dess2007/05/21 07:00 PM
  Barcelona Article OnlinePeter Lund2007/05/17 12:25 PM
  macro-op vs. micro-opdess2007/05/21 07:24 AM
    macro-op vs. micro-opDavid Kanter2007/05/21 04:38 PM
      macro-op vs. micro-opdess2007/05/21 06:15 PM
        macro-op vs. micro-opDavid Kanter2007/05/22 12:11 AM
          macro-op vs. micro-opdess2007/05/22 03:56 AM
            macro-op vs. micro-opGipsel2007/05/22 05:05 AM
              macro-op vs. micro-opdess2007/05/22 05:52 AM
                macro-op vs. micro-opanonymous2007/05/22 06:14 AM
                  macro-op vs. micro-opdess2007/05/22 06:44 AM
                    macro-op vs. micro-opEduardoS2007/05/22 02:19 PM
                      macro-op vs. micro-opdess2007/05/24 08:52 AM
            Stop comparing apples to orangesEduardoS2007/05/22 02:30 PM
              Stop comparing apples to orangesdess2007/05/22 04:09 PM
                Stop comparing apples to orangesdess2007/05/22 04:30 PM
                Stop comparing apples to orangesEduardoS2007/05/22 04:31 PM
                  Stop comparing... apples to oranges?dess2007/05/24 09:30 AM
              Stop comparing apples to orangesanonymous2007/05/22 08:12 PM
                Stop comparing apples to orangesEduardoS2007/05/23 02:50 PM
          macro-op vs. micro-opanonymous2007/05/22 06:08 AM
            macro-op vs. micro-opdess2007/05/22 06:40 AM
              macro-op vs. micro-opanonymous2007/05/22 06:48 AM
      macro-op vs. micro-opdess2007/05/21 08:30 PM
        macro-op vs. micro-opanonymous2007/05/22 06:44 AM
          macro-op vs. micro-opdess2007/05/24 09:38 AM
      macro-op vs. micro-opMichael S2007/05/22 05:26 AM
Reply to this Topic
Name:
Email:
Topic:
Body: No Text
How do you spell avocado?