From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Mon,  1 Dec 2008 08:24:58 -0700
From: plalonde@telus.net
To: 9fans@9fans.net
Cc: 9fans@9fans.net
Message-ID: <27049442.3521101228145098701.JavaMail.nitido@priv-edtnes95>
MIME-Version: 1.0
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Subject: Re: [9fans] image/memimage speed
Topicbox-Message-UUID: 54f7cbce-ead4-11e9-9d60-3106f5b1d025

<br/>Intel integrated and derivatives have an actual "frame buffer", in tha=
t you set a base pointer to a contiguous block of memory visible to the ada=
pter and scan-out happens from there.  <br/><br/>I think the real performan=
ce issue for hardware where the frame buffer is in the PCIe shared memory a=
pperture is that writes are write-through/coalesced on their way across the=
 PCIe, but reads can't be, and so incur huge stalls.<br/><br/>Paul<br/><br/=
><br/>Dec 1, 2008 06:34:59 AM, 9fans@9fans.net wrote:<br/><br/>=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D<br/><br/><br/>> <br/>> Very true=
, the only exception to this I know of is some of the modern<br/>> Dual PCI=
Express cards which use a bus in each direction.<br/>> <br/><br/>do you hav=
e a reference for "dual pciexpress"? =C2=A0as far as i know,<br/>pcie/agp/p=
ci cards only have a single bus that goes both ways.<br/><br/>my limited un=
derstanding was that the reason that reading the<br/>framebuffer was slow, =
was that there is no framebuffer. =C2=A0it's an<br/>illusion that the card =
provides that's easy to write but not so<br/>easy to fake on reads.<br/><br=
/>someone with real understanding of graphics please correct me.<br/><br/>-=
 erik<br/><br/>