Re: [9fans] nvidia scrolling performance

Re: [9fans] nvidia scrolling performance

[erik quanstrom]

i only have a pci card, but someone with an agp card and a machine that
allows bios-controlled agp bandwidth could eliminate some possibilities.
if it's bus limited, then the total performance should be linear in agp
bus speed, right?  of course we still wouldn't know which direction on
the bus was limiting.

perhaps it's time to add another machine to the second-hand hardware collection.
;-)

pci-x (1.066G/s) has the same bandwidth as PCIe x4 (1G/s).
PCIe SLI = 2 * 16x = 8G/s, which ought to be enough for just about anyone.

- erik

On Sun Apr 30 13:13:21 CDT 2006, steve@quintile.net wrote:
> > Frame buffer memory is very very slow to read from,
> > and not just on nvidia.  When I did some timings six years
> > ago, I found that reading from frame buffer memory
> > was slower than reading from disk.  I'm sure the situation
> > hasn't gotten better.  It's not on the fast path for any
> > other system, so the vendors just don't care.
>
> I may be talking rubbish but I understood this is a fundamental
> problem with reading VGA memory over the PCI bus. VGA cards are
> designed for fast writes and not fast reads.
>
> People have been very interested in using the GCPUs in graphics cards
> to do video processing (to disk rather than for display) but the limiting
> factor seems to have been the speed at which data can be read back.
> I do hear that some cards are appearing with dual PCIX which will allow
> symetric access speeds to the frame buffer.
>
> -Steve
>

Re: [9fans] nvidia scrolling performance

[Paul Lalonde]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

8G/s? Nowhere near enough.  Enough for text, but try doing real
computation using that GPU...
PS3 is running 25G/s bi-directional.  Those bits move.

Paul

On 30-Apr-06, at 5:52 PM, erik quanstrom wrote:

> i only have a pci card, but someone with an agp card and a machine
> that
> allows bios-controlled agp bandwidth could eliminate some
> possibilities.
> if it's bus limited, then the total performance should be linear in
> agp
> bus speed, right?  of course we still wouldn't know which direction on
> the bus was limiting.
>
> perhaps it's time to add another machine to the second-hand
> hardware collection.
> ;-)
>
> pci-x (1.066G/s) has the same bandwidth as PCIe x4 (1G/s).
> PCIe SLI = 2 * 16x = 8G/s, which ought to be enough for just about
> anyone.
>
> - erik
>
> On Sun Apr 30 13:13:21 CDT 2006, steve@quintile.net wrote:
>>> Frame buffer memory is very very slow to read from,
>>> and not just on nvidia.  When I did some timings six years
>>> ago, I found that reading from frame buffer memory
>>> was slower than reading from disk.  I'm sure the situation
>>> hasn't gotten better.  It's not on the fast path for any
>>> other system, so the vendors just don't care.
>>
>> I may be talking rubbish but I understood this is a fundamental
>> problem with reading VGA memory over the PCI bus. VGA cards are
>> designed for fast writes and not fast reads.
>>
>> People have been very interested in using the GCPUs in graphics cards
>> to do video processing (to disk rather than for display) but the
>> limiting
>> factor seems to have been the speed at which data can be read back.
>> I do hear that some cards are appearing with dual PCIX which will
>> allow
>> symetric access speeds to the frame buffer.
>>
>> -Steve
>>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (Darwin)

iD8DBQFEVV2gpJeHo/Fbu1wRAowpAJ0SaqAQvQO7LWvyWvbDkAsa9FY6lACgyZaL
A9qxHIBefBlanunFaLXl9Nc=
=PktC
-----END PGP SIGNATURE-----

[9fans] Re: nvidia scrolling performance

[Paweł Lasek]

On 5/1/06, erik quanstrom <quanstro@quanstro.net> wrote:
>....
> PCIe SLI = 2 * 16x = 8G/s, which ought to be enough for just about anyone.

Looks like I have to disrupt some pleasant assumptions....

PCI-Express cards don't give a damn about the number of lines that are
connected ("1x" means one pci-express line), the only thing it changes
is the transfer speed.

And except for workstation (traditional sense of specialized mb and so
on) or special gaming motherboards I have yet to hear about a SLI
motherboard with 32 PCI-express lines.... (Standard PC motherboards
have 20 lines, latest PowerMacs have at least 32 lines - 1*16x 1*8x
2*4x + 4x ones have 8x physical connectors to use 8x cards in 4x mode,
maybe more - need to visit an Apple Store...)

Actually, when you use the switch (hardware or software) to change
into sli mode, you split 8 lines from "normal" port to "second",
making them 2* 8x ports :)

Of course, using special chips (like workstation line of Nforce, hard
to get from what I heard, allows you to build up to *LOT* of lines
:P), you can get the required 32 lines to get both cards working at
2*16x (PCI-Express defines single ports with up to 32x lines)

Of course, each line is 250 MiB/s bi-directional full-duplex (On
PCI-Express, reading from video memory is fast enough, shortly
speaking - it's the core factor for nVidia/ATI low-budget cards like
"Turbo Cache" and so on....)

> - erik

Over 4 years of trying to buy a new computer - At least I learned a
little from it :D
(Basically all my hardware is n-hand.... and people ask why it
sometimes it just doesn't work!? Not to mention that it's hard to get
i440bx compatible sdrams these days...)

--
Paul Lasek