PCI Express

[daytripper]

And good luck on getting your hands on one of those
boards or a processor for one of those boards.

I know someone who is desperate for a Socket T
mobo\cpu combo and the vendors are telling him he'll
be lucky to get one before 2005. All production -
what little there is so far - is apparently going
only to the big computer manufacturers for the next
few months.

otoh, maybe it's the plethora of errata - and some pretty scary ones at that -
that are delaying deployment?

/daytripper (some of those bugs are effing ugly...)

 

[AJ]

And good luck on getting your hands on one of those
boards or a processor for one of those boards.

I know someone who is desperate for a Socket T
mobo\cpu combo and the vendors are telling him he'll
be lucky to get one before 2005. All production -
what little there is so far - is apparently going
only to the big computer manufacturers for the next
few months.

As far as I care, all the changes are to be swallowed at once:
BTX, 90 nm, PCIe, SATA II. I won't care for a couple of years
at least. Unless the old stuff runs out. I'm not an early adopter.
The point is though that the current generation of boards has
already nixed AGP.

AJ

 

[dhs]

There are both PCI and AGP versions for the Parhelia,
but AFAIK only AGP for the P650 and P750.

Matrox just introduced a low profile PCI version of the P650.
It should come with both low profile and standard size brackets.
You can use it with either dual DVI or dual VGA monitors.
(And no, I don't know why there isn't an AGP version ...)

 

[AJ]

Matrox just introduced a low profile PCI version of the P650.
Cool!

It should come with both low profile and standard size brackets.
You can use it with either dual DVI or dual VGA monitors.
(And no, I don't know why there isn't an AGP version ...)

Probably because AGP is dead going forward?

AJ

 

[emmanuel poirier]

Kevin C. a écrit :

Since this is Intel's new chipset, would I be correct in assuming that a
pcie mainboard would only support P4s (no AMD)? At least, before Via gets a
clone out?

Also, have there been any tests to confirm real-world performance gains?
Something comparing AGP 8x to PCIE 16x, with comparable GPUs? If so, what is
the margin of difference like?

Basically, I am in the market for a new mainboard/cpu/video and while PCIE
is obviously the forward-thinking choice, it seems like it is just barely
emerging and I don't know how this will affect cost-effectiveness. (My
budget is a concern.)

Hi !

There are tests but which I known are in french, use google.
The test I've read said that it doesn't worth the cost for now : in fact
even AGP 4x isn't used at it's full potential, years after its release,
AGP 8x the same and even the same for PCI Express which was just
released a month ago. I have even see that you loose perf with DDR-2 !

 

[assaarpa]

/*
The test I've read said that it doesn't worth the cost for now : in fact
even AGP 4x isn't used at it's full potential, years after its release,
AGP 8x the same and even the same for PCI Express which was just
released a month ago. I have even see that you loose perf with DDR-2 !
*/

It does not come as a surprise that AGP x, 4x et cetera are NOT used to
their "full potential" as it is much more efficient to keep the working set
in GPU local memory. Initially, when AGP was launched some vendors promoted
use of "DIME" feature of AGP, including Matrox Graphics with their G200
series of products. While DIME was "okay" speed it still was a lot slower
than rendering from local memory. During the years the gap between the two
has just grown wider.. using AGP 8x to it's "full potential" just doesn't
make any sense, unless there is data constantly generated by the CPU and
pushed to the GPU.. even this is less practical now that GPU's are much more
programmable than the fixed pipeline model that we have to deal with just a
few years ago.. basicly we can do with a much smaller seed data and
synthesize data in the GPU.

Introduction of VS 3.0 with vertex samplers is important step since now we
can synthesize data with pixel program, render into texture and use that as
input to next pass vertex program. Latencies these techniques introduce can
be amortized in practise.

What PCI Express brings to the developers is a window to completely
different kind of uses of GPU the most important factor in this is that the
bus is full bandwidth to both directions, finally it is feasible to read
back from GPU memory to system memory. Thinking PCI Express only as
increased bandwidth is very narrow-minded and shows that whoever makes these
claims does have shallow understanding of modern GPU programming. Propably
some journalist without technical background, just a lucky guess..

 

[assaarpa]

Uh, in short: PCI Express doesn't bring any new value to contemporary
applications, they wouldn't use the extra bandwidth anyway. What PCI Express
does is that it allows a completely new breed of applications to be
developed which would be completely unthinkable due to performance reasons
in curreny bus architechtures on PC, such as AGP. In this light for the
first year or two PCI Express is something that developers will fall in love
with and the labours of love should be apparent 9 months later.

This is to say.. consumers shouldn't look forward to immediate benefits,
hardware still needs software to dance.