Autor: Bronek Kozicki (brok_at_rubikon.pl)
Data: Fri 10 May 2002 - 23:19:46 MET DST
Obszerny artykuł o nowym, bardzo poważnym konkurencie nVidii i ATI -
chipie P10 VPU z 3DLabs został opublikowany tutaj
http://www.anandtech.com/video/showdoc.html?i=1614
Poniżej wybrane fragmenty:
- - -
While 3DLabs hasn't released any card specs yet, they are claiming over
20GB/s of memory bandwidth is possible with the P10 meaning that they'd
need at least 312.5MHz DDR SDRAM. Considering that the current GeForce4
Ti 4600 uses 325MHz DDR SDRAM, it's very possible that you'll see cards
with over 20GB/s of memory bandwidth. [ . . . ] Other than the 256-bit
memory bus and the addition of a block labeled as the Command Processor
(which we'll get to later), the P10 looks pretty familiar; here are some
quick specs:
- 0.15-micron manufacturing process (same process as the GeForce4)
- 76M transistors [ . . . ]
- 256-bit DDR memory interface (up to 20GB/s of memory bandwidth w/
312.5MHz DDR)
- up to 256MB of memory on-board
- AGP 4X support
- Full DX8 pixel and vertex shader support
We've been deliberately light on the specs here as the rest must be
explained in greater detail; it wouldn't do them justice to put them in
a simple list.
- - -
By far the most powerful desktop GPU when it comes to vertex shaders is
NVIDIA's GeForce4 which has two vertex shaders. The 3DLabs P10 on the
other hand has 16 32-bit floating-point geometry processors that handle
vertex processing, does this mean that the P10 has 8 times the vertex
throughput of the GeForce4? Of course not, let's start off by making
this an apples to apples comparison. [ . . . ] If you are comparing
theoretical vertex throughput between the P10 and the GeForce4's dual
vertex shaders you'd end up with a little more than a 2x advantage in
favor of the P10.
- - -
Another fairly major feature that is brought to the table with the P10
is what 3DLabs calls the P10's Virtual Memory System (VMS). The way VMS
works is by storing all textures in main memory and treating the memory
on the graphics card itself as a very large cache. When a texture is
requested, the entire texture isn't downloaded instead a 256x256 block
of 32-bit pixels can be pulled in locally and accessed. [ . . . ] From
the standpoint of game developers, VMS is quite attractive as it enables
the use of far more textures than is currently possible. Currently game
developers are very cautious to use more textures that there is
available video memory because swapping out to main memory results in a
huge performance hit. But with VMS you get the benefits of an extremely
high bandwidth caching effect and can deal with much larger textures
than ever before.
- - -
The P10 will be first made available very soon on a 3DLabs card aimed at
the professional market. But the chip obviously has a lot of potential
in the consumer/gaming market and indeed it will make its way down
there. [ . . . ] The P10 will then be found on Creative Labs branded
boards that will sell at prices competitive with the GeForce4 (or
whatever NVIDIA's high-end card is at the time). While 3DLabs wouldn't
give us an indication of exactly when we could expect consumer/gaming
cards from Creative, they did say it would be before the end of the
year. [ . . . ] In terms of the viability of the P10's architecture,
it's definitely a very powerful chip. [ . . . ] It will be interesting
to see how their next-generation VPU on a 0.13/0.10-micron process turns
out with a full DX9 compliant pipeline.
- - --
no i ostatnie zdanie z artykułu:
As unexpected as 3DLabs' announcement was (we weren't anticipating it
for another few weeks) it's just the first in a line of new technologies
in the graphics sector.
- - -
Zainteresowanym polecam całość; nie zacytowałem wielu innych rewelacji
(obsługa wielowątkowości; programowalny pixel shader; programowany AA i
końcowe przetwarzanie w 64 bitowym kolorze).
Zdaje się, że szykuje się kolejna rewolucja (po T&L) w kartach
graficznych :)))
B.
To archiwum zostało wygenerowane przez hypermail 2.1.7 : Wed 19 May 2004 - 00:48:49 MET DST