Not logged in
Log in now
Create an account
Subscribe to LWN
LWN.net Weekly Edition for May 16, 2013
A look at the PyPy 2.0 release
PostgreSQL 9.3 beta: Federated databases and more
LWN.net Weekly Edition for May 9, 2013
(Nearly) full tickless operation in 3.10
We must put the pressure on it. Novell has already made a start by no
longer distributing proprietary modules.
ATI, AMD, and free drivers
Posted Aug 3, 2006 10:30 UTC (Thu) by NRArnot (subscriber, #3033)
At present, Intel's graphics chips are not anywhere near the standards set by NVidia and ATI, so this pressure is not intense. However, don't expect it to stay that way much longer. For one thing, the "entry level" requirements of MS Vista are quite high. For another, the direction of technological development is towards closer and closer integration, and more and more bandwidth, between the GPU and the CPU. I expect that in the not very distant future, entry-level will be on the same chip as the CPU, and high-end graphics will plug into one [or more] of the other CPU socket[s] on a dual- [or multi]-CPU motherboard.
The thing that worries me most is Hollywood, DRM and laws. We could end up with a situation where it is actually *illegal* to have a GPU supported by fully open source, just as (in some countries) it's already illegal to have a fully open-source Wireless NIC driver.
Posted Aug 3, 2006 22:59 UTC (Thu) by bluefoxicy (guest, #25366)
Having CPU + GPU on the same chip means a strong lack of flexibility. The die has to be redesigned to reposition the transistors to fit either A) A new CPU core; or B) A new GPU core; this is -very- -expensive-. The north bridge usually connects the GPU to the PCI-Express or AGP bus; in effect, the north bridge stays the same, the GPU is effectively run to a controller chip for a PCI-E card which is just wired straight into the north bridge. This gives cheap and easy flexibility because you swap two chips on the board if you want to release a new board with a better GPU.
Besides that, both GPU and CPU activities cause heat. You don't want to concentrate that in the same chip. Imagine a Barton running twice as hot; instead of a 2.0GHz Barton you have a 1.2GHz Barton to control the heat. Worse, GPU and CPU are independent, so you would have temperature fluctuations based on two separate models.
As for plugging into another CPU socket, CPU sockets are all proprietary. You don't want to search for an ATi Socket A vs ATi Socket 754 vs ATi Socket 939; this puts more manufacture stress on the company to satisfy the same market. They either have to manufacture all kinds of chips; or reduce their target market. Both are bad for business and will increase costs and thus consumer prices. Then you also have the issue of changing connectors, i.e. D-SUB vs DVI vs whatever they come up with nex; how do you change those? Buy a new mobo?
Single-chip CPU/GPU is too expensive and inflexible to work. It's not that it brings substantial gains but has a few hurdles; it's that it brings substantial problems and has negligible gains. Huge bandwidth between the GPU and the CPU is needed to load textures; besides that we're sending just a ton of simple commands. Video memory and video card DMA (which is kind of what AGP did) handles the textures and models, since they can be loaded in and sit waiting to be used; CPU-GPU integration will make the commands go more realtime by maybe 1/10000 FPS, not useful.
Posted Aug 5, 2006 1:04 UTC (Sat) by dlang (✭ supporter ✭, #313)
they have opened up the hyperchannel interface and there are several companies out there selling FPGA's and other custom co-processors that drop into opteron sockets
Posted Aug 7, 2006 17:35 UTC (Mon) by job (guest, #670)
The trend is to move more and more logic to the CPU die. I wouldn't at
all be surprised if the PCI(E) controller moves on-die along with the
memory controller. Every year there is room for more and more
transistors, you can't expect that to only be used for an ever larger
cache, and very complex instruction decoders has limited use as we've
seen lately with Intel.
Yes, having separate components is more
flexible but also more expensive. But I certainly don't miss the external
FPU. I probably won't miss the external GPU either, but I don't expect
that transition within the next five years. I don't think a certain
component "belongs" anywhere. It all depends on your process
Posted Aug 10, 2006 9:14 UTC (Thu) by Wol (guest, #4433)
Repeat after me "A GPU is a CPU" one hundred times - there isn't actually much difference between the two - as should be obvious looking at the PS3 or Xbox 360 - both use the Cell processor, and both divide the CPU and GPU functions up amongst the (presumably identical) cores.
If you read The Inquirer, Charlie has written that AMD *HAD* to buy a video company, because GPUs and CPUs are moving ever closer, and they needed the GPU technology to put it into the CPU, otherwise Intel's "today's blue skies" projects would be eating AMDs lunch in five years time.
Posted Aug 15, 2006 11:24 UTC (Tue) by NRArnot (subscriber, #3033)
Posted Aug 3, 2006 16:06 UTC (Thu) by smoogen (subscriber, #97)
Posted Aug 3, 2006 19:15 UTC (Thu) by wilck (subscriber, #29844)
Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds