Announcement

Collapse
No announcement yet.

NVIDIA PR Responds To Torvalds' Harsh Words

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • V!NCENT
    replied
    Originally posted by johnc View Post
    That may be true... but probably for different reasons than you think it sucks.
    Like what? Not for closed driver reasons? Let's hear what the Android crowd has to say about that:


    And then what about John Carmack, who cries for low level, direct access to graphics cards on PC's? He really raged (pun intended) with the last id's release about the abstractions of graphics card drivers slowing down performance to such an extend that he favors Intel graphics because Intel lets him actually access the low level memmory management and memmory blitting, without having the graphics driver needing to load an entire tile of texture memmory before a single pixel can be read out (see also Battlefiel 3 PC horrors):
    I have high hopes that because it is all integrated memory, Intel will be able to lead the way with surfacing and direct access. That will give them the opportunity to sometimes take console developers who are used to this lower level access and maybe have something -shock of shocks- run better on Intel’s Integrated Graphics part than the much more expensive NVIDIA or AMD card that has all the layers of driver overhead.
    John Carmack Interview: GPU Race, Intel Graphics, Ray Tracing, Voxels and more! Last week we were in Dallas, Texas covering Quakecon 2011 as well as hosting

    And now that AMD allows direct access with Fusion and documentation, where does that leave nVidia?

    Face it man, nVidia realy does SUCK. PS: They are the new 3dfx. I hope you will not clamp onto a GeForce like the 3dfx fans clamp onto their Voodoo's, and just move on.
    Last edited by V!NCENT; 21 June 2012, 04:47 PM.

    Leave a comment:


  • johnc
    replied
    It would depend on the driver / firmware blob they're distributing and the license associated with it I would think.

    Leave a comment:


  • johnc
    replied
    Originally posted by asdx
    I understand the firmware blobs aren't part of the kernel and that they run directly in the card. But why does the kernel has to ship them in the same project? Why couldn't they create another project for those blobs?

    Some people think now that Linux isn't a free project anymore for this reason, what about that? Aren't they infringing with the GPLv2 by shipping Linux with these blobs?

    All this stuff is pissing me off.
    Where does the GPLv2 say that you can't ship a Linux kernel with a binary driver?

    That would be a major problem for Android.

    Leave a comment:


  • johnc
    replied
    Originally posted by V!NCENT View Post
    It appears that the market agrees that nVidia SUCKS
    http://vr-zone.com/articles/jon-pedd...ced/16037.html
    That may be true... but probably for different reasons than you think it sucks.

    And NVIDIA is going to have a much harder time in the desktop / laptop markets since IGP products are going to increase in popularity. (And Intel refused to license x86 to NVIDIA, so they have no hope in the IGP market.) However, getting in on the ground floor of the ARM SoC business was a good move for them.

    Leave a comment:


  • V!NCENT
    replied
    Originally posted by johnc View Post
    lol... you're totally out there.

    Don't buy NVIDIA.

    For those who don't place a priority on the things you do, they can purchase NVIDIA if they want.

    The consumer choices made in the market will guide the way.
    It appears that the market agrees that nVidia SUCKS

    Leave a comment:


  • johnc
    replied
    Originally posted by V!NCENT View Post
    Hey we sell this car, but you can only drive routes we provided you clearance for. You want to use this Jeep on the highway? No realy, you can only take it ofroading because we advertise it to be usefull for ofroading. Oh you don't like that? But it's so good for ofroading! No realy, why should we tell you how to put it in eco mode, while it's so great for ofroading! Yes you could use it to drive on roads, but we're not goi to tell you where the controlls are for that. Why? Because we're too busy SUCKING, ofcourse!
    lol... you're totally out there.

    Don't buy NVIDIA.

    For those who don't place a priority on the things you do, they can purchase NVIDIA if they want.

    The consumer choices made in the market will guide the way.

    Leave a comment:


  • V!NCENT
    replied
    Originally posted by johnc View Post
    Seriously... what are you talking about?

    The function of the GPU is what the manufacturer advertises and sells, and nothing more. That means if you want to do CUDA programming then you have to do it through the means that the manufacturer provided.
    Hey we sell this car, but you can only drive routes we provided you clearance for. You want to use this Jeep on the highway? No realy, you can only take it ofroading because we advertise it to be usefull for ofroading. Oh you don't like that? But it's so good for ofroading! No realy, why should we tell you how to put it in eco mode, while it's so great for ofroading! Yes you could use it to drive on roads, but we're not goi to tell you where the controlls are for that. Why? Because we're too busy SUCKING, ofcourse!

    Leave a comment:


  • V!NCENT
    replied
    Originally posted by kotakotakota View Post
    Isn't clean room reverse engineering legal?
    Hey we sell a product for you to compute with! How? No we're not selling you this so you can compute with it... Documentation, you say? No we're to busy SUCKING to give you that! You didn't expect us to let you know how? Figure it out yourself by "clean room reverse engineering", which should take you about two years, while you could just buy AMD and Intel instead. Great deal, no?

    Leave a comment:


  • johnc
    replied
    Originally posted by V!NCENT View Post
    It is the perfect analogy; a car is a transport vehicle, not a computer. A parallel microprocessor, is not just a graphics accelerator anymore; it's also meant to be programmeable (CUDA, hello?) guess what? It can't be programmed. The DRIVER needs to be programmed. And THAT is as useless as programming a car.
    Seriously... what are you talking about?

    The function of the GPU is what the manufacturer advertises and sells, and nothing more. That means if you want to do CUDA programming then you have to do it through the means that the manufacturer provided.

    Leave a comment:


  • V!NCENT
    replied
    Originally posted by johnc View Post
    This is such an awful analogy. Every car today is loaded up with closed-source, proprietary software used to drive the engine, powertrain and emission systems. And, at least in the freedom-loving USA, it is illegal for anything but the manufacturer to re-program this module.

    Without this software you can't even start your damn car.
    It is the perfect analogy; a car is a transport vehicle, not a computer. A parallel microprocessor, is not just a graphics accelerator anymore; it's also meant to be programmeable (CUDA, hello?) guess what? It can't be programmed. The DRIVER needs to be programmed. And THAT is as useless as programming a car.

    Leave a comment:

Working...
X