Announcement

Collapse
No announcement yet.

NVIDIA Unveils The GeForce RTX 20 Series, Linux Benchmarks Should Be Coming

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

  • Panix
    replied
    Originally posted by pcxmac View Post
    There are consumers who will swallow what they are given, and there are users who appreciate what open source means for building one's own operating system. If you don't understand the value of what amdgpu means in terms of being able to build your own user environment, you probably should stick to a 'consumer friendly' (user-abuser/vendor-dicker) brand like NVIDIA.
    What a weak reply. Good one, ignoring my point entirely. Kudos. *Shrugs*

    Leave a comment:


  • Panix
    replied
    Originally posted by torsionbar28 View Post
    ^ This. Here in 2018, there is no reason for any self-respecting Linux geek to consider Nvidia.
    Yeah, I want a massively power hungry AMD card in my computer because they do a few FOSS things.

    Leave a comment:


  • pcxmac
    replied
    Originally posted by fuzz View Post

    The new dies are pretty big, especially the 2080 Ti. The costs for the large dies are getting more and more expensive as the the production processes get smaller. It's reflected in the costs of the cards. They are really only getting away with it because some people think, due to brainwashing, there is no other option.

    I'm surprised we aren't starting to see chiplets in GPUs on AMDs end, but who knows what's coming.
    I am hoping AMD in revealing that they were/are having their cpu heads come over and 'assist' means that maybe they will be breaking open the graphics pipeline and manufacturing several smaller chiplets which represent the pipeline. On top of some sort of infinity fabric along side HBM or GDDR6, maybe even making a memory controller chiplet that can handle requests from the various pipeline aspects and handle either a GDDR6 array of memory modules or HBM stacks, (4+)

    Leave a comment:


  • pcxmac
    replied
    There are consumers who will swallow what they are given, and there are users who appreciate what open source means for building one's own operating system. If you don't understand the value of what amdgpu means in terms of being able to build your own user environment, you probably should stick to a 'consumer friendly' (user-abuser/vendor-dicker) brand like NVIDIA.

    Leave a comment:


  • Panix
    replied
    Originally posted by pcxmac View Post
    as someone who compiles all the software I use (Gentoo), and someone who appreciates Open Source and the GPL, I have to say I don't think I will ever oblige NVIDIA on the Linux platform unless they stop being like they are and offer competitive open source alternatives. There are plenty of reasons to want to have an open source graphics driver, even if it has issues keeping up with the windows proprietary variant.

    That said, I fully expect AMD to have a cost effective graphics chip that is very competitive with the 2080TI in terms of compute/graphics performance (not ray tracing) early next year, just by going to TSMC the process node will be superior to GloFo's and with the die shrink/new process, I expect big things until Nvidia upsets with it's new offering, probably next year too. If I were most people, I would ignore the 2080 graphics cards, they are new, and probably will be seriously outdated when 7nm and AMD come around, next year.
    Name some? These $500+ cards are gamer cards, aren't they? Both Windows and Linux gamers buy Nvidia.

    AMD Radeon cards STILL can't compete with regards to power and temps - as AMD cards are still power hogs.

    I would buy an AMD card but that alone prevents me. What good is the open source if the power consumption is way above the equivalent/comparable Nvidia card? No thanks. It means their cards are not efficient and the 'open source' qualities are still limited/restricted.

    Leave a comment:


  • fuzz
    replied
    Originally posted by milkylainen View Post

    Claiming that Nvidia wasted a year in building a high throughput compute core is just a severe lack of insight.
    The new dies are pretty big, especially the 2080 Ti. The costs for the large dies are getting more and more expensive as the the production processes get smaller. It's reflected in the costs of the cards. They are really only getting away with it because some people think, due to brainwashing, there is no other option.

    I'm surprised we aren't starting to see chiplets in GPUs on AMDs end, but who knows what's coming.

    Leave a comment:


  • fuzz
    replied
    Originally posted by TemplarGR View Post
    I am waiting for Navi myself. Been an AMD certified fanboi for more than a decade, and now that AMD supports open source, there is no alternative. LOL.
    True, i think Intel will be more of a competitor in the open source world once they launch their dedicated GPUs.

    Leave a comment:


  • pcxmac
    replied
    as someone who compiles all the software I use (Gentoo), and someone who appreciates Open Source and the GPL, I have to say I don't think I will ever oblige NVIDIA on the Linux platform unless they stop being like they are and offer competitive open source alternatives. There are plenty of reasons to want to have an open source graphics driver, even if it has issues keeping up with the windows proprietary variant.

    That said, I fully expect AMD to have a cost effective graphics chip that is very competitive with the 2080TI in terms of compute/graphics performance (not ray tracing) early next year, just by going to TSMC the process node will be superior to GloFo's and with the die shrink/new process, I expect big things until Nvidia upsets with it's new offering, probably next year too. If I were most people, I would ignore the 2080 graphics cards, they are new, and probably will be seriously outdated when 7nm and AMD come around, next year.

    Leave a comment:


  • vein
    replied
    Originally posted by torsionbar28 View Post
    ^ This. Here in 2018, there is no reason for any self-respecting Linux geek to consider Nvidia.
    I agree completely. My fury card was my best buy ever...not a single problem, everything I use it for just works.

    I'll never go back to Nvidia and their proprietary problems.

    Leave a comment:


  • vein
    replied
    Originally posted by vegabook View Post

    Vega 64 has almost the same raw 32-bit TFLOPS as the 2080ti, and is much cheaper.

    Actually, I'm wondering if this launch, considering the pricing, isn't actually good for AMD. Some Vega 64s are running at around 500 USD on Amazon right now which is a steal. Not to mention 580s which are at 250 area now.
    Now I really really want AMD to release Vega 20 with 20 TFLOPS

    Leave a comment:

Working...
X