Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×
Graphics Games Hardware

NVIDIA's Job Listings Reveal 'Game Remastering' Studio, New Interest In RISC-V (forbes.com) 40

An anonymous reader quotes Forbes: Nvidia has a lot riding on the success of its GeForce RTX cards. The Santa Clara, California company is beating the real-time ray tracing drum loudly, adamant on being known as a champion of the technology before AMD steals some of its thunder next year with the PlayStation 5 and its own inevitable release of ray-tracing enabled PC graphics cards.

Nvidia has shown that, with ray tracing, it can breathe new life into a decades-old PC shooter like id Software's Quake 2, so why not dedicate an entire game studio to remastering timeless PC classics? A new job listing spotted by DSOGaming confirms that's exactly what Nvidia is cooking up.

The ad says NVIDIA's new game remastering program is "cherry-picking some of the greatest titles from the past decades and bringing them into the ray tracing age, giving them state-of-the-art visuals while keeping the gameplay that made them great." (And it adds that the initiative is "starting with a title that you know and love but we can't talk about here!")

Meanwhile, a China-based industry watcher on Medium reports that "six RISC-V positions have been advertised by NVIDIA, based in Shanghai and pertaining to architecture, design, and verification."
This discussion has been archived. No new comments can be posted.

NVIDIA's Job Listings Reveal 'Game Remastering' Studio, New Interest In RISC-V

Comments Filter:
  • by ShanghaiBill ( 739463 ) on Sunday October 13, 2019 @06:06PM (#59304086)

    RISC-V is the future. China has learned from the Huawei debacle and trade sanctions that they can't rely on America for critical technology. They are putting big resources into RISC-V as part of Made in China 2025 [wikipedia.org].

    • That's a tall order. The best China has done so far with RISC-V is the XuanTie 910, which is intended for IOT. It was developed by Alibaba, which is not even a hardware company and it has not intentions of actually producing it. merely licensing the design, probably to some Taiwanese or South Korean company with the capacity to actually mass produce it
      • There are only three companies in the world capable of fabbing top end ICs: Intel, TSMC, and Samsung. For now, the RISC-Vs will be fabbed in Taiwan by TSMC. But a lot can change in six years. Mainland fabs are improving rapidly.

        • They don't even have to be top-end, if you can throw enough of them at the problem. Remember all that stuff we did back in the day with single-digit MHz computers?

          • Remember all that stuff we did back in the day with single-digit MHz computers?

            You mean programming in assembly, since compilers were inadequate?

            • Or C with inlined ASM, since usually you spend most of the time in a small portion of the code. And back then we had architectures you'd want to write assembly for. Well, some of us.

    • by AHuxley ( 892839 )
      1+ for this. China will never allow its advancement to be held back by other nations again.
      They are going full RISC factory as a Communist nation building project.
    • RISC-V is the past. The core ISA is the MIPS ISA from about 1987. It's practically useless unless you go with M, A, D, and B extensions, and that's to get to a MIPS ISA from about 1990. Even then, assuming you invest the vast amount of manpower and capital to turn it into a running CPU, you've now got maybe an R4000 from close to 30 years ago.

      I just don't understand the fuss over RISC-V. There's simply no value to it, no matter what axis you look at it on. Obviously nVidia agrees, since they've assigne

      • I just don't understand the fuss over RISC-V. There's simply no value to it, no matter what axis you look at it on.

        Patents and licensing. RISC-V presents an option for domestic fabrication for a country without having to muck with international IP. Heck, you don't even need domestic production to still reap the benefits, but in the case of China it makes sense because they willfully violate all kinds of other people's IP.

        they already have a decade of experience building Loongson's, which are also MIPS ISA, why would they ditch that for a different MIPS-style ISA?

        Because the fabrication is done by STMicroelectronics which is based in Geneva, Switzerland. If the US wanted a big enough push, they can indicate to them that they can continue fabrication and trade

      • I'm feeling the same. I wanted some ideas for a hobby 8-bit CPU I'm developing, so I decided to check out the RISC-V specification. Almost every section talks about some advanced features that were considered, but rejected because they were "too complex". All in all, I don't find the architecture to be very interesting, and other than a few ideas regarding compressed instructions, wasn't very useful to study.

        The feeling I get is that RISC-V is squarely designed to only meet the needs of the embedded mark

        • Almost every section talks about some advanced features that were considered, but rejected because they were "too complex". All in all, I don't find the architecture to be very interesting ...

          You're missing the key point of RISC. "Advanced features" and complex instructions look good on paper, but end not being used by compiled code. Research over the last 30+ years has confirmed this time and time again. You think RISC-V is boring, and you're not wrong. Their goal was to be useful, not exiting.

          If you want to see an interesting architecture, check out The Mill [millcomputing.com]. Unfortunately, they're five years away from having real chips, and always will be :-(. Same with compilers that can fully exploit

      • RISC-V is organized into layered standards which makes it a bit easier for ASIC designers than MIPS ever was. And you can drive RISC-V to a lower end spec than MIPS, which is useful if you're prioritizing area for an embedded microcontroller. The modularity parallels a lot of the advantages found in ARC (Synopsys), so I'd draw the comparisons there rather than to MIPS.

        Obviously nVidia agrees, since they've assigned the bare minimum number of people you possibly can to it. Once they have a 200-person business unit and design team working on it, you'll know they're serious about it.

        There is a nice presentation on why NVIDIA is using RISC-V [riscv.org]. And it's landed in some of their GPUs, which means a significant number of people

    • So where are the RISC-V development boards?

      • Maybe try here. [sifive.com]

        • Their fancy site takes you to this crowdfunding page. Still 100% vaporware. https://www.crowdsupply.com/si... [crowdsupply.com]

          • You'll notice that all the fundraising is set to $1. That's because they're using the fundraising site as a storefront as opposed to actually raising funds. Now if that sounds sketchy to you, well that's not what you asked. You asked who is selling boards. These people are and there are people who've written reviews on these boards. Yes, they're using a crowdfunding site to avoid having to set up an actual storefront and that seems highly suspect, but again you just strictly asked who was selling boards.
            • But they still aren't selling anything. How do I get one? They don't even show a picture of a real product just a rendering and supposedly real reviews. Vaporware.

      • I've been using an FPGA as a development kit. (Terasic DE0-Nano, but just about anything will work)

        I suspect most implementations of R5 will be by big ASIC vendors who embed them, and we'll probably not see too many developer-friendly dev kits.

  • by DontBeAMoran ( 4843879 ) on Sunday October 13, 2019 @06:30PM (#59304110)

    Leisure Suit Larry with real-time ray tracing! Fuck yeah!

  • I wanna see dynamic realtime lighting on both pixels.

  • It seems that since the early 2000s, all we get is sequels, prequels, reimagineds, or even blatant plain remakes.

    And all "new" games are just rehashs of the same old same old.
    Like the bazillionth "open world mmo military shooter". Entire "genres" consisting of nothing but rehashs of the first game like that from ages ago.

    How about you come up with something NEW for a change? Like "risking" *actual* innovation!

    • by e3m4n ( 947977 )

      Theres a few problems that drove this.

      1) people were only buying FPS like quake2. Games that older generations like myself played, like RTS, adventure games, and turned based strategy were not keeping the lights on.

      2) quake3 arena and unreal tournament opened the door to PvP in a way that showed other players was a better substitute for machine NPC. Only RTS like starcraft yielded much benefit from the same idea.

      3) ever since the first forum flame war, people have been addicted to being some of the biggest

      • 3) ever since the first forum flame war, people have been addicted to being some of the biggest assholes they can possibly be without literally getting their face crushed in for doing/saying the shit they do. There are few repercussions except for the occasional swatting that makes the news. Free license to be so inhuman Adolf Hitler himself would receive s Nobel Peace prize if it were a choice between the two. MMOs give an endless venue to ruin peoples day and some people just cant get off any other way.

        Ahh Usenet, I remember thee well.

    • It seems that since the early 2000s, all we get is sequels, prequels, reimagineds, or even blatant plain remakes.

      Why is it every time I see an incredibly stupid comment it's posted by you? Seriously just because you don't seem to bother and investigate new IP doesn't mean it doesn't exist.

  • by fintux ( 798480 ) on Monday October 14, 2019 @04:02AM (#59304828)

    Nvidia is acting as if their raytracing is all the rage in the gaming. However, there's only a handful of new games that support it, the benefits are very small and the performance hit is substantial. Until there's some better performing technology and a wide enough adaptation so that game makers don't need to make the same effects twice (with and without raytracing), I'm not really that convinced.

    I think the reason why Nvidia is pushing this on the old games is that they are so lightweight that even having the burden of raytracing does not slow them down too much. It still almost certainly isn't the easiest or most efficient way of improving the visuals of old games, but of course it suits Nvidias purposes of portraying raytracing as the miraculous solution to improving game graphics. "See how this game looked 20 years ago? Now with raytracing, it looks this great (small print: also we have remastered the whole game graphically and adding raytracing was just a tiny portion of what was done)."

    • Sadly raytracing DOES slow down those old games. Look at the RT port of Quake 2. Pathetic.

      • Pathetic

        What's pathetic about applying a hugely computationally intensive process that likely requires changes in hardware in order to make things look good?
        Should we go back to playing the original Doom, or Quake since software 3D is a "pathetic" performance impact just because you don't own a 3D accelerator?

        Surprise surprise, applying new technology to old games causes the same performance hit as that technology has on any other game.

        • I remember looking at two computers Quake side-by-side: one doing software rendering and the other running a 3Dfx. The difference was night and day, just like someone looking at QuakeIIRTX on a GTX vs RTX.

          Don't compare looking at screenshots (software/3Dfx or GTX/RTX), you need to watch videos to really grasp how much of a difference it makes in both cases.

          • And? Running on a Voodoo1 actually sped up the game and improved image quality at the same time versus software rendering. What is your point?

        • Quake 2 requires so little computational power that a 2080Ti should be able to run it at less than 1% utilization. Applying RT to the game should be doable at at least 144 fps easily. Instead, 60 fps or so? Maybe 70?

          Seriously?

          No, applying "any new technology" to old games will not cause such a disaster. If I apply recent image sharpening tech (like Radeon Image Sharpening) to Quake 2, it does not lower fps by such an absurd amount. Ditto for modern AA techniques and other things. It's only RT.

    • So they can explain just how amazing Zork will be on the new processors :-)

    • Nvidia is acting as if their raytracing is all the rage in the gaming. However, there's only a handful of new games that support it

      The same can be said for any introduction of a new graphical concept ever. PhysX was the same now physics is the norm, hell 3D acceleration itself also followed this arc. It was many years after the introduction of the 3D accelerator that we normalised it to the point where software rendering was depreciated.

      I'm not sure what you expect from any new technology that is supposed to show off hardware that is less than a year old. I mean it already has several AAA titles implementing the features.

      Raytracing *IS

      • by fintux ( 798480 )

        Nvidia is acting as if their raytracing is all the rage in the gaming. However, there's only a handful of new games that support it

        The same can be said for any introduction of a new graphical concept ever. PhysX was the same now physics is the norm, hell 3D acceleration itself also followed this arc.

        PhysX was done so that it either used NVidia GPU, or a very dummy implementation on the CPU. I think the current implementation is usually used on the CPU after it has been optimized. It was another example of NVidia GameWorks kind of effort to make NVidia look better simply by focusing on making others look worse.

        It was many years after the introduction of the 3D accelerator that we normalised it to the point where software rendering was depreciated.

        But it wasn't due to the performance hit of 3D acceleration finally getting to an acceptable level - 3D acceleration improved the performance, and also graphical fidelity. Which can't be said abou

  • by e3m4n ( 947977 ) on Monday October 14, 2019 @07:59AM (#59305134)

    Iâ(TM)m sure Leisure Suit Larry will look pretty amazing but Infocom games work pretty well even on a Busybox embedded platform. Iâ(TM)m deducting 4 mod points if any of you had to look any of those titles up LOL

  • And Sony's plans to remaster by emulation their entire back catalog.

    https://gearnuke.com/sony-plan... [gearnuke.com]

  • And the Ghost Bear addon. Compelling story, good game play, just needs better graphics.

  • The problem is that it's too risky for a large studio to target technology that doesn't have a large market base, but do quake3 nvidia, it will make the internet happy

What is research but a blind date with knowledge? -- Will Harvey

Working...