Rumor: new desktop and mobile GeForce RTX chips break cover

It's a pretty quiet news day, but over yonder in the Plains of Rumors there have been echoes of the words "Nvidia" and "RTX." Two rumors in particular made our ears prick up, so we figure you'd like a look too. Please, do keep the salt shaker handy.

The first rumor we've spotted is probably the juiciest one of the two. It appears that laptops packing Nvidia RTX graphics chips might be peeking just around the corner. Twitter leaker @TUM_APISAK was browsing benchmark databases and spotted 3DMark scores for what looks like a laptop fitted with an RTX 2060 graphics card. That RTX chip apparently goes along with 6 GB of RAM and shows a core clock of 960 MHz. The memory speed is apparently set at 1750 MHz.

The tweet above actually contains information about a second machine packing a chip whose graphics driver specifically notes "RTX 2060 with Max-Q design." Interestingly enough, the core clock for this variation is 975 MHz (so 15 MHz over the regular chip), while the memory bus clock tops out at 1500 MHz. TUM_APISAK helpfully points out that the 3DMark scores obtained by the RTX 2060 laptop put it somewhere between the laptop GTX 1070 and GTX 1070 Max-Q. A single benchmark doesn't mean much, but given how desktop RTX cards have shaken out relative to their 10-series predecessors, that level of performance wouldn't be a surprise.

Additionally, TUM_APISAK also spotted other laptops with mobile RTXs in them. The Geekbench database contains results for what looks like a Lenovo unit packing an RTX 2070 Max-Q, another one with an RTX 2080 Max-Q, a nondescript machine with an RTX 2080, and even an Asus ROG Zephyrus M notebook with a purported mobile RTX 2080 Ti in it. The latter machine in particular would offer a rather impressive amount of horsepower in a mobile form factor, though we shudder at imagining what the battery life would be like.

By now, the sight of the numeral "2060" may have you wondering if you missed any bit of recent news. You're good, as Nvidia has yet to officially announce an RTX 2060 graphics card at all. However, and somewhat predictably if the rumors above hold any truth, desktop RTX 2060s may be incoming soon. Videocardz has posted what it says are pictures of a Gigabyte RTX 2060 OC desktop card.

Source: VideoCardz

According to the site, the chip underneath the heatsink has 6 GB of GDDR6 (the same amount as the purported laptops above), and there are 1920 CUDA cores ready to handle rendering work. The rumor mongers also remark that the card in question is overclocked from the factory, but there are no Megahertz measurements to be had at this point. As usual, there's no telling whether any of these rumors will pan out, but at least they all look reasonable enough.

Comments closed
    • Freon
    • 12 months ago

    I feel like a remake involving transistors of Monty Python’s “Every Sperm is Sacred” is in order as a love letter to all the transistors being spilled this generation on ray tracing that will never be used.

    • Neutronbeam
    • 12 months ago

    I didn’t believe ANY of this until I saw the word “purported.” As they say in the TR forums: “If it’s purported, it’s got to be true!”

    • Thresher
    • 12 months ago

    If these don’t price out at the same price as a 1060, I don’t really see a point in it. Just get one with a 1070 Max-Q if it’s more expensive. Most gaming laptops sold are 1920×1080 and just aren’t going to benefit THAT much from the speed bump. As for ray tracing, the desktop 2070 already struggles with that. A 2060 Max Q would probably make it pointless.

      • Chrispy_
      • 12 months ago

      Honestly, RTX is expensive (in monetary, thermal, battery, and performance terms) so even with the latest RTX performance patch, the concept of RTX features in a laptop is a little bit silly.

      The desktop 2070 with RTX enables still performs worse than a $149 RX 570, and it does so whilst guzzling down 185W.
      I’m sorry, but that is not laptop-friendly.

        • Srsly_Bro
        • 12 months ago

        And the time you waste staring at the two reflections per map so you feel like you spent your money well.

          • Chrispy_
          • 12 months ago

          Heh, yeah.

          DXR features don’t really have much of an opportunity to distinguish themselves over the traditional rendering methods in outdoor games with a largely non-reflective environment.

          I imagine that as DXR becomes more widely-adopted, some games (sci-fi games with that glossy sheen to everything) may really make the differences stand out more obviously.

    • BIF
    • 12 months ago

    Wait a minute. You “shudder” at battery life?

    One doesn’t buy a computer with these and expect to go off the grid for a week without bringing along the power adapter. That’s the only reason to put these in a laptop, to allow guys like me to render from a hotel room if I so choose, not the back woods or the Sahara.

    The battery serves as a nice UPS to help me from losing my work if the hotel power goes out, but that’s about all.

      • derFunkenstein
      • 12 months ago

      Agreed. The battery is not the point of a gaming notebook.

      • auxy
      • 12 months ago

      It’s just a turn of phrase…

        • Srsly_Bro
        • 12 months ago

        And also used in accordance with his personal values regarding battery life. BIF started off strong with a red herring that could clear a forest.

        A reasonable person wouldn’t expect to go off the grid for a week with any laptop.

        Then the argument is finished nicely with “my values are x and this is why you are wrong.”

        Sometimes you visit clients, and you only need the battery to last longer than your demo/meeting. It’s not always practical to have a charger, and it’s nice to have some battery life.

        Don’t be the “There is one way to use a computer. Change my mind.”

        Your bro, srsly.

          • derFunkenstein
          • 12 months ago

          I think [url=https://www.youtube.com/watch?v=RcsFpztaJIc<]Sammy said it best[/url<]. (btw the tone of that lead guitar is impossibly fat)

            • Srsly_Bro
            • 12 months ago

            LMAO. David lee Roth > Sammy

            I was trying to see if people caught my reference.

            • derFunkenstein
            • 12 months ago

            No idea what your reference was. But part of your math is wrong. That’s Sammy solo and Sammy solo > Dave solo.

            Oh wait. Your reference is the guy sitting at a table picture meme. I like that one, it still has legs.

            • Srsly_Bro
            • 12 months ago

            I was makinga MP reference with the herring usage. It didn’t work as planned.

            And what you said but that was secondary.

            • derFunkenstein
            • 12 months ago

            ohhhh, clear a forest. Got it. That’s good, and it was subtle.

          • DavidC1
          • 12 months ago

          And how many people meeting their clients are going to use a gaming grade laptop to do so? You’d likely have an ultrabook with Intel iGPU or at best, MX150.

          If the laptop has a 2080 Ti then that’s going to be a brick that you’ll have as a separate gaming device at home, won’t you?

            • Srsly_Bro
            • 12 months ago

            He made the claim using it to render with the included GPU. Gaming grade or not, he made the claim.

    • tipoo
    • 12 months ago

    At the point where we get down to RTX 2060 with max-q, I have to wonder how useful the RTX bits will be given the performance loss on high end chips, and what could have been if they aimed at pure raster performance with these except on flagship parts. Then again that would stymie adaptation.

      • MOSFET
      • 12 months ago

      No idea why the DVs, I completely agree with you. It’s a thought that occurred to me once before as well.

      • Thresher
      • 12 months ago

      I agree.

      Ray-tracing will be pointless on a card with less performance than a 2070.

        • Leader952
        • 12 months ago

        Lower cards are for lower resolution displays not everyone expects that the RTX 1060 will drive 4K displays at max settings.

          • Spunjji
          • 12 months ago

          Lower than 1080p? It would have to be for RTX to make sense on anything less than a 2070.

      • Leader952
      • 12 months ago

      You must be behind in the news:

      [url<]https://forum.beyond3d.com/threads/ray-tracing-meaningful-performance-metrics-and-alternatives-spawn.60873/page-28[/url<] Battlefield 5 Massive Performance Improvement [url<]https://youtu.be/QKV8VdhZuW4[/url<]

        • sweatshopking
        • 12 months ago

        Still WAYYYY to slow

        • tipoo
        • 12 months ago

        I was not, and a 50% smaller loss is still a substantial loss. Even looking at the new numbers with the setting on “low”, it would appear to be an unappealing feature for a large step down from the 2070, which the 2060 max-q is.

        [url<]https://www.extremetech.com/wp-content/uploads/2018/12/2070-1080p.png[/url<] [url<]https://www.extremetech.com/gaming/281996-new-dxr-patch-substantially-improves-nvidia-rtx-ray-tracing-performance-in-battlefield-v[/url<]

    • Star Brood
    • 12 months ago

    Yawn. Wake me up when there’s an announcement that they’ll be priced low. I’ve seen way too many extremely pricey NVidia and Radeons to hold out much hope.

      • Spunjji
      • 12 months ago

      Yup. Been tired with laptop GPU pricing since the uncontested GTX 900 series *really* knocked it out of the park. High-end parts priced at the same performance per dollar as the outgoing generation… not cool.

    • Krogoth
    • 12 months ago

    Bloody double posting due to networking issues/packet loss. 😛

      • chuckula
      • 12 months ago

      I’m Krogothed with Krogoth’s ISP.

    • Krogoth
    • 12 months ago

    Hopefully it is binned TU106 silicon not rebranded GP104/GP106 silicon. It is too tempting for Nvidia not to rebrand old stock as “new”.

      • techguy
      • 12 months ago

      If the product is branded RTX then it surely will be Turing.

        • Krogoth
        • 12 months ago

        That’s assuming Nvidia doesn’t pull a “Geforce 4 MX” from its infamous marketing playbook.

        ATI did the same thing with Radeon 9xxx-92xx which were rebranded R2xx chips not R3xx chips.

          • auxy
          • 12 months ago

          Don’t worry Krogoth. You might get downvoted for being Krogoth (much as I get downvoted for being auxy) but I know you’re right. ( ;∀;)

      • psuedonymous
      • 12 months ago

      Could be a ‘half TU104’ in a similar way that the TU106 is conceptually half a TU102. Either heavily binned or dedicated silicon.

      I’d have normally dismissed the dedicated silicon option, but the Turing lineup is already utterly bizarre in launching with [b<]three[/b<] discrete (and pretty damn massive) dies right out the gate, so maybe Nvidia have lucked the hell out with 12nm FFN - well, that and having beaten the process into submission with Pascal and Volta, 16FFN and 12FFN are damn near the same proces - and get such ludicrously good yields they they can rely on size-to-chip rather than needing to bin down.

Pin It on Pinterest

Share This