• The_Zura@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    The 1050 Ti has not aged well, so a low bar to clear. If one wishes to play older games and not get a neutered experience, the 1050 Ti will probably crush it, like in GTA5.

    He mentions desktop pcs as being different, but ignore the comparison to laptops. Those two have roughly the same volume, with the laptop coming with a screen, keyboard, portability. There are $800 4060 laptops that will run circles around the sub 1650 performance. Needing this much cpu performance and poor graphics is a niche within a niche. I find the fascination with these mini-pcs to be odd. They remind me of the tunnel vision that reviewers seem to have with handhelds where they don’t look broadly at the market. Integrated graphics are always going to be turds. Sub $300, maybe it has a place. $630 is a hell no.

  • prophetmuhammad@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    that’s just insane. even more so considering i went from a GTX 1050 (given to me for free in 2019) to an RTX 2060 Super just several months ago. so this integrated graphics is faster than my last video card.

    • Healthy_BrAd6254@alien.topB
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      EXPO RAM is not faster than what you get with these mobile chips. The Z1 Extreme in the Legion Go for example uses 7500MHz LPDDR5X. Meanwhile desktop Ryzen CPUs struggle to do 6200-6400MHz.

    • uzzi38@alien.topB
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      Probably because it hits a higher frame rate.

      The higher you go, the more CPU you need, and the more CPU you need, the less power and memory bandwidth you have for the iGPU.

  • mateoboudoir@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    It’s always neat to see occasional IGP vs. dGPU comparisons like this, if for nothing else then as a sort of barometer of IGP progress.

    As always, the numbers indicate that the major Achilles’ heel of IGPs remains memory bandwidth. The 780M even seems like it could perform a tier better than implied here - in the ballpark of the legendary GTX 1060/RX 480 GPUs - if only it could fetch more data more quickly. The revelation isn’t exactly surprising, as there already exists basically a 780M in dGPU form in the RX 6400 (also 12 CUs, just paired with GDDR6 memory) that performs right in that range.

    • stilljustacatinacage@alien.topB
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      Hmm. This is the first time it’s occurred to me, but I wonder what sorts of performance we might see from iGPUs once Direct Storage is more mature. I know pulling texture data from SSD could never be as fast as from RAM - but I wonder if reducing SSD -> RAM -> GPU to just SSD -> GPU will have meaningful savings.

  • dracolnyte@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    theres still more headroom since this is TDP limited to 45W right? what if a full desktop APU with 65W, would it perform better?

    • Healthy_BrAd6254@alien.topB
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      Doesn’t make a difference. The issue is memory bandwidth, which is only around 100GB/s on DDR5. And it’s shared with the CPU, which makes it effectively even worse.

  • AwesomeFly96@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Back in the day the 1050 Ti was a great bang for the buck GPU. This shows how times are changing.

  • Fuck_the_fascists@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    I had a gtx1050 mobile up to 6 months ago. The integrated graphics of my ryzen 7 5800H does better in 3D capabilities, while using much less power. It’s only worst for VRAM (only has 512MB ddr3), so the 780M doing much better isn’t surprising

  • DuhAlGhul-@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Rx 6400 is the theoretical maximum performance to come out of a 780m APU (Both have 12 rdna 2.0 cores), right?

    Speculating on a hyper overclocked 8000g isn’t that great if you think like this.

  • colbyshores@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    What is even more mind blowing is that the 10 Terraflop architecture in the PS5 slim is an APU. AMD could go so much further with APU gpu performance if they really wanted to.

      • Healthy_BrAd6254@alien.topB
        link
        fedilink
        English
        arrow-up
        1
        ·
        1 year ago

        8 TFLOPS on RDNA 3 is comparable to 4 TFLOPS on RDNA 2 (PS5).
        The PS5 uses 36 compute units. The 780M has 12.
        The PS5 has 450GB/s memory bandwidth. The 780M has around 100.

        The 780M runs at significantly lower power and shares the memory bandwidth with the CPU, but even ignoring that it’s like 3-4x worse than a PS5.

        • Pillokun@alien.topB
          link
          fedilink
          English
          arrow-up
          1
          ·
          1 year ago

          it does not work like that… 8tflop rdna3 is becaue they count the fpu perf like it is done on ampere/ada ie both the int and fpu alu can to fp while older u arch had fixed alu for int and fp.

          And there is a problem with the rdna3 where this function is not working correctly so we only get the additional 30% fpu perf from the additional fp alu units.

          and like u sait the memory is holding back the apus but it will still not perform like an 8Tflop gpu.

    • Pillokun@alien.topB
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      does not really matter if they dont do it. u can look at a fiat or an ferrari, both are pretty much owned by the same company yet they are vastly different products. same here, does not matter if the apu in the consoles(ferrari) is so much faster if the desktop apu is like a fiat.

      • colbyshores@alien.topB
        link
        fedilink
        English
        arrow-up
        1
        ·
        1 year ago

        Most likely they just don’t want to cannibalize their mid tier gpu sales since there wouldn’t be a need for one.