Nvidia's latest DLSS revision reduces VRAM usage by 20% for upscaling — optimizations reduce overhead of more powerful transformer model

DLSS versions compared
(Image credit: Nvidia)

Besides moving DLSS 4 out of beta, Nvidia has also optimized VRAM usage in its latest DLSS SDK release. Initially discovered by VideoCardz, DLSS 310.3.0 improves the Transformer model VRAM usage by 20%.

DLSS's updated memory footprint brings its overall memory impact closer to that of the older CNN model. Previously, the DLSS transformer model consumed nearly twice as much memory as the CNN model it replaces. Still, with Nvidia's latest VRAM optimizations, the transformer model now consumes just 40% more memory than the CNN model.

At 1080p, the old transformer model consumed 106.9MB of VRAM; the new one consumes just 85.77MB. For comparison, the CNN model consumes 60.83MB of VRAM. DLSS's memory consumption increases or decreases linearly based on resolution; the larger the resolution, the more VRAM DLSS requires. For example, running DLSS at 4K increases the transformer model's memory footprint by 3.5X, consuming 307.37MB of VRAM.

Swipe to scroll horizontally
Header Cell - Column 0

1080p

1440p

4k

8k

New Transformer Model

85.77 MB

143.54 MB

307.37 MB

1,225.17 MB (1.2GB

Old Transformer Model

106.9 MB

181.11 MB

387.21 MB

1,517.60 MB (1.5GB)

CNN Model

60.83 MB

97.79 MB

199.65 MB

778.3 MB

Despite the advantages, you probably won't notice the VRAM improvements in real-world applications. A 20% reduction in memory capacity for DLSS only translates to an 80MB reduction at 4k (not to mention 1440p or 1080p). 80MB (or less) is almost negligible compared to the VRAM capacity of modern graphics cards. The only exception might be at 8k resolution, where the DLSS transformer model consumes upwards of a gigabyte of memory (even with Nvidia's new VRAM optimizations).

Regardless, a 20% reduction in memory footprint is impressive considering the capabilities of the DLSS transformer model. If history repeats itself, we can expect DLSS 4's DLL file to get trimmed down in the future. Nvidia performed the same technique with DLSS 3.8.10. DLSS 4's transformer model boasts twice the number of parameters as its CNN-based predecessor, allowing the upscaler to achieve native-like image quality.

Nvidia's VRAM improvements apply strictly to its transformer upscaling model, and not frame generation. DLSS 4 launched with substantial frame generation improvements, reducing VRAM consumption by 30%. Frame generation consumes significantly more VRAM than the upscaling portion of DLSS. For instance, Nvidia claimed in one of its blog posts that Warhammer 40,000: Darktide uses 400MB less VRAM at 4k with DLSS 4 frame generation compared to DLSS 3 frame generation.

Follow Tom's Hardware on Google News to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button.

Aaron Klotz
Contributing Writer

Aaron Klotz is a contributing writer for Tom’s Hardware, covering news related to computer hardware such as CPUs, and graphics cards.

  • palladin9479
    Slowly chiping away at the ungodly VRAM requirements for DLSS / MFG.
    Reply
  • fiyz
    Eh? I'm still adamant about giving us realistic audio. Visual effects is how the wizard controlled oz, but the audio effects gave him away.
    Reply
  • usertests
    fiyz said:
    Eh? I'm still adamant about giving us realistic audio. Visual effects is how the wizard controlled oz, but the audio effects gave him away.
    Audio has long been a pain point in game development. It's harder to market than fancy graphics. We see things like AMD TrueAudio, or Tempest in the PS5, but we don't hear about them. ;)

    Ideally, sounds could be simulated in real time (e.g. below 5 milliseconds). For example, if you throw a metal wrench at a granite surface at 2 m/s, at a certain angle, it starts clattering, and the sounds are accurately reflected or absorbed off nearby walls, objects, the hallway around the corner, etc. Sounds difficult, maybe it isn't anymore but I don't know.

    The only games that can easily justify spending a lot of attention to sound design are stealth games. Games like Thief come to mind, but RPGs like The Elder Scrolls also use stealth. VR games are also likely to benefit from realistic audio, but it's still a small market.
    Reply
  • hotaru251
    give me the stats on their 60 tier base model.

    dlss/MFG use more vram than normal...so can the base model of their "entry" gpu even play games well w/ it given their gimped vram/bandwidth in AAA 2025 games?
    Reply
  • TheyStoppedit
    It's easy to see what's happening in the R&D department over at NVidia. See, the problem here is that.... AI GPUs require tons and tons of VRAM. VRAM is finite, so they are trying to put as little VRAM as possible on GeForce cards so there is more for AI cards, because that's where all the money is. So they are trying to find a way to make it so games can use less VRAM. So..... fake frames it is!!!.... You're gonna see more and more different variants and addons to DLSS/MFG. It will get to the point where GPUs will just stop increasing VRAM and do very little rendering, and it will all be AI/MFG/DLSS based and optimized to be VRAM efficient. What this means is that gen over gen uplifts in VRAM will be very little.

    By the time you get to the RTX 9000 series, VRAM won't be a whole lot better than the 50 series. VRAM Capacities on RTX 9000 cards in 6 years from now:

    9090 = 48GB
    9080 = 32GB
    9070Ti = 24GB
    9070 = 20GB
    9060Ti = 16GB
    9060 = 16GB
    9050 = 12GB

    You read it here first. That's 4 gens ahead of what we have now. It looks awful for a 4-gen jump, but that's the way it will be. Why? Because VRAM will become scarce as it all gets absorbed by the AI industry, and DLSS/MFG will be designed to be optimized to use very little to no VRAM for games. This means less VRAM on GeForce cards for more VRAM on AI cards. All your frames will be fake frames by this point. Why? Because AI is taking over and that's where the VRAM is needed. Let's jump ahead even further........

    15090 = 64GB
    15080 = 48GB
    15070Ti = 40GB
    15070 = 32GB
    15060Ti = 24GB
    15060 = 24GB
    15050 = 16GB

    That's 20 years from now. Some of us reading this will probably be dead by then. Look at those numbers, those are close to 2x what we have now. We will have MFGx120 & DLSS 14.0. Before you call me crazy with these predictions, just remember, the 2060 comes in a 12GB variant and the 5060Ti we have now comes in a 8GB variant. That means in some VRAM-heavy games, the 3 gen old GPU thats 1 tier lower will perform better.

    And then when customers cry about the 15090 that they paid $10K for with so little VRAM, NVidia will just come out and say "64GB is enough for todays games, we dont need anymore than that" ..... that's because we aren't actually rendering/rasterizing the games anymore. Its all fake frames and AI.

    It's all because of AI. AI is where the money is and that's where the VRAM is needed. Those values I gave for the 15XXX series, with true, full native rendering should be 4x those values in 20 years. But they won't be. Real rendering/rasterization will be history and it will all be MFG/DLSS/AI based
    Reply
  • bolweval
    TheyStoppedit said:
    It's easy to see what's happening in the R&D department over at NVidia. See, the problem here is that.... AI GPUs require tons and tons of VRAM. VRAM is finite, so they are trying to put as little VRAM as possible on GeForce cards so there is more for AI cards, because that's where all the money is. So they are trying to find a way to make it so games can use less VRAM. So..... fake frames it is!!!.... You're gonna see more and more different variants and addons to DLSS/MFG. It will get to the point where GPUs will just stop increasing VRAM and do very little rendering, and it will all be AI/MFG/DLSS based and optimized to be VRAM efficient. What this means is that gen over gen uplifts in VRAM will be very little.

    By the time you get to the RTX 9000 series, VRAM won't be a whole lot better than the 50 series. VRAM Capacities on RTX 9000 cards in 6 years from now:

    9090 = 48GB
    9080 = 32GB
    9070Ti = 24GB
    9070 = 20GB
    9060Ti = 16GB
    9060 = 16GB
    9050 = 12GB

    You read it here first. That's 4 gens ahead of what we have now. It looks awful for a 4-gen jump, but that's the way it will be. Why? Because VRAM will become scarce as it all gets absorbed by the AI industry, and DLSS/MFG will be designed to be optimized to use very little to no VRAM for games. This means less VRAM on GeForce cards for more VRAM on AI cards. All your frames will be fake frames by this point. Why? Because AI is taking over and that's where the VRAM is needed. Let's jump ahead even further........

    15090 = 64GB
    15080 = 48GB
    15070Ti = 40GB
    15070 = 32GB
    15060Ti = 24GB
    15060 = 24GB
    15050 = 16GB

    That's 20 years from now. Some of us reading this will probably be dead by then. Look at those numbers, those are close to 2x what we have now. We will have MFGx120 & DLSS 14.0. Before you call me crazy with these predictions, just remember, the 2060 comes in a 12GB variant and the 5060Ti we have now comes in a 8GB variant. That means in some VRAM-heavy games, the 3 gen old GPU thats 1 tier lower will perform better.

    And then when customers cry about the 15090 that they paid $10K for with so little VRAM, NVidia will just come out and say "64GB is enough for todays games, we dont need anymore than that" ..... that's because we aren't actually rendering/rasterizing the games anymore. Its all fake frames and AI.

    It's all because of AI. AI is where the money is and that's where the VRAM is needed. Those values I gave for the 15XXX series, with true, full native rendering should be 4x those values in 20 years. But they won't be. Real rendering/rasterization will be history and it will all be MFG/DLSS/AI based
    Conjecture!
    Reply
  • BTM18
    Still a hard pass on fake frames and snake oil from these clowns.
    Stuff your fake frames up your green a**.
    Reply
  • NightKnight1337
    palladin9479 said:
    Slowly chiping away at the ungodly VRAM requirements for DLSS / MFG.
    "running DLSS at 4K increases the transformer model's memory footprint by 3.5X, consuming 307.37MB of VRAM."

    300mb is ungodly? wtf
    Reply
  • palladin9479
    NightKnight1337 said:
    "running DLSS at 4K increases the transformer model's memory footprint by 3.5X, consuming 307.37MB of VRAM."

    300mb is ungodly? wtf

    That's in addition to anything else that is running and only for the upscaler. MFG consumes about 1.5GB of VRAM currently.
    Reply
  • Amdlova
    Vram for consumer it's just a dream... why the nvidia will looser profit over quadro cards.
    If you want vram pay for it.


    :)
    Reply