Results 1 to 11 of 11

Thread: 3DMark adds DLSS test to Port Royal benchmark

  1. #1
    HEXUS.admin
    Join Date
    Apr 2005
    Posts
    31,709
    Thanks
    0
    Thanked
    2,073 times in 719 posts

    3DMark adds DLSS test to Port Royal benchmark

    And Nvidia releases a 'Game Ready' driver for it (as well as the new GeForce RTX laptops).
    Read more.

  2. #2
    Senior Member
    Join Date
    Apr 2004
    Location
    Geneva, Switzerland
    Posts
    374
    Thanks
    0
    Thanked
    26 times in 15 posts

    Re: 3DMark adds DLSS test to Port Royal benchmark

    Does this means that they are pushing AMD to implement DLSS?
    Isn't the benchmark a tool that should compare products on the common ground?
    We all know that 2080ti>2080>2070>2060.

    I do not get what's the point?
    The more you live, less you die. More you play, more you die. Isn't it great.

  3. #3
    Two Places At Once Ozaron's Avatar
    Join Date
    Jan 2017
    Location
    Sometimes UK
    Posts
    638
    Thanks
    86
    Thanked
    34 times in 33 posts
    • Ozaron's system
      • Motherboard:
      • MSI X570 Unify
      • CPU:
      • Ryzen 3700X
      • Memory:
      • 32GB Patriot Blackout @ 3800 CL16
      • Storage:
      • Toshiba X300 4TB (2), Samsung 850 Evo 500GB
      • Graphics card(s):
      • Sapphire 5700XT, Sapphire R9 Fury Nitro
      • PSU:
      • Seasonic M12-II 620w
      • Case:
      • Corsair Obsidian 500D
      • Operating System:
      • W10 Enterprise 64bit
      • Monitor(s):
      • Gigabyte G27QC
      • Internet:
      • 2.5 MB/s ↓ 0.86 MB/s ↑ ~20ms

    Re: 3DMark adds DLSS test to Port Royal benchmark

    Quote Originally Posted by darcotech View Post
    I do not get what's the point?
    Well, *someone* has to make NVidia's shiny new features relevant, somehow. 3DMark team clearly have nothing else to do.

  4. #4
    Senior Member
    Join Date
    Dec 2013
    Posts
    3,526
    Thanks
    504
    Thanked
    468 times in 326 posts

    Re: 3DMark adds DLSS test to Port Royal benchmark

    While DLSS undeniably looks better than TAA using a fixed run benchmark isn't a very fair comparison, the 'AI' is dealing with known knows.

  5. Received thanks from:

    Pleiades (05-02-2019)

  6. #5
    Anti-Viral Pleiades's Avatar
    Join Date
    Dec 2004
    Location
    Amongst barbarians
    Posts
    959
    Thanks
    1,839
    Thanked
    62 times in 50 posts
    • Pleiades's system
      • Motherboard:
      • MSI Z-97 Gaming 5
      • CPU:
      • 4690K @ 4GHz / Phanteks TC-12DX
      • Memory:
      • 16 GiB HyperX
      • Storage:
      • Sandisk 480GiB; Transcend M.2 256GiB; Velociraptor 300GB
      • Graphics card(s):
      • Sapphire Pulse RX580 8GiB
      • PSU:
      • BeQuiet Straight Power 800w
      • Case:
      • CoolerMaster HAF932
      • Operating System:
      • Win 10
      • Monitor(s):
      • AOC 31.5" WQHD 144Hz; Samsung 49" 9500 HDR UHD TV
      • Internet:
      • Virgin 100 Mibs

    Re: 3DMark adds DLSS test to Port Royal benchmark

    Quote Originally Posted by Corky34 View Post
    While DLSS undeniably looks better than TAA using a fixed run benchmark isn't a very fair comparison, the 'AI' is dealing with known knows.
    And isn't TAA, even when done correctly (and I'm not convinced it was in this demo) one of the poorer implementations of AA (along with FXAA)?

    Would like to see further comparisons with MSAA etc...
    ------------------

    Valar Morghulis

  7. #6
    Senior Member
    Join Date
    Sep 2014
    Posts
    400
    Thanks
    0
    Thanked
    9 times in 9 posts

    Re: 3DMark adds DLSS test to Port Royal benchmark

    Well, on the movie is hard for me to see a difference unless it is stopped. Therefore free performance is good. Yet still not in this pricepoint.

  8. #7
    Registered+
    Join Date
    Nov 2018
    Posts
    34
    Thanks
    0
    Thanked
    2 times in 2 posts

    Re: 3DMark adds DLSS test to Port Royal benchmark

    As a feature, it's probably more relevant than the ray tracing. Better image quality at higher framerate seems to be a no brainer though as Corky34 points out, how it works with a 'normal' game implementation remains to be seen.

    I couldn't afford to upgrade from my 980Ti anyway so it's just a future technology that's interesting for now. Disclaimer - I also own Radeon cards.

  9. #8
    Old Geezer
    Join Date
    Jul 2016
    Location
    Under a rusty bucket
    Posts
    540
    Thanks
    53
    Thanked
    42 times in 31 posts

    Re: 3DMark adds DLSS test to Port Royal benchmark

    Quote Originally Posted by darcotech View Post
    Does this means that they are pushing AMD to implement DLSS?
    Isn't the benchmark a tool that should compare products on the common ground?
    We all know that 2080ti>2080>2070>2060.

    I do not get what's the point?
    The point is simple, the kind of people who want to spend money on the latest benchmarking suite, want the latest features and, they are prepared to extra for it, as in most things aimed at the early adopters.

  10. #9
    Hooning about Hoonigan's Avatar
    Join Date
    Sep 2011
    Posts
    2,308
    Thanks
    171
    Thanked
    442 times in 316 posts
    • Hoonigan's system
      • Motherboard:
      • MSI MEG X570 ACE
      • CPU:
      • AMD Ryzen 7 5800X3D
      • Memory:
      • 32GB Corsair Dominator Platinum RGB
      • Storage:
      • 2x 2TB Gigabyte NVMe 4.0
      • Graphics card(s):
      • PALIT NVIDIA RTX 3070Ti Gaming Pro
      • PSU:
      • be quiet! Straight Power 11 Platinum 750W
      • Case:
      • Corsair Crystal Series 680X
      • Operating System:
      • Windows 11 x64
      • Monitor(s):
      • Acer Predator Z35P + ASUS ROG PG279Q
      • Internet:
      • Giganet (City Fibre) 900/900

    Re: 3DMark adds DLSS test to Port Royal benchmark

    I've had a little play around with 3D Mark this morning, running the DLSS test, comparing screenshots etc.

    There's not a great deal to say, especially with it being a benchmark, so it's incredibly easy to teach the AI how to render it, but screenshots of with and without are difficult to differentiate, most of the time, but performance sees a decent rise. We're seeing that the lower end cards see the biggest impact, which is good news.

  11. #10
    Senior Member
    Join Date
    May 2015
    Posts
    359
    Thanks
    0
    Thanked
    7 times in 7 posts

    Re: 3DMark adds DLSS test to Port Royal benchmark

    The point is showing how much you can expect from each card tested. The point is comparing each card to itself. IE, how much can you expect to gain just from games turning this on? The answer is in the charts. If you already bought the card this shows what you might expect this xmas etc as more and more games roll out with this on. If you haven't bought one but are thinking about it, this is a pretty good reason as I'll be shocked (like many devs) if it isn't on the next consoles. This could be very useful for weak consoles attempting 4k. Clearly it will help all at least some, but as many have noted, it is best on weak systems perf wise, although everyone enjoys better quality especially when it comes with a boost to perf. In BF5, this gave you RT for free. So for weaker systems maybe you only get more perf with equal or better quality, but for stronger system, you clearly could get RT for free (BF5 lost 2 fps IIRC of ~100, so very nice).

    It should be noted that the dev said BF5 should get faster as they learn more about the RTX features, as they only had a few weeks with it (and some hardware wasn't even working when they first got drivers). This bodes well for future games MADE for the tech, rather than patched in. BF5 went from "why would I turn on RT with this perf hit?", to "why the heck wouldn't I want this on?"...LOL. I will be buying NV if AMD doesn't have this by xmas (7nm black friday...ROFL), along with RT and VRS (I could live without this one I think, but I want RT+DLSS at worst AMD!). I wonder what happens if you turn on all 3 in a game made for this stuff? I'm pretty sure I read they work on different paths, so it is doable if a devs wants to use it all at once.

    https://www.youtube.com/watch?v=9mxjV3cuB-c
    This vid covers how RTX features work in a few games. Not sure why others can't beta test to show what is coming when DF did it ages ago.

    https://www.youtube.com/watch?v=edYiCE0b8-c
    More VRS info in wolfenstein.

    https://www.youtube.com/watch?v=QKV8VdhZuW4
    1st good patch with 120% improvement possible & down to about ~40% depending on settings etc for higher res stuff, down to ~10% or more at 1080p it seems (higher res more to save in this case it seems at least). Reflections improved, and cost less too and this vid is 2 months old. Not sure what the latest patch is and if they're even better now. 60fps 2080ti 1440p flat now (8:40 or so) on ULTRA. It will take 7nm versions to do 4k in everything IMHO which is precisely why I'm waiting No plans for 4k for the next monitor anyway, but it will enable me to live in 1600p (1440 if forced) for that much longer on new games. DF has many good vids on their channel discussing features of many things in depth. Then again, maybe BF5 will hit 4k at some point with a few more patches (game hasn't been out long), but I'd still prefer 7nm whatever. I live in a VERY hot state.

  12. #11
    Senior Member
    Join Date
    May 2015
    Posts
    359
    Thanks
    0
    Thanked
    7 times in 7 posts

    Re: 3DMark adds DLSS test to Port Royal benchmark

    One more point, this should be included in the regular benchmark if it is BETTER looking AND faster. Any NV person would turn the feature on given the chance if this is the case. If I'm better looking than you AND faster, that is how you should test it. If it degraded quality, OK, not the same. But if it is BETTER, WTH would you test it OFF for? To make the other side that is lacking seem better than it is?

    It is only unfair to test it if they are CHEATING quality wise. Why the heck should we wait to test something BETTER until the other guy catches up? Is it NV's fault AMD can't afford to do more R&D yet (start charging more, DUH)? It's still not a game benchmark, but it IS representative of the future to some extent at least, or why does ANYONE use it at all? You either call the benchmark complete BS (it is, IMHO, but idiots like synthetics...LOL), or test it and take it for what it is. I don't know anyone with one of these cards that has ever paid for a benchmark unless they are a review site. Not quite sure what that comment meant (who buys benchmarks at home to show new features of their cards?...ROFL). AMD better get this stuff soon as I'm pretty sure there will be at least a few games available to test when NV 7nm reviews hit. You will separate the men from the boys at xmas for sure and that will cause AMD to have to drop prices yet again. AMD got lucky this launch, with nothing out of beta to test for reviews (dumb NV, this was very dumb, 2-3 months could have fixed this issue), but that won't be the xmas case. There will be NO excuse to NOT test this stuff for the next launch.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •