Seven games benchmarked before and after Denuvo got removed

Discussion in 'Frontpage news' started by Hilbert Hagedoorn, Aug 9, 2018.

  1. Kaaskop

    Kaaskop Member

    Messages:
    31
    Likes Received:
    2
    GPU:
    ASUS STRIX 1080A8G
    I'm sorry but I actually do see it as being rather negative about the CPU in this article.
    But let's just agree to disagree on the subject of it being negative or not.

    Like I said above, the CPU is (still) amazing and unless the next has 8c/16t as the €350-400 i7 (or i9 if they're going a different route), I will still keep it for as long as I can.
    For 1080p60 this CPU can still hold it's ground against most games. However, at higher refreshrates (and thus framerates), I do indeed think it will bottleneck quite a bit more than I'm experiencing.

    I'm glad that you didn't treat me as an complete idiot. As for me it's quite difficult to portrait what I mean to say into actual text and come across as rude or childish sometimes. Even though I try to back-up my words with as much evidence as I can to show as to why my standpoint is like that.
    Anyway, it was a good one and I had fun trying to convince you of my point while you did the same for yours [​IMG]
     
  2. RealNC

    RealNC Ancient Guru

    Messages:
    3,108
    Likes Received:
    1,339
    GPU:
    EVGA GTX 980 Ti FTW
    It didn't seem there was anything said about the CPU that was unreasonable.

    However, it did seem that the perf issue was downplayed a bit. The only acceptable perf loss due to DRM would have been 0%. Because otherwise, it's like you're saying "ah well, it's only 3-10%, a small price to pay SO THAT WE CAN HAVE DRM, YEY."
     
    Prince Valiant likes this.
  3. MaCk0y

    MaCk0y Master Guru

    Messages:
    533
    Likes Received:
    156
    GPU:
    GB RX 5700 XT OC
    It's on a case by case basis. No need to classify the 2600K as either being a bottleneck or not because it depends on the circumstances.

    Do you play at 1440p/2160p resolutions? The 2600K will not be a bottleneck. Do you play at 1080p to try to achieve 120FPS+ even at the expense of lowering graphics quality? The 2600K is most likely going to be a bottleneck.

    No need to fight because everybody is right. :)
     
    carnivore and alanm like this.
  4. Koniakki

    Koniakki Ancient Guru

    Messages:
    2,843
    Likes Received:
    443
    GPU:
    ZOTAC GTX 1080Ti FE
    I also would like @Kaaskop to know that I didn't mean any disrespect when I was being a bit defensive, since from time to time the forums get a few random new members behaving in "certain" ways.

    Clearly you are not in that category, based on your previous comment. Welcome to Guru3D. :)


    p.s: Also offering a proper point for discussion is one of many things forums are for, so please don't be deferred to bring up any in the future. I'm sure many would be happy to discuss it.
     
    Last edited: Aug 12, 2018
    airbud7 likes this.

  5. gx-x

    gx-x Maha Guru

    Messages:
    1,281
    Likes Received:
    101
    GPU:
    1070 G1
    This needs to be said:
    Less performance =/= bottleneck

    Bottleneck does not mean slower CPU. It means something entirely different. But in most cases people use "bottleneck" to describe a slower processor/gpu/whatever. Which is wrong.
    CPU would be a bottleneck for GPU, , to put it real simple, IF you were to put 1080Ti instead of 1060 and got the same performance.
     
    386SX, carnivore and alanm like this.
  6. Denial

    Denial Ancient Guru

    Messages:
    12,386
    Likes Received:
    1,625
    GPU:
    EVGA 1080Ti
    I don't entirely agree - for example in the video I posted the 2600K is remarkably slower than the 6700K at the same frequency - up to 25% in specific games. Which means there is an IPC and/or platform advantage Skylake has over Sandybridge. That difference also indicates that Sandybridge, at least for specific instances during the test, is indeed bottlenecking the GPU to a far higher degree than Skylake is. Enough to affect the overall score by 25%. That bottleneck might only be temporary, during specific scenes, but it's still occuring at various points and affecting the overall score to a large degree.There is nothing about bottleneck that indicates it has to be during the full duration of the test or meet some arbitrary GPU comparison test.

    You'd could argue that the overall score impact could weigh on usage of the term - for example if it was only 2% you could argue margin of error, or too statistically insignificant to be considered a "bottleneck" - but in the case of 25%, the CPU is clearly bottlenecking performance.

    Unrelated to your post:

    It seems to me people want to move goal posts around and create ultra specific conditions to form arguments on.. for example only consider 1080p @ ~60fps and not higher framerates, or only consider the CPU itself and not the platform.. idk. The original topic is about Denuvo affecting performance and I think Hilbert's correct that a more modern processor, implying a modern platform, may be less affected by the impact seen here. That being said, there are a lot of people that run lower end configurations or older machines and there are other games that may be more CPU dependant, that run Denuvo, that weren't tested. So it kind of swings both ways. I'm going to agree with RealNC and say that the only acceptable performance loss from DRM is 0%.
     
  7. Kaaskop

    Kaaskop Member

    Messages:
    31
    Likes Received:
    2
    GPU:
    ASUS STRIX 1080A8G
    Yeah no worries, the way I replied was also very defensive as I was frustrated that I couldn't get my point across.


    The difference you're seeing in that video is most likely due to RAM bandwidth.
    What else would affect it like that, when they have the same cores, same threads, same freq, same instruction set. (I'm genuinely curious as to what specifically is causing it.)

    It's more about, what is considered "standard" these days. Because that's what most people have and most people care about.
    Little snowflakes with top tier stuff at 4k/120hz aren't important. Yeah everyone is jealous (even me), but they make up 0.0001% of the market, so they ain't important and considered in a discussion like this.
     
  8. prazola

    prazola Member Guru

    Messages:
    171
    Likes Received:
    10
    GPU:
    R9390XSOC / R9290DCU2OC
    A test with a core 2 quad or an FX8350 would be better.
     
  9. vbetts

    vbetts Don Vincenzo Staff Member

    Messages:
    14,714
    Likes Received:
    1,296
    GPU:
    RTX 2070FE
    At this point in time seeing how titles are utilizing more than 4 threads, the FX 8350 results might surprise us all I bet at higher resolutions. But I wouldn't put a Core 2 quad and an 8 core FX cpu at the same level.
     
  10. gx-x

    gx-x Maha Guru

    Messages:
    1,281
    Likes Received:
    101
    GPU:
    1070 G1
    ...you are just sticking to the claim that slower = bottleneck. It is not. It's just slower. Like i5 or i3 sandy or FX8xxx is slower than i7 2600. Not bottleneck, just slower.
    Also, there are instances where game with denuvo works better than without denuvo. It is clear: Denuvo increases performance in some cases /sarcasm
    It's called BAD CODING.
     

  11. Denial

    Denial Ancient Guru

    Messages:
    12,386
    Likes Received:
    1,625
    GPU:
    EVGA 1080Ti
    How can you say "same instruction set" when AVX2 (256bit), gather, FMA and most importantly TSX (fixed with broadwell/stepping) was added literally with the next architecture after Sandy, Haswell? Not to mention improvements to out of order execution in the scheduler, increases in dispatch/execution resources, cache bandwidth doubling, decoupling the frequency of last level cache to ring bus, improvements to microcode for frequency ramping, etc. And that's all with just Haswell. Even if it's just RAM bandwidth - who cares? Memory controller is what governs memory support and is part of the CPU. Like I said we're talking about the platform - if you're just going to decouple random things from the processor you might as well say a Pentium 1 wouldn't be a bottleneck if literally the processor was different.


    No one is saying 4K 120hz.. in fact 4K 120hz would work in your favor because the CPU would probably make no difference. 1080p 120hz on the other hand is rather popular - especially on Guru3D, which is the target audience for Hilbert. Just go to any monitor thread that's 60hz and look at all the people asking why it's even being posted or why it exists.

    ...

    Bottleneck Definition: a situation that causes delay in a process or system.

    The process here is 'rendering the frame' - the CPU is slow, thus delaying the process of 'rendering the frame' and thus the CPU is a what? A bottleneck.
     
  12. mbk1969

    mbk1969 Ancient Guru

    Messages:
    8,061
    Likes Received:
    4,792
    GPU:
    GeForce GTX 1070
    ... And to seven games the Rings Of Protection were given. But one by one they lost them...
     
  13. gx-x

    gx-x Maha Guru

    Messages:
    1,281
    Likes Received:
    101
    GPU:
    1070 G1
    Nope. That is not it. When you try to push a large amount of something that cannot fit through something, etc. It's pretty simple really, you just refuse to admit it. If you can put a faster GPU on that i7 in question, and the card runs faster then the previous one you had - there is no bottleneck.
    What you are claiming is silly.
    It's like if I have a pair of very good tires on a 120hp car, and same pair of tires on 500hp car and then claim that tires are bottleneck for the 120hp car, because they go faster on a 500hp car.
     
    Last edited: Aug 10, 2018
  14. fry178

    fry178 Maha Guru

    Messages:
    1,187
    Likes Received:
    133
    GPU:
    MSI 1080 X@2GHz
    Still, its not a 1080ti.
    Not talking about performance, but hardware .
    I have yet to see a rig with a 1080ti that runs a 2nd gen intel.
    Either its something smaller, or the cpu is newer.
    The few that actually might (as no one knows ALL the rigs out there), are a very small percentage thats not relevant when looking at common/biggest share of the pc gamers.
     
  15. vbetts

    vbetts Don Vincenzo Staff Member

    Messages:
    14,714
    Likes Received:
    1,296
    GPU:
    RTX 2070FE
    If you can't max the gpu out because of the CPU, then that's a bottlneck. Simple as that. If a Core i7 920 is maxed at 100%, but a GTX 1080 TI is sitting at 70%, that's a bottleneck. In the case of say you upgrade from an FX8350 to an i9-7900x and your cpu is sitting at say 90% but your gpu can hit 90% usage as well, then that bottlneck is gone.

    If you are only getting say 40fps on system A, then upgrade your platform and with the same GPU, you're getting double the FPS then you are eliminating the bottleneck.
     
    KissSh0t likes this.

  16. fry178

    fry178 Maha Guru

    Messages:
    1,187
    Likes Received:
    133
    GPU:
    MSI 1080 X@2GHz
    Except that there is a difference between load.
    I can play nfs rivals at 4k and only have 60-70% on the gpu, but not because im running a bottleneck on the cpu.
    Outside synthetic benchmarks you wont see 100% gpu load continously all the time.
     
  17. Killian38

    Killian38 Master Guru

    Messages:
    312
    Likes Received:
    88
    GPU:
    1060
    I wonder how my old 8350 black would do with a 1080TI. Well, it's my son's. Thinking about upgrading his PC and taking his old one. Think it would suck with a 1080ti?
     
  18. Fender178

    Fender178 Ancient Guru

    Messages:
    3,779
    Likes Received:
    92
    GPU:
    GTX 1070 | GTX 1060
    No you are the one acting like a fanboy because you're complaining about the negativity that isn't there about the 2600k. HH said the 2600k is a good CPU. All HH was doing was gathering facts and the 2600k CPU is old so it will suffer in performance vs a 6700k or 7700k. Also in a game like FFXV older CPUs are having a heck of a time keeping up it's the way it is at the end of the day.
     
  19. fry178

    fry178 Maha Guru

    Messages:
    1,187
    Likes Received:
    133
    GPU:
    MSI 1080 X@2GHz
    not if you play above 1080p and set nv/game settings towards max, that usually keeps load on the gpu.
     
    Killian38 likes this.
  20. HeavyHemi

    HeavyHemi Ancient Guru

    Messages:
    6,278
    Likes Received:
    601
    GPU:
    GTX1080Ti
    "Bottleneck":
    Subjective term loosely defined as whatever subsystem in your PC is impacting your pleasure the most. Cured whenever the pain of the "bottleneck" becomes severe enough to force the victim to pry open their wallet. This short term relief will soon be dampened by the inevitable and too soon discovered following.... "bottleneck".
    It's a perpetual cycle.
     
    Koniakki likes this.

Share This Page