Show your FIRE STRIKE ULTRA scores

Discussion in 'nVidia Flavor' started by UnrealCpu, Feb 6, 2016.

Should this benchmark competition include Multi Core Processors more than 4 physical cores?

Poll closed Jul 15, 2016.
  1. YES

    75.0%
  2. NO , because Physics score is unfair and in games it makes no real difference

    12.5%
  3. Include but add a separate spreadsheet for 4 physical core since games dont use more than 4 cores

    12.5%
  1. Armenius

    Armenius [H]ardForum Junkie

    Messages:
    10,098
    Joined:
    Jan 28, 2014
    Graphics score was 5312. That is 6% higher than my Titan X run at 5004, overclocked to +100 core and +500 memory.
     
  2. sk3tch

    sk3tch [H]ard|Gawd

    Messages:
    1,268
    Joined:
    Sep 5, 2008
    Yeah - so far it is not as big of a jump as I had anticipated. Based off of the 980 Ti benchmarks in reviews I thought I'd be getting about a 20% jump. My 2nd card arrives today...that will give me a lot of apples to apples comparos.

    I posted a single card versus TITAN X SLI run Nvidia GeForce GTX TITAN X Owners Club. 1080 does 61% of TITAN X SLI.
     
    {NG}Fidel likes this.
  3. Dullard

    Dullard [H]ard|Gawd

    Messages:
    1,044
    Joined:
    Jun 28, 2012
    Looking forward to your SLI score, my TX SLI graphics score posted earlier was 10,305 so you are more than halfway there. SLI scaling is supposed to be better with Pascal, so it'll be a nail biter!
     
  4. Armenius

    Armenius [H]ardForum Junkie

    Messages:
    10,098
    Joined:
    Jan 28, 2014
    I'll also be interested in seeing if the new HB bridge will make any significant impact. Unfortunately there is no availability yet, as far as I know.
     
  5. sk3tch

    sk3tch [H]ard|Gawd

    Messages:
    1,268
    Joined:
    Sep 5, 2008
    I will at least be able to show how much "better" the newer, light-up bridges are (I have an NVIDIA 2015 bridge).

    5,449
    +200 GPU, +350 MEM run
    NVIDIA GeForce GTX 1080 video card benchmark result - Intel Core i7-5960X,MSI X99S XPOWER AC (MS-7881)

    5,549
    +200 GPU, +400 MEM run
    NVIDIA GeForce GTX 1080 video card benchmark result - Intel Core i7-5960X,MSI X99S XPOWER AC (MS-7881)
     
    Last edited: Jun 1, 2016
  6. Armenius

    Armenius [H]ardForum Junkie

    Messages:
    10,098
    Joined:
    Jan 28, 2014
  7. sk3tch

    sk3tch [H]ard|Gawd

    Messages:
    1,268
    Joined:
    Sep 5, 2008
  8. Armenius

    Armenius [H]ardForum Junkie

    Messages:
    10,098
    Joined:
    Jan 28, 2014
    Not seeing as large a boost from overclocking the memory as you did on big Maxwell. Could the GTX 1080 actually be bandwidth starved over the 256-bit bus?
     
  9. sk3tch

    sk3tch [H]ard|Gawd

    Messages:
    1,268
    Joined:
    Sep 5, 2008
    I'm not sure if something is wrong with MSI Afterburner (4.3.0 Beta 3), GPU-Z, or both...but I did a run with +200 core, +200 mem and my score actually went DOWN from the memory OC (9,713 to 9,567).

    GPU-Z doesn't register memory changes...although with a single card I *did* see gains from memory OC. Maybe this is why the HB SLI bridge is needed??

    [​IMG]
     
  10. cageymaru

    cageymaru [H]ard|News Staff Member

    Messages:
    15,504
    Joined:
    Apr 10, 2003
    If I OC the memory on my HD 7950 or my current R9 290 too far it wouldn't artifact or act strange. What it would do is start thrashing and my frame rate would decrease significantly. Each card I have owned has been unique when this occurred. One HD 7950 I could take to 1800 on the memory and the other would thrash at +50.

    Maybe that bridge will help if the memory isn't stressed. :)
     
  11. KickAssCop

    KickAssCop [H]ardness Supreme

    Messages:
    6,181
    Joined:
    Mar 19, 2003
    Thanks for posting the benches. I hit 9003 with my 980 Ti SLi setup with ofcourse only a 5930K. Doesn't seem like worth investing in these puppies just yet. Also will have to dismantle my water setup which sucks.
     
  12. skypine27

    skypine27 Gawd

    Messages:
    656
    Joined:
    Apr 18, 2008
    Couple of points:

    1. For what its worth, nivida itself has said the dual tabled LED SLI bridges will run at the same 650mhz as the "new" HB SLI bridges:

    "Nvidia adds that its SLI HB bridges aren’t the only ones able to support dual-link mode. Existing LED-lit bridges may also run at up to 650MHz if you use them on Pascal-based cards."

    Source is here: The Display Pipeline, SLI And GPU Boost 3.0 - Nvidia GeForce GTX 1080 Pascal Review

    So it sounds to me as long as you are using a bridge that connects both SLI tabs, like this 3 way card (which you would use in 2 way-spaced SLI)

    [​IMG]

    I would think you would be getting the exact same SLI performance as the new SLI bridges.

    2. The reference 1080s are really getting a "bad" rap for the single 8 pin power port, most sites are saying throttling due to power pull is occurring rapidly when overclocked. The MSI Gaming model (with 1 x 8 and 1 x 6 pin plug) is getting really good over clocking results:

    Hardware.Info: MSI GeForce GTX 1080 GAMING X is much better than Founders Edition | VideoCardz.com

    and

    MSI GeForce GTX 1080 GAMING X 8G review

    3. EK is making full blocks and backplates for the MSI model. I cant wait for 2 x of them :)
    Review and select water blocks | MSI GeForce GTX 1080 Gaming X 8G 8GB GDDR5X | CoolingConfigurator.com


    Thanks for posting your results bro! Hopefully your score goes up with some more mature drivers.
     
  13. cageymaru

    cageymaru [H]ard|News Staff Member

    Messages:
    15,504
    Joined:
    Apr 10, 2003
    A lot of the testing sites are using 80% - 100% fan speeds to get their numbers. Might want to ramp up the fans and turn up the juice to the cards. For science of course.

    skypine27 I read the same thing about the bridges.
     
  14. sk3tch

    sk3tch [H]ard|Gawd

    Messages:
    1,268
    Joined:
    Sep 5, 2008
    I did you guys one better...I have both types of NVIDIA 2015 LED-lit bridges (2-way and 3-way) so I was able to test with the "two finger" connector in place. SPOILER: no performance difference. I guess we'll have to wait for the SLI HB bridges to be 100% sure...but a bit disappointing.

    [​IMG]

    2-way SLI bridge:
    9,200 [stock] - NVIDIA GeForce GTX 1080 video card benchmark result - Intel Core i7-5960X,MSI X99S XPOWER AC (MS-7881)
    9,592 [+150 core, +0 mem] - NVIDIA GeForce GTX 1080 video card benchmark result - Intel Core i7-5960X,MSI X99S XPOWER AC (MS-7881)
    9,713 [+200 core, +0 mem] - NVIDIA GeForce GTX 1080 video card benchmark result - Intel Core i7-5960X,MSI X99S XPOWER AC (MS-7881)

    3-way SLI bridge:
    9,112 [stock] - NVIDIA GeForce GTX 1080 video card benchmark result - Intel Core i7-5960X,MSI X99S XPOWER AC (MS-7881)
    9,580 [+150 core, +0 mem]] - NVIDIA GeForce GTX 1080 video card benchmark result - Intel Core i7-5960X,MSI X99S XPOWER AC (MS-7881)
    9,626 [+200 core, +0 mem] - NVIDIA GeForce GTX 1080 video card benchmark result - Intel Core i7-5960X,MSI X99S XPOWER AC (MS-7881)

    Details here:
    NVIDIA SLI HB Bridges and Pascal GTX 1080 SLI Performance
     
  15. KickAssCop

    KickAssCop [H]ardness Supreme

    Messages:
    6,181
    Joined:
    Mar 19, 2003
    Damn and I had some hope for even better results! How far is your processor overclocked?
     
  16. sk3tch

    sk3tch [H]ard|Gawd

    Messages:
    1,268
    Joined:
    Sep 5, 2008
    4GHz
     
  17. Dullard

    Dullard [H]ard|Gawd

    Messages:
    1,044
    Joined:
    Jun 28, 2012
    You're north of 10K graphics score, so that's still pretty good, and what you should focus on. Comparing the total scores is kinda hard on you without OCing the CPU more, you could get 22K+ physics with a higher OC on a 5960X. That'd likely get your total score over 10K as well.
     
    Armenius likes this.
  18. Armenius

    Armenius [H]ardForum Junkie

    Messages:
    10,098
    Joined:
    Jan 28, 2014
    2. Kind of misleading. While the MSi GAMING X was able to hold a consistent Boost clock compared to the FE its highest overclock was still around 2.1 GHz, which is what we have been mostly seeing from the FE. The voltage on the 1080 might have a hard limit imposed by NVIDIA. We'll have to wait and see what BIOS tweaking will show us and if this will be the same story across boards from other AIB manufacturers. Could explain why ASUS put 4-pin fan headers on the STRIX; to put a use to the extra power they're feeding the board.
     
  19. skypine27

    skypine27 Gawd

    Messages:
    656
    Joined:
    Apr 18, 2008
    Thank you so much! Yeah it seems like the results are within the 2-3% margin of error for the 3 way vs the 2 way.

    One thing you could do, if you still have the cheap SLI ribbon cable that comes with most motherboards, that at +200 on on the cards (no need to test it 3 times)

    Thx
     
  20. sk3tch

    sk3tch [H]ard|Gawd

    Messages:
    1,268
    Joined:
    Sep 5, 2008
    No need, homie has us covered here: Doom on my 2 Gtx 1080s in Sli is no bueno - it is, in fact, a huge issue if you want to play beyond 1440P 60Hz...
     
  21. UnrealCpu

    UnrealCpu 2[H]4U

    Messages:
    2,647
    Joined:
    Jun 20, 2003
    I scored 9 164 in Fire Strike Ultra

    my previous 980 ti scores overclocked was roughly 8600 3d marks
    not much of a change. Seeing how people with multi core cpus x99 can achieve 10k due to physics score I would like to hear peoples opinions if we should include those scores


    9164 4k ultra score

    stock MSI gaming X in SLI not overclocked " old sli bridge"

    haswell i7 at 4.8ghz

    will be upgrading to skylake this week


    Would like everyones personal opinion about scores coming from X99 users with multi cores taking advantage of physics to get a higher 3d mark score.. Do you think we should include them or exclude them in a separate score sheet on the front page?

    Since X99 vs 6700k does not have any real world benefits in gaming vs the higher IPC 6700k and most games only use 4 physical cores anyway..
    If i am wrong and you think multi core processors are faster at gaming vs a 4790k or 6700k with a particular video card let me know
    As far as i know multi core processors do not help at all with gaming benchmarks and think this should be everyone's goal to get higher FPS in games .
     
    Last edited: Jul 5, 2016
  22. UnrealCpu

    UnrealCpu 2[H]4U

    Messages:
    2,647
    Joined:
    Jun 20, 2003
    added voting poll for 10 days on first page ,will make changes and update accordingly around July 15th
     
  23. Dullard

    Dullard [H]ard|Gawd

    Messages:
    1,044
    Joined:
    Jun 28, 2012
    The graphics score is the true measure of the GPU, but the overall score is a measure of the system. I wouldn't disqualify 4+ core processors just because they score higher, it just means the user has a more powerful overall system - you can still compare the graphics score in the screenshot/link regardless of the overall score.

    Firestrike includes graphics, physics, and overall scores for a reason. Want a better score, build a more powerful system. Just because games use X numbers of cores is irrelevant in regards to Firestrike.

    Also, games may well start using more than 4 cores as DX12 becomes more widely adopted - that was one of the supposed key features of DX12.

    Firestrike is what it is. If you want a thread for "game" benchmarks, then do that. Don't neuter a widely used benchmark to cater to the lowest common denominator.

    And as my FSU score was not suitable for addition to the scoreboard, anyway, I don't say this to try to get to the top - SLI 1080s are kicking SLI TX scores to the curb already. I just find the whole premise of the reason for change to be flawed. But as more people own lower core chips than X99, they are of course going to vote to disallow the competition from higher core count chips. Slam dunk win for "neuter".

    ETA:

    Looks like you might want to disallow TITAN X scores, as well. I have several hundred higher point graphics score than yours without even figuring the physics score into the equation.
     
    Last edited: Jul 5, 2016
  24. Supercharged_Z06

    Supercharged_Z06 [H]ard|Gawd

    Messages:
    1,395
    Joined:
    Nov 13, 2006
    Pretty much this. Firestrike is a system benchmark, not a GPU benchmark. One can easily go look at anyone's linked score to see exactly how it all stacks up and what their graphics score pulls.
     
    Armenius likes this.
  25. Supercharged_Z06

    Supercharged_Z06 [H]ard|Gawd

    Messages:
    1,395
    Joined:
    Nov 13, 2006
    Yes, the jump from a dialed in pair of 980Ti's to 1080's is very underwhelming where 4k benches and 4k gaming is concerned.

    Did a comparison of your stock 1080 SLI setup scores against my 980Ti OC'ed setup scores and they are very, very close:
    Code:
                          Your 1080 SLI        My 980Ti SLI
                          --------------       -------------
    Graphics Score             9793               9784
    Physics Score             12963              12410
    Combined Score             4770               4646
    Graphics Test 1           55.63 fps          54.95 fps
    Graphics Test 2           34.49 fps          34.70 fps
    Physics Test              41.15 fps          39.40 fps
    Combined Test             22.19 fps          21.61 fps
    Glad I didn't make the jump yet - holding out for the arrival of a 1080Ti or 1080 Titan variant to make it worth the hassle of messing with my loop.
     
    Last edited: Jul 5, 2016
  26. heatlesssun

    heatlesssun Pick your own.....you deserve it.

    Messages:
    45,066
    Joined:
    Nov 5, 2005
  27. UnrealCpu

    UnrealCpu 2[H]4U

    Messages:
    2,647
    Joined:
    Jun 20, 2003
    i am really liking the MSI gaming X 8gb

    they are even more quieter and cooler.. I wanted to dump them asap to get my money out of them while i can . Now with RX 480 4gb that can be flashed to 8gb for free we might even have a harder time getting rid of those 980s , 980 ti 4gb

    Btw i have not overclocked them and yes the 980ti is good card however they are hot and pull more electricity.
    some people dont care about those things but i do. Going to over clock them to see how they perform. I still need to flash them to the OC bios so might be a few days until i get around to it
     
    Last edited: Jul 5, 2016
  28. UnrealCpu

    UnrealCpu 2[H]4U

    Messages:
    2,647
    Joined:
    Jun 20, 2003
    thanks for the input.
    I think we will just stick to graphics scores to see who has the best gaming system.. what do you think?
     
  29. Supercharged_Z06

    Supercharged_Z06 [H]ard|Gawd

    Messages:
    1,395
    Joined:
    Nov 13, 2006
    Hey, it's your thread, but in my book, best graphics scores /= best gaming system. I'll say it again, Firestrike is a system benchmark, not a GPU benchmark. The graphics score is just one component.
     
  30. Nytegard

    Nytegard 2[H]4U

    Messages:
    2,455
    Joined:
    Jan 8, 2004
    But as he mentioned, best gaming system is not exactly defined. Running a game which can use multiple cores will be different than running something that sticks with one core. And with DX12 starting to leak into games, this will change. It's not really worth having a list of "Best Single Core Gaming Systems" and "Best Multi Core Gaming Systems". At that point, one could make an argument that many games don't support SLI, so multi GPU systems might not be the best gaming computer.
     
    {NG}Fidel likes this.
  31. cageymaru

    cageymaru [H]ard|News Staff Member

    Messages:
    15,504
    Joined:
    Apr 10, 2003
    Hope nobody minds if I add my score to the list.
    I scored 2 997 in Fire Strike Ultra

    2 997 with AMD Radeon R9 290(1x) and AMD FX-9370
    Graphics Score 3010
    Physics Score 8336
    Combined Score 1505
     
    Armenius likes this.
  32. skypine27

    skypine27 Gawd

    Messages:
    656
    Joined:
    Apr 18, 2008
    Thanks!!

    Finally this is what I was waiting to see, 2 x 1080s OC'ed a bit and in SLI.

    I didnt pull the trigger on a pair like I said I was going to. Not because Im unimpressed by the card, but I refuse to buy the reference style card with a single 8 pin and I refuse to run them until I cool them under custom water. Last I checked, EK STILL isn't selling water blocks for the custom cards with an 8 + 6 or dual 8 pin power plugs. And just the other day the leak came out that the Titan P (as they are calling it) will be announced at Gamescon in August and might be shipping in September. Gonna wait for them. The big diff (to me) is on the Titan P or whatever its actually called, I dont mind the reference card version since it no doubly will have an 8+6 pin power setup from the factory and EK should be pretty quick in getting blocks out for them.

    Thanks for your score!
     
  33. Armenius

    Armenius [H]ardForum Junkie

    Messages:
    10,098
    Joined:
    Jan 28, 2014
    As already mentioned, Firestrike is a system benchmark. I think the rankings should stay ordered by combined score, but maybe highlight what the graphics score was in parentheses or something. This is a video card subforum, after all.
     
  34. Supercharged_Z06

    Supercharged_Z06 [H]ard|Gawd

    Messages:
    1,395
    Joined:
    Nov 13, 2006
    Good suggestion - would make the pure graphics score readily apparent and easy to compare.
     
  35. wareyore

    wareyore [H]Lite

    Messages:
    86
    Joined:
    Jan 1, 2014
    I ran single card tests on my SR2 with 2 x X5690's with a total of 12/24 cores/threads and a 3770k with 4/8 cores/threads. Obviously, they are different platforms but might be useful to compare the graphics and physics scores and overall scores. Both systems are at stock settings.

    3770k - I scored 4 566 in Fire Strike Ultra

    SR2 - I scored 4 519 in Fire Strike Ultra

    Anyway, 3dmark is just one data point I use to determine the overall performance of a system.
     
  36. cybereality

    cybereality [H]ard|Gawd

    Messages:
    2,046
    Joined:
    Mar 22, 2008
  37. quikt

    quikt Limp Gawd

    Messages:
    166
    Joined:
    Sep 11, 2006
  38. KickAssCop

    KickAssCop [H]ardness Supreme

    Messages:
    6,181
    Joined:
    Mar 19, 2003
    My supremacy reign hath ended. Time to order some 1080s so I can post higher scores here :p.
     
    Armenius, cageymaru and cybereality like this.
  39. cybereality

    cybereality [H]ard|Gawd

    Messages:
    2,046
    Joined:
    Mar 22, 2008
    But, please, not before the OP edits the first post with the new scores.
     
    Armenius likes this.
  40. {NG}Fidel

    {NG}Fidel [H]ardness Supreme

    Messages:
    5,419
    Joined:
    Jan 17, 2005
    3dmark.PNG
    5335, could go higher but my clocks had been lowered recently and I forgot to change them for this test!