-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
** The Official Nvidia GeForce 'Pascal' Thread - for general gossip and discussions **
- Thread starter Boomstick777
- Start date
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
in Hitman Nvidia cards get a significant boost if you test anywhere apart from the defined benchmark. I wonder why that is
Because you're basing that off the gain from doing stock cards in one benchmark to heavily overclocked Nvidia cards in the other, thus you are attributing the gain from AMD purely on being in game rather than the benchmark and completely ignoring the clock gain.
But wait for layte to post that every few pages, have that pointed out by multiple people and then use that as a basis to make such a claim.
For the most part almost all game benchmarks show different performance from the game in the benchmark but they usually do with ALL cards. You can't use all stock cards in the benchmark and overclocked vs stock being benched in the game as proof that Nvidia are handicapped in the benchmark.
Find a review that benches both companies stock cards(or at least stock Nvidia cards alongside overclocked Nvidia cards) in game then you'd have a point, currently you're making a baseless claim on extremely dodgy benchmarking.
I'm absolutely fine with benchmarking overclocked cards, but bench a stock 980ti and a heavily overclocked 980ti to show the difference. Showing no reference Nvidia cards vs only reference AMD cards is entirely misleading. Okay Fury doesn't overclock great but it does overclock and the other cards shown like 390/380 overclock just fine.
It's only 20% slower. You youngsters and your maths.
And the gen before that. And the gen before that.
High end pricing is never the best value in just about any product line.
But then it doesn't really make sense to me why they price the X80 Ti so close to the X80 considering its performance is usually so much better. They could milk it for a lot more (here's hoping they don't for the 10 series)
Where they've priced the x80 compared to the x80Ti has varied. Especially since what they call the x80 or the x80Ti has varied. In the case of Maxwell, the 980 was always the black sheep of the lineup. It was always the case that the 970 and 980Ti were the proper 'buys' of the group. The 980 was not a bad card and may have met certain people's needs, but the pricing was not quite as aggressive for it. And then you have a card like the 780 which was like $650 or something on launch, but it was actually a cut-down GK110.
But ultimately, it's been their practice to create a good value enthusiast card along with a pricey, 'gotta have the best' sort of card for those who insist and will pay what they need to for it. The 980 at the time was that 'best of the best' card. And they charged a premium for it. Gotta remember there was a good 6 months or so before the Titan X/980Ti came out.
Yeah I get that. To me it seems they're doing the exact same with Pascal cards too. The pricing for the 1080 is not very aggressive, meanwhile the 1070 is looking (if these benchmarks are anything to go by) like a pretty good deal. I'd expect the 1080Ti to be pretty decently priced relative to its performance also, considering we'll have Vega cards available.
I'm only really interested because I need a new card within the next month but I also plan to sell and upgrade to a Ti. To me it looks like the 1070 is most likely the best to hold its value over the months.
Originally Posted by ICDP View Post
I'm not trying to be argumentative but my point is not that 1080 FE is a slow GPU, it's that it's overall a pretty poor overclocker.
Here is the TPU review, again showing an average of ~13% actual performance gain with overclocking. Though only in Battlefield 3 do not an ideal source.
https://www.techpowerup.com/reviews/...X_1080/30.html
Techspot got an average of ~15% over 3x games. This is the highest average performance gain I have seen.
http://www.techspot.com/review/1174-...080/page8.html
Yes 1080 is an excellent GPU but I have been reading reviews (and some comments in this thread) where they ooh and aah over how great the 1080 FE overclocks but the actual numbers tell a different story.
Last edited:
Its not particularly a big overclock either really. I'd expect 2.2-2.3gz+ for AIB cards.
Which would be a similar % increase again. Being as the techspot one went from 1827mhz to 2054mhz in their testing.
Time will tell if AIB are better overclockers (I hope so) because it looks like the FE is TDP constrained.
Read back a few pages - Overclock3d did a comprehensive comparison and someone worked out that the average over all games and resolutions put the 1080 as 22% faster overall. I think the 1080 was at 2.0ghz and the 980Ti was at 1450mhz.
LMFAO! I spat out my protein shake!
One other thing the 980Ti reference cards and custom cooled AIB cards all seemed hit a similar max OC potential of around 1500 core (give or take 50MHz). The advantage of the AIB custom cooled versions was that they could maintain that OC without throttling.
If we have the same thing with the 1080 then ~2.1-2.2GHz could be a best case scenario but the AIB custom cooled versions will just be able to prevent throttling better.
If we have the same thing with the 1080 then ~2.1-2.2GHz could be a best case scenario but the AIB custom cooled versions will just be able to prevent throttling better.
Why not just go from the numbers in the overclockersclub review, they have 10+ games, multiple benchmark programs, and say exactly the clock range and exactly what compared to etc.
Yes I agree the FE is not a great overclocker, its good, not great, I never said it was... but the 1080 chip, the GPU used, looks like it will be good, when it has better cooler etc. I have no plans to get an FE card.
Last edited:
I'm not disputing that the 1080 OC vs 980Ti OC was ~22% faster. I'm pointing out that the OC potential of the Founders Edition 1080 is actually rather poor. This is confirmed by the fact that stock vs stock the performance difference is closer to 30%.
Reading some of the reviews you would get the impression the OC result were stunning but they aren't in all honesty. ~13% is really nothing special if we look at 980Ti.
Yes exactly what he said.
I have no idea where you got this from, me and most of the people in this thread have universally said we are waiting for AIB coolers not the FE.
Last edited:
Reading the reviews for the 1080
Actual overclocking MHz was good
Actual performance increase was poor
This is hinting that there is something amiss with the power delivery or cooling or both.
The AIB upmarket cards should fix the above.
Please don't assume it was just you I am debating with. It is the general tone in here (and especially reviews) that imply the 1080 FE is an awesome overclocker. I rolled my eyed when I read Hardwarecanucks use the term, "going beyond insanity" in reference to getting a whopping 13% average actual in game performance boost from overclocking.
Last edited:
Nope I am not sure anyone said that. Reviews always hype everything, I think what everyone here agreed is that it would be good, ifit did not have the throttling on the FE.
Last edited: