-
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.
Nvidia FY 2015 (yes 2015) Q3 financial results
- Thread starter layte
- Start date
Didnt Simsang do the same with apple ? They seem to be the king of court orders.
Considering the PowerVR SoC is over triple the speed with the Metal API and half the energy drain NVidia are in a bad position to be claiming highest-performing mobile application processor. They are also wrong to say they are the first mobile compute SoC. In reality there where not only the last to do mobile GPU compute but last by years, its about time someone did something about all there false advertisements.
Yet again somebody is going to call you out on this. Please give us some concrete proof of these claims, and I don't mean an Apple PR event.
The Anandtech iPad Air 2 had the performance per watt as 21.64 for the Ipad and the Shield tablet gets 12.05 performance per watt.
None of the Metal API benchmarks are publicly out yet but various devs have gone on record saying they have seen a x10 to x20 speed increase major devs like Epic games. Metal API games also look better with more detail proving the increase is real.
I believe the fact we can see the difference in Metal API apps and games added with the dev comments is enough to prove Metal API is faster. We can argue precisely how much faster later when the benchmarks support metal API but right now I am picking the lower end of the numbers being used by devs. Not got time to search for videos right now as my break is almost over but there are a number of dev interviews on youtube about the subject.
In short PowerVR are faster in benchmarks and we know the metal API is even faster. So adding the two together how can NVidia say they have the fastest most powerful SoC? How can NVidia say they have the first GPU compute SoC?
Last edited:
That’s not how I see it.
My claims on benchmarks without Metal API have confirmed PowerVR is faster.
My claims on Apps and games with Metal API get a large increase in speed have been confirmed. Adding the two together is basic logic.
How is there any wild performance claims from me? If anything the performance claims I am saying are massively smaller then what the game and app devs are reporting.
Oh look there's one....
Certainly not triple the performance......using your own source.
And before you comeback with 'oh but metal will make it so much faster yada yada yada' yes metal will make things faster. but we have heard this all before from the pr people at Nvidia and AMD with gameworks and Mantle, 'it can do things hundreds of times faster yada yada yada. And yet in the end it makes 10% difference at best.
Yes Metal will make a difference, but not another 100% difference, so triple the performance is just a wild claim from you.
Oh dear... Nvidia threatening to sue Qualcomm for IP infringement, Samsung accuse Nvidia of IP infringement and lying about Tegra K1 claims....
I said Nvidia would #### ### the industry in behaving like a bull in a china shop.
Last edited:
That is true the source I told you to look for said a 10 fold increase in rendering efficiency which is better then what I said. Some devs are reporting a 20 fold speed increase depending in what they are doing.
The difference here is it’s not just NVidia or AMD talking about the performance increase that we never see in real world games. Metal API has been checked out by lots of independent devs and devs from major company’s all confirming it. Backed up by actual real games with a leap forward in graphics like Asphalt 8, Beach Buggy Racing, Defenders, Epic Zen Garden, and Modern Combat 5 just as a few for example.
http://appadvice.com/appnn/2014/09/...at-5-blackout-with-metal-enabled-enhancements
Frostbite, "Metal has created possibilities previously out of reach and for the first time we can include both high visual fidelity and a large number of objects."
http://www.redmondpie.com/frostbite...-8-demo-using-apple-metal-graphics-framework/
Devs said over triple the amount of objects allowed on screen without slowdown. We have now seen games and apps with triple the amount of objects. So I stand by in saying Metal API is at least triple as good as the old API. There is no way the Metal API games are looking as good as they do with a 10% difference at best speed increase.
This is not a case of ”'because I said so',” and this is not a PR stunt. The Metal API has been proven by so many devs and games I don't see how anyone can say its not good. Even if I have the numbers wrong it doesn't change the point of the argument. NVidia are going around saying they are the fastest when clearly they are noticeable slower. Which is one of the reasons Samsung are taking NVidia to court.
Triple the amount of objects on screen does not equate to triple the speed.
Your claim was that the PowerVR SOC was triple the speed of the K1, which it clearly isn't.
Show us a benchmark that shows the PowerVR SOC being triple the speed of the K1 and I will agree with you, until then it is in fact just another wild claim.
Your claim was that the PowerVR SOC was triple the speed of the K1, which it clearly isn't.
Show us a benchmark that shows the PowerVR SOC being triple the speed of the K1 and I will agree with you, until then it is in fact just another wild claim.
Pottsey, your last few thread rants (especially the nVidia Tablet) had some wild claims from you and everyone you made was shown to be false but you still insisted that you was right and this time you are doing it again... I am sure the PowerVR SoC is a fast chip but you need to back your wild claims up, which yet again, you haven't... From your supposed back-ups:
Nada there about being 3x faster than the Tegra K1
And nada there...
Terrible.
Nada there about being 3x faster than the Tegra K1
And nada there...
Terrible.
If you can render x3 more detail/objects then before for no speed loss then you have effectively gained triple the speed. If the new API can render the 2500+ objects x10 faster then the old API I think it’s fair to say the new API is triple the speed. Or are you saying all the games and devs are wrong?
As I showed PowerVR is almost double performance per watt over K1 without the Metal API. Add in the Metal API and we see triple the performance per watt over K1.
EDIt:
“Pottsey, your last few thread rants (especially the nVidia Tablet) had some wild claims from you and everyone you made was shown to be false but you still insisted that you was right and this time you are doing it again...”
Looking back it turned out for the most part I was right and you and the others where wrong. What I posted was not false but in fact correct.
The backups I posted prove the Metal API is massively faster. If it wasn't those graphics improvements I posted would be impossible. If the Metal API isn't massively faster how can the real world games and apps have those graphic improvements implanted? Why are the devs talking about a x10 fold increase and showing us massively improved scenes if the improvements are not real?
Look that x10 and x20 speed comments are not from me. They are from various game devs. No one ever said everything is x3 faster. Draw Cell rates for example are x10 faster. Epic games Tim Sweeney said x10 rendering efficiency. The apps and game prove the speed increase is massive even if we don’t have precise numbers. Even after we get precise number different thing will see difference speed increases.
Bru I already explained there are no public benchmarks yet. All we have are games that are way better than before and game devs talking about over x3 rendering efficiency. The games alone prove it’s not a mere 10% improvement. The precise numbers don’t even matter for my point. Tegra K1 is massively slower for games and apps while NVidia are saying K1 is faster which is not true.
I don’t want to derail this thread arguing precisely how much faster PowerVR are all that matters is its faster. In reality Tegra K1 is slower at games, has much worse graphics in games while NVidia are going around saying we have the best speed and best graphics and first mobile GPU compute all 3 of which are not true. How will that play out in the lawsuit?
Last edited:
Once again, prove it. Show us the numbers, show us the benchmarks. If you cant then hush.
(from another thread where Potty keeps unning his mouth off)
The Apple A8X, barely faster in one test, humiliated in the other. Imagine the kurb stomp that will occur when Nvidia move K1 to 20nm like the Apple SoC.