NVIDIA GeForce Ampere 3DMark Time Spy Benchmarks show it to be 30 percent faster than RTX 2080 Ti

...And opengl, they did save their day with Mantle and Vulkan, but they never fixed bugs and issues. I believe as a high level programmer language not low level (drivers, etc), there must be made bad decisions from engineers and instead of improving, they abandoned everything, bypassing them, so they never cared the code, either dx9-11 or ogl, unfortunately. "If ain't broke, don't fix it", isn't smart.

 

There is no point of fixing OpenGL, Vulkan is better than it in every single way possible, even from a high level.

DirectX 11 probably worthwhile as Microsoft intends DX11 and DX12 development to run concurrently.

 

Like others have said before, is this is the 3080Ti, then it´s very disappointing unless they also drop the price by 20 or 30%. If this is the 3080, then the results are not bad but not amazing at the same time. We need a big performance jump.

 

Like Duh, ofc 100% leap is not gonna happen, ever. It is not how business/market is working - yeah wipe your entire previous generation and have a brutal stock of GPUs no one would buy. And if you expect 100% performance gain expect at least 70-80% price uplift as well - you do not want that. So if it is RTX 3080 it would be a nice surprise but i bet it is the 3080Ti. So good luck with your dreams.I bet Zen 3 CPUs will be 5-10% better gaming perf than Intel 10th gen and then Intel 11th Gen 5-10% better gaming perf to AMD Zen 3. So that's that.Companies would always want to leave headroom so they can sell us 10-30% max perf uplift in their next gen. Whether we like it or not, i guess we can call that a roadmap strategy. Hope i am wrong about it, we'll see.

 

pfftt! Can't be 3080ti, better not be 30% performance increase over 2080ti, gotta be 3080 instead.

 

I think they will focus on raytracing perfomance therefore the dedicated chip for that the leak mentioned.

Current games like Control would be finally be playable.

 

guys you can't compare the performance jumps from 10-20y ago where semi-conductors had a ton of improvement left to nowadays where we are at the limits
my 2002 config : 1.4Ghz processor 512Mbs of ram a 64Mbs videocard at 215Mhz (420$ for that videocard)
yes back then there was a huge margin of improvements but not anymore

I and many judging by the polls I've seen have skipped the 2080Ti and will buy the 3080Ti even if it's only +30%

 

...just wow at the crying about an alleged 30 percent.
This is a rumor, any drivers would be preliminary and imho, the change in nodes alone (in the real world) would give us at least this uptick.
but when you add in uArch, which is Nvidia's specialty, I just don't believe the rumor.

 

until the sales numbers on mobile 2xxx came in then the investors were delighted

 

Well, if there would be +30% performance increase for same model line 2070/3070 and some big TDP decrease and still Windows 7 drivers, i would be fine with it..Nothing exciting, but good enough.

 

Think about 3060 having the same raytracing capabilities to 2080 super. It will give rtx super users to upgrade even they overpayed their cards not long ago.

 

I completely disagree. Like DX11, OpenGL will never go away, and NVidia supports it just fine. AMD should have no excuse.

 

I don't think we'll get anything exciting until AMD can get their act together on the GPU side.

I know a lot of people wanted a more significant improvement than 30%, but for 10x0 owners, jumping to 30x0 with a cumulative ~60% from 10x0 to 30x0 isn't the worst assuming prices aren't hiked again. Plus, RT might be a bit more worthwhile this time around. I though RT was a smart move even if it still seems half-baked.

 

I have no intention on buying this but I don't see what the negative fuss is about. We don't know what the power consumption is. A 30% improvement is much more significant if there is also (for example) a 30% decrease in power consumption. Nvidia has been refining the architecture and moving to a smaller node.
We also don't know if this is the 3080 Ti.

There's more to a GPU than raw performance.

 

It all depends on what card you are upgrading from. I'm still rocking a GTX 980 G1 and was waiting on nVidia 2xxx cards that ended up being RTX.

So because it was RTX i thought i'd wait for the second generation of that tech plus i thought 8gb of vram was too low for such a card.

So for me 3080 will be a Massive upgrade. I'm hoping for 16gb of vram and a UK price of around £600. we'll see though soon enough.

 

Nvidia supports it fine because they've been building extensions and working with developers to get said extensions utilized for decades. There is no point for AMD to build that kind of infrastructure now to support an API that's being supplanted by Vulkan. At some point Khronos said they would be doing OpenGL/Vulkan side by side, but it's pretty clear that's no longer the case. For professional apps, AMD's support is mostly fine. For games it sucks, but the number of games shipping with OpenGL is basically nil outside of emulators - most of which are starting to get Vulkan support. Even the various GDI's are moving from EGL (Including Apple/Android) - but even in those cases AMD's support is generally fine, it's not like their drivers are bottlenecking desktop rendering.

The cost of supporting it the way Nvidia has is increasing and the amount of software that's utilizing it is decreasing. Just doesn't make sense to me for AMD to build that out and I'm sure AMD weighed that cost and decided similarly.

 

Indeed. Price.

Nobody gives a rat's ass about 30% reduced power consumption...

... however, if it has 30% more performance for 30% less money (than Turing), hell yea, I'll take it !

 

and we no how that went lol

 

I agree, but didn't I read a few month back that AMD had hired new staff specifically for their driver team. Hopefully it will help if true.

I don't think so, AMD need to make a big splash in the GPU space which is the area they are still lagging behind Nvidia. AMD has done the unthinkable in the CPU space, pushing hard on IPC, core/thread counts, and most all price to performance as well.

AMD simply can't throw all that made them successful in the CPU space out the window and just price match Nvidia. If they do they are pretty much screwed as people see the PC graphics space and the first name the comes to most peoples minds is Nvidia. AMD need to change peoples minds, and not just the core PC gamer either but the mass market.

AMD saying that RDNA2 is their halo product has not been heard from them in a very very long time if not ever.

New more gaming focused architecture with all remnants of GCN finally out, new node which they have already worked with in their CPU's, both major consoles using all AMD and using the latest/unreleased technology too, and a MUCH stronger market position than they had 3 years ago before Zen 1 launched.

Remember this moment?


All signs point towards the same thing happening with RDNA2 release. Like I have said in previous posts, AMD don't need to beat the 3080Ti/3090 they need to demolish the 2080Ti in terms of rasterization and ray tracing performance. If they can come in at 30-40% better performance for around the $600 (that more or less being half the price of a 2080Ti) and with strong drivers they have a winner.

If they don't do this and they just price match Nvidia then the GPU space will stagnate for more years to come.

 

Hey, I care about power consumption. Less power means less heat, less heat means a cooler environment, a cooler environment is good for hardware.