Yeah... once you enable TSSAA you end up with this..
Because the AA is being handled as a compute task in parallel with the Graphics etc tasks.
So you gain quite a bit.
Oh... and it is clearly not just about AMDs bad OpenGL drivers. How many other games can you name where a FuryX slaughters...
It is not just the Vulkan results... but the various postings some of us have been doing on this subject which garnered the attention of scores of gamers. When you couple the two... folks are pausing and coming to realize that... AMD are not going away anytime soon.
Like I said back on...
Oh... so like everything we were talking about is now starting to come into play? Gee Whiz... who would have thought?
When you see a FuryX demolishing a GTX 1070 under Vulkan in Doom... you probably start to realize that those of us who were saying that this was coming were probably not...
The clock speed and additional ROPs are the most likely culprits.
Some dude just overclocked a reference RX 480 to 1.5GHz..
He obtained near Fury-X performance. Add another 200MHz or so and you should be near GTX 980 Ti performance... with only 32 ROPs.
So it is not that impressive. What is...
Well it appears to be true... look at the ALU performance (proving it is running at 1898MHz).
2(1898 x 1920) = 7.2 TFlops and the figure that PCGameshardware obtain is 6.9 TFlops.
If it were running at a nomial boost clock of 1683MHz then we would get a theoretical TFlops MAX rating of 6.4...
Which only further proves that it is a 48 ROPs part. In the test above they have set the boost speed to 1898MHz (read the picture posted by Silverforce carefully).
I think that the TechReport (once they post their GTX 1070 review and their "sizing them up" section) may prove to be enlightening...
Actually... Silverforce11 appears to be right.
Take the clock speed and multiply it by the number of ROPs and you get the theoretical Pixel Fillrate.
Take 1898MHz x 48 = 91 GPixels/s. If you look at the GTX 1070s performance it is around 85 GPixels/s. In other words... it only has 48 active...
Not in the slightest.
With all due respect, I am recommending the Oxford books for you because they are great starting points for people struggling with English. Take your sentence formation for example, "Oh, touchy. I hit a nerve?" would be the proper way to communicate your message...
Here is one of many sources available to anyone who does a simple Google Search...
Stock issues will not improve until sometimes in July or August..
http://www.tweaktown.com/news/52650/nvidia-geforce-gtx-1080-stock-improve-until-late-july/index.html
You have a history of misquoting me or making stuff up about things I have said. If you do not have time to search my posts then it would be common decency to simply not attribute claims to me of which you are not certain.
Umm... it is an educated guess based on these rumors holding true. If these rumors are not true then my guess is not valid. I have predicated my statements on these rumors holding true... re-read my posts.
If you have issues comprehending how the English language works I suggest you get...
Can you quote me plz? On wait.. you cannot because you are making stuff up.
I did state that Pascal would be slower than Vega (if you find my quote you will see that this is what I stated). This appears to be true no?
As for AMD launching first... do you have a quote on me stating this? I...
You will need a very high clocked 1080 Ti and that chip is much larger than the 1080 which has issues keeping heat under control beyond 2000 MHz. The overclocking headroom on Pascal is not there on the smaller chips therefore it likely will not be there on the larger chips.
AMD on the other...
Umm... I think AMD are pretty much confirmed dominance with Vega 10/11 now.
If a Polaris 10 can match a 1070 due to this massive overclocking headroom while still running lower clocks than the 1070 then we have confirmation that the performance per clk figures of AMDs Polaris and Vega cards are...
Overclocking doesn't matter anymore either. Oh and heat doesn't matter too.
I swear, you can't make this up. I'd like to create a collage of folks who make changing claims as generations of GPUs release in an effort to back their favorite side. No doubt I'd get banned though.
Bingo.
The $599 cards will actually turn out to be superior with the AIB vendors absorbing the extra costs associated with high quality electrical components.
That's not true. NVIDIA claimed top engineering from the FE.
http://forums.anandtech.com/showpost.php?p=38242302&postcount=87
Watch it yourself.
Clearly, the AIB vendors will be the ones to offer that once the FE reference run is over.
The GTX 1080 FE is selling for more than the GTX 980 Ti did. So compare it with a board of similar pricing..
http://www.techspot.com/photos/article/1011-nvidia-geforce-gtx-980-ti/
See, 6 VRM phases.
So it's not FUD.
That's an Asus GTX 980 Ti. Not only does it contain 12 VRM phases but they're higher quality phases.
Another example is the MSI Lightning boards..
Reference designs have historically contained 6 high quality phases. NVIDIA have just dropped this down to 5.
VRM Phases are like tiny power filters which clean up variances in the power delivery to the GPU. Basically, the GPU board is fed 12v of power but the GPU needs far less than that (usually between 1-1.3v). What the VRMs do is phase the power down from 12v to the required 1.xv. Each phase step...
What??? You're grasping at straws at this point mate.
Pixel level preemption is finer grained preemption. The terms are interchangeable. It simply means instruction level preemption.
Preemption requires a context switch, which Pascal can do quickly, like GCN. This is what I stated before...
They're trying to make money before AMD strikes. That's why the release is filled with marketing and hype while the product isn't really that impressive imo.
Charging the $100 premium before AMD launches Polaris. Polaris won't be competing with the likes of a GTX 1080 and maybe not even a GTX...
Under DX12, small Vega will decimate the GTX 1080.
Big Vega will beat big Pascal. Polaris will cover the rest of the market.
Under DX11, harder to tell because AMD haven't released a lot of info regarding their IPC improvements (single threaded performance). If the improvements are...
Pascal is still executing the code in serial but the gap between executing the graphic and compute context is shorter, less latency, due to improved context switching over Maxwell (switching at the instruction level rather than thread block level).
The kicker is that while the GTX 1080 is 30%...
Async Compute + graphics is a subject I've studied closely. I know EXACTLY how it works. Folks at Beyond3D and elsewhere don't seem to have a full grasp of the feature.
What Pascal does is exactly what I stated above. End of story.
Pascal has the same performance in lower resolutions as higher resolutions under AotS.
Each AotS run is different. See Pascal has the same performance at 1080p, slightly higher at 1440p, and a loss at 4K. This is just due to the slight differences in every run of the benchmark.
Pascal...
I'll add more...
What is preemption used for and why is a context switch involved?
Preemption is used to pause/stop currently executing work in order to push through a higher priority workload.
When is preemption used? VR and more specifically when an Asynchronous Time Warp is required...
Oh and its not necessarily preemption that is involved but a context switch. Preemption is based on allowing for smooth context switching without incurring a performance loss from latency involved in the switch.
Asynchronous Compute has nothing to do with context switching. Asynchronous...
Sontin, don't opine on topics you're not well versed in.
Preemption is not tied to Async Compute directly but comes into play when switching between Graphics and Compute loads. What Async Compute + Graphics does is execute both a Compute and a Graphics task in parallel. So a switch is...
As for Kyle, he's an interesting character. Seems to me that some folks, who defended him as he bashed AMD, are now attacking him and questioning his credibility after he goes on the offensive against NVIDIA. I see Hypocrisy, as per usual, is strong with partisan folks.
I don't like Kyle. He...
What Pascal has fixed is preemption, so switching between compute and graphics tasks no longer requires a flush of the SM.
This minimizes idle time and as such minimizes the performance impact of synchronization points (fences). Due to this, performance between having Async turned on and off...
Renderstate, I have to disagree.
What you're saying would require that a GPU be capable completing compute and graphics tasks at "EXACTLY" the same time. Meaning that if a compute task took 5ms so would a graphics task. This happening without fences (synchronization points).
This would...
Ditto.
There's no way I'm falling for a 1070 or 1080. The pricing is ridiculous and the DX12 performance boost, outside of maybe Tomb Raider, won't be convincing for me. I'm looking at DX12 performance from this point on.
Basically, what NVIDIA have done with Pascal is overclocked Maxwell and split the SMs into two in order to boost compute efficiency. This will help push Pascal past the FuryX with ease under DX12 scenarios.
The question is, what about Vega? Vega will be quite a boost over Fiji.
We know...
Maxwell could, under CUDA, make use of an onboard ARM chip for management of Graphics and Compute loads. This doesn't work for DX12. So that graph is wrong.
NVIDIA are the ones having yield issues hence the "Founder's Edition". AMD, on the other hand, don't have yield issues. They stated that they would launch in mid-2016 and that's what they're doing.
I'm willing to bet that once Polaris launches, the market will be flooded with various products...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.