"We have yet to see the benefits of building dx12 into the hardware."

Yeah, yeah, yeah...We've heard it all before. ESRAM was also going to be a magical performance boost as well we just needed to wait and see! *yawn* This is coming from what was an Xbox fan, this whole wait and see marketing crap is all talk. The simple fact of the matter is they put in a flashy GPU that is still going to be held back by a trash CPU. You need 2 parts to the equation for 60 FPS in everything and MS chose to only upgrade one of them knowing that they could sit on top of a mountain beating their chest about Teraflops.

Colby lack of bandwidth hurt 360. But the lesser GPU seemed capable of keeping pace. All indications are Microsoft is getting the most out of their architecture with X and that is because they simulated the architecture before making any chips. So if building dx12 into the architecture doesn't make a difference I'm not sure why they would go through the trouble of adding it once fabrication was about to start. This fear of Microsoft doing something right is said. Get over it. This is a new era and the old days when architectures were theoretical and mistakes didn't show till it was too late are over. If you plan to just go with premature Microsoft hate you're in for a long and frustrating gaming future. Good luck with that.

@MegamamXXX lol x1x has the exclusives probably not your type but still have some dam good exclusives
and you speak of the cpu bottle necks in pro & x1x but pro will bottle neck 1st in every game
x1x also absolutely destroys the pro in gpu and memory department not even close
this means even when x1x is not hitting native 4k it will still be hitting ultra pc setting with all it's 4k assets
and unlike the pro witch scales up from a lower resolution like1440p 1800p and checkerboards up to hit 4k
the x1x always starts at native 4k and will scale down when the games get to demanding
and thats a big difference from the pro so lets all be realistic the x1x is far better then the pro
im embarrassed for sony to even having to compete power wise with x1x because realisticly
it is no competition and well worth the extra 100$.

I love it.

MegamanXXX in the comment-section at N4G.com knows more about building systems and customised chipsets than the good people at Sony and Xbox.

You know, the ones with the Master's degree at the wall and the $10,000 paycheck per month.

It's not a laptop cpu(I keep having to say this). It's a netbook cpu, which is worse.

@dcbronco

"We have yet to see the benefits of building dx12 into the hardware."

Oh you mean the same one thats in the regular X1?

https://twitter.com/digital...

Read'em and weep, but people's ignorance doesnt stop them from believing in bs.

To all that dont know, the so called built in Dx12 hardware also exist is the reg. X1

OB1Biker Microsoft have said they are not doing gens any more so we would see the Xbox one being dropped in terms of new games the X replaces the one and the Y will replace the X in terms of price and power where Sony have said they will carry on with gens so ide expect the pro to carry on getting limited support like the PS3 did. Only time will tell though

lol DC did super high res textures. DC ports on GCN and Xbox kept the high res but ports to PS2 they had to muddy them down. PS2 did some things better but DC did other things better and some things PS2 just had real issues with DC didn't.

Can we stop using Tflops like we know what it means? Or like it's in any way an accurate measure of performance?

The *only* time Tflops actually matters as a performance metric is when it's a direct comparison between GPUs using the same or very similar architecture. So, between current gen consoles and the current RX lines using AMDs Polaris, yes. Using that same metric as a comparison to literally anything else, no.

Example. The R9 390 has 5.1 Tflops. The RX 480 Caps at 6.0. The RX 580 caps at 6.2. The GTX 1070 caps at 6.4. When actually plaing games however, the 390 is consistently within ~ 5 FPS (in either direction) of the 480 and 580, despite their large "power" advantage. The GTX 1070 consistently plays games at about a 20% higher framerate than the RX 580, despite the miniscule power advantage.

while the switch is an incredible portable console and worth the price perfectly, its not significant more thn wii u while docked. the article weirdly listed tFLOPS of switch in fp16 and other consoles in fp32. switch is around 512gFLOPS if it can run to max possible clocks but switch even in dock mode is underclocked so its much less thn 512gFLOPS.

@Princess_Pilfer - comparing tFLOPS between nvidia and amd is useless while comparing tFLOPS between same vendor makes more sense.

@CartoonX1

No, it doesn't, for the reason I just described. Using TFLOPS as a measure, the RX 480 is a 20% power increase over the R9 390. Except performance wise, they're about the same, because despite both being AMD is the actual archiceture of the card that matters, so using TFLOPS to compare between archetectures will give a highly inaccurate picture of the performance of the parts.

It's like saying your mustang has 500 horse power and there for your cars is better while ignoring torque and gear ratios and the efficency of delivering that power to the wheels and the weight of the car and the shape of the car

I'm not even tech savvy, this information isn't hard to come by. Google gaming benchmarks and then check the giga flops of the card on wikipedia (1000= 1 TFLOP) and compare. It becomes quite obvious.

@Princess_Pilfer; Teraflops is a way of estimating the processing power of a unit or if you are talking about consoles the "raw power" of that unit. The reasson why you see Nvidia taking the lead in almost all benchmarks is because they use their tech in a "smarter way" than Amd.

Nvidia optimization is historically better, as well as driver support and access to unique enterprise-wide technologies that make AMD users' lives more complicated. A classic example is the PhysX technology offered by Nvidia. Take a look at old Arkham series videos about the subject if you want to know more about it.

Now something you said is true Teraflops alone do not make an console. The company's development kit and range of supported tools and engines can make a brutal difference. And as I said before that what the Switch truly have over the Wii U.

Way to actually read my comment. I specifically compared AMD cards with wildly different levels of TFLOPS but similar levels of performance. IE the R9 390 with it's 5.1 and the RX 480 with it's 6.0, yet similar performance.

So no, you're wrong. Nvidia may or may not do things more efficently, that's irrelivant, what is relevant is even between parts made by the same manufacturer it's not a good comparison because architecture changes change the TFLOP/real Performance ratio dramatically.

@dcbronco

No, we are not reaching the upper limit of what will make a visual difference. Not even close. Now, it's good to remember that true photo realism is a pipe dream that, barring some big advancements in how computers work, is never going to happen, but you go play The Witcher 3 and then hold a screenshot up next to a photo of a person and you'll see exactly how obvious it is that one of those is real and one of them isn't. That's not even getting into things like resolution (we're at less than half the resolution it would take to be unable to tell the difference between an image at that resolution and any higher resolution image) and framerate (30 FPS is passable at best, 60+ is ideal, and many games simply don't run at 4k/60 even with 2 of the most powerful consumer GPUs in the world running at the same time.)