Vigilant Walrus
Member
How does Polaris stack up to Pascal?
I wonder if Sony and Microsoft made a mistake by going with AMD.
I wonder if Sony and Microsoft made a mistake by going with AMD.
Nah... They'd never turn a profit paying Nvidia licensing fees. AMD was the best choice.How does Polaris stack up to Pascal?
I wonder if Sony and Microsoft made a mistake by going with AMD.
How does Polaris stack up to Pascal?
I wonder if Sony and Microsoft made a mistake by going with AMD.
How does Polaris stack up to Pascal?
How does Polaris stack up to Pascal?
I wonder if Sony and Microsoft made a mistake by going with AMD.
Dev kits already shipped out, VR this fall... I'd say Neo this year for sure.
Polaris 10 is a 390 replacement and Polaris 11 is some low end card. AMD haven't got dick. Fury was merely playing catchup to Maxwell and Pascal just leapfrogged their less than a year old flagship card.
It does, except that if they pulled it off, it could possibly make backwards compatibility for whatever the PS5 ends up being much easier.
What kind of BS answer is this? We don't know anything about Polaris yet.
We have estimates from Polaris 10 and 11, which are the two PC GPU's set to release in 2016. We know they are intended to be the mainstream products, with an enthusiast part scheduled for sometime in 2017.
How that relates to what Sony is doing is completely unknown. It's not one of the retail parts.
Polaris 10 is a 390 replacement and Polaris 11 is some low end card. AMD haven't got dick. Fury was merely playing catchup to Maxwell and Pascal just leapfrogged their less than a year old flagship card.
It would be really STUPID from AMD to release a new 300USD card that performs like an old 300USD (390)
Polaris 10 is around 230mm^2 and 130+ Watts based on rumours. If that only performs at the 390 level then the claimed 2.5x perf/watt figure and the approximate 2.2x perf/mm figure are both bogus and AMD has utterly failed.
The last 200mm chip AMD released on a new node was around 20% faster than their previous ~390mm chip. There is no reason that trend won't be followed again and the other two metrics also indicate that level of performance.
So unless the die is smaller than the rumours and the power usage is lower than the rumours I don't think it is possible for the top card to only be as fast as the 390.
Polaris 10 is around 230mm^2 and 130+ Watts based on rumours. If that only performs at the 390 level then the claimed 2.5x perf/watt figure and the approximate 2.2x perf/mm figure are both bogus and AMD has utterly failed.
The last 200mm chip AMD released on a new node was around 20% faster than their previous ~390mm chip. There is no reason that trend won't be followed again and the other two metrics also indicate that level of performance.
So unless the die is smaller than the rumours and the power usage is lower than the rumours I don't think it is possible for the top card to only be as fast as the 390.
Since Hitman is out, do we know what is needed from old GPU hardware to push it to 1440p60 with DX12? Polaris 10 managed to render that.
How does Polaris stack up to Pascal?
I wonder if Sony and Microsoft made a mistake by going with AMD.
polaris 10 : 130w 230mm²
r9 390 : 275W, 438mm²
it really isn't far off.
Especially if amd is being cheeky with their marketing and using peak power draw for comparisons or something because then you're looking at 330+W power draw for an r9 390
That said : I still have some hope that they'll be able to approach fury x performance from pure clockspeed gains despite the tiny die size
ARM, Qualcomm, Samsung or Imagination Technologies most likely. Mobile tech can be scaled up.I wonder who Sony/MS/Nintendo turn to if AMD were to go bankrupt.
No one. Console gaming will be done.I wonder who Sony/MS/Nintendo turn to if AMD were to go bankrupt.
AMD was basically the only choice. Going nVidia would have meant having to use an ARM CPU, and going Intel wouldn't have had enough graphical horsepower. Not to mention that licensing from either wouldn't have been cheap/easy, and AMD was desperate enough that it would have under-bid them even if they had made offers. As long as AMD doesn't fold before the next set of consoles comes out, they made the right choice. And even if they did fold, it would just mean they'd have a tougher time with the PS5/Xbox Two.
No one. Console gaming will be done.
Intel and Nvidia either won't play ball or will be prohibitively expensive, PowerVR currently into mobile or HPC.
Nvidia in particular will sue anyone else that tries to do performance GPUs.
What would be wrong with an ARM CPU?
The 390X also has the same die size as the 390 and performs better. I tend to think using a similar size chip is better so something like R9 270.
The 2.5x perf/watt was not from a marketing slide iirc.
AMD has high debt load, no one will buy into getting brutally sandwiched by Intel and Nvidia (two legendarily competitive companies), console business is not worth this type of capexCouldn't Nintendo or Microsoft just use their Billions of dollars in cash and Buy AMD out and make their own Hardware in house?
So the Neo is a PS5 in everything but name only.
Combine that with a short lifecycle (3years for the ps4) and Next gen I will definitely be looking at another platform to game on. Nintendo + a Windows PC here I come
So the Neo is a PS5 in everything but name only.
Combine that with a short lifecycle (3years for the ps4) and Next gen I will definitely be looking at another platform to game on. Nintendo + a Windows PC here I come
The article reminds me of Jeff Rigby threads
So the Neo is a PS5 in everything but name only.
Combine that with a short lifecycle (3years for the ps4) and Next gen I will definitely be looking at another platform to game on. Nintendo + a Windows PC here I come
I wonder who Sony/MS/Nintendo turn to if AMD were to go bankrupt.
About that cheaping out, I'm not sure how much performance bump that, say, a $100 BOM extra will buy with the available tech and TDP circa 2013.LOL at die shrinking to 14nm Finfet being pricier than buying a better GPU. It was almost like both companies were forced into upgrading their consoles because they cheaped out on hardware at launch. I feel for early console adopters..No, I don't.
LOL at die shrinking to 14nm Finfet being pricier than buying a better GPU. It was almost like both companies were forced into upgrading their consoles because they cheaped out on hardware at launch. I feel for early console adopters..No, I don't.
true but the 390x also has even higher power consumption.
A 2x gpu upgrade is not something i would call a new gen.
Clearly not.It would cause Sony/MS/Nintendo huge problems with BC & Forward Compatibility after finally gaining this luxury by going x86/64.
Why not switch now if it upsets you so much? You just gonna ride out the next 2-3 years of inferior ports(this is pessimists view not mine)? Funny how no seems to have problem with Nintendo basically ending the Wii U after 3 years even it was because of shitty sales.
I bought the Wii U at launch, and have bought a few games every year of its lifecycle. Most of which I have loved. So for me 5 years and a total of 10-20+ great exclusive titles has really renewed my faith in Nintendo after the Wii. I think the Wii U taught Nintendo a lot and I'm completely confident that whatever comes next will be something that gives me personally good value for money.
In comparison I've had a PS4 for close to 2 years and it's spent most of that time in its box. I own a couple of games, mostly found in the bargain bin, and nothing that has made owning a PS4 worth owning compared to any other machine that plays 3rd party games. For the most part the PS4s lineup has felt redundant, even including 3rd party games and poor value for money. Especially if you owned a PS3. Which had a really great run late in its lifespan. Why buy a Neo when first gen VR will suck and the lineup of games is mostly empty promises, that will just be ported to the PSNext and will be cheaply and easily accessible to a late adopter.
Why not just sell up and buy a new PC now?
Money. Building a good PC can be expensive and with Polaris, Zen and Vulcan all about to launch I'd rather wait until next year and optimize a build that takes advantage of what the new tech does, and at the very least it should reduce costs of parts that can current games at 1080p/Ultra or Even Downsampled 4K
A 2x gpu upgrade is not something i would call a new gen.
Point is to choose a card where perf/watt and perf/mm2 scaling gives a similar answer as it is more likely to be correct. R9 270 does this and seems to indicate somewhere around Fury X, might be a bit lower but it is probably ahead of the 390X.
I can imagine a cut version of P10 at 390 perf levels though.