Scaling is 151% of a single card.
Shintai, you are nitpicking to prove your agenda. Nothing else. The affects your credibility.
From the history, Nvidia usually understate their TDP while AMD usually overstated it.
Its 250W TDP.
Credi-what now?The affects your credibility.
AMD's first Polaris-based graphics card is here: the Radeon RX 480. Rather than launch a high-end card to compete with the likes of Nvidia's GTX 1080 or 1070, AMD's RX 480 is pitched at the wider mainstream market, offering just over five teraflops of performance for a mere $199—about half the price of a GTX 1070. UK pricing is currently TBC, but it'll probably be about £160. The RX 480 will be available to buy on June 29.
"In the past we used to reference our competition, but now it's just about us."
Details on the Polaris architecture—which is based on AMD's fourth-generation GCN architecture and a new 14nm FinFET manufacturing process—were thin on the ground during the RX 480's reveal at Computex 2016 in Taiwan, but the company did divulge a few key specs. The RX 480 will feature 36 compute units (CUs)—that's eight more than the R9 380 and just shy of the 40 of the R9 390—along with some fast GDDR5 memory attached to a 256-bit memory bus for 256GB/s of bandwidth.
The RX 480 will come in both 4GB and 8GB configurations (the former being the £160/$200 model) and will support AMD FreeSync and HDR video via its DisplayPort 1.3/1.4 and HDMI 2.0b outputs. Best of all, it has an average power draw of just 150W, which should make it cooler and quieter than AMD's previous-generation graphics cards.
When it comes to the matter of performance, AMD isn't quite ready to give up exact numbers. Instead the company says the RX 480's "VR capability" compares to that of £400/$500 GPUs. Indeed, VR is a key part of AMD's pitch for the card, with the company hoping to "jumpstart the growth of the addressable market for PC VR" and "accelerate the rate at which VR headsets drop in price" across both desktops and laptops.
While it's unlikely a £160/$200 graphics card from any company is going to help speed up price drops on VR headsets, the RX 480's specs are appealing for the price and in theory would put its performance somewhere around an R9 390 or Nvidia's now discontinued GTX 980, depending on the RX 480's clock speed. Considering all of those cards sell for well above £250/$300, that's impressive.
That said, the tech heads amongst you may have noticed that Nvidia is squeezing a lot more out of its Pascal architecture than AMD is seemingly getting out of Polaris. The GTX 1070 sports the same TDP of 150W, yet according to recent reviews offers performance just slightly above that of the GTX 980 Ti and Titan X. After all the power efficiency and performance-per-watt improvements touted by the move to 14nm, that's a wee bit disappointing.
But for the mainstream crowd that couldn't give a monkeys about TDP and process improvements, a £160 graphics card that plays high-end games at 1440p on the desktop and spits out VR visuals without fuss is a very attractive proposition—one that Nvidia simply doesn't have a response to right now. Here's hoping that RX 480 can deliver when it's released on June 29.
Good post.
For better or worse, AMD hitched their higher-end options to HBM/HBM2. Unfortunately that meant a limit of 4GB on their Fury line (although it performed pretty well regardless) and is a prerequisite for Vega. On the positive, it gave some nice efficiency gains and worked well.
NV went a different route and mitigated somewhat by sticking with high-speed GDDR5, using GDDRX and only using HBM2 for P100 (so far). GP102 likely will have the option for HMB2.
Because of this, NV likely released mid/mid-high first and will follow with lower-end and the high, high end later. AMD went with mid/mid-low and is waiting to add the high end with HBM2 (most likely).
Different strategies, but both can work. It is important to not focus too much on the technology solution, but rather performance and constraints (if applicable). Sometimes I think the forum gets so set that 'only one company can be right'. It is not always that way...every decision has it's pluses and minuses.
I wasn't the one claiming AMD was under TDP while Nvidia is over. I simply said it wasn't true. And that's a fact
Its 250W TDP.
Or what do you think RX 480 power consumption will be, since the 150W TDP number from AMD must be meaningless?
Nope, It's an outdated NDA slide. It's 290W according to final presentation.Its 250W TDP.
No kidding!
The author and contributors obviously don't understand the market play here. They applaud AMD a few weeks ago for console deals (lower ASP vs. P10 for sure) and then discount this? They just don't get it...
If P10 doesn't sell well, AMD cannot blame themselves! They made a great card at a great price point...
A lot of these contributors have personal interests and bias. They could either be Nvidia investors or they could be playing the market by shorting AMD. Or simply being paid to shill and cause market hysteria.
That article in particular was timed perfectly to try and cause an AMD panic sell of. And even though AMD stock is down by 3% for the day, considering the market was in turmoil today, it doesn't look like it worked.
AMD's stock seems to holding its new gained ground well.
Mark Hibben for instance is another shill that timed his article around the same time: http://seekingalpha.com/article/3979206-amd-exposes-lack-competitiveness-computex
This guy has a history of downplaying any positive AMD news. He did it when the Chinese JV deal was announced, and he does it now. Every single one of his AMD articles has a very negative spin. And surprise he's an Nvidia investor.
I think the reason AMD does not show benchmarks yet, is that they are still improving the driver to gpu performance.
They have a pretty large gaming catalogue to test and optimuize the driver settings for each game, i assume.
So it makes sense that in the light that a driver improvement for doom alone delivered up to 35% increase in performance,
that AMD will make sure that a sloppy driver does not make the RX480 look extra ordinary bad.
The elephant in the room:
Ashes uses procedural generation based on a randomized seed at launch. The benchmark does look slightly different every time it is run. But that, many have noted, does not fully explain the quality difference people noticed.
At present the GTX 1080 is incorrectly executing the terrain shaders responsible for populating the environment with the appropriate amount of snow. The GTX 1080 is doing less work to render AOTS than it otherwise would if the shader were being run properly. Snow is somewhat flat and boring in color compared to shiny rocks, which gives the illusion that less is being rendered, but this is an incorrect interpretation of how the terrain shaders are functioning in this title.
The content being rendered by the RX 480--the one with greater snow coverage in the side-by-side (the left in these images)--is the correct execution of the terrain shaders.
So, even with fudgy image quality on the GTX 1080 that could improve their performance a few percent, dual RX 480 still came out ahead.
As a parting note, I will mention we ran this test 10x prior to going on-stage to confirm the performance delta was accurate. Moving up to 1440p at the same settings maintains the same performance delta within +/-1%.
For one gpu, and that's the only case where hawaii was at the perfromance level of a tonga chip so don't get your hopes high with drivers alone.
I want to wait for reviews, however... It Appears that Polaris 10 may be very close to AMD Fury GPU.
3dMark 11 benchmarks indicate that, and presentation on Doom.
This is freakin 199$ GPU... O_O