MD unveiled its new Radeon RX 5700 line of graphics cards with 7nm chips at E3 last month, and with just days to go before they launch on July 7th, the company has announced new pricing. In the "spirit" of competition that it says is "heating up" in the graphics market -- specifically NVIDIA's "Super" new RTX cards -- all three versions of the graphics card will be cheaper than we thought.
The standard Radeon RX 5700 with 36 compute units and speeds of up to 1.7GHz was originally announced at $379, but will instead hit shelves at $349 -- the same price as NVIDIA's RTX 2060. The 5700 XT card that brings 40 compute units and up to 1.9GHz speed will be $50 cheaper than expected, launching at $399. The same goes for the 50th Anniversary with a slightly higher boost speed and stylish gold trim that will cost $449 instead of $499.
That's enough to keep them both cheaper than the $499 RTX 2070 Super -- we'll have to wait for the performance reviews to find out if it's enough to make sure they're still relevant.
AMD fires back at 'Super' NVIDIA with Radeon RX 5700 price cuts
Yeah clearly when you work with Blender, Updates are not important. So unimportant that AMD are happy for their Polaris GPUs to run slower or not even be able to complete a benchmark at all. 
T77979 Blender Open Data Benchmark "Victor" fails on AMD Polaris GPUs. T77095 Artefacts when rendering with the GPU in Blender 2.90.2 and 2.83 release
colesdav wrote:
Yeah clearly when you work with Blender, Updates are not important. So unimportant that AMD are happy for their Polaris GPUs to run slower or not even be able to complete a benchmark at all.
I do not use Blender for benchmarks
I think nVidia may be going the route that all the cards will only be offered in one memory capacity aside from the lowest end, so there won't be a price premium for a memory capacity bump.
If the board IDs are going to be used to designate different models, PG142-10 may be akin to the vanilla 2070 and PG142-2 may be akin to the 2070 Super, same take with the PG132 boards, 15% performance differences between the two before memory is taken into account.
Also today nVidia started the countdown timer until Ampere's (likely soft) launch: August 31st. nVidia is really good at hyping things up, and they've built up enough of it over the last few months to build it into a monster. AMD's going to have to do SOMETHING to try and steal some thunder...
https://wccftech.com/nvidia-teases-next-generation-ampere-gpus-begins-countdown-to-august-31/
More news out of the green camp today courtesy of Micron not being able to keep their mouths shut. The RTX 3090, or whatever nVidia will call it, will ship with GDDR6X and will have over 1Tb of bandwidth, nearly the same as the Radeon Instinct MI50, and they are developing the double capacity chips for use, presumably, in the RTX 3000 series based Titan or RTX 3090 Ti, whatever nVidia will call it.
Do note how there is NO "Big Navi" RDNA2 card listed on the application type...
https://www.tomshardware.com/news/micron-confirms-rtx-3090-will-have-over-1-tbs-gddr6x-bandwidth
More news out of what I guess now will be Team Blue from their Hot Chips presentation. Assuming their 2T (2 tile) performance is being rated the same way AMD rates theirs, can't go with nVidia because they use a different rating, it is looking as if Intel Xe-HP will launch with about the same power as the 5700 XT, or maybe even more since these are still engineering samples. Depending on their launch price, as well as AMD's Navi Refresh prices, consumers could be looking at a very, very competitive market for cards the power of the 2070 Super, since it's not difficult to believe Intel is going to come in very low (relatively) to sap market share, and AMD cannot afford to leach what little market share they have (relative to nVidia) for another generation to Team Blue or Green, especially since it's going to be 2022 or 2023 before RDNA3 and its "revolutionary" design hits the shelves. I'll go out on a limb and say 2070 Super levels of performance for $250-$300 from both Team Red and Blue, with Team Green banking on Ampere's higher compute performance and being $50-$100 higher. That's great for the consumer, a price war with very powerful cards, easily 2560x1440 60hz or 1920x1080 120hz capable in most titles, along with much greater ray tracing capabilities than Turing.
The article is mostly about their 4 tile part which could bring a petaflop of compute power, as it contains tensor cores, and be quite the benefit for datacenters where space is at a premium.
https://www.tomshardware.com/news/raja-koduri-petaflops-scale-4-tile-xe-hp-gpu-hot-chips
Personally I think the RTX 3090 will end up being a Titan model, since nVidia really is up the creek with the Titan naming convention, so designating it as the RTX xx90 makes a lot of sense. Given how Threadripper has made HEDT so much more affordable and accessible, the market for a Titan level GPU is growing, especially since Covid-19 might permanently shift some jobs to work at home. Also given the fact it's supposed to pack 24GB VRAM lends credence to this fact, since I haven't even heard any memory issues at 4K with an 11GB 2080 Ti, and there will supposedly be a RTX 3080 20GB model as well.
If that's the case then it really does make sense at that price level given its target audience, and I would expect Ampere price levels to be about $100-$150 more than current Turing prices ASSUMING AMD and Intel decide to play along with the price fixing and keep prices high, but that's not in either of their best interests, both because of expensive new GDDR6X, but also because they're going to be the only ones on the market for at least two months, and it wouldn't make sense for them NOT to do it. Right now they're still having to target the GTX 1000 series and earlier crowd for upgrades given the high prices to low performance increases for Turing, and we know from years back when AMD dropped the bomb with the HD 4870 that nVidia will not hesitate to slash prices drastically overnight if RDNA2 turns out to be 50% faster than the 2080Ti as initially rumored. The Steam Hardware Survey backs this up to an extent, with non-RTX cards having a much larger share than Turning (and Navi having a 0.87% share with the 5700 XT, nothing higher), and reflects what most of the reputable review sites have said since Turing launched, that if you already had a GTX 1000 series there was little reason to upgrade.
In any case the more I think about it, it's going to be up to Intel to set the prices. nVidia can do whatever they want for 2 months and they're going to set the bar crazy high, AMD likely will have RNDA2 cards that will be able to compete with at least the RTX 3070 and will likely price match since they will have ray tracing, but Intel we know from their Hot Chips presentation will tap out at RTX 2070 Super levels of performance with their 2T Xe-HP, and being fresh to the market, and considering their CPU division isn't going to get back on track until about this time next year, they need to score a win, and like a fresh contender into any market who lacks products to swing with the upper levels, they'll come in low and attract a following which will force AMD to lower their prices since they're shortstacked against nVidia 8:1 essentially in share and no doubt ticking off loads of Navi owners with persistent issues, which will force nVidia to -slightly- lower their prices since they do have tricks up their sleeve, such as DLSS 2.0, which neither AMD nor Intel will be able to match.
The wildcard, if there is one, are the new consoles. Not only are they a trickle of revenue AMD can use to offset lower margins on GPUs, but they may start causing a shift back to consoles if they truly are as powerful as they are being hyped up to be. If they do then that's going to hit the midrange market the hardest which will force lower prices all around.
Navi and Adrenalin 2020 GUI/UI and Drivers was clearly a big success:
NVIDIA GeForce GPUs Gain Major Market Share Versus AMD Radeon in Q2 2020
I cannot even get Adrenalin 2020 20.8.3 to install without BSOD on non overclocked systems.
Hybrid Adrenalin 2019 19.12.1 + 20.8.3 drivers via Windows Update works o.k.
Example:
What are AMD doing?
Why did they move from the above GUI/UI?
Any more than that and they won't attract any upgrade crowd. No matter how powerful it is it's still a mid range card that won't be able to do 4K in new games, and still have to take detail hits in some of the more recent games to hit 60fps.
Even at 50% faster than a 2080Ti it still puts in sub-60 FPS performance on Microsoft Flight Simulator at full details at 4K. Granted it's a bit of an extreme case, but it does seem to be a game which is very well optimized, though high clock speeds definitely help in the overall average vs core count, vs something like Horizon Zero Dawn, and is clearly the direction games are going to be heading in in the near term given consoles will be outputting 4K.
Even though custom models will likely push the $600 mark, it's still going to put AMD behind the 8-ball unless they return to Polaris levels of pricing, since even a card which matches 2080 Ti performance at $300 will be an epic seller, assuming of course AMD sorts out their software issues...
