Graphics Cards - Page 175

Sony still hasn't given us much details on the inner workings of its next-gen PlayStation 5 console, but a new patent Sony filed in 2019 is interesting.

The new patent was filed in 2019 but only published a few days ago on July 23, and discuses image reconstructing technology that uses reference images through machine learning. You know, kinda like how NVIDIA's own Deep Learning Super Sampling -- or DLSS, works.

This is exciting as NVIDIA DLSS 2.0 provides double the performance in games that support it, while not looking any worse -- and in some cases they look better.

NVIDIA has just released a new hotfix driver, something that fixes some of the issues that were found in the recent GeForce 451.67 WHQL Game Ready drivers.

The new GeForce 451.85 drivers solve some problems in Hideo Kojima's masterpiece Death Stranding, something that I checked out in 8K with DLSS 2.0 on with some mind-blowing results.

Not just that, but the new hotfix drivers provide stability to Shadow of the Tomb Raider when you're running it in DX12 with hardware-accelerated GPU scheduled enabled. Death Stranding gets some texure corruption problems fixed on GeForce RTX 20 series and GeForce GTX 16 series graphics cards.

NVIDIA has its new Ampere GPU architecture powering the new benchmark crushing Ampere A100 accelerator, and soon to be Ampere-based GeForce RTX 3000 series -- AMD has its RDNA 2 architecture right around the corner and APUs inside of next-gen PlayStation 5 and Xbox Series X... while Intel has, well lots of pluses -- and no GPU.

Intel is expected to finally have something to share on its Xe graphics according to a quickly-deleted tweet. Intel posted the tweet to its official Intel Graphics Twitter account, where it said: "You've waited. You've wondered. We'll deliver. In 20 days, expect more details on Xe graphics".

Keep in mind that Intel has a few events it will be (virtually) attending, with David Blythe, Senior Fellow and Director of Graphics Architecture at Intel delivering a speech at Hot Chips on August 17 all about the Xe GPU architecture.

NVIDIA announced its Ampere GPU architecture with the introduction of the Ampere A100 accelerator earlier this year, the company's first 7nm GPU -- and also its first PCIe 4.0 card too. But what good is a next generation GPU if it hasn't been benchmarked yet?

This is where Jules Urbach, the CEO of OTOY -- a cloud graphics company famous for its Octane Render software, has benchmarked NVIDIA's new A100 accelerator. The new Ampere-based NVIDIA A100 accelerator was benchmarked on OctaneBench -- a benchmark designed to test the performance in OctaneRender.

OTOY benchmarked the GA100-based card which consists of 6912 CUDA cores, and 40GB of super-fast HBM2 memory.

I thought it was known from previous reports, but yeah -- NVIDIA will be unveiling its new Ampere-based GeForce RTX 3000 range of graphics cards in August, with a September launch into gamers' hands.

The news is coming once again from Overclocking.com and other sources, but it's something that I revealed back in March 2020 in an article titled "NVIDIA GeForce RTX 3080: August 2020 reveal, launch at Computex 2020" which you can read here.

Videocardz points out that there were people after me saying that an August reveal and September launch would happen, and now even more outlets are reporting it so it seems like we're ramping right into a September launch for NVIDIA's next-gen GeForce RTX 3000 series graphics cards -- exciting times, people!

I have a feeling this is how it's going to be for the next few months: leak after leak, drip after drip of information on AMD and NVIDIA's next generation graphics cards.

The first one I'm writing today is some new details on AMD's upcoming Big Navi aka RDNA 2 but aka as an NVIDIA killer. AMD's upcoming flagship Big Navi graphics card is expected to make its big debut with a huge 16GB of VRAM -- but we don't know if it's GDDR6, or HBM2 / HBM2e memory.

AMD's new flagship Big Navi card should have 16GB of VRAM on a 384-bit bus according to leaker 'Wjm47196' suggesting AMD will go with GDDR6 over HBM2 or the newer HBM2e standard.

We have another new rumor to share on Ampere GPU performance, this time with the GeForce RTX 3080 and insider KatCorgi on Twitter who tweeted:

This about lines up with previous estimates, with the new rumor suggesting the GeForce RTX 3080 is around 20% faster than the GeForce RTX 2080 Ti. There is nothing specific to go from here, but it lines up with the 40-50% upgrade that the RTX 3080 Ti / RTX 3090 is meant to bring over the RTX 2080 Ti.

If it's a raw 20% improvement over the RTX 2080 Ti without running magic DLSS 3.0 technology or something, then that will be a nice upgrade over the RTX 2080 Ti -- except I'm sure the RTX 3080 will deploy with 10GB of GDDR6 (and a much higher TDP at 300W or over).

A bunch of hot RDNA 2 information dropped today from Tom on Moore's Law is Dead, with his sources telling him that AMD's next-gen Big Navi will offer up to 225% more performance over RDNA... the GPU that powered the Radeon RX 5700 XT.

According to these sources estimates are that RDNA is a "major leap forward for them 195% to 225% of the current available cards. But their internal estimates are still projecting the performance per dollar cards up to the upper end of the range".

Another tease is that the "top RDNA 2" would feature 72 compute units, which is where the leaks on the flagship RDNA 2 graphics card being 40-50% faster than NVIDIA's current-gen flagship GeForce RTX 2080 Ti graphics card.

We heard about NVCache back in May 2020 with some Ampere leaks, with NVIDIA's purported NVCache acting like HBCC -- HBCC being AMD's own High Bandwidth Cache Controller that debuted with Vega.

Sony is using some truly next-gen storage technology inside of its PlayStation 5 console, with a deep dive on that here. Microsoft will be doing the same with their Xbox Series X console, but what about the PC? AMD will surely have something nifty up their sleeve with RDNA 2 -- while NVIDIA has NVCache.

Well, back in the May leaks about Ampere and NVCache we heard that it "leverages both your DDR & SSD for enhanced load times & VRAM". The new leaks from Moore's Law is Dead tease that NVIDIA's upcoming NVCache on Ampere cards in the GeForce RTX 3000 series family will "dynamically utilize bandwidth from SSDs, VRAM, and DDR for multiple tasks at the same time".

NVIDIA has some near GPU cheat codes when it comes to its DLSS technology (Deep Learning Super Sampling), with DLSS 2.0 providing double the performance in Hideo Kojima's Death Stranding.

But according to Moore's Law is Dead, DLSS 3.0 is going to be an even bigger game changer. In his latest video, Tom talks about some of the other Ampere GPU features (you can read more on the 3-4x ray tracing performance on the GeForce RTX 3080 Ti right here) and one of them is DLSS 3.0 technology.

DLSS 3.0 will reportedly "work in any game with TAA" but it will require a Game Ready driver to do so, meaning developers will have to do some "specific programming per game to get it to work, but it should be easier than before".

AMD is having one of the best years in its history, with the successful launches of its new Zen 2-powered CPUs in the Ryzen 3000 series desktop CPUs, Ryzen 4000 series mobile CPUs, new Ryzen Threadripper PRO chips, new EPYC processors, and so much more.

But the GPU dominance hasn't started just yet, and while AMD might have shaken the GPU tree with the RDNA-powered Radeon RX 5000 series and its flagship Radeon RX 5700 XT -- we ain't seen nothing yet. The new rumors from Moore's Law is Dead suggests we should expect AMD to outright dominate 2020 in more ways than one.

We are already expecting a huge 40-50% increase in performance from the flagship RDNA 2 card over the GeForce RTX 2080 Ti, which would make the -- let's call it the Radeon RX 6700 XT -- twice as fast as the Radeon RX 5700 XT.

The rumors for NVIDIA and AMD's next-generation GPUs is getting hot and heavy, with Moore's Law is Dead posting a new video with some updates from his sources for July 2020. Check it out:

There's a lot to cover with this video, with a July 2020 update confirming that we're still to expect a huge 3-4x performance upgrade in ray tracing performance with the Ampere GPU architecture over the current-gen Turing GPU architecture.

If this is true, the next-gen Ampere-based GeForce RTX 3000 series cards will indeed makes "Turing look like Kepler trying to run DX12". This is some exciting stuff, as the next wave of games coming through next-gen consoles that have hardware ray tracing thanks to RDNA 2 -- it means NVIDIA wants to go for the throat when it comes to ray tracing performance dominance.

Xiaomi has just unleashed its new 34-inch 21:9 UltraWide gaming monitor, in the new Mi Curved Gaming Monitor.

Xiaomi's new Mi Curved Gaming Monitor features a 34-inch panel with a native 3440 x 1440 resolution, and super-slick 144Hz refresh rate. We have a 4ms response time for gamers, and a 1500R curvature for the 34-inch 21:9 aspect ratio panel.

The company is going with a barely-there bezel design, which also looks really nice as well. We have 121% coverage of the sRGB color space, 300 nits of brightness, and a 3000:1 contrast ratio.

AMD has just released its new Radeon Software Adrenalin drivers, with the new 20.7.2 beta providing optimizations for Hideo Kojima's masterpiece Death Stranding, which just dropped on the PC.

There's an additional 8% performance using a Navi-based Radeon RX 5700 XT graphics card with these new Radeon Software Adrenalin 20.7.2 beta drivers in Death Stranding. F1 2020 also gets some optimizations on the Ultra High preset with the Radeon RX 5700 XT with 5% more performance.

Ubisoft just launched its new battle royale game Hyper Scape, with the new Hyper Scape beta recdeiving support from AMD inside of the new Radeon Software Adrenalin 20.7.2 beta drivers.

NVIDIA is expected to have many large design changes over previous-gen graphics cards, and not just design changes over previous-gen GeForce graphics cards -- but an entire new power connector.

A new rumor has surfaced from Chinese tech site FCPOWERUP, which claims NVIDIA will be using a new power connector on its next-gen Ampere GeForce RTX 3000 series graphics cards. I have heard something similar from sources, but thought it would come down to something different than this.

It looks like we're expected to see a strange new 12-pin PCIe power connector on the reference GeForce RTX 3000 series graphics cards. Not just that, but power will exceed 300W -- so with a 12-pin PCIe power connector and radically new design cooling (fans on the front, and back according to previous reports) big changes are coming with Ampere.

The spotlight has been on NVIDIA's next-gen Ampere GPU architecture and the upcoming GeForce RTX 3000 series graphics cards for a while now... so it's now time for AMD's next-gen RDNA 2 architecture to enjoy that light.

A new rumor for Big Navi is teasing some truly huge performance, with YouTuber Moore's Law Is Dead saying that AMD's flagship RDNA 2-based graphics card will be 40-50% faster than NVIDIA's current flagship GeForce RTX 2080 Ti graphics card.

Big Navi being 40-50% faster than the RTX 2080 Ti is a huge, huge deal.

Ubisoft detailed Watch Dogs: Legion recently, where the London-based open world game was run on the flagship NVIDIA GeForce RTX 2080 Ti graphics card by the good folks at Digital Foundry. Check it out:

In their hands-on time with Watch Dogs: Legion, Digital Foundry may have been using the most powerful graphics card you can buy in the GeForce RTX 2080 Ti -- yet it couldn't handle the game at 1080p and anything over 30FPS. Ubisoft had locked Watch Dogs: Legion's preview build to 30FPS, for some stupid reason.

Nothing Digital Foundry could do would bump it above 30FPS into the much-more-playable 60FPS region. The outlet notes Ubisoft is using SSR (Screen Space Reflections) on certain types of surfaces, like rivers, etc but they're using SSR in the preview build of Watch Dogs: Legion -- but this might change in the final build with ray traced reflections.

We all remember how insane the first wave of cryptocurrency mining was, with graphics cards sold out across the world and gamers being very, very mad about it.

Well, it looks like we could be preparing to see a second wave hit us very soon. The folks at MyDrivers are reporting that NVIDIA send out an internal notice to their AIB partners telling them to expect a boost in demand as cryptocurrency mining is growing, again.

We also just heard that NVIDIA is reporting ending their GeForce RTX 2080 Ti, GeForce RTX 2080 SUPER, GeForce RTX 2070 SUPER, and GeForce RTX 2070 -- all leading into the warm, welcoming arms of the Ampere-based GeForce RTX 30 series which should drop in September.

NVIDIA is reportedly discontinuing a bunch of its highest-end GeForce RTX series graphics cards based on the Turing architecture.

The new rumors come from ITHome, which is reporting NVIDIA is going to discontinue and delist the following graphics cards: GeForce RTX 2080 Ti, GeForce RTX 2080 SUPER, GeForce RTX 2070 SUPER, GeForce RTX 2070.

I've been noticing that it's been harder and harder to source GeForce RTX 20 series graphics cards on Amazon as it is, and checking today was no different. There's not many GeForce RTX 20 series cards in stock on Amazon right now, and it has been an issue for weeks and even a couple of months now.

NVIDIA has just released their latest GeForce 451.67 WHQL drivers that are optimized for Death Stranding, Hideo Kojima's masterpiece that was until recently a PlayStation 4 exclusive.

Death Stranding isn't the only PlayStation 4 exclusive getting some PC lovin', with Horizon: Zero Dawn about to launch and optimizations baked into these new drivers, too. We also have optimizations for F1 2020 which also just launched.

NVIDIA also added in some new G-SYNC compatible monitors in the new GeForce 451.67 WHQL drivers, with the following monitors now getting Variable Refresh Rate (VRR) on NVIDIA GeForce graphics cards: