Graphics Cards - Page 174

A fresh new rumor on AMD's next-gen RDNA 2 aka Big Navi GPU is here, with a new post on Chiphell suggesting AMD will have two different Navi 21-based cards.

The new report suggests AMD will have Big Navi GPUs with both 12GB and 16GB of GDDR6, but AMD could possibly use the full fat Navi 21 GPU with the huge 80 Compute Units which means we'd be seeing 5120 stream processors.

We only just recently heard that AMD's big bad Big Navi "NVIDIA Killer" flagship card had 16GB of VRAM.

NVIDIA announced the "Ultimate Countdown" towards its next-gen Ampere GeForce graphics cards, but then there were the hints using numbers and more specifically: 21.

After posting the #UltimateCountdown tweet, NVIDIA added: "21 Days. 21 Years" which signifies August 31 (21 days after the countdown was announced on August 10) and it has been 21 years since the first GeForce, the first-ever GPU, the GeForce 256.

So much around the number 21 that if you take '256' and times it by '21' you get 5376, which is the purported number of CUDA cores on NVIDIA's upcoming GA102 GPU.

NVIDIA found itself in a spot of trouble earlier this year, when reports surfaced that NVIDIA wanted to "leap frog" AMD in order to "prevent them" from getting bigger.

What happened? NVIDIA reportedly "underestimated the effect of AMD embracing TSMC, and made some errors in its own plans for migrating to advanced nodes. Moving to Samsung and wavering in its TSMC strategy, only to be go back to TSMC later, NVIDIA was unable to secure enough 7nm capacity, resulting in AMD grabbing the limelight, eroding NVIDIA's brand value, and increase its own market share for GPUs".

Well, that might be the case this year -- where AMD's next-gen RDNA 2-based 'Big Navi' graphics card possibly toppling the best Ampere-based GeForce RTX 30 series graphics card (more on that here). At least, for this year -- but don't underestimate NVIDIA and its unrelenting thirst for being the best.

Oh boy, a super juicy rumor is sizzling like chili on my tongue -- NVIDIA could software lock the GeForce RTX 3080 and its VRAM at launch, but why?

Because NVIDIA want to see how good the flagship RDNA 2-based graphics card is, that's why. NVIDIA could hold back and gimp the GeForce RTX 3080 with its purported 12GB of VRAM on a 384-bit bus, down to just 10GB on a 320-bit bus.

Once AMD details its flagship RDNA 2 graphics card, then NVIDIA would -- at least according to Moore's Law is Dead's sources -- un-gimp its card and unleash the proper 12GB of VRAM on its original 384-bit memory bus. Tom does note that there is a "slim chance" of this happening, but still the very suggestion this could happen is actually interesting.

We have had some incredibly juicy rumors to deal with in the last few days, and one of those is about the RDNA 4-based Navi 41 GPU from AMD.

Komachi has tweeted out some delicious tidbits on the purported the Navi 41 GPU, simply saying "there is Navi 41". We don't know anything about Navi 41, but if we connect some of those dots from what Navi 10 was, what Navi 21 will be very soon in RDNA 2 aka Big Navi aka Navi 2X -- and then after that the super-exciting RDNA 3 architecture.

AMD's next-gen RDNA 2 architecture will power the new Navi 21-powered Radeon RX 6000 series graphics cards, but after that will be RDNA 3 with the Navi 31 GPU. Komachi reports: "NV3X GCD/MCD means probably Graphics Complex Die/Memory Complex Die of Navi 3X (Navi 31)".

The rumor mill is churning for NVIDIA's next-gen Ampere-based GeForce RTX 30 series graphics cards, with fresh info stating the flagship card will feature a huge 24GB of VRAM.

NVIDIA is rumored to be launching 3 different variants based on the PG132 reference board, which should fill out the 3 higher-end GeForce RTX 30 series graphics cards. This is most likely going to be something like the GeForce RTX 3080, GeForce RTX 3080 Ti, and GeForce RTX 3090.

These PG132 boards with 24GB of VRAM isn't new, with rumors from June 2020 stating that the flagship GeForce RTX 3090 would feature the GA102 GPU, 350W TGP, and 24GB of super-fast GDDR6X memory.

The last few months have been some of the most insane months of the GPU industry, between NVIDIA's upcoming Ampere launch and AMD's next stomp with RDNA 2 and Big Navi -- there's no time for a break.

But then a source reached out to me a few days ago and told me something very interesting, something that I think could be a real game changer for NVIDIA's big launch of Ampere-based GeForce RTX 30 series graphics cards: custom models will be launching alongside the Founders Edition models... or at the very least, very shortly after the FE cards ship.

I had a source tell me that custom AIB models will be on offer starting in September, so with current rumors of an Ampere GeForce RTX 30 series announcement on September 9 and then a purported launch on September 17 -- custom cards being made available "in September" would be huge.

I'm currently on-board the Ampere Hype Train right now, with NVIDIA officially kicking off the marketing campaign for its next-gen GeForce graphics cards.

As a long time GeForce fan (my online handle has been 'anthony256' right after I got my hands-on the GeForce 256 back in 1999) it makes the hairs on my arms stand up to even write about this. NVIDIA followed up its #UltimateCountdown tease with "21 Days. 21 Years".

Now, 21 days from now is August 31, 2020 -- why is that important?

NVIDIA has officially kicked off the marketing hype train for its next-gen Ampere-based GeForce RTX 30 series graphics cards with a simple teaser on Twitter.

All the tweet says is #UltimateCountdown which has got me personally thinking that NVIDIA could change the nomenclature like I've wanted for the new GeForce RTX graphics cards -- I wanted the GeForce RTX 3090 ULTRA, for the return of ULTRA -- but maybe we're looking at the GeForce RTX 3090 ULTIMATE?

ULTIMATE is an important branding for NVIDIA right now, and I'm sure even more so going forward -- they have their G-SYNC ULTIMATE monitors and TVs, so it makes sense to have the "Ultimate Countdown" being a countdown to something Ultimate-related, right?

The last we heard on the launch of NVIDIA's next-gen GeForce RTX 3000 series was that they were launching on September 17, but now an announcement date is floating around: September 9.

According to GamersNexus, NVIDIA will be revealing its next-gen Ampere-based GeForce RTX 3000 series cards on September 9. If this is the case, I'd expect tech media -- including myself, to start hearing official news from NVIDIA, NDAs will be signed, and samples will begin to arrive in the next couple of weeks.

If NVIDIA did announce the new GeForce RTX 3000 series cards on September 9, and then launched on September 17 -- that makes sense. It also means NVIDIA will begin the hype train for its new cards very, very soon.

ASUS has just announced an unnamed ROG gaming monitor, the world's first HDMI 2.1-certified gaming monitor that has HDMI 2.1 bandwidth pushing 48Gbps. This makes it support 4K at up to 120Hz, previously exclusive to DisplayPort.

This is the perfect gaming monitor for next-gen consoles in the Sony PlayStation 5 and Microsoft Xbox Series X, or upcoming next-gen NVIDIA Ampere GeForce RTX 3000 series and AMD RDNA 2 aka Big Navi graphics cards with HDMI 2.1 when they launch over the coming months and into 2021.

ASUS has baked in Auto Low Latency that allows consoles to control the processing mode of the display, where it will be able to prioritize either low latency or processing quality depending on what you're doing at the time.

We are another step closer to Ampere, with NVIDIA reportedlyt discontinuing the GeForce RTX 2070 SUPER graphics card. This isn't the first time we heard about the GeForce RTX 20 series being discontinued, with the first rumors popping up back in July 2020.

The news is coming from Chinese site MyDrivers, which reports that NVIDIA has "completely discontinued" the Turing-based GeForce RTX 2070 SUPER graphics card.

The last batch of Turing-based GPUs has been dispatched, with no further orders on the GeForce RTX 2070 SUPER to be accepted. GeForce RTX 2070 SUPER is now EOL = end of life. This list will eventually, where I'm sure the wheels are already rolling, to discontinue the GeForce RTX 2080 Ti, GeForce RTX 2080 SUPER, GeForce RTX 2080 and GeForce RTX 2070.

The hype is beginning to feel real about AMD's next-generation Big Navi GPU, otherwise known as Navi 2X on the RDNA 2 architecture. But what about RDNA 3? Yeah, let's dive right into it.

A fresh batch of rumors suggest that AMD is currently in the early design phase of the RDNA 3 architecture, and it could represent the most radical change of GPU design ever. AMD could shift towards using a chiplet design for its RDNA 3 architecture, similar to the chiplet design used on the revolutionary Zen architecture.

This would let AMD mix and match many different GPU IPs together, where we could see I/O like memory controllers and other parts of the chips being put on different nodes. The main GPU dies (in this case, RDNA 3 chiplets) could be on the 'Advanced Node' (5nm with TSMC) while the other parts like the memory controller, can be on a different node.

AMD has just released a new set of graphics cards that have support for Horizon Zero Dawn Complete Edition, which is about to hit the PC.

Not only that, but the new AMD Radeon Adrenalin 2020 20.8.1 drivers include optimizations for Ubisoft's new battle royale game Hyper Scape, and Grounded -- which is pretty much the worlds of battle royale and the movie Honey, I Shrunk the Kids blended together -- the game has been popular, racking up 1 million players in 48 hours.

If you're gaming on a Radeon RX 5700 XT graphics card, the new AMD Radeon Adrenalin 2020 20.8.1 drivers will provide up to 9% more performance in Grounded on the Epic preset over the previous Radeon drivers.

AMD's upcoming Big Navi has been in the rumor mill the entire year, with whispers that it would use 24GB of HBM2e memory with 2TB/sec of memory bandwidth -- but new reports suggest AMD won't be using HBM2 or HBM2e memory on the Navi 2X cards at all.

It has been rumored that AMD would use HBM2 memory or a 2.5D design on Big Navi, but according to the latest information that is wrong. We should see 16GB of GDDR6 (we heard that in a rumor just two weeks ago) on a whopping 512-bit memory bus, which is double the memory and bandwidth offered on the Radeon RX 5700 XT.

This doesn't mean AMD couldn't use HBM2 memory on another Big Navi GPU, so we could expect a higher-end workstation version with HBM2 memory. The consumer Radeon gaming graphics cards, they will be powered by the mainstream GDDR6 standard.

ASRock has just unveiled a new custom Radeon RX 5700 XT graphics card, with the introduction of the new Radeon RX 5700 XT Challenger Pro 8G OC.

What is new with ASRock's Radeon RX 5700 XT Challenger Pro 8G OC? The triple-fan cooler, where ASRock is using 3 x 80mm fans that have 0dB silent cooling tech. It's a larger 2.7-slot card that should stay a few degrees cooler than the dual-fan Challenger.

It's a strange move, for ASRock to unveil a new custom Radeon RX 5700 XT right before AMD is about to unleash its next-gen RDNA 2 architecture... but it's also good to see ASRock continue to push into the custom Radeon RX graphics card business.

We are getting closer and closer to AMD's big reveal of Big Navi / RDNA 2 and now we have yet another rumor to throw onto the ever-growing pile of hype.

The new rumor comes from Komachi and _rogame on Twitter, which have been teasing an 80 compute unit beast for AMD's next-gen Navi 21 GPU. With 80 compute units under the hood, we're looking at a huge 5120 stream processors which would really put the Big in Big Navi.

If AMD runs these 80 compute units and 5120 stream processors at 1700MHz, we'd be looking at a monstrous 17.5 TFLOPs of power.

A huge leak on AMD's purported Radeon Instinct MI100 acclerator teases that AMD could be serious about not just contending against NVIDIA in the 4K gaming market with a new Radeon to combat the new GeForce RTX cards coming -- but also in the pro market, too.

AdoredTV is reporting from some leaked slides that slid into their inbox that AMD's upcoming Radeon Instinct MI100 accelerator is quite the beast. It is being teased in the slides in a dual socket AMD EPYC Rome/Milan system with 8 x MI100 PCIe-based accelerators.

In this system, we'd be looking at 256GB of HBM2 memory with 9.8TB/sec (yes, that's terabytes per second) of bandwidth and an insane 272 TFLOPs of FP32 compute performance using just 3kW of power.

We should expect some solid confirmation and an announcement of NVIDIA's next-gen GeForce RTX 30 series graphics cards in the next month or so -- but until then, all we have are leaks:

The one today is from KatCorgi once again, who has tweeted: "RTX 3090 may up to 50% PERFORMANCE INCREASE in TimeSpy Extreme". We've heard that the GeForce RTX 3080 Ti or GeForce RTX 3090 (sort that naming out before launch, NVIDIA) could be up to 60-90% faster in previous rumors.

But remember that AMD's next-gen Big Navi / RDNA 2-based flagship Radeon RX 6000 series graphics card, which has been seen around as the purported Radeon RX 6900 XT -- being 40-50% faster than the GeForce RTX 2080 Ti. If this is true, then AMD will indeed have a new card that could meet the performance of NVIDIA's best Ampere GeForce RTX 30 series card, at least in 2020.

NVIDIA is bundling copies of Rainbow Six: Siege Gold Edition with their GeForce RTX 20 series graphics carrds, as part of their Frames Win Games bundle.

Rainbow Six: Siege launched back in 2015, and has over 50 million people playing the game as of September 2019. The game has recently had its Year 5 pass launch, with a bunch of DLC included in the Gold Edition -- and thus, for free when you purchased an NVIDIA GeForce RTX 20 series graphics card.

You can buy any GeForce RTX 20 series card between the regular GeForce RTX 2060 and GeForce RTX 2060 SUPER, right through to the flagship GeForce RTX 2080 Ti graphics card.