Graphics Cards - Page 189

NVIDIA is currently preparing its GeForce RTX SUPER series of graphics cards in the coming weeks, with WCCFTech stirring the pot today with some exclusive news on pricing of the new SUPER cards. This is what we're told to expect:

I posted my own GeForce RTX SUPER series price prediction a little under two weeks ago now and seemed to be pretty much on the money with WCCFTech's exclusive sources. I expect the pricing to be a little more competitive, but NVIDIA can do whatever they want with the high-end/enthusiast market when they're that far ahead.

We're just getting 7nm parts in the form of AMD's upcoming Ryzen 3000 series and Radeon RX 5700 series products, but TSMC has decided that it is time to announce R&D has begun for the 2nm node.

We will soon be hearing more about the 5nm and 3nm nodes from the likes of Samsung and TSMC as research into those processes has already begun, but the news on 2nm is super new. Zhuang Zishou, a senior director at TSMC, explained to Taiwanese media that the 2nm plant will join the other next-gen node plants in Hsichu, Taiwan.

Hsichu is where the huge Science Park is located where 400 tech companies reside, with TSMC hoping to get the 2nm plant operational by 2024. The road to 2nm is pretty damn insane, as we're quickly shifting from 14/16nm to 7/10nm, while 5/3nm are in the works and now R&D on 2nm has begun.

E3 2019 - NVIDIA has just had a super announcement at E3 2019 with news that the company has teamed with CD Projekt RED on bringing RTX effects into Cyberpunk 2077, which was recently detailed at E3 2019 and features Keanu Reeves in a starring role.

Cyberpunk 2077 is yet another major AAA title to receive RTX abilities, with the recently-announced Call of Duty: Modern Warfare rebooted with RTX abilities and it looks beyond gorgeous. There's also Control from Remedy, as well as the new Wolfenstein and many other games in the pipeline with RTX abilities coming. Cyberpunk 2077 with RTX is a big, big deal as it is one of, if not the most-hyped game in many years.

NVIDIA securing Cyberpunk 2077 as an RTX-capable game will definitely be a huge hit for 2020, and it'll be a SUPER time for gamers. Cyberpunk 2077 drops on April 16, 2020 and stars Keanu Reeves (!!!).

So... I'm sitting here in my hotel room the morning after a massive 10-hour information download session from AMD and their Next Horizon Gaming event and wanted to write it all up. The problem? There's too much to talk about.

So before I give you the basic run down of Navi and the new Radeon RX 5700 and RX 5700 XT, I'll tell you one of two things. First, I will detail much more of this when I get back home next week in much greater detail. Secondly, AMD has never hosted an event with so many new things to show us and I don't think I've ever been this excited for the company, ever.

We all know about Navi by now: AMD's next-gen GPU architecture on the new 7nm node. It is here to compete with NVIDIA's current GeForce RTX 2060 and RTX 2070 (before things get SUPER). It is not, and was never meant to, compete or beat the flagship RTX 2080 Ti and that is not a bad thing.

It looks like the hype train for NVIDIA's upcoming GeForce RTX 20 series refresh/super release is gaining traction with leaks of the purported cards arriving on Chinese forum Weibo, confirming my story that there will be not one, but three new 'SUPER' GeForce RTX 20 series graphics cards.

The new graphics cards will still be based on the Turing GPU architecture and should replace the current GeForce RTX 20 series card price points, and then drop the current price of RTX 20 series cards. We can expect the new SUPER range to be led by the flagship GeForce RTX 2080 SUPER which will be the card featuring overclocked GDDR6 memory.

The full specs on the purported GeForce RTX 2080 SUPER include the full TU104 GPU, 3072 CUDA cores, 8GB of GDDR6 at 16Gbps on a 256-bit memory bus. We already know most of the leaked info on the RTX 2070 SUPER but the RTX 2060 SUPER sounds like it'll be the most interesting of the bunch.

NVIDIA is reportedly launching three new GeForce RTX series cards next week according to our sources at Computex 2019, and it looks like things are beginning to take shape with rumors of a purported GeForce RTX 2070 Ti graphics card.

The new purported RTX 2070 Ti would be 10-15% faster than the stock RTX 2070, with its 2560 CUDA cores (vs 2304 in RTX 2070) while the RTX 2080 has 2944 CUDA cores. So it is much closer to an RTX 2070 than an RTX 2080, and also packs 8GB of GDDR6 at 14Gbps, with its GPU boost at 1770MHz, while the boost clock on the RTX 2070 is at 1710MHz.

Previous rumors stated that the new cards would include faster GDDR6 memory that would be up from 14Gbps to 16Gbps, so we need to wait a little while to see what rumors pan out and what don't. For now, this is the latest on the RTX 2070 Ti.

Apple launched its new Mac Pro and inside the machine that will cost up to $50,000 was the surprise release of the new Radeon Pro Vega II and Radeon Pro Vega II Duo graphics cards from AMD. Both of these new cards are based on the Vega GPU on the new 7nm node.

The new Radeon Pro Vega II and Radeon Pro Vega II Duo graphics cards are not for gamers in any shape, way or form -- they are pure workhorses that are perfect for professionals that need crazy amounts of GPU horsepower. Tasks like video editing native 8K resolution content, rendering, and other compute-intensive tasks the Vega 20 GPU on 7nm with 32GB of HBM2 per GPU is up to the task.

Each of the Radeon VII Pro Vega II graphics cards are quad-slot cards, because they're passively cooled with a unique cooling system deployed by Apple into the new Mac Pro. The new Radeon Pro Vega II has 14.2 TFLOPs of FP32 performance and 28.4 TFLOPs of FP16 performance while the Radeon Pro Vega II Duo dual-GPU card has 28.4 TFLOPs and 56.8 TFLOPs, respectively. The single card solution has 32GB of HBM2 while the dual-GPU solution rocks 64GB.

AMD has just inked a new deal with Samsung where the two companies are entering a new multi-year strategic partnership with the smartphone giant gaining access to AMD's large range of patents and IP in video and graphics accelerators.

Samsung's own LSI division, the part of the company that handles chip design and fabrication is contracting with AMD to use its new Navi-based RDNA architecture to build its future mobile GPUs for next-gen smartphones and other mobile devices. The companies didn't go into depth about what this relationship will entail when it comes to the future of mobile GPUs, but this is seriously exciting news.

It means that Radeon technology, and possibly even the Navi GPU architecture, is scalable enough to go right down into mobile with this new RDNA architecture. As well as up into Radeon RX 5700 series and hopefully even higher with next-gen Navi in 2020 and beyond.

NVIDIA didn't hide the fact that 'something Super is coming' before Computex 2019 but it seems plans changed, and AMD took the spotlight with its reveal of the Radeon RX 5700 series graphics card during keynote.

Well, a little birdie confirmed with us that are to expect not one, not two, but three new GeForce RTX 20 series graphics cards to be unveiled at E3 2019. I've been told that there will be a new RTX 2060, RTX 2070, and RTX 2080 with souped-up, faster GDDR6 memory.

The source said there won't be faster version of the RTX 2080 Ti (it's already fast enough) and after that I began discussing the news with industry friends. Some agreed, and some expected that we could see something faster than the GeForce RTX 2080 Ti that will sit between the RTX 2080 Ti and the TITAN RTX. I don't think that will happen, but if I were NVIDIA -- I wouldn't be using the word 'Super' without it actually being you know... SUPER.

Computex 2019 - NVIDIA didn't unveil any new graphics cards at Computex this year, so ZOTAC didn't have too much new in the graphics card department to unveil but that doesn't mean they were silent.

The company brings eyes to their booth in Nangang with what I would call the world's largest graphics card, which looked freaking awesome with its RGB lighting glowing and flowing.

ZOTAC was showing off their new single slot GeForce RTX 2080 Ti ArcticStorm graphics card, which looked super slick in person and super slim with its single slot design.

Computex 2019 - MSI officially unveiled their GeForce RTX 2080 Ti LIGHTNING Z at Computex 2019 but something went a little further under the radar: a new pink variant of the LIGHTNING Z.

It looks pretty damn cool in person, and I'm sure any fans of pink and girl gamers out there that want to spend some serious cash on a graphics card are going to dive at this card.

The backplate also looks cool with its pink textured styling.

Computex 2019 - AMD has just officially unveild its next-gen Navi GPU architecture here at Computex 2019, with some detials on the upcoming Radeon RX 5000 series.

Yes, it's not the Radeon RX 3000 series like we previously thought -- with AMD using its 50th anniversary as a jumping point to launch the new Radeon RX 5000 series. AMD unveiled the new Radeon RX 5700 versus NVIDIA's current GeForce RTX 2070 with Strange Brigade running and it outperforming the Turing-based RTX 2070 in early form.

AMD committed to a July release window for Radeon RX 5000 series cards but didn't detail any other cards, so we can expect that the previously rumored 7/7 release to take place. We will hear everything about Navi and the new Radeon RX 5000 series at AMD's upcoming Next Horizon Gaming event next month.

NVIDIA has just filed for trademarks of 3080, 4080 and 5080 with the European Union Intellectual Property Ofice in an effort to slow the hype train for AMD's next-gen Navi-based graphics cards that are rumored to launch as the Radeon RX 3000 series.

These new Navi-based cards will reportedly include the Radeon RX 3070 and Radeon RX 3080 which, these new patent grabs, would hurt AMD big time. NVIDIA currently sells the GeForce RTX series of cards including the RTX 2060, 2070, 2080, and 2080 Ti.

We all know how this works: most unsuspecting (not all) people will see a higher number (RX 3080) and think by default it is faster than the smaller number (RTX 2080). I worked retail IT sales selling these exact types of computer products for 10 years before I began at TweakTown, and it happened all the time. The issue here is, will NVIDIA call its next-gen GeForce cards RTX 3080, followed by RTX 4080, and then RTX 5080? Or is this the company trying to meddle in the launch of its competitors' next-gen graphics cards?

We are not far away from the unveiling of AMD's next-gen Navi GPU architecture and a new family of Radeon Navi graphics cards in the form of the purported Radeon RX 3000 series, and now we have leaked specs on the Radeon RX 3080.

AMD is expected to use what should be the Navi 12 GPU inside of the Radeon RX 3080 leaving some room for larger Navi cards in the future, with the new specs teasing 8GB of VRAM, shader engine designs, and improved compute unit utilization efficiency, and more.

Navi 10 or 12 (whatever this is) should feature 40 compute units each of which will contain 64 stream processors, as well as 8 shader engines. Each of these shader engines will pack 5 compute units and a raster back end, which is a big deal. With previous GCN-based GPU architectures the raster back end would comprise of 16 ROPs, so we don't know if Navi 10/12 will feature 128 ROPs, or if things have changed under the hood with Navi (which I think has happened).

I'm literally in an Uber on the way to the airport typing this news up as it truly did come as a surprise, but NVIDIA has just teased 'something Super is coming' on its official Twitter and YouTube -- check out the teaser:

Now, what can we expect? Ampere is the next generation GPU architecture post-Turing but I think we're a little early for that, so what about a refreshed GeForce card on 7nm? Again, I can't see that happening but it's definitely a possibility. Maybe we can expect a refreshed high-end GeForce GTX graphics card (think RTX 2080/2080 Ti power without ray tracing abilities, similar to the GeForce GTX 16 series).

The new 'Super' GeForce in without is RTX abilities would fight the rumored AMD Radeon Navi graphics cards which will be unveiled just before E3 2019 after Computex. Rumor has it the new Radeon RX 3000 series graphics cards will NOT have real-time ray tracing abilities so a juiced up Super GeForce unveil in the coming weeks makes sense.

MSI has been making LIGHTNING series graphics cards for 10 years now, and in order to celebrate the milestone the company has unveiled that a new special edition of its GeForce TX 2080 Ti LIGHTNING graphics card will debut at Computex 2019.

There's not much known about the new special edition MSI GeForce RTX 2080 Ti LIGHTNING Z graphics card, but if the normal version is anything to go by the 10th anniversary special edition card should be something very, very special indeed. We should expect nothing but the very best from MSI -- something we'll definitely be checking out on day one of Computex 2019 next week.

The very first LIGHTNING series card was the MSI GeForce GTX 2060 LIGHTNING back in 2009, the first graphics card to feature 5 heat pipes and the first with a metal cover and dual-fan design. It also featured the first integrated touch OC panel, and kicked off the entire LIGHTNING series family of cards from MSI.

SAPPHIRE has revealed some next-gen Radeon Navi graphics card details at a new Chinese press event, with a SAPPHIRE product manager and PR director talking about two Radeon Navi cards at $399 and $499, as well as a next-gen TOXIC version of Navi.

You might not remember the TOXIC brand from SAPPHIRE but it used to be a kick ass super-custom Radeon graphics card until AMD started to slow down with enthusiast-level Radeon cards. Well, TOXIC is making a come back according to the SAPPHIRE PR manager who said that the company has designed the TOXIC version of Radeon Navi and that it will be watercooled.

That is some good news for gamers and SAPPHIRE, as the company has had its hands tied behind its back with recent Radeon releases. There were no custom Radeon VII graphcis cards as I exclusively revealed, with all roads leading to Navi and now a TOXIC-infused SAPPHIRE card that is watercooled. We're in for some great times when it comes to Navi excitement in the coming months.

AMD will be detailing its next-gen Navi GPU architecture at its Computex 2019 keynote and Next Horizon Gaming event before E3 2019, but now VideoCardz is reporting major AIB partner SAPPHIRE is launching not one, but two Radeon Navi graphics cards on July 7.

The news is directly from a PR manager for SAPPHIRE during a chat with Chinese press, with the rep saying AMD is working on two Navi SKUs: the first is going to cost $399 while the higher-end Navi will cost $499. The $399 card will reportedly compete with NVIDIA's GeForce RTX 2060 while the $499 model will fight with the RTX 2070. This new rumor blows out the $330-priced Radeon RX 3080 XT that would be a RTX 2070 competitor, but there's no mention of the name here.

We can expect the previous rumors of AMD calling its new Navi-based cards the Radeon RX 3000 series with these two cards most likely breaking down into the Radeon RX 3080/XT and the 3090/XT depending on what AMD brand these new cards as. But for $399 and $499 is it going to be enough to pull people away from GeForce RTX cards?

We are only weeks away from AMD unveiling its next-gen Navi GPU architecture in full, with two new Navi-based graphics cards turning up on 3DMark and Ashes of the Singularity benchmarks. TUM_APISAK found the benchmarks which look like engineering samples, and not finalized Radeon Navi-based graphics card.

The device ID for the first card is '731F:C1' which sees the Navi GPU clocked at just 1GHz and its 8GB of VRAM clocked at 1250MHz -- I'm expecting this to be GDDR6 memory but AMD could be using GDDR5 memory on its lower-end Navi parts. If AMD is using GDDR6 then we should see effective clock speeds of 10Gbps and on a 256-bit memory bus we're looking at 320GB/sec of memory bandwidth, beating out the Radeon RX 590.

The Navi GPU clocked at 1GHz is probably because it's an early sample, so the GPU clocks might be increased for final retail cards or this is a lower-end card with the GPU clocks firmly around 1GHz.

NVIDIA is reportedly working on suped-up GeForce RTX graphics cards with faster GDDR6 memory, with rumors suggesting a near 15% improvement in memory clocks from 14Gbps to 16Gbps, which would boost memory bandwidth as well.

The new rumors are coming from RedGamingTech on YouTube, which if true would see NVIDIA releasing a new GeForce RTX 2080 with its 8GB of GDDR6 at 16Gbps resulting in 512GB/sec (from 448GB/sec on the original RTX 2080). NVIDIA will be needing to muscle up to fight AMD and its soon-to-be-released Navi-based Radeon RX 3000 series graphics cards, at least according to the new rumors. We are waiting for more details on Navi at AMD's upcoming Next Horizon Gaming event just before E3 2019.

What about GPU clocks? We could expect higher clocks on new GeForce RTX graphics cards that would replace the current price of the RTX 2070/2080 with higher-end versions of themselves, and then price drops of the current RTX 2070/2080 so they can compete with Navi.