Graphics Cards - Page 179

NVIDIA is expected to unveil its new Ampere GPU architecture during its GTC 2020 keynote, where it wants you to GET AMPED about it between now and especially on May 14, at 6AM PST.

We know that NVIDIA is one of TSMC's largest clients for its 7nm process node, while Samsung announced it was to begin production of its new 5nm (5LPE) EUV process in Q2 2020. Remember the mysterious 5nm chip NVIDIA was working on? Yeah, I would say it has something to do with that.

Samsung could be making NVIDIA's next-next-gen Hopper GPU on its new 5nm EUV node, with Ampere finding a home on the 7nm node by TSMC. We should expect NVIDIA to unveil its new GeForce RTX 3080 and RTX 3070 graphics cards sometime in August 2020, with a launch at Computex 2020 -- nothing is in concrete right now, but that seems like a viable launch window.

NVIDIA most likely would've unveiled its Ampere GPU architecture and this new DGX A100 system at its GPU Technology Conference (GTC) in March, but it was canceled over the COVID-19 coronavirus outbreak. We are now told to GET AMPED for a GTC 2020 keynote presentation from NVIDIA founder and CEO Jensen Huang on May 14, at 6AM PST.

Still, we're now being teased with the new Ampere-based DGX A100 system that packs Ampere GPUs that, right now -- we don't know much about but can expect them to be monsters. A new entry has popped up in the Justia trademark database, with NVIDIA filing an application for the DGX A100 trademark.

The first DGX systems were named DGX-1 and DGX-2 but it seems that NVIDIA won't be naming the Ampere-based system DGX-3 but rather DGX A100. If this is true, we now have the GA100 teased -- the first GPU based on the new Ampere architecture.

It was barely 24 hours ago that I reported on Intel teasing the 'Father of All' GPUs that is the Xe-HP, but now Raja Koduri is teasing the successor to that -- something he calls the 'Superman of All' GPUs.

In a follow-up tweet, Koduri explained "This was called the "baap of all" by our team... The "Baahubali of all" is baking as well. Let's hope the wait will be shorter than what @ssrajamouli put us through for @BaahubaliMovie". Wccftech decoded the post, translating it into the "Superman of All".

Any regular readers of TweakTown will know I'm one of the biggest Superman fans on the planet, as well as a gigantic geek and GPU nerd so this news is righit up my alley. In this tweet, Koduri is referring to the Baahubali movie, which is a superhero movie that Raja was involved with.

AMD partnered with Samsung in August 2019 where it would allow its RDNA architecture would power future Samsung devices, after announcing a multi-year strategic partnership in June 2019.

That new RDNA-powered design seems like it's on the right track, with new rumors from South Korean tech forum Clien which says the Radeon-powered chip "crushed" the Qualcomm Adreno 650 GPU in GFXBench. There's quick a smack down from the AMD/Samsung chip versus the Adreno 650, check it out:

If these scores are real, then we're in for a gigantic surprise when Samsung releases its new phones. It could mean that a huge uptick in performance will come from Qualcomm's next-gen chips, with mobile graphics a big deal when it comes to next-gen gaming smartphones.

NVIDIA might have stolen the headlines with a hot new rumor of some pretty crazy GeForce RTX 3080 Ti specs, but now Intel is getting in on the fun and shoving AMD out of the way with some Xe-HP news.

It was barely 24 hours ago that the official Intel Graphics team tweeted out some photos, but the tweet was quickly deleted. Not because it showed anything important where Intel lawyers would tell them to quickly delete the tweet because it showed something it shouldn't -- but rather Intel was slammed for the incorrect use of face masks by major Intel staff like Raja Koduri and Jim Keller.

But don't worry, nothing is forever with the internet because I took a screenshot. Here you can see the original tweet from the official Intel Graphics Twitter account.

NVIDIA's next-gen GeForce RTX 3080 Ti could be an absolute world-tearing graphics card, with new rumors suggesting it will be beyond anything previous rumors have said.

A new rumor is coming from a now deleted tweet from 'chiakokhua' who has a history of predicting big movement of tech companies, where he said AMD would use a chiplet design for its new Zen design. He also drew up a block diagram of EPYC processors receiving 64 cores over 8+1 dies.

With that said, check the specs above: we have the GA100 powering the GeForce RTX 3080 Ti with up to 8192 CUDA cores which is absolutely insane. These rumored specs see GA100 have up to 256 RT cores, up to 1024 Tensor cores, and GPU boost clocks of up to 1750MHz. All of the new Ampere-based GeForce RTX 30 series cards will have GDDR6 at 16Gbps.

GALAX has just announced an interesting new custom GeForce RTX 2070 SUPER graphics card, with a white and pink design in the new EX Pink Edition.

The new GALAX GeForce RTX 2070 SUPER EX Pink Edition continues the tradition behind the normal EX model, with an identical cooler and PCB design -- but dips them in girl-gamer-worthy white and pink.

The new EX Pink Edition graphics card has RGB fans that glow pink, while the entire backplate of the card is also pink -- I kinda dig it, and I don't know why.

NVIDIA has pushed out its new GeForce hot fix drivers in the form of the new 445.98 drivers, which fix up some of the issues that weer in the GeForce 445.87 Game Ready drivers.

Those drivers were released with optimizations and support for games like Minecraft RTX beta, Saints Row: The Third Remastered, Snowrunner, and Call of Duty: Modern Warfare 2 Campaign Remastered. The new drivers fix up problems with the shadow cache that was being purged incorrectly, brightness problems with HDR-enabled games, and more.

Overwatch fans might enjoy some more "improved game stability", while F1 2019 gets an issue with randomly crashing to the desktop with the new drivers.

So I laid my head down last night after a huge day, to read an email from NVIDIA that opens with "Get Amped for Latest Platform Breakthroughs in AI, Deep Learning, Autonomous Vehicles, Robotics and Professional Graphics".

Yes everyone -- GET AMPED for the Ampere GPU architecture unveil at the now-heavily-delayed GPU Technology Conference with NVIDIA boss and CEO Jensen Huang. The company will be hosting its GTC 2020 keynote with Huang and his legendary opening keynote on May 14, at 6AM PST -- all on YouTube thanks to coronavirus.

Huang will "highlight the company's latest innovations in AI, high performance computing, data science, autonomous machines, healthcare and graphics during the recorded keynote".

So, NVIDIA is making something on a fresh new 5nm node -- ordering TSMC to make it for the GeForce giant.

Digitimes is reporting from its "industry sources" that NVIDIA is tapping TSMC's fresh new 5nm process node -- and it won't be for NVIDIA's next-gen Ampere GPU architecture or the GeForce RTX 3080 and GeForce RTX 3070 graphics cards.

There is absolutely nothing else to this news other than Digitimes reporting that its sourcesh ave told them, that NVIDIA is making something new and mysterious on TSMC's new 5nm process node. It could be something for the Tegra side of things, or here's a wild rumor -- so prepare your salt shakers peeps -- the Nintendo Switch 2 SoC. Just throwing that out there.

NVIDIA and AMD are both ramping up into the world of 7nm GPUs with the companies buying up the spare capacity of 7nm at TSMC for next-gen GPUs and graphics cards.

Digitimes is reporting that a "ramp-up in short lead-time orders placed by AMD and NVIDIA will allow TSMC to post relatively strong results compared to other foundries in the first half of 2020, according to industry sources". Yes my fellow nerds and gamers, NVIDIA and AMD are ramping up for next-gen graphics cards.

NVIDIA is expected to unleash its new Ampere-based GeForce RTX 3080 graphics card later this year, where a recently-delayed Computex 2020 in September would make sense for Ampere's big reveal for next-gen GeForce.

ASRock has just unveiled the newest member of its Challenger series of graphics cards, with the new Radeon RX 5500 XT Challenger ITX 8G graphics card. Check out our review on the ASRock Radeon RX 5500 XT Phantom Gaming series card.

Where ASRock does something different here with the Challenger ITX graphics card, is that it shrinks down the Radeon RX 5500 XT into ITX form for super-small form factore gaming. You still get the 7nm RDNA chip with 8GB of GDDR6 memory, but it has been shrunken down with ASRock's in-house Shrink Ray.

ASRock provides a boost GPU clock of up to 1845MHz on the card, with 1080p 60FPS gaming being no issue on the Radeon RX 5500 XT Challenger ITX. The card is just 17.8cm long and fits into small form factor and mini-ITX chassis, with a 10cm fan and 2 x copper heat pipes keeping things cool.

MSI has just unveiled the fastest-ever GeForce RTX 2080 Ti graphics card, with the company using new, and faster 16Gbps GDDR6 memory dies.

The new GeForce RTX 2080 Ti GAMING Z TRIO is just like the GeForce RTX 2080 Ti GAMING X, except that it packs the faster GDDR6 memory. MSI's flagship GeForce RTX 2080 Ti LIGHTNING Z (my review here) still packs higher GPU clocks than the new RTX 2080 Ti GAMING Z TRIO, but the GAMING Z TRIO has the faster 16Gbps GDDR6 memory. The LIGHTNING Z has its 11GB of GDDR6 at 14Gbps.

This isn't the first time MSI has added faster memory to one of its graphics cards in the GAMING Z TRIO model, with our recent review on MSI's GeForce RTX 2070 SUPER GAMING Z TRIO equipped with faster 15.5Gbps GDDR6 (up from 14Gbps).

Huawei is reportedly about to announce that it has been working on a new discrete GPU for the server and HPC market, according to a new report from a South Korean publication.

The Chinese giant has opened a new division in China called the Cloud and AI Business Group, where Huawei will begin its battle against NVIDIA and AMD -- and once Xe launches it will have Intel join the fight, for the GPU server and HPC businesses. Huawei is reportedly looking at a late 2020 release for its first-gen GPU, but it could have a gaming-focused GPU in the works and released in the next few years.

Huawei is pushing its new GPU-focused business with a home base in South Korea, with the Chinese giant acquiring local talent as well as acquiring a number of ex-NVIDIA engineers.

Remember when NVIDIA unveiled that ridiculously amazing custom GeForce RTX 2080 Ti Cyberpunk 2077 Edition graphics card? Yeah, well someone is selling one of them on eBay for $5000.

NVIDIA made less than 200 of the custom GeForce RTX 2080 Ti Cyberpunk 2077 Edition graphics cards, where they were mostly given away during NVIDIA's huge competition. The company hosted a popular Twitter competition where they took in over 130,000 entries -- and picked people in pairs of two.

Some of the custom GeForce RTX 2080 Ti Cyberpunk 2077 Edition graphics cards have been selling on eBay for anywhere between $3000 and $4300, but at the time of writing there was one on eBay for a wallet-busting $5000. For that, I'd want Keanu Reeves to personally call me up and call me "breathtaking"... right? Keanu?? Call me later, kay?

ASUS has just launched its new GeForce GTX 1650 GDDR6 graphics cards which if you couldn't already guess from the first sentence, have GDDR6 memory -- the original GeForce GTX 1650 launched with GDDR5 memory. And yes, I also had fun writing that headline... and ASUS really did use "space-grade lubricant".

The new GDDR6-powered GeForce GTX 1650 has memory bandwidth over 50% faster than the GDDR5-based GeForce GTX 1650. This means that the new GDDR6-powered GeForce GTX 1650 is much faster option for 1080p and 1440p gamers at a lower price point than the GeForce GTX 1660/GTX 1660 SUPER.

The original GeForce GTX 1650 with its GDDR5 memory was clocked at 8Gbps and offered 128GB/sec of memory bandwidth, but the new GDDR6 memory is clocked much higher at 12Gbps and offers 192GB/sec of memory bandwidth -- a win-win situation for gamers.

NVIDIA has just released the most powerful notebook graphics solutions in the world, with the introduction of the GeForce RTX 2080 SUPER and GeForce RTX 2070 SUPER mobile graphics cards.

The new GeForce RTX 2080 SUPER and GeForce RTX 2070 SUPER are offered in Max-Q configurations, which offer unrivaled performance and efficiency. NVIDIA will have GeForce RTX 2060 SUPER-powered gaming notebooks from partners like Acer, ASUS, HP, Lenovo, MSI, and others starting from $999. Meanwhile, RTX 2070-powered gaming notebooks will start from $1199.

Enabling DLSS 2.0 will give you a massive chunk of a performance gain, if you are running games that have DLSS 2.0 -- these include games like Control and Wolfenstein: Youngblood, and others.

AMD has announced that a hacker has stolen some of its graphics IP, and is demanding that the company pay them $100 million for its return.

Back in December 2019, the company says it was contacted by someone who said they had access to "test files" for AMD's current, and future graphics products. A handful of these GPU "test files" were then uploaded to the graphics forum GitHub, but AMD quickly reacted and slapped it with a DCMA takedown notice.

No, the AMD GPU IP leak cannot be used as a source code to build, and deisgn some custom RDNA graphics solution -- don't go expecting that at all. It can't be used to build products, it can't be used to reverse engineer a Radeon GPU design, or anything close to that.

NVIDIA was all systems go for its Ampere GPU architecture (which I exclusively unveiled all the way back in April 2018) unveiling at its own GPU Technology Conference (GTC) last week, but it was canceled over the coronavirus COVID-19 outbreak. This meant the Ampere announcement was also delayed.

But now Bloomberg's Tai Kim has written a new article titled 'NVIDIA's Biggest Pleasant Surprise May Be Yet to Come', where he goes into the fact that NVIDIA is kinda flourishing over the coronavirus outbreak. In positive news about coronavirus, more gamers are at home, and more games are being played right now than pretty much ever before in the history of gaming.

NVIDIA's gaming segment (GeForce graphics cards) represents close to half of its revenue, so with more gamers playing games at home that is only a good thing for NVIDIA. Kim writes that NVIDIA management said they've seen a 50% surge in "total gaming hours from its installed base as many students and workers were staying-at-home".

Microsoft peeled the wrapping off of its new DirectX 12 Ultimate API today, with AMD actively invovled in the middle of the new DirectX 12 Ultimate API with a new ray tracing video running on RDNA 2 silicon. Check it out:

AMD says it has been working "very closely with Microsoft to help move gaming graphics to a new level of photorealism and smoothness thanks to the four key DirectX 12 Ultimate graphics features -- DirectX Raytracing (DXR), Variable Rate Shading (VRS), Mesh Shaders, and Sampler Feedback".

AMD's new RDNA 2 architecture powers both the Xbox Series X and PlayStation 5 consoles, as well as the next-gen Radeon graphics cards coming out later this year. There will be hardware-accelerated DXR 1.1 ray tracing support built into the new RDNA 2 architecture, which will allow next-gen console and next-gen Radeon gamers to enjoy some of the best visuals available.