Graphics Cards - Page 155

Look -- you might have tried to buy one of NVIDIA's new GeForce RTX 30 series graphics cards in the past 6+ months to no avail, but that doesn't mean NVIDIA is shouting from the rooftops at how good the Ampere launch has been.

NVIDIA said during its latest investor presentation that the Ampere GPU architecture was their best launch ever, seeing double the amount of end market sales of its new GeForce RTX 30 series graphics cards. NVIDIA shows that in the first 18 weeks of the new Ampere GeForce RTX 30 series being launched that demand for cards was insane... and it has only gotten stronger.

When compared to the end market sales of the previous-gen Pascal and Turing GPU architectures, the new Ampere GPU architecture and kick ass GeForce RTX 30 series graphics cards are wanted more than Pascal and Turing cards ever were in the first 18 weeks of their release -- but both of those cards also didn't have to contend with an out-of-nowhere, never-before-seen GPU shortage like we have witnessed through the last 6+ months.

NVIDIA has just pushed out some new drivers with the GeForce Game Ready 466.11 WHQL drivers arriving as a 640MB download, with day-one support for Mortal Shell's new RTX update and so much more.

NVIDIA is adding more than just DLSS support for Mortal Shell's new RTX update but also support for 6 new G-SYNC compatible gaming monitors, a new NVIDIA Reflex feature in Valorant, as well as support for the AI-powered NVIDIA Noise Removal technology inside of popular streaming software OBS Studio.

The new Mortal Shell RTX update adds DLSS (Deep Learning Super Sampling) support that radically improves performance in the game. NVIDIA provides some benchmark charts for Mortal Shell and its new RTX update -- turning ray tracing on and setting DLSS to "Performance Mode" we have performance on the GeForce RTX 3060 skyrocketing from an unplayable 15FPS at 4K to 36FPS average.

Sorry everyone -- if you have been having a hard time (everyone is) trying to buy a new GeForce RTX 30 series graphics card -- it's not going to get better this year, at all.

NVIDIA is now saying that 2021 isn't going to get better for PC gamers and that GPU shortages will continue through the entire of 2021 and into 2022. NVIDIA CFO Colette Kress explained during NVIDIA's annual investors day: "We expect demand to continue to exceed supply for much of this year. Our operations team is agile and executing fantastically. We expect our supplies to increase as the year progresses".

The company is noticing a lot of its graphics cards being scooped up by not just scalpers but crypto miners, with NVIDIA reacting to crypto miners buying GPUs in huge numbers with the introduction of the new Crypto Mining Processor (CMP) series of mining GPUs.

NVIDIA will have its kick ass DLSS (Deep Learning Super Sampling) technology powering Unity-based games with news that the Unity Engine 2021.2 engine will natively support DLSS by the end of 2021.

Game developers that are already using the Unity Engine can now use DLSS through HDRP (the High Definition Rendering Pipeline) while gamers will have to wait for those developers to do their magic while Unity-powered games will receive DLSS support by the end of the year and into 2022 and beyond.

It comes at a perfect time, with Intel launching a new GPU later this year with its purported XeSS -- their answer to DLSS while AMD will have their own DLSS competitor in FSR or FidelityFX Super Resolution technology for Radeon RX gamers of the future.

PowerColor has just released its new Radeon RX 6900 XT Red Devil Ultimate graphics card, with some devilish tricks of its own against competing custom Radeon RX 6900 XT graphics cards.

The new PowerColor Radeon RX 6900 XT Red Devil Ultimate graphics card packs a higher-binned Navi 21 XTXH GPU, with out-of-the-box game GPU clocks set at 2235MHz while boost GPU clocks are at 2425MHz. But some reviewers have got their cards already (I was blacklisted by PowerColor years ago for leaking out a card that I triple-checked I could at a previous Computex -- that's for another day).

HotHardware received their new custom PowerColor Radeon RX 6900 XT Red Devil Ultimate graphics card and were able to hit GPU speeds that couldn't be done on current Radeon RX 6900 XT cards. This is because AMD is using higher-binned GPUs on the new PowerColor Radeon RX 6900 XT Red Devil Ultimate graphics card which uses a new PCI Device ID versus the original Radeon RX 6900 XT cards, and isn't recognized by GPU-Z.

NVIDIA couldn't host a GPU Technology Conference (GTC 2021) event without announcing some new GPUs with the introduction of the new NVIDIA RTX A5000 and RTX A4000 GPUs.

The new NVIDIA RTX A5000 and A4000 like the flagship RTX A6000 pack new RT Cores, Tensor Cores, and CUDA cores that all super-speed AI, graphics, and real-time rendering. The new NVIDIA RTX A5000 packs the GA102 GPU with 28 billion transistors on Samsung's custom 8nm node, with 8192 CUDA cores and up to 27.8 TFLOPs of single precision compute performance.

We have 24GB of GDDR6 with ECC memory on a 384-bit memory bus with up to 768GB/sec of memory bandwidth that can be doubled when you put 2 x NVIDIA RTX A5000 GPUs together with NVLink -- you cannot use the NVIDIA RTX A4000 in NVLink so you're limited to just 16GB of GDDR6 on the RTX A4000 versus up to 48GB on the RTX A5000 in NVLink.

GIGABYTE has just announced 3 new high-end AORUS gaming monitors, with the introduction of the FI32U, FV43U, and FO48U gaming monitors with 32-inch, 43-inch, and 48-inch sizes respectively.

The new flagship AORUS FO48U is the big daddy of them all rocking a huge 48-inch panel, and still rocking 4K 120Hz -- meanwhile the smaller siblings crank up a bit higher to 4K 144Hz. All of the monitors have HDMI 2.1 connectivity that allows 4K 120Hz while GIGABYTE goes all-out using OLED panels on the new AORUS gaming monitors.

OLED panels deliver the best blacks, deep beautiful colors and instant response -- making them one of, if not the best choice for gamers. With the new AORUS gaming monitors coming in 32-inch, 43-inch, and 48-inch sizes all offering 4K 120Hz and HDMI 2.1 at a minimum, the huge 48-inch AORUS FO48U now joins the LG 48-inch 4K 120Hz OLED TV as an OLED-based 48-inch display with 4K 120Hz for the ultimate in gaming.

NVIDIA has just announced its super-beast new DGX Station 320G which packs the latest and greatest NVIDIA A100 Tensor Core GPUs which in this 4-way GPU system pack an insane 320GB of super-fast HBM2e memory.

This isn't a system for gamers but for the much higher-end AI and HPC (high-performance computing) markets, which will super-speed machine learning and data science performance where it's needed. Think about laboratories, home offices, VFX studios, research facilities and everything in between.

NVIDIA's new DGX Station 320G is the fastest-ever server-in-a-box designed for AI research.

NVIDIA hasn't teased its next-gen GeForce RTX 40 series graphics cards, but it has just teased its next two GPU architectures codenamed Ampere Next and Ampere Next Next.

Yeah -- Ampere Next should be the Lovelace GPU architecture we've been hearing about and drops in 2022 meanwhile NVIDIA's next GPU architecture after that is codenamed Ampere Next Next and could be the MCM (multi-chip module) or GPU chiplet-based Hopper GPU architecture.

During the GTC 2021 keynote NVIDIA founder and CEO Jensen Huang explained that the company will focus on x86 architecture one year, while the next year they'll focus on ARM -- with NVIDIA having a "2-year rhythm with a kicker in-between".

NVIDIA has just announced its new A30 Tensor Core GPU which the company calls its "most versatile mainstream compute GPU for AI inference and mainstream enterprise workloads".

The new NVIDIA A30 Tensor Core GPU was built for AI inference at scale, packing an Ampere GPU with 24GB of HBM2 memory with 933GB/sec of memory bandwidth available. NVIDIA includes third-gen NVLink support here with 200GB/sec of bandwidth available between GPUs in multi-GPU situations.

We have peak FP32 performance of 10.3 TFLOPs, peak FP64 performance of 5.2 TFLOPs and peak FP64 Tensor Core performance of 10.3 TFLOPs -- all within a 165W low-power TDP. Not only that, but the NVIDIA A30 Tensor Core GPU runs on a regular PCIe 4.0 card which will slot into mainstream servers.

AMD is working on a next-gen Radeon PRO workstation graphics card, with a new leak teasing the first RDNA 2-powered workstation card. Check it out:

We are being told from the new rumors that AMD's new Radeon PRO workstation card packs 16GB of GDDR6 memory -- just like the flagship Radeon RX 6800 XT and Radeon RX 6900 XT graphics cards. One of the labels on the cards teases that this model we're being shown is the "Full Secure TT GLXL" model and the "TT" definitely doesn't stand for TweakTown... or does it?!

AMD is using a blower-style cooler here with a dual-slot design, which is different to the thicker dual-fan reference Radeon RX 6800 XT and Radeon RX 6900 XT graphics cards. We should expect lowered GPU clocks on the Radeon PRO versus the Radeon RX gaming graphics cards, with under 250W TBP expected.

It's not the first time it has happened, and it might not be the last time it'll happen -- but NVIDIA CEO Jensen Huang will be hosting the GPU Technology Conference 2021 keynote from his personal kitchen.

NVIDIA has announced: "The keynote, delivered again from the kitchen in Huang's home, will kick off a conference with more than 1,500 sessions covering just about every innovation - from quantum computing to AI - that benefits from moving faster".

GTC 2021 will kick off on April 12 @ 8:30 AM PST, where we should hopefully see NVIDIA unveil a bunch of new products, talk about their current successes, and everything in between. GTC isn't a gaming-focused event so there shouldn't be a new GeForce graphics card, but rather new workstation-focused Quadro graphics cards, new Tesla accelerators, and more.

NVIDIA's upcoming GPU Technology Conference 2021 is about to kick off, with the company teasing people with a GTC 2021 Scavenger Hunt with some "hidden treasures" to be found.

The first discovery of NVIDIA's surprise "hidden treasures" for GTC 2021 were unearthed by Wccftech, confirmed by VideoCardz where there is a hidden message inside of the NVIDIA GTC 2021 marketing videos. The light bulbs on the video above are attached to a neural network pattern blink in a pattern... which is of course Morse Code.

The NVIDIA hidden treasure was worked out, and was a webpage that NVIDIA congratulates you for finding, saying "You've discovered our first hidden treasure. We love rewarding curious minds. On Monday, April 12 (8:30 a.m. Pacific), the NVIDIA GTC keynote will go live and we thought we'd have a little bit of fun by adding a few "Easter eggs." These are clues & hidden treasures for those with eagle eyes to discover".

We saw our first tease of the new custom Lenovo Radeon RX 6900 XT LEGION graphics card in February, but the card is now here and being tested in Lenovo LEGION gaming PCs. Check out the Lenovo Radeon RX 6900 XT LEGION:

Lenovo is heavily inspired by AMD's own wicked Radeon VII reference graphics card design, with a triple-fan, triple-slot card that has a full custom cooling solution and non-reference PCB. Lenovo is tapping dual 8-pin PCIe power connectors, unlike some of the custom Radeon RX 6900 XT graphics cards which require another 6 or 8-pin PCIe power connector. Not bad, Lenovo.

The company uses a different dark design with their Lenovo Radeon RX 6900 XT LEGION graphics card, compared to AMD using a silver design theme with its reference Radeon VII graphics card. Lenovo will be using this design for both the Radeon RX 6800 XT and Radeon RX 6900 XT LEGION graphics cards.

AMD surprise launched a new Radeon RX 6800 XT Midnight Black graphics card today, which sold out as quickly as it was announced.

The new AMD Radeon RX 6800 XT Midnight Black edition is an all-black Big Navi graphics card launched on the AMD.com webstore, with super limited quantities -- we don't know if it'll come back since it sold out so quickly, but I wouldn't be holding my breath for that one.

As for the card itself, it's the same Radeon RX 6800 XT reference graphics card but dipped in the all-black "Midnight Black" theme. We still have the Navi 21 XT GPU with 4608 Stream Processors, 16GB of GDDR6 memory and 128MB of Infinity Cache, dual 8-pin PCIe power connectors using up to 300W of power, and a 2.5-sloit design with a cost of $649.

What? Is that headline real? April Fools is over, right?!

Yes, you read that right -- NVIDIA is rumored to bring the flagship Pascal-based GeForce GTX 1080 Ti graphics card because of the huge GPU shortages worldwide. You can't buy an Ampere GPU right now with most GeForce RTX 30 series graphics cards, and the GeForce GTX 1080 Ti is still a killer graphics card so it kinda makes sense... right?

The news is coming from someone who returned their EVGA GeForce GTX 1080 Ti SC graphics card under warranty, with the returned GPU saying it was manufactured in 2021. Yeah... a GeForce GTX 1080 Ti graphics card manufactured this year. This actually could make a lot of sense given the GeForce GTX 1080 Ti still kicks so much ass today in 2021, but what about the price if it re-launched?

We just got a decent information dump on Intel's next-gen Xe-HPG graphics card, the upcoming gaming GPU that will pack 16GB of GDDR6 and performance that should target the NVIDIA GeForce RTX 3070 Ti graphics card. But what about a DLSS competitor?

Now we're hearing about Intel XeSS which has been mentioned as an internal codename for Intel's upcoming NVIDIA DLSS competitor, which is already magic in itself. XeSS makes sense and I actually love the name, so I'm hoping that Intel keeps this XeSS codename for its upcoming Xe-based graphics cards for gamers.

We don't know anything else about Intel XeSS apart from its codename reveal today by Tom @ Moore's Law is Dead, but if it's anything like DLSS (and more specifically like NVIDIA's newer DLSS 2.0 technology) then that is going to be a great thing for Intel to have going forward as it moves into the graphics card market (again).

Intel's upcoming Xe-HPG gaming graphics card is back in the spotlight with Tom at Moore's Law is Dead posting a new video with some juicy new information. Check it out:

In the video we're told that Intel's upcoming Xe-HPG graphics card will have a 512 EU (Execution Units) model that is "initially designed" for TSMC's N6 node -- this could be changed, however -- packing 16GB of GDDR6 memory on a 256-bit memory bus. As for power, Tom reports that his sources tease Intel Xe-HPG's top model will have a 225-250W TDP but they have "since pushed it harder" up to 275W.

As for performance we're looking at somewhere between NVIDIA's new Ampere-based GeForce RTX 3070 and GeForce RTX 3080 graphics cards with Tom pointing out that his sources said "Top Xe should be treated like an RTX 3070 Ti. That is where the majority of samples are performing". As for 3DMark TimeSpy benchmark results the new Intel Xe-HPG in early samples is throwing out scores that are between the GeForce RTX 2080 and the flagship GeForce RTX 3090.

We all know about the GPU shortages but I'm sure none of us expected NVIDIA to ramp up Turing GPU production with rumors that the company is increasing the supply of its GeForce GTX 1650 graphics card.

The rumors are coming Chinese site Channel Gate, which teases that NVIDIA is bringing in the big guns -- well, not-so-big-guns in GPU power but one of their most popular GeForce GTX 10 series graphics cards. NVIDIA never really stopped making the GeForce GTX 1650 but has since had its Ampere GPUs filling in the GeForce RTX 30 series family -- but they're so hard to find right now.

We could expect to see supplies of NVIDIA's mid-range GeForce GTX 1650 graphics card sometime in April-May 2021, so over the course of the next few weeks. The card debuted at a price of around $149 but with the current GPU shortages a new GeForce GTX 1650 will cost you around $400 today.

We've already had a decent look at the patent for AMD's upcoming GPU chiplet technology, but a new patent that was published on April 1 teases a few very interesting new things.

AMD's new GPU chiplet technology patent has an inter-chiplet bridge that has its own integrated cache, which is the main course here versus the previous tease of AMD's upcoming GPU chiplet tech. The active bridge would be a high-bandwidth die-to-die interconnect between the GPU chiplet dies, depending on how many there are.

The active bridge chiplet would handle inter-chiplet communications and route cross-chiplet synchronization signals, where the LLC here means L3 cache -- AKA the new Infinity Cache found inside of the new RDNA 2-based Radeon RX 6000 series graphics cards.