Graphics Cards - Page 117

NVIDIA has just updated the specifications for its new Hopper H100 datacenter GPU, with the updated specs coming just two weeks after the company detailed its huge Hopper H100 GPU.

The new NVIDIA Hopper H100 GPU has improved single-precision and double-precision compute performance numbers, where during the announcement we heard it would have 30 TFLOPS of FP64 compute performance, but Hopper H100 now has 34 TFLOPS of FP64 compute performance.

NVIDIA has up to 16896 CUDA cores inside of its Hopper H100 GPU, which will be clocked at a higher GPU clock of at least 1982MHz -- not the 1775MHz GPU clocks that were previously reported -- NVIDIA is tapping the same custom TSMC N4 process node as its Ada Lovelace GPU architecture inside of the upcoming GeForce RTX 40 series graphics cards.

You've got to give it to Acer, they've crafted one of the best-looking custom Intel Arc graphics cards so far, with a very brief tease of the "Predator Bi Frost" graphics card.

The new Acer Arc A770 Predator BiFrost will feature an interesting design: a hybrid-style blower + open-air fan design that hasn't been done before. It looks fantastic, and I'm sure even better inside of your PC. Inside, the Acer Arc A770 Predator BiFrost will feature the full ACM-G10 GPU with 32 Xe-Cores, but we don't know if it will have 8GB or 16GB of GDDR6 (Intel's in-house Arc A770 Limited Edition has 16GB, but AIB-made Arc A770s can have 8GB).

Acer didn't provide any details of its new Arc A770 Predator BiFrost graphics card, but it will require dual 8-pin PCIe power connectors, while the Intel Graphics Twitter account joining in the launch fun saying "Bravo, team! We're excited to have you on board". Jarred Walton of Tom's Hardware tweeted, asking if it were the 8GB or 16GB model, and what was the TBP, the GPU and GDDR6 clocks.

The release of NVIDIA's next-gen GeForce RTX 4090 and Intel's first-gen Arc A770 and Arc A750 graphics cards will all release on October 3, with GPU v2.50.0 here to support all three of the cards.

GPU-Z v2.50.0 supports everything you need to know about your GPU, especially if you've got a new GeForce RTX 4090 graphics card: detailed BIOS information, GPU clock and GPU overclocking information, power limits, and a very cool and important new feature: a built-in NVIDIA DLSS scanner. This scans your PC for all of the games installed that use DLSS, detailing which exact version of DLSS the game uses (which is very useful for reviewers).

You can download GPU-Z 2.50.0 here.

GIGABYTE has a monster graphics card in its behemoth AORUS GeForce RTX 4090 MASTER graphics card, but on the other side of the scale we have the small -- at least for the card -- powerhouse card with the upcoming AORUS GeForce RTX 4090 WATERFORCE graphics card.

The new AORUS GeForce RTX 4090 WATERFORCE has been teased with a tiny RTX 4090 card, but a large 360mm AIO cooler that will surely keep the next-gen Ada Lovelace GPU nice and cool.

GIGABYTE is re-using the same look, design, and size as its current-gen Ampere-based AORUS GeForce RTX 3090 Ti WATERFORCE with the new AORUS RTX 4090 WATERFORCE. That's not necessarily a bad thing, because it's a good-looking card... and you get all that RTX 4090 horsepower in a smaller card (sans the huge 360mm AIO cooler).

Intel has finally detailed its new Arc A770 and Arc A750 graphics cards with everything we need to know: full specs, pricing, launch dates, the works.

The company has confirmed it will be offering its Arc A770 in both 8GB and 16GB variants, with Intel offering the Arc A770 with 8GB of GDDR6 memory starting at $329, while the Arc A770 with 16GB will cost $349. Intel will be offering its Arc A770 Limited Edition graphics card with 16GB of memory for the $349 price.

Intel provided some more details on its new Arc A750 graphics card, which will cost $289 -- $30 cheaper than the Arc A770 -- but maintains its 8GB of GDDR6 memory.

Well, I guess it's confirmed: ASUS is really upgrading things across the board for NVIDIA's new flagship GeForce RTX 4090, with its custom ROG Strix RTX 4090 retail packaging teased... and it's way more than a handful.

The new ASUS ROG Strix GeForce RTX 4090 OC Edition graphics card has arrived in the hands of the folks of Chiphell, with a shot of the retail packaging and man... there are just no words.

It's absolutely huge, and from the photo alone it looks far bigger than the retail packaging for the ASUS ROG Strix LC GeForce RTX 3090 Ti OC Edition graphics card that I reviewed not too long ago now. The packaging for that I said was absolutely huge, and would command a large portion of your shelf.

Intel is now shipping out its Arc A770 Limited Edition and Arc A750 Limited Edition review samples to tech media, ready for the review embargo on October 5... just days away from now.

The new Intel Arc A770 graphics card review embargo lifts on October 5 while the cards will be officially launched on October 12 -- the very same day NVIDIA's new flagship Ada Lovelace-powered GeForce RTX 4090 graphics card launches -- so it's going to be a busy few months for GPU reviewers, that's for sure.

Intel hasn't been clear with a full list of partners and retailers that will be selling Arc A-series desktop GPUs at launch, even though the promised Q3 2022 launch window has been missed, we STILL don't know which partners + retailers are participating in the launch with Intel. But don't worry, special communication is being done with China as there is a slick Intel Arc A750 Gold Edition graphics card teased.

Intel hosted its recent Innovation 2022 event in San Jose this week, unveiling the new 13th Gen Core "Raptor Lake" CPUs, announcing the release date and price of its new Arc A770 graphics card, and had custom Arc A770 and Arc A750 graphics cards on display.

We can see that the GUNNIR Intel Arc A770 while ASRock had two cards teased: the Phantom Gaming Arc A770, and Challenger Arc A750 graphics cards. We should expect a little bit of overclocking fun on the custom Arc graphics cards, with one of them an "OC" model that should have GPU clocks higher than Intel's own Arc A770 Limited Edition graphics card.

GUNNIR's new custom Intel Arc A770 graphics card was on display, teased with a triple-fan cooler and 2.2-slot design with dual 8-pin PCIe power connectors (up from the 8+6-pin PCIe power connectors on the Intel Arc A770). I really dig the design here, as it feels like it's the best of the AMD Radeon Vega Frontier Edition + Radeon VII graphics cards.

Intel's new Arc A310 graphics card has just been teased on the official Intel website, with some specifications on Intel's new entry-level desktop Arc GPU.

The specifications for the Intel Arc A310 point to the entry-level GPU featuring 6 Xe-Cores clocked at either 2100MHz or 2000MHz -- as there's some conflicting information here -- with Intel making the Arc A310 graphics card on TSMC's N6 process node.

Intel's entry-level Arc A310 graphics car has 4GB of GDDR6 memory clocked at 15.5Gbps on a 64-bit memory bus, with up to 124GB/sec of memory bandwidth and a 75W TDP. Intel's own website says that the Arc A310 launches in Q3 2022, meaning we only have a couple of days left for that to happen or someone is going to have to go in and ninja-edit that entry.

NVIDIA's next-gen Ada Lovelace GPU architecture will deliver the world DLSS 3 technology, exclusive to the new GeForce RTX 40 series graphics cards... and now Marvel's Spider-Man Remastered has been put through its paces with the Ada-powered AI upscaling tech.

We've seen NVIDIA tease the performance of its new GeForce RTX 4090 with DLSS 3 in games like Marvel's Spider-Man Remastered, Microsoft Flight Simulator, Cyberpunk 2077, and more... but now the folks at Digital Foundry are testing with the RTX 4090, DLSS 3, and Spider-Man Remastered.

Digital Foundry has already played around with Cyberpunk 2077 and its update that will usher in Overdrive RT + DLSS 3 support, but also early drivers to the GeForce RTX 4090. The team at DF posted an article and large 30-minute video, showing off DLSS 3 with side-by-side comparisons in Marvel's Spider-Man Remastered against DLSS 2 and native 4K.

Here we are again: Intel is gearing up for the launch of its Arc A770 graphics card... we guess... as VideoCardz is reporting that new embargo guidelines are going out to GPU reviewers for the Arc A770.

Intel is reportedly launching its new Arc A770 and Arc A750 graphics card at the same time, but we don't know if they'll be launching at the same time. But for the flagship Intel Arc A770 graphics card, unboxings of the Arc A770 take place on September 30, while Arc A770 reviews will reportedly go live on October 5.

Astute readers will realize that Intel is months and months late for the launch of its Arc GPU... but hey, it's only two more weeks, and this time we're not flattening the curve. Intel GPUs should be detailed during the Intel Innovation event being held today in San Francisco for the unveiling of its 13th Gen Core "Raptor Lake" CPUs.

Intel has finally launched its new Arc A770 graphics card during its Innovation event, with the Intel Arc A770 Limited Edition launching on October 12 starting at $329.

The new Intel Arc A770 graphics card was detailed during Intel Innovation by company CEO Pat Gelsinger, costing $329 and finding itself in a market of graphics cards that are far more expensive than the Arc A770. Inside, the Intel Arc A770 features the full ACM-G10 GPU with 32 Xe-Cores and 16GB of GDDR6 memory (on the Arc A770 Limited Edition) and 8GB of GDDR6 on AIB custom models of the A770.

I don't know why, but Intel took the time to talk about ray tracing domination with its Intel Arc A770 graphics card, where the company claims it has 65% better peak performance versus the competition in ray tracing. Intel, your cards aren't competing with the high-end, and no one buying a $329 is running out to the shops to buy a lower-end GPU for playing games with ray tracing enabled, but hey... investors will love it.

We know that NVIDIA has a silicon monster with its flagship new GeForce RTX 4090 graphics card, built for 4K gaming and beyond -- and when we say beyond, we mean 8K gaming -- which the RTX 4090 is tuned perfectly for.

First off you've got an unbelievable amount of GPU horsepower with the AD102 GPU inside of the GeForce RTX 4090, joined by a large 24GB framebuffer, with super-fast GDDR6X memory with 1TB/sec+ of memory bandwidth at your disposal.

But in a new post on NVIDIA's own landing page for 8K HDR gaming on its website, the company explains that the GeForce RTX 4090 lets you play, capture, and watch 8K HDR. If you don't know how good 8K is, it's 4x as many pixels as 4K and 16x as many pixels as 1080p.

Corsair is preparing its customers for the world of NVIDIA's new GeForce RTX 40 series cards, and in particular, the flagship GeForce RTX 4090. Corsair has new 600W 12VHPWR cables and new water blocks on the way to chill down that Ada beast.

NVIDIA's new GeForce RTX 4090 uses the new 12+4-pin 12VHPWR power connector, with all RTX 40 series graphics cards shipping with a 3 x or 4 x 8-pin PCIe power to 12VHPWR power adapter. If you didn't want to wait for future PSUs and new ATX 3.0 support, then Corsair is ready for you with their existing PSUs that provide everything you need to drive next-gen hardware.

Corsair says its current RM1000 PSU right through to a brand new HX1500i PSU will power graphics cards into the future. The company is already shipping new "Corsair 12VHPWR 600W cables" that are compatible with all Corsair Type-4 PSUs, where the new 16-pin power connector plugs directly into the PSU with existing PSU-side connectors, and then directly into your beasty new Ada Lovelace-powered GeForce RTX 4090 graphics card.

NVIDIA only just unveiled their next-gen GeForce RTX 4090 graphics card, but the official release isn't for another couple of weeks -- October 12 to be precise -- but this hasn't stopped GIGABYTE's custom RTX 4090 graphics cards from hitting the market in Hong Kong early.

GIGABYTE's upcoming custom GeForce RTX 4090 GAMING OC 24G graphics card has turned up in Hong Kong for around 20,000 HK (around $2548 USD or so) which is quite the premium: nearly $1000 USD over the MSRP of the RTX 4090 at $1599.

But there's something to remember here: there are no GeForce drivers that are compatible with the GeForce RTX 40 series graphics cards just yet, so this would be an expensive (but very pretty) paperweight, at least for now. NVIDIA will be pushing out Ada Lovelace compatible GeForce drivers soon, while the public release of those drivers should take place on October 12, the day that NVIDIA launches the GeForce RTX 4090.

NVIDIA has another generation of GPU arriving, so it's time for another custom GeForce RTX graphics card: this time it's Cyberpunk 2077 -- again, actually.

NVIDIA and CD PROJEKT RED are teaming up for a big contest where they're giving away 3 x GeForce RTX 4090 graphics cards with a custom Cyberpunk 2077 backplate. Cyberpunk 2077 and Witcher developer CD PROJEKT RED will be sharing clues on the official Cyberpunk 2077 Twitter account, with the tasks including taking special in-game screenshots.

Those who are the most creative with their photo-snapping skills in Cyberpunk 2077 will be up for one of three of the next-gen NVIDIA GeForce RTX 4090 graphics cards with the custom Cyberpunk 2077 backplate. NVIDIA has some super-powered silicon inside of the RTX 4090, with the new AD102 GPU pumping away with 24GB of GDDR6X memory and a 450W TDP.

NVIDIA might have just announced its next-gen Ada Lovelace GPU architecture, while the Hopper GPU architecture lives alongside it in the data center... and now there's a new Hopper H100 GPU with 120GB of HBM2e (or HBM3, it's not clear from the picture) memory floating around in the world.

A new PCIe-based graphics card called the "NVIDIA H100 120GB PCIe" has surfaced, looking like it should be the beefed-up NVIDIA Hopper H100 GPU with 120GB of HBM2e/HBM3 memory. NVIDIA makes a few different SKUs of its Hopper H100 GPU, with SXM and PCIe models but the SXM variant rocks higher-end HBM3 memory.

The model teased here in the screenshot reportedly has 120GB of HBM memory (HBM2e or HBM3 it's not clear) which is a sign of things to come for NVIDIA customers. For now, 80GB of memory is "all" you get... with 120GB variants in PCIe form at least, launching in the coming months it seems.

We are only weeks away from AMD's big next-gen Radeon RX 7000 series "RDNA 3" graphics cards reveal on November 3, but ahead of that there's Radeon Monster Profile.

Radeon Monster Profile (RMP) is made by developer Yuri "1usmus" Bubliy, who is the creator of Hydra, and last teased Hydra 1.2 and its new OC tool that would auto-tune your RDNA 2 GPU, as well as including Ryzen 7000 series "Zen 4" CPU support.

But the new Radeon Monster Profile is something exciting: it combines the unique settings and a new volt-frequency curve for AMD's current-gen RDNA 2 GPUs, which will let the graphics card you own "reveal its maximum potential". Yuri explains that he paid particular attention to the memory subsystem, where the operating voltage was reduced by 6% which reduced the heat of the GDDR6 memory modules.

PNY has outed their new custom GeForce RTX 4090 that will drop into their XLR8 Gaming family of GPUs, with the introduction of the VERTO GeForce RTX 4090, GeForce RTX 4080 16GB, and GeForce RTX 4080 12GB graphics cards.

The new PNY VERTO GeForce RTX 4090 features a base vapor chamber, 8 heat pipes, a triple-fan cooler, and durable twin ball bearing fan hubs, with it all build into a sleek aluminum backplate that protects the important (and very sensitive) components on the card, as well as act as additional cooling.

PNY's new VERTO GeForce RTX 4080 16GB graphics card has the same 8 heat pipes but drops the base vapor chamber, but retains the rest: triple-fan cooler, and more. PNY does kick things down a notch or two with the VERTO GeForce RTX 4080 12GB graphics card, with 7 heat pipes compared to 8 heat pipes on the RTX 4090 and RTX 4080 16GB.

NVIDIA's next-gen GeForce RTX 40 series GPUs do not support NVLink, and have no NVLink physical connector on the cards whatsoever... disappointing, but expected.

The news is coming directly from NVIDIA founder and CEO Jensen Huang, who on a call with the tech press he said that Ada Lovelace drops the NVLink connector so that the I/O room for "something else". We do know that NVIDIA has crammed a helluva lot more into Ada Lovelace GPUs, with transistors, CUDA core, and ROP counts going through the roof.

Jensen said that NVIDIA engineers wanted to use as much of the silicon area they had their hands-on to "cram in as much AI processing as we could". We also now know: Ada Lovelace does NOT support PCIe 5.0, but NVIDIA included the Gen5 power connector.