Graphics Cards - Page 73

The HBM market is one of the most important markets in the world right now, with the meteoric rise of AI powered by vast amounts of AI GPUs powered by HBM memory... including HBM3, HBM3e, and soon the next-gen HBM4 memory standard.

Because of the dominance of AI GPUs, the HBM memory business is expected to double in market revenue in 2025 as there are more and more next-gen AI GPUs released. We've got the just-announced AMD Instinct MI300X AI GPU, as well as NVIDIA's beefed-up Hopper H200 AI GPU and their next-gen Blackwell B100 AI GPU coming out this year... all of which are powered by HBM.

Market research outlet Gartner reports that the HBM market will rocket up to $4.976 billion in revenue by 2025, which is close to double the HBM revenue of 2023. Right now, these numbers are based purely on current and anticipated demand from the industry, but things can change very quickly. Just look at the AI GPU demand right now; it's insatiable, and the big driver behind AI GPUs getting faster and faster is high-speed HBM memory.

MaxSun's RTX 4070 Super iCraft graphics card is notably different to all the other 4070 Super models now out there in that it sticks with a dual 8-pin power connector.

As HXL, a regular leaker on X (formerly Twitter), pointed out, the iCraft version of the RTX 4070 Super has an 8-pin + 8-pin setup rather than the 16-pin (12VHPWR) power connector seen on every other RTX 4070 Super model (at least to our knowledge).

Indeed, as VideoCardz, which spotted the above tweet, made clear, the thinking was NVIDIA had made it compulsory for third-party graphics card makers to use the 16-pin connector on their RTX 4070 Super boards, and they all do. Well, save for this one apparent exception from MaxSun, of course.

NVIDIA's current-generation GeForce RTX 4090 Laptop GPU and GeForce RTX 4080 Laptop GPU can now be found inside desktop GPUs, at least in China.

According to a new article on VideoCardz, Chinese companies that specialize in marketing mobile GPUs as desktop GPUs with a low operating margin by using less expensive mobile GPUs, budget cooling designs, and simpler PCBs, with their power limits capped at 175W because they are mobile GPUs.

This isn't all sunshine and rainbows, as there are things like warranty that gamers have to worry about, as well as possible issues when it comes to driver support. VideoCardz reports that they get emails from readers who have "difficulties" with using these cards. Consumers of these mobile-now-desktop GPUs have to use modded drivers or seller-provided drivers that we don't know where they come from.

GALAX has introduced its new GeForce RTX 4070 SUPER HOF OC Lab Master-X and Master Edition graphics cards, with a beautiful all-white look GALAX is known for... check them out:

The GALAX HOF (Hall of Fame) graphics cards are now rolling out in new RTX 40 SUPER series cards now that NVIDIA has introduced its new GeForce RTX 4070 SUPER, RTX 4070 Ti SUPER, and RTX 4080 SUPER graphics cards into the wild. GALAX is introducing a new look for its HOF OC Lab series graphics cards, with plans to release all RTX 40 SUPER cards, starting with the RTX 4070 SUPER.

Inside, the new GeForce RTX 4070 SUPER features the AD104-350 GPU with 7168 CUDA cores, which is a 20% bump over the RTX 4070, with the same 12GB of GDDR6 memory as the RTX 4070 non-SUPER. NVIDIA has a reference boost GPU clock of 2475MHz for its RTX 4070 SUPER, while GALAX is offering the fastest GPU boosts for an RTX 4070 SUPER so far with its new GeForce RTX 4070 SUPER HOF OC Lab Master Edition graphics card.

NVIDIA's RTX 4070 Super is only just out - literally yesterday - but this graphics card has already sunk below its MSRP, although not in the US, mind.

In Europe we've seen price tags drop below the recommended pricing set by NVIDIA, and as VideoCardz (and Wccftech, 3DCenter) made clear, this has happened in three countries.

Firstly, in Poland, an online retailer has dropped the price of an RTX 4070 Super by 5% under the local MSRP - that's for a Gainward model (Ghost).

NVIDIA's RTX 4070 Ti Super could be difficult to get your hands on, or at least stock levels might be shaky according to a new rumor, so a decent level of demand may see the graphics cards soon vanish from shelves when it's released.

That's the theory floated by Moore's Law is Dead (MLID) in his latest YouTube video which explores the RTX 4070 Super launch primarily - spoiler alert: it went badly with sluggish sales if retail and distributor sources are to be believed - but it also touches on the other new RTX Super refreshes which aren't out yet.

We're focusing on what MLID says about the RTX 4070 Ti Super in particular here, which is the next of NVIDIA's refreshes to arrive, landing next week on January 24 (to be followed by the RTX 4080 Super at the very end of the month).

SPARKLE has added a pair of new Intel Arc GPUs to its line-up, this time in the form of a few low-profile offerings that would suit SFF builds or something like a home theater PC. The SPARKLE Intel Arc A310 ECO is a low-profile, single-slot, single-fan GPU with a reduced 50W power profile, while the SPARKLE Intel Arc A380 GENIE is a 75W low-profile, dual-slot, dual-fan GPU.

Outside of the awesome all-blue SPARKLE Arc aesthetic, what's impressive is that both don't need anything in the way of external power from a PSU while offering three display outputs with 1 x HDMI and 2 x mini-DisplayPort connectors on the back.

As a 4GB card with only 124 GB/s of memory bandwidth on a 64-bit bus, the SPARKLE Intel Arc A310 ECO isn't designed for gaming, but with AV1 hardware encoding and AI hardware, it should be decent for light workloads. The SPARKLE Intel Arc A380 GENIE includes 6GB of memory on a faster 96-bit bus to handle some low-fidelity gaming.

With our reviews of the GeForce RTX 4070 SUPER Founders Edition and the INNO3D GeForce RTX 4070 SUPER TWIN X2, we got a good look at what the reference design delivered over the existing GeForce RTX 4070. And yes, the pre-release hype was real - 1440p gaming performance faster than the previous generation flagship GeForce RTX 3090.

Throw in no price increase over the RTX 4070 launch MSRP, and the compact, super-efficient designs make the first SUPER GPU for the 40 Series pretty exciting.

As with all GPUs, there are multiple models, cards with more premium cooling, overclocked specs, and more. Today, our full review of the ASUS TUF Gaming GeForce RTX 4070 SUPER OC Edition has gone live, and it's like getting a bonus chapter or SUPER-hero post-credits scene that adds to the story.

NVIDIA's RTX 4070 Super is on sale and reviews of different models are everywhere (we've got a write-up of the Founders Edition and a third-party INNO3D version of this graphics card too).

TechPowerUp turned up an interesting finding, though (hat tip to VideoCardz), namely that there are two RTX 4070 Super models that sport a TDP of 245W at stock - GIGABYTE's AORUS Master and Palit's Jetstream cards.

The Founders Edition is pitched at 220W (up from 200W with the RTX 4070), and the max power usage cap for NVIDIA's own variant is 240W, so by default, these third-party boards are slightly exceeding that.

NVIDIA has a fresh Game Ready driver out which introduces support for the new RTX 4070 Super that just hit shelves today, plus it does a fair bit more besides.

Driver version 546.65 ensures that owners of a shiny new RTX 4070 Super GPU are catered for as mentioned, and also optimizes performance for Palworld, introducing DLSS 2 for this game.

Palworld (pictured top of this article) is an open world multiplayer survival game built around collecting monsters and crafting. You can capture these beasts, fight with them, force them to work, sell them, or indeed slaughter them for food, so, yes - there are options for your poor monster minions.

After new US sanctions on GPU hardware effectively banned NVIDIA's flagship GeForce RTX 4090 from being sold in China, it didn't take long for the company to quietly announce a special cut-down model for the region called the GeForce RTX 4090D.

With fewer CUDA Cores, RT Cores, and Tensor Cores compared to the baseline GeForce RTX 4090 but the same price, we now know the impact on performance. And it's not too bad, and certainly not a massive nerf.

With a new GALAX RTX 4090 D Metal Master card put to the test by Expreview, it turns out that it's only 5.5% slower on average for rasterized gaming, which increases slightly to 5.8% when turning on ray-tracing. However, this gap drops to 5% when DLSS 3 Frame Generation is enabled. Either way, 5% makes the new GeForce RTX 4090 D similar to the RTX 4090 - with the difference barely noticeable for gaming workloads.

ASUS has just unveiled its latest low-profile graphics card, with the introduction of the GeForce RTX 4060 LP BRK graphics card; check it out:

Inside, the new ASUS GeForce RTX 4060 LP BRK graphics card features the AD107-400 GPU that has 3072 CUDA cores and 8GB of GDDR6 memory. There's no factory overclocking on the ASUS RTX 4060 LP BRK graphics card, which is aimed at small form factor gaming PCs, so you want to keep those thermals contained.

ASUS uses a custom cooling design with a triple-fan cooler mounted on a pretty big -- at least for the low profile card -- heatsink. ASUS keeps the RTX 4060 LP BRK nice and short but maintains the chunkier dual-slot design. I think that's a good idea, having the card as a dual-slot design but keeping the fantastic cooling and shorted SFF design.

AMD and NVIDIA are the two biggest players when it comes to discrete gaming GPUs, but with its dominant market share and recent innovations like DLSS Super Resolution, Reflex latency reduction, and DLSS 3 Frame Generation, the consensus by many is that NVIDIA leads while AMD follows.

In the case of DLSS, AMD's FSR or FidelityFX Super Resolution took a while to hit the market - with FSR 2 proving to be a success regarding image quality. In the case of FSR 3, which includes frame generation and super-resolution upscaling, it took almost a year after the arrival of DLSS 3 for AMD's equivalent tech to appear in a few select games.

Unlike NVIDIA's DLSS 3, FSR 3 works on all GPUs - Radeon and GeForce alike - but still requires per-game implementation by developers. The company's latest endeavor, HYPR-RX, brings upscaling and frame generation versions to drivers - meaning they can be instantly enabled in hundreds of games. And with this, AMD believes it now has the edge over NVIDIA.

ASUS has its flagship ROG Strix GeForce RTX 4080 SUPER graphics card listed by European retailers for upwards of £1,350 in the UK, which works out to $1700+ USD.

OverclockersUK has five different options when it comes to new GeForce RTX 4070 SUPER graphics cards, while there are four each of the RTX 4070 Ti SUPER and RTX 4080 SUPER when it comes to new custom models from AIB partners.

There's no real sticking to NVIDIA's tighter MSRP, but you can get the ASUS RTX 4070 SUPER DUAL graphics card for £599 at OverclockersUK. If you want the faster RTX 4070 Ti SUPER, there's the TUF variant for £899, while if you want the best the new SUPER cards have to offer, there's the ASUS TUF Gaming RTX 4080 SUPER for £1199.

NVIDIA's RTX Super refreshes might be grabbing all the limelight right now, but there's a budget move afoot (in theory) that's much more under the radar with the old RTX 3050.

Rumors have been circulating for a while now about a new take on Ampere's RTX 3050 which has 6GB of VRAM (rather than 8GB), but a new leak suggests that the spec could be cut down more than just having the video memory reduced.

This isn't the first time we've heard theorizing that the RTX 3050 6GB might be reined in quite a bit, disappointingly for gamers on a budget, but the new specs here are quite a drop down indeed.

NVIDIA released its interesting, almost concept-style GeForce GTX 295 graphics card 15 years ago now... with sandwiched PCBs with dual GPUs enabling SLI on a single card back in 2009.

The NVIDIA GeForce GTX 295 with dual GPUs cost just $499 back in the day, which isn't bad for a ridiculous dual-GPU monster card. It isn't a monster today at all, with its GT200 GPU made on the older 55nm process node. Each GT200 GPU featured 240 unified processors for a total of 480 unified processors through the dual-GPU design, twice that of the GeForce GTX 280 and its single GPU.

At the time, the GTX 280 had its benefits -- from being a single-GPU card -- but the GTX 295 dual-GPU graphics card wasn't just two GTX 280s together. NVIDIA had a monster dual GT200 GPU sitting on the GeForce GTX 295, but the rest of the specs were knocked down between the GTX 260 and GTX 280 single-GPU graphics cards.

MSI had a few new GPUs to show off at CES 2024 in Las Vegas, Nevada. The good-looking chap above is the RTX 4080 Super 16G EXPERT, and we are in love with it. It just might be a shame or a crime to put a waterblock on such a beautiful GPU. Now, this GPU has MSI's ZERO FROZR tech to keep the fans from spinning until a load is present. When there is a load, the fans spin much like the Founders Edition of the RTX 4080/RTX4090 does, but with more static pressure.

This GPU also features a vapor chamber that covers both the GPU core and the memory modules, and then the heat is brought out via heat pipes to the fin array.

Now it's time to discuss the RTX 4070 Super 12G Ventus 2x White OC. Featuring a dual fan dual slot heatsink, which is quite a rarity nowadays. Featuring two of the Torx Fan 4.0, a copper baseplate, core pipes, and a reinforcing backplate rounds out list of features.

A fresh rumor has sprung up around NVIDIA's imminent RTX Super refreshes for its Lovelace range that suggests the RTX 4080 Super might have a little more pep in its step than previously thought.

The spillage was brought to the attention of X (formerly Twitter) users by I_Leak_VN and is supposedly from a presentation to Vietnamese retailers (as flagged by Wccftech).

The slide shown contains the full range of RTX 4000 models and pricing along with comparative performance and value calculations - add your own seasoning and don't go light with the saltshaker.

Here's some good news for PC gamers with a graphics card from Team Red - AMD's Fluid Motion Frames tech is officially debuting on January 24.

AFMF for short, also known as frame generation (as first seen with NVIDIA DLSS 3), the technology has been in preview for quite some time - since September 2023 in fact - but it's about to graduate to full release.

AFMF is frame generation at the driver level, so it can work across all games as a result, with caveats. Firstly, it's not going to give you the same results as FSR 3 (which incorporates frame generation, too, but does a lot more than just this - though games have to be coded to support it, of course).

While at CES 2024 in Las Vegas, we were invited to take a look at what ASUS had on display. Conveniently, both AMD and NVIDIA just announced new GPUs while at CES 2024. AMD launched the RX 7600 XT and NVIDIA the RTX Super series consisting of the RTX 4070 Super, RTX 4070 Ti Super, and the RTX 4080 Super. So, let's take a look at how ASUS has up its sleeves.

The big boy, the Strix RTX 4090 BTF OC Edition, was a GPU that was not really talked about but is welcomed anyway. This version has a patented vapor chamber with a milled heat spreader for lower GPU temperatures. The real story here is the BTF (back to front) power connectors on this high-end GPU, which is exciting here with no visible need for the traditional 8-pin or 16-pin external power connections. The release date and the price are unknown.

Next up is the new TUF Gaming GeForce RTX 4070 Ti Super BTF White Edition. It's all about the white this year at CES 2024, and this GPU is certainly no exception. The TUF gaming cooler seems to just be a superficial paint job over the original TUF cooler. The BTF (back to front) part is what is exciting here again. Available Jan 24th, 2024. There is no official word on the price as of yet.