Graphics Cards - Page 108

NVIDIA CEO Jensen Huang flew out to Taiwan last week to meet with AIB partners and fan the (hopefully not explosive) waves surrounding the 12VHPWR power connector on its new GeForce RTX 40 series "Ada Lovelace" GPUs... but he had other things on his mind.

The billionaire CEO and founder of NVIDIA were walking the streets of Taiwan, where he walked up and into the opening night of the Taiwan esports carnival "WirForce 2022" and right into the BYOC (bring your own cocktail, probably) and began talking, and joking around with gamers and live streamers.

The Commercial Times reports that Jensen Huang was definitely recognized by some people, but others (like the girls in the embedded video where he's singing Lady Gaga tunes) had no idea who they were singing with. Of course, esports players were getting out their boxes of NVIDIA GeForce graphics cards and getting Jensen to sign them, as well as the obligatory selfies with the NVIDIA founder and CEO.

NVIDIA has cut down its A100 Tensor Core GPU to meet the demands of US export controls to China, with the introduction of the new A800 Tensor Core GPU that is exclusive to the Chinese market.

The new NVIDIA A800 was revealed by Reuters and confirmed by an NVIDIA spokesperson who confirmed that the new NVIDIA A800 GPU went into production in Q3 2022, acting as "another alternative product" to the NVIDIA A100 GPU for customers in China.

The NVIDIA spokesperson told Reuters: "The NVIDIA A800 GPU, which went into production in Q3, is another alternative product to the NVIDIA A100 GPU for customers in China. The A800 meets the U.S. Government's clear test for reduced export control and cannot be programmed to exceed it".

NVIDIA is now not the only one that is using the new 12VHPWR power connector on its graphics cards, with Intel announcing its new Data Center GPU Max 1100 GPU series.

Intel's new Data Center GPU Max 1100 GPU series GPU packs the new 16-pin PCIe power connector, which will power its new Ponte Vecchio GPU architecture. The company will be offering up its new Data Center GPU Max 1000 series GPUs in multiple SKUs using the GPU architecture, either in OAM form factor, on in regular PCIe.

The company explains that its new Data Center GPU Max Series was formally codenamed Ponte Vecchio, which is now the industry's highest-density processor that packs over 100 billion transistors that are spread across 47 active tiles and feature up to 128 Xe-HPC cores.

AMD has just announced its latest Raise the Game promotion, where any purchase of an AMD Radeon RX 6000 series graphics card will include free copies of The Callisto Protocol and Dead Island 2.

Anyone who purchases an AMD Radeon RX 6950 XT, 6900 XT, 6800 XT, 6800, 6750 XT, 6700 XT, 6700, 6650 XT, 6600 XT, or 6600 will receive a free copy of The Callisto Protocol and Dead Island 2. AMD's last two graphics cards in the Radeon RX 6000 series in the Radeon RX 6500 XT or 6400 won't get a copy of The Callisto Protocol, and instead, only get a free copy of Dead Island 2 (because they're much cheaper cards).

In the US, retailers including the official AMD Store, Amazon, Newegg, and Micro Center will have AMD's new Raise the Game bundle for any new Radeon RX 6000 series GPU purchase. Across the pond, European retailers include the AMD Store, Infomax, Topachat, MediaMarkt, and others.

NVIDIA's new GeForce RTX 4090 graphics card has been in the headlines for many reasons over the last few months, but one of the more controversial reasons is it uses the new 12VHPWR power connector... which has been melting and destroying products that should never have this issue.

But now, the organization behind the PCI-Express standard itself -- PCI-SIG -- is issued an engineering change notification for the 12VHPWR connector. This isn't exactly public yet, but the internet being the internet we have the latest leaks... with revision updates for the 12VHPWR connector posted by organization members including Amphenol.

PCI-SIG's proposed changes to the 12VHPWR connector would include more shroud, and a gripping feature for the additional 4 data pins (the 16-pin connector has 12 pins for power, and 4 pins for data). These proposed changes wouldn't be a solution to the issues surrounding NVIDIA's new flagship GeForce RTX 4090 graphics card seeing their 12VHPWR connector melting... because the additional 4 pins aren't carrying high amounts of electrical load.

COLORFUL has just unveiled its new iGame GeForce RTX 4080 Advanced graphics card, a new custom design that packs up to 400W of power.

The new COLORFUL iGame GeForce RTX 4080 Advanced graphics card was teased back in September, with the company unleashing the iGame GeForce RTX 4090 Advanced OC-V (which I reviewed at launch) but the iGame GeForce RTX 4090 Advanced hasn't been released so far.

However, COLORFUL isn't launching one, but rather two iGame GeForce RTX 4080 Advanced graphics cards... the first is at stock, and the second is factory overclocked. COLORFUL teases up to 2625MHz GPU boost clocks on the factory overclocked iGame GeForce RTX 4080 Advanced through hits OC BIOS... chewing down up to 400W of power, too.

GIGABYTE has just announced a rather large fleet of GeForce RTX 4080 graphics cards, with the AORUS XTREME WATERFORCE series, AORUS WINDFORCE series, AORUS MASTER, AERO, EAGLE, and GAMING series GeForce RTX 4080 16GB graphics cards.

Starting with the flagship AORUS GeForce RTX 4080 XTREME WATERFORCE WB graphics card, for users who have an existing open-loop liquid cooling PC. GIGABYTE has improved the thermal system on its new AORUS GeForce RTX 4080 XTREME WATERFORCE WB graphics card, with optimal channel spacing between the micro fins for improved heat transfer from the AD104 GPU through stable water flows.

The company has sunk-designed copper micro fins that shorten the heat conduction path from the GPU, so that heat can be transferred into the water channel area quickly. GIGABYTE is using its patented "Leak Detection" technology here, which in reality is a built-in detection circuit that covers the entire fitting and water block and will alert users by flashing lights if there is any leak. This means you can jump to it, fixing the leak and stopping further damage from occurring to your system.

NVIDIA has unleashed its new GeForce RTX 4090 graphics card and is barely a week away from the release of its second Ada Lovelace GPU, the upcoming GeForce RTX 4080 graphics card.

But we all know the story... NVIDIA had not one but two GeForce RTX 4080 graphics cards -- the GeForce RTX 4080 16GB and GeForce RTX 4080 12GB -- but then NVIDIA "unlaunched" the GeForce RTX 4080 12GB variant. The GeForce RTX 4080 16GB is still marching ahead for release on November 16, but now we've got some more information on what NVIDIA will be doing with the Ada Lovelace GPU release after that.

We've now got leaker "kopite7kimi" that tweeted that the GeForce RTX 4080 12GB will "become" the GeForce RTX 4070 Ti, which could be cut down on the original GeForce RTX 4080 12GB specs that we were hearing about months ago.

EK Water Blocks has just unveiled its new EK-Quantum Vector² water blocks for AMD's just-announced Radeon RX 7900 XTX reference graphics card.

The water cooling giant makes note that because AMD's new Radeon RX 7900 series graphics cards include special memory cache dies (MCDs) next to the main chip -- or the Graphics Compute Die (GCD) -- EK had to make some modifications to the cooling engine to keep the MCDs, as well as the regular GPU components, cool (the VRMs, VRAM, Voltage Controllers, and the chip die).

EK points out that AMD is tapping TSMC's new 5nm process node for its GCD, which pushes for up to 54% more performance per watt, and that even if the GPU is efficient the power consumption is higher than the previous-gen RDNA 2-based GPUs. AMD is sucking up to 355W of total board power for its flagship Radeon RX 7900 XTX graphics card and has put it into its smallest package ever: 300mm² GCD that contains 165% more transistors per mm² than the previous generation.

NVIDIA's new GeForce RTX 4080 graphics card isn't too much further away now, with the second Ada Lovelace card releasing November 16 starting at $1199... and now we're getting a look under the covers of the NVIDIA GeForce RTX 4080 Founders Edition graphics card.

The upcoming NVIDIA GeForce RTX 4080 Founders Edition graphics card has been torn down by @KittyYyuko on Twitter, with a cleverly placed EK sticker and even an Intel CPU placed on top of the board and GPU... so we don't know how many VRMs that NVIDIA is using on its in-house GeForce RTX 4080 Founders Edition graphics card.

You can see the AD103 GPU that NVIDIA is using for the GeForce RTX 4080 in the center of the card, while there's 16GB of GDDR6X memory surrounding it. There are 8 x 2GB modules of GDDR6X memory, adding up to 16GB of GDDR6X which sits on a 256-bit memory bus that provides up to 716GB/sec of memory bandwidth.

GAINWARD has one of the most impressive custom GeForce RTX 4090 graphics cards on the market, and now the company is reportedly replacing faulty GeForce RTX 4090 power cables.

According to Australian retailer TechFast, GAINWARD is delaying shipments of its GeForce RTX 4090 family of graphics cards for another week, spotted by Redditor 'varunbabu008". TechFast has reached out to customers providing them with some updates on their GeForce RTX 4090 graphics card orders Down Under.

TechFast customer support explained to -- I'm sure every Australian gamer who ordered a GAINWARD GeForce RTX 4090 graphics card -- that GAINWARD explained to them that the cables that were shipped with their cards are being replaced. Because of this, the company is "holding shopping on all cards until this has taken place".

EVGA may have terminated its relationship with NVIDIA which means we won't see any GeForce RTX 40 series graphics cards, at all, from EVGA... except YouTuber "JayzTwoCents" has gotten his mitts on a prototype card.

The unreleased EVGA GeForce RTX 4090 FTW3 graphics card was sent to JayzTwoCents in engineering form, which would've been close to the retail version of the card I'm sure, and it is a big boi graphics card. It's thicker than its Ampere-based predecessor in the EVGA GeForce RTX 3090 Ti FTW3, which he has for comparison against the prototype EVGA GeForce RTX 4090 FTW3 graphics card.

EVGA was still using a triple-fan cooler on the GeForce RTX 4090 FTW3 engineering sample, with the in-house ICX cooling system, and something that I love: the rear-placed 16-pin PCIe power connector.

NVIDIA has reportedly shipped over 100,000 of its AD102 GPUs that power its monster new flagship GeForce RTX 4090 graphics card.

The source of this information is coming directly from our friends over at Wccftech, who reported that they have received the figure of the total number of AD102 GPUs shipped: 100,000 so far. Wccftech along with more confirmation from "Chi11eddog" have a considerable amount of that beautiful 4N process node at TSMC for NVIDIA's new AD102 GPU.

There were rumors of the limited stock of the new GeForce RTX 4090 graphics cards, but I guess we know that's not a problem given there are over 100,000 of the bad boys in the world now. NVIDIA has its own in-house GeForce RTX 4090 Founders Edition graphics card, but then we've also got an absolute avalanche of custom GeForce RTX 4090 graphics cards from AIB partners like ASUS, COLORFUL, MSI, and more.

COLORFUL has unleashed more details on its upcoming iGame GeForce RTX 4080 Neptune graphics card, which will have a maximum TDP of 380W... higher than AMD's upcoming Navi 31-powered Radeon RX 7900 XTX graphics card.

The upcoming COLORFUL iGame GeForce RTX 4080 Neptune graphics card will ship in a very tight and small dual-slot design, but a much larger triple-fan 360mm AIO cooler that will ensure that the AD103-300 GPU inside and its 16GB of GDDR6X memory stay nice and cool. Very cool, given its AIO cooler.

But now that the company is detailing its custom iGame GeForce RTX 4080 Neptune graphics card, COLORFUL has unveiled that it will have a 380W TDP. This is a decent whack up the power scale, given that AMD's new flagship Radeon RX 7900 XTX graphics card has a 355W TDP, meaning COLORFUL's beasty custom iGame GeForce RTX 4080 Neptune blasts past the RX 7900 XTX by 35W. Not a huge amount, but considerable.

AMD's upcoming Radeon RX 7900 XTX and Radeon RX 7900 XT reference graphics cards will both have hidden ambient temperature sensors, something GamersNexus has discovered.

The sensor finds itself hidden behind the last fan on the card, towards the end, letting the reference Navi 31-powered Radeon RX 7900 XTX and Radeon RX 7900 XT graphics cards measure the temperature of the ambient temperature inside of your case. Very cool for the reference cards from AMD.

It's not something that's overly new on custom AIB designs that have similar sensors, as that's how they get the "0 RPM" and "0dB" features... but now AMD's in-house reference Radeon RX 7900 series graphics cards will have this, meaning when there's no load on the GPU, the fans will turn off. Better yet, low GPU loads on some games will not have the fans spooling up at all.

Intel can't send out its Arc A750 or Arc A750 graphics card here to TweakTown for review, but they sure do seem to have enough cards to tweak into a "Gold" edition variant for the Chinese market.

The new Intel Arc A750 Limited "Gold" Edition graphics card is only sold in China and isn't any different inside than the regular Intel Arc A750 graphics card. Intel is using a slightly tweaked retail package, and a slightly tweaked backplate which I must admit does look rather cool... other than that, it's still a regular Arc A750.

Inside, the Intel Arc A750 graphics card -- and this new Arc A750 Limited "Gold" Edition -- we have the ACM-G10 GPU with 28 Xe-Cores, and 8GB of GDDR6 memory which spits out up to 512GB/sec of memory bandwidth.

AMD's next-gen Radeon RX 7900 series graphics cards are now official, with the Radeon RX 7900 XTX and Radeon RX 7900 XT both dropping on December 13... but now we're hearing a little more about where the Radeon RX 7900 XTX lines up against its competitor.

NVIDIA unleashed its GeForce RTX 4090 with the biggest generational leap ever, absolutely demolishing the GeForce RTX 3090 (and GeForce RTX 3090 Ti) with most people expecting AMD's new chiplet-based Navi 31-powered Radeon RX 7900 XTX would topple it.

But it won't, and AMD has confirmed that.

Remember those super-exciting, crazy rumors of 4GHz+ and beyond GPU clocks on Navi 31? Yeah, well, I've got some disappointing news for you.

AMD's new Radeon RX 7900 XTX and Radeon RX 7900 XT have now been unveiled and detailed (in some ways) but now we're hearing more information about Navi 31, the GPU inside of the two new Radeon RX 7900 series graphics cards.

PC World attended the RDNA 3 launch event in Las Vegas, where AMD confirmed with them that the company designed the new Navi 31 GPU to scale up to 3.0GHz... much higher than the GPU boost clocks on Navi 21, which is inside of the Radeon RX 6900 XT and Radeon RX 6950 XT graphics cards.

GALAX has smashed some world records with its flagship GeForce RTX 4090 HOF graphics card, breaking not one but two world records with the monster custom GeForce RTX 4090.

If you haven't heard about the GALAX GeForce RTX 4090 HOF, it is a monster custom GeForce RTX 4090 with dual 16-pin PCIe power connectors... yeah not one, but two 16-pin power connectors. Every other GeForce RTX 4090 graphics card has a single 16-pin power connector, but GALAX is driving some serious power into their GeForce RTX 4090 HOF.

Overclocker "OGS" cranked up the GPU frequency on the GALAX GeForce RTX 4090 HOF to an absolutely incredible 3705MHz throughout the GPUPI benchmark (which is a part of the official HWBOT competition).

AMD's new Radeon RX 7900 XTX and Radeon RX 7900 XT have now been detailed, with the company also unveiling its new FSR 3.0 technology during its RDNA 3 reveal event in Las Vegas.

The company announced the new FidelityFX Super Resolution 3 technology will drop in 2023, offering twice the performance over FSR 2. AMD showed off the new Radeon RX 7000 series GPU running an Unreal Engine 5 demo, with FSR 2.0 capable of around 60FPS or so, but with FSR 3.0 enabled the performance jumped up to around 112FPS... impressive stuff.

AMD's fresh new FSR 3.0 upscaling tech also includes Fluid Motion Frames Technology, but the company wasn't clear on what it actually does... but we do know that you'll get more performance in games that support the technology with the most demanding games on the market.