Graphics Cards - Page 125

NVIDIA has announced its Computex 2022 keynote, penciled in for May 23 @ 8-9PM PDT with 6 speakers getting ready to talk your ear off.

We are to expect Ian Buck (Vice President of Accelerated Computing), Brian Kelleher (Senior Vice President, Hardware Engineering), Ying Yin Shih (Director of Product Management, Accelerated Computing), Michael Kagan (Chief Technology Officer), Deepu Talla (Vice President of Embedded and Edge Computing) and Jeff Fisher (Senior Vice President of GeForce).

I wouldn't be expecting any new GPU announcements, but rather NVIDIA will most likely highlight everything the company has achieved in the last 6 months (new GPUs including the delayed GeForce RTX 3090 Ti, as well as their next-gen Hopper GPU architecture). The last time we saw Jeff was at CES 2022 earlier this year, when he announced the new GeForce RTX 3090 Ti graphics card.

NVIDIA's new GeForce RTX 4090 is once again in the headlines, with new leaks on the GPU teasing that the next-gen AD102-300 GPU will have 24GB of GDDR6X memory @ 21Gbps.

In new leaks from "kopite7kimi" we're being told that the GeForce RTX 4090 will use the same AD102 GPU as the higher-end GeForce RTX 4090 Ti, but will have some things cut down of course. Even in its cut down form, AD102 will still offer a 71% increase in CUDA core count over GA102 (the GPU that powers the RTX 3080 Ti, RTX 3090, RTX 3090 Ti).

kopite7kimi says that the RTX 4090 that is getting more leaked details is the "PG137/139-SKU330" which uses the "AD102-300" GPU, has 24GB of GDDR6X memory at 21Gbps, and of course: 600W of power. But the new RTX 4090 isn't the only GPU that the leaker is providing more details on, but also the upcoming RTX 4070, too.

Intel has detailed its Arctic Sound M data center GPU, announcing that it will be available in Q3 2022 -- but no one in their right mind would be trusting their GPU shipment dates after all their Arc GPU delays -- still, let's get this Arctic Sound M GPU show on the road.

There are two different designs of the Arctic Sound M, the first is the flagship based on a single ACM-G10 GPU and a 150W design. The second card is packing not one but two ACM-G11 GPUs and has a smaller 75W TDP, but it's aimed at high-density multi-purpose workloads. Intel's new Arctic Sound M are "super flexible data center GPUs aimed at cloud gaming, media processing and delivery, virtual desktop infrastructure, and inference.

Arctic Sound GPUs can reportedly handle 30+ 1080p streams, 40+ game streams, up to 62 Virtualized Functions, and up to 150 AI TOPs. Intel says its new Arctic Sound M GPUs will have the latest technologies like AV1 HW Encode/Decode, and XMX AI-Accelerators built-in.

AMD has released its new Radeon Software Adrenalin 22.5.1 drivers, which have built-in support for the just-released Radeon RX 6950 XT, Radeon RX 6750 XT, and Radeon RX 6650 XT graphics cards.

If you've just picked up a new RDNA 2 refresh GPU then you'll want to download the new v22.5.1 drivers obviously. AMD has also fixed issues with some Radeon GPU owners using Windows 10 noticing the "absence of the windows transparency aero effect".

There's still a list of known issues, ranging from performance drops when playing Fortnite with Multithreaded Rendering and DX11 on some Radeon GPUs, including the Radeon RX 6900 XT. GPU utilization being stuck at 100% in Radeon performance metrics after closing games on some Radeon GPUs with AMD noting the Radeon RX 570.

AMD has officially launched its new Radeon RX 6950 XT graphics card, and with it we have a slew of custom cards from AIB partners... ASRock is the latest to announce its new Radeon RX 6950 XT OC Formula 16G graphics card.

The new ASRock Radeon RX 6950 XT OC Formula graphics card has a GPU Boost CLock of up to 2495MHz, but we'll see it breach 2800MHz and I'm sure 3000MHz+ with professional overclockers and LN2 cooling. ASRock says that its newly designed Striped Ring Fan "further enhances the airflow performance with its striped fan blades and ring structure".

Underneath we have AMD's latest Navi 21 KXTX GPU, upgraded 16GB of GDDR6 memory that's clocked at 18Gbps (up from 16Gbps) while ASRock is including some built-in ARGB LED lighting that syncs with ASRock motherboards through Polychrome SYNC support.

Intel has completely screwed the launch of its Arc GPU, with multiple delays, a whimpered appearance on select notebooks in Arc mobile GPU form, and now Intel has absolutely confirmed Arc desktop GPUs will be a China exclusive at launch. Okay.

It was only a few hours ago that I wrote a news article that talked about Intel delaying the launch of its desktop-focused Arc Alchemist GPU, which Igor from Igor's Lab reported from his sources. But now... now this is from the horse's mouth -- and that horse's mouth... is Intel -- delaying Arc Alchemist GPUs on desktop.

Not only that, but Intel Arc Alchemist GPUs on the desktop will be exclusive to China... yeah, exclusive to China at launch. This is a mess and should've been explained to the press in detail, months ago. Nope. Instead, it comes out in a tiny little story today, which I hope truly blows up. I do love that Intel uses a picture that says: "Intel Arc, a new player has entered the game"... except that player hasn't entered the game, and when it does it'll be exclusive to China for god knows how long.

Intel has had a troubled time getting its Arc Alchemist GPU out the door, with delay after delay, and now yet another delay that is reportedly pushing out Arc Alchemist desktop GPUs into mid-summer.

In a new post, Igor from Igor's Lab explained: "various sources told me (some of them quite annoyed), the release of the already so often announced Alchemist desktop graphics cards could be postponed even further towards late summer. A somewhat vague period is mentioned here, which extends from July 1, 2022 to August 31, 2022. Why one is then already more than half a year behind schedule can of course also be speculated creatively".

The launch has been a mess so far, and doesn't give confidence whatsoever. But in Intel's defence: you can't just build, and easily launch an entire family of graphics cards and then keep up with the two very established GPU companies (NVIDIA and AMD) in a single year, or even a few.

AMD isn't too far away from the launch of its RDNA 2 refresh GPUs with the flagship Radeon RX 6950 XT, and the new Radeon RX 6750 XT and Radeon RX 6650 XT graphics cards... now it's time for some leaks on GIGABYTE's upcoming graphics cards.

But what you'll notice immediately is how damn big the GIGABYTE Radeon RX 6950 XT GAMING OC graphics card is, with GIGABYTE adding some chunk to the card over its custom Radeon RX 6900 XT GAMING OC. THe new GIGABYTE RX 6950 XT GAMING OC is a huge, thick card... with a thicker and larger cooling system.

It looks like GIGABYTE is using the same PCB design for its new Radeon RX 6950 XT GAMING OC as it did for the Radeon RX 6900 XT GAMING OC -- where you'll still need 3 x 8-pin PCIe power connectors -- and then 2 x HDMI 2.1 and 2 x DisplayPort 1.4 connectors. That's non-standard display I/O, and welcomed -- the second HDMI 2.1 port that is.

Well, the LHR gimping has been cracked by the NBMiner (NebuMiner) and NiceHash teams, with NBMiner available right now with a 100% LHR unlock for NVIDIA's current-gen GeForce RTX 30 series GPUs.

The crypto teams seem to be running crypto tools that are unlocking NVIDIA LHR gimping, almost at the same time, which seems suspicious. How? Are they communicating with each other? Did NVIDIA help them out? Maybe. AMD is launching its RDNA 2 refresh -- the new Radeon RX 6950 XT, Radeon RX 6750 XT, and Radeon RX 6650 XT cards in the next 12 hours -- and then there's this news of the un-gimping of LHR-locked GPUs.

If you're sitting there with a GeForce RTX 30 series GPU that is LHR locked, and you're crypto mining, this will be music to your ears. If you're a crypto miner that has fleets of LHR-locked GPUs... then this is Christmas coming early in a huge way.

The interesting ASUS x Noctua GeForce RTX 3080 graphics card has been teased for a while now, but there's some official photos of the card and it gives us some greater details on the custom RTX 3080.

Interestingly, the company is using the RTX 3080 10GB variant -- not the RTX 3080 12GB -- for the new ASUS x Noctua GeForce RTX 3080. For the cooler, the custom Noctua cooler uses a dual 12mm fan cooler that should keep the card very cool.

Very cool, because it's a gigantic quad-slot design... from the front it doesn't look thick but man, it is thick. There are a few other thick RTX 3070 and RTX 3080 graphics cards out there, but nothing like this with a Noctua cooling solution. It also looks cool, and should also be cool.

NVIDIA's new flagship GeForce RTX 3090 Ti graphics card has a high TDP by default, with 450W at default... but now we have professional overclockers pushing it up to a huge 890W power limit.

MegaSizeGPU has uploaded a screenshot to Twitter from GPU-Z teasing the insane 890W power limit on the GeForce RTX 3090 Ti. The screenshot is proof that the GPU had a power limit of 890W, which is virtually double the 450W default TDP... crazy stuff, all from a custom XOC BIOS.

This isn't for you or I, but rather extreme overclockers who are pushing their modified GeForce RTX 3090 Ti graphics cards to the absolute limits with a special XOC BIOS. MegaSizeGPU was rocking the custom ASUS TUF Gaming GeForce RTX 3090 Ti OC Edition graphics card, with the special XOC BIOS and insane 890W unlocked power limit.

NVIDIA's next-gen Hopper H100 SXM has been spotted in the flesh, and until now... we've only seen renders, so this is the first real pics of the H100 GPU.

The next-gen Hopper H100 SXM GPU was pictured in the flesh by Patrick Kennedy from ServeTheHome, with a device with a board model of "PG520". Inside, the NVIDIA H100 GPU is using TSMC's very latest CoWoS packaging technology, with a huge 814mm2 H100 GPU die with 6 memory modules around it: 80GB of ultra-fast HBM3 memory to be exact.

NVIDIA's new H100 GPU itself comes in two forms with two counts of CUDA cores: 16896 FP32 CUDA cores for the SXM variant and 14592 cores for the PCIe-based model. The model that we have here is the SXM variant, so we have the full 16896 FP32 CUDA cores and 80GB of HBM3 memory.

AMD is still cooking its next-gen RDNA 3 architecture and flagship Navi 31 GPU, where we should see the flagship Navi 31-based Radeon 7000 series graphics card, which will have next-gen PCIe 5.0 x16 support.

I think it's a foregone conclusion even before a leaker "Kepler_L2" giving the Chuck Norris thumbs up to PCIe 5.0 x16 support on Navi 31. AMD will be launching its next-gen Zen 4 architecture for the CPU side of its business later this year, with new X670-based motherboards rocking DDR5 + PCIe 5.0 support.

Intel had DDR5 first on its new Z690-based motherboards ready for its new 12th Gen Core "Alder Lake" CPUs, and now AMD will have not only catch up with DDR5 but also introduced PCIe 5.0. Yes... Intel also has PCIe 5.0 connectivity on its flagship Z690-based motherboards, but they have no PCIe 5.0-ready GPU. AMD will have a PCIe 5.0-ready GPU (Navi 31) and a PCIe 5.0-ready chipset, right in time for the holidays.

PowerColor has just teased its new custom Radeon RX 6650XT Hellhound White graphics card, which differs from the regular RX 6650 XT Hellhound in a big way: its pink + white color theme.

The custom PowerColor Radeon RX 6650XT Hellhound White graphics card arrives with a white theme, with a pink accent that was influenced by cherry blossom. The announcement of PowerColor's new Radeon RX 6650XT Hellhound White came directly from the Japanese PowerColor social media account.

We should expect PowerColor to be using the same cooling design as its slightly lower-end Radeon RX 6600 XT Hellhound graphics card, with the same thickness and same single 8-pin PCIe power connector. We should expect the same GPU boost clocks as the Black SKU of the Radeon RX 6650 XT Hellhound GPU, so we should expect up to 2689MHz GPU clocks and the 8GB of GDDR6 memory at 17.5Gbps bandwidth.

NVIDIA's next-gen Ada Lovelace GPU architecture will be fighting AMD's next-gen RDNA 3 GPU architecture later this year, with both GPUs rumored to provide some rather exciting upgrades in performance over the current generation GPUs.

The new NVIDIA AD102 GPU should be the chip that powers the new flagship GeForce RTX 4090 Ti and GeForce RTX 4090 graphics card -- or, according to some rumors the RTX 5090 Ti and RTX 5090 -- with leaker "kopite7kimi" tweeting that AD102 could have over 100 TFLOPs of compute performance.

100 TFLOPs of power from AD102 would make it 2.5x faster than the RTX 3090 Ti, which spits out 40 TFLOPs and 2.8x better than the RTX 3090 with its 35.5 TFLOPs of FP32 compute performance. Remember: pure TFLOPs doesn't mean = gaming performance, but kopite7kimi does say "Lisa and Jensen's competition will give us a 100 TFLOPs gaming war in a few months" with a key takeaway from that: 100 TFLOPs + gaming war.

The GPU rumor game has changed considerably over the last couple of years, in a pandemic-fueled Twitter and YouTuber frenzy that has seen the hype, specs, hype, rumors, and hype reach new bounds. Navi 31 is in the middle of it, with some fresh new leaks that water down AMD's next-gen MCD-based GPU design.

Industry leaker "Greymon55" has tweeted out a fresh rumor, as well as deleting his previous tweet referring to AMD's next-gen Navi 31 GPU having ~92 TFLOPs of compute performance. Greymon55 has tweeted just now, with a simple:

What does that mean? GPU cores on AMD's next-gen RDNA 3-based Navi 3X GPUs, in order: Navi 31, Navi 32, and Navi 33 with 12288, 8192, and 4096 GPU cores respectively.

ASUS has just launched its ROG Strix GeForce RTX 3090 EVANGELION Edition, a new custom GeForce RTX 3090 -- non-Ti -- themed on anime-mecha story EVANGELION.

The new ASUS ROG Strix GeForce RTX 3090 EVANGELION Edition graphics card uses the ROG Strix OC design and is the now fourth variant of the card. ASUS is going with a wicked-looking green/purple design, with black/red throughout -- themed like EVANGELION -- with fantastic results.

You're not going to be able to buy the card anywhere, with CoolPC store the first to take preorders. CoolPC says that they will only have 15 of the new ASUS ROG Strix GeForce RTX 3090 EVANGELION Edition graphics cards for sale, with the Taiwanese store selling them for 63,990 TWD which works out to around $2173 in the US.

EVGA must be having some serious fun in its OC lab, with overclocking legend Vince "KINGPIN" Lucido, playing around with not just one... but two of its crazy custom GeForce RTX 3090 Ti KINGPIN graphics cards.

EVGA is using a very custom design with its GeForce RTX 3090 Ti KINGPIN, with a TGP rated at 516W and a 1200W power limit.

With KINGPIN using the two cards each with 2 x 16-pin PCIe 5.0 power connectors mean for 4 x 16-pin PCIe power connectors into 12 x 8-pin PCIe power adapters. The 1200W+ per GPU means EVGA was using a separate PSU per GPU, something that's required at this (absolutely bonkers) stage of 2400W of power from two GPUs and all those power adapters.

NVIDIA first announced its next-gen H100 Hopper GPU at GTC 2022 (GPU Technology Conference) earlier this year, its next-gen GPU is built on the new 4nm process node at TSMC and rocking a huge 80 billion transistors. Well, that next-gen GPU is now for sale... and it'll cost ya.

Japanese HPC retailer GDEP Advance has the NVIDIA H100 GPU listed for sale, which costs a whopping 4,745,950 yen (or around $36,550) which includes taxes and shipping: the card on its own costs 4,313,000 yen (around $33,200). This is also for the PCIe-based model with passive cooling for server enclosures, so the price of other SKUs is unknown at this point.

As for the NVIDIA H100 GPU, it's based on the new Hopper GPU architecture with NVIDIA using TSMC's 4nm process technology and packing 80 billion transistor into 814mm2 of silicon. The PCIe-based version of the NVIDIA H100 rocks 80GB of HBM2e memory with 2TB/sec of memory bandwidth, rocking on the PCIe 5.0 x16 connector.

Update: Ahhh, the leaker has deleted his tweet... which said "92 TFLOPs" and now is saying that there are toned down specifications of Navi 31 that would radically reduce the FP32 compute power down to around 73 TFLOPs (a 20% reduction).

AMD's next-gen Navi 31 GPU has been shaping up to be a monster so far, but with every rumor the RDNA 3 architecture is looking better and better... to the point where it could topple NVIDIA and its next-gen Ada Lovelace GPU architecture.

In a new tweet, leaker "Greymon55" tweeted something rather simple: "~92T FP32" as a reply to his Navi 31 leaks from November 9, 2021. In that tweet, rumors had Navi 31 at 75T FP32 compute performance, but now the leaker has upped those specs by a considerable 28%.