Video Cards & GPUs News - Page 9
AMD will debut its next-gen RDNA 3 architecture later this year, where rumor has it that the big boy flagship Navi 31-based offering will feature a GPU chiplet-based design... and NVIDIA won't have anything with GPU chiplets, and AMD knows that.
In a recent interview with VentureBeat, AMD senior vice president, corporate fellow and product technology architect, Sam Naffziger talked about GPUs using up to 600W of power by 2025, and so much more. But then VentureBeat's last question was a doozy "Compared to NVIDIA and Intel, do you feel like we're in a state of divergence when it comes to designs, or some kind of convergence?"
Naffziger replied, explaining: "It's hard to speculate. NVIDIA certainly hasn't jumped on the chiplet bandwagon yet. We have a big lead there and we see big opportunities with that. They'll be forced to do so. We'll see when they deploy it. Intel certainly has jumped on that. Ponte Vecchio is the poster child for chiplet extremes. I would say that there's more convergence than divergence.
We all know that the next-gen GPUs from both AMD and NVIDIA will be consuming much more power, especially for NVIDIA's highest-end Ada Lovelace GPUs, with AMD expecting to see GPUs using up to 600W of power by 2025.
AMD set an aggressive goal of 25x20 (hitting 25x better energy efficiency for its CPUs and GPUs by 2020) and has smashed it, so much so that the company has set a brand new 30x25 goal (hitting 30x better energy efficiency) by 2025 in the machine learning, and high-performance computing (HPC) markets in data centers.
The news is coming from VentureBeat, which spoke with AMD senior vice president, corporate fellow and product technology architect, Sam Naffziger. Naffziger said: "We've driven the frequency up, and that is something unique to AMD. Our GPU frequencies are 2.5 GHz plus now, which is hitting levels not before achieved. It's not that the process technology is that much faster, but we've systematically gone through the design, re-architected the critical paths at a low level, the things that get in the way of high frequency, and done that in a power-efficient way".
You can't buy Intel's new Arc Alchemist desktop GPU in the USA right now, but that is for us... the mere mortals... but the China-exclusive GUNNIR Arc A380 Photon 6G OC graphics card has made it into the Intel labs, and into the hands of Ryan Shrout.
The new GUNNIR Arc A380 Photon 6G OC graphics card packs the full Intel ACM-G11 GPU with 8 Xe-Cores, and 6GB of GDDR6 memory. Ryan Shrout, who does "graphics marketing" for the Intel GPU department, used to own and operate PC Perspective, has some hands-on with the GUNNIR Arc A380 Photon 6G OC in a video posted to his personal Twitter above.
He is of course talking it up, in its brief showing -- even though Intel's new Arc A380 GPU is disappointing compared to its competitors in their low-end SKUs -- with a very short demo of PlayerUnknown's Battlegrounds (PUBG) at a strange 2560 x 1080 (why... why...) on medium settings. The results seem alright, but we don't know any FPS details at all. Sad, given how Ryan had to do that when he ran PCPer and knows the tech press would pull this apart. #TeamIntel, I guess.
AMD has just launched another Radeon Adrenalin Edition driver set, after the launch of the latest AMD Adrenalin Edition driver just last week. The new AMD Software: Adrenalin Edition 22.6.1 includes some F1 2022 performance optimizations. You can read more, and download the new AMD Software: Adrenalin Edition 22.6.1 drivers right here.
The new AMD Software: Adrenalin Edition 22.6.1 also includes some performance and stuttering fixes to both Fortnite, and my personal favorite: Overwatch. Using the new AMD Software: Adrenalin Edition 22.6.1 and AMD's flagship Radeon RX 6950 XT graphics card, you're going to get up to 5% more performance in F1 2022 at 4K on Ultra High settings versus the previous software driver version 22.5.2.
There's up to a 4% increase in performance in F1 2022 at 4K on Ultra High settings using the new AMD Software: Adrenalin Edition 22.6.1 on the Radeon RX 6800 XT, versus the previous software driver version 22.5.2. Finally, there's up to 6% more performance in F1 2022 at 4K and Ultra High settings using the new AMD Software: Adrenalin Edition 22.6.1 on the Radeon RX 6750 XT, versus the previous software driver version 22.5.2.
ELSA has just released its new GeForce RTX 3090 LC graphics card, a new custom GeForce RTX 3090 graphics card that offers something a little different.
The new custom ELSA GeForce RTX 3090 LC graphics card has a full-cover waterblock that keeps the GPU, 24GB of GDDR6X memory, and VRM sections chilled. The Japanese company teamed with AlphaCool on developing the water block, pump, and external 360mm radiator for the ELSA GeForce RTX 3090 LC graphics card.
ELSA's new GeForce RTX 3090 LC graphics card looks like it's using a custom PCB with dual 8-pin PCIe power connectors, so it's very close to the size of the reference RTX 3090 board, except it measures in longer at 26.4mm.
The rumored specifications on NVIDIA's next-gen GeForce RTX 40 series graphics cards are running hot today, with the latest leaks pulling up the GDDR6X memory bandwidth on the GeForce RTX 4080.
NVIDIA's new GeForce RTX 4080 rumored specs haven't changed much, with leaker "kopite7kimi" teasing the AD103-300-A1 GPU and 420W of TDP -- unchanged rumored specs here -- but the 16GB of GDDR6X is reportedly being cranked up to 21Gbps, up from previous rumors pegging it at 18Gbps.
The current-gen Ampere-based GeForce RTX 3080 in its original 10GB GDDR6X (not in the second "LHR" Lite Hash Rate version) form saw the 10GB of GDDR6X memory clocked at 19Gbps. NVIDIA's new GeForce RTX 4080 graphics card with 16GB of GDDR6X memory at 21Gbps is a better rumor to hear, up from the 18Gbps GDDR6X (lower clocks than the RTX 3080).
The week is starting off well with the next-gen NVIDIA Ada Lovelace GPU rumor mill, with leaker "kopite7kimi" teasing the continuously updating rumored specs of NVIDIA's upcoming GeForce RTX 40 series graphics cards.
NVIDIA's new flagship GeForce RTX 4090 graphics card has a little more details from kopite7kimi, with the AD102-300-1 GPU clocking at over 2.75GHz. The leaker says that we can expect a 2235MHz base GPU clock, 2520MHz boost GPU clock, and 2750MHz "actual max" GPU clock. Impressive, very impressive... considering 2250MHz is about the limit for the GeForce RTX 3090 Ti outside of insane LN2 cooling and mods for days.
If we're expecting "actual max" GPU clocks of 2.75GHz then we could see custom GeForce RTX 4090 graphics cards from the likes of ASUS and MSI to reach 2.8GHz and above... while the heights of 3.0GHz+ on an NVIDIA GPU would be awesome to see.
The price of AMD Radeon and NVIDIA GeForce graphics cards have been dropping for a while now, being a combination of a few things... but now we have some firmer numbers: GPU prices have dropped by around 57% since January 2022 alone.
In a new piece, Jarrod Walton @ Tom's Hardware reports that graphics card prices are dropping all over the place in both retail and second-hand markets. There's a bunch of data there that shows that retailers are offering their usual deals and promotions (free games) which saw graphics card prices dropping by an average of 3% over the last month.
But when we look to eBay, a much larger 14% drop in graphics card prices has been noted... and then for previous-gen graphics cards that drop is bigger at 17% in the last month. Graphics cards are now selling at below MSRP, but with GPUs being used in crypto mining farms... it's a tricky position to be in. Still, lower graphics card prices are a fantastic thing to see.
EVGA's new custom GeForce RTX 3090 Ti KINGPIN graphics card has finally been unleashed, which is now available on the EVGA website in the US for $2500... but you'll also get a free 1600W power supply.
If you decide to slap down $2500 for the EVGA GeForce RTX 3090 Ti KINGPIN graphics card, well, EVGA is giving customers a free EVGA SuperNOVA 1600 P2 power supply with an 80+ PLATINUM 1600W rating. Not too bad, considering you need so many PCIe power adapters to plug into the dual 16-pin PCIe power connectors on the EVGA RTX 3090 Ti KINGPIN graphics card.
EVGA's highly custom GeForce RTX 3090 Ti KINGPIN and its dual 16-pin PCIe power connectors translate into 6 x 8-pin PCIe power adapters (3 x 8-pin PCIe power adapters per 16-pin power connector on the card). It's a crazy amount of adapters, but the free EVGA SuperNOVA 1600 P2 PSU rocks 9 x PCIe power connectors inside, so EVGA needs to supply the PSU or your new GPU is kinda useless out of the box (but you should know that before you buy a monster graphics card like this).
TSMC is seeing its 3 biggest customers revising their orders all at the same time: AMD, Apple, and NVIDIA for their next-gen RDNA 3, iPhone 14, and Ada Lovelace products, respectively.
NVIDIA is the big one that we'll tackle in this article because they're in a real precarious spot: NVIDIA has been bedding Samsung for a while now, leaving the South Korean arms for the arms of Taiwan and back to its old faithful: TSMC (Taiwan Semiconductor Manufacturing Company) for its next-gen Ada Lovelace GPU architecture and upcoming GeForce RTX 40 series graphics cards... well, they've hit a snag.
According to the latest report out of DigiTimes and translated by @RetiredEngineer on Twitter: "Faced with abrupt drop-off in PC demand and customers' inventory adjustment, AMD and NVIDIA has no choice by to reduce orders".