NVIDIA hosted the GeForce RTX: Game On event on Tuesday as part of CES 2021. As expected, several new features and hardware were unveiled for gamers. In addition to the mobile GeForce RTX 30-series GPUs for notebooks, Nvidia introduces a new GeForce RTX 3060 graphics card for desks, the ability to expand the Reflex Latency Analyzer (RLA) display, and the performance enhancements AMD introduced with AMD Smart Access Memory (SAM) on the Radeon RX 6000 series It also announced plans to support the PCIe resizable BAR function similar to the function.
In addition, we announced the addition of the next game, Call of Duty: Warzone and FNAF (Five Nights at Freddy), as Nvidia’s DLSS 2.0 technology-enabled games. Freddie also supports real-time ray tracing.
GeForce RTX 30 Series for Notebooks
Let’s start with the mobile GeForce RTX 30-series GPUs, which are likely to be featured in several gaming laptops announced at CES, along with new notebook processors from AMD and Intel. NVIDIA unveiled the mobile chips for the GeForce RTX 3080, 3070 and 3060, based on the next-generation Ampere GPU architecture for desktops.
The 6GB GeForce RTX 3060 is the cheapest, but Nvidia’s Jeff Fisher says it’s faster than any other gaming laptop in existence, and has the same performance as the previous generation flagship RTX 2080 Super. Nvidia says the mobile chip is 30 percent faster than Sony’s new PlayStation 5 console and can record 90 frames per second during 1080p gaming.
But the latest graphics cards go beyond 1080p resolution. NVIDIA announced that it is working closely with panel and notebook manufacturers to distribute 1440p-capable displays. Two graphics cards that were already unveiled at the end of 2020 were developed to address this. The 8GB GeForce RTX 3070 is up to 50% faster than its replacement RTX 2070 chip and achieves 90 fps at 1440p. The flagship RTX 3080 is said to be able to play at 100 fps at 1440p with a large 16GB of onboard memory.
Interestingly, the number of CUDA cores in this mobile family is different from the desktop version. It also changed the standards established over the previous generations. The laptop RTX 3080 has 6,144 CUDA cores (desktop model is about 9,000), the RTX 3070 mobile chip lacks 768 cores than its desktop cousin, and it’s hard to believe, but the mobile RTX 3060 is actually the version of the laptop released just a while ago. It has 3,840 more CUDA cores than (3,584).
In other words, the laptop version of the RTX 3080 has hundreds of cores more than the desktop 3070, which can be used to predict the level of performance. Ampere uses a large die at its maximum size in the desktop, but its performance doesn’t seem to be fully applied to the notebook world where energy efficiency is critical.
NVIDIA announced that the implementation of the third-generation Max-Q technology in this new GPU is twice as efficient as the previous version. This argument is difficult to compare because it relates to the DLSS feature, which is only available in some games. However, other Max-Q improvements are also noticeable, using 2nd-generation Dynamic Boost and AI algorithms to intelligently switch power between CPU and GPU as well as GPU memory to adjust components to achieve user-specified noise levels. There is also an upgraded version of Whisper Mode to match. Resizable BAR support is also possible with the 3rd generation Max-Q, which will be discussed in more detail later.
Notebooks powered by NVIDIA’s next-generation GeForce mobile GPUs will be available on January 26th, with over 70 models available to all major OEMs. NVIDIA said GeForce RTX 3060 laptops will start at $999, RTX 3070 laptops will start at $1,299 and RTX 3080 laptops will start at $1,999.
GeForce RTX 3060
It also unveiled a new toy for desktop gamers. After launching high-performance products such as the GeForce RTX 3090, 3080, 3070, and 3060 Ti, the amperage-based graphics card through the GeForce RTX 3060 will be released at the end of February at a more reasonable recommended retail price of $329.
In reality, the RTX 3060, like all other next-generation graphics cards unveiled so far, is likely to sell out immediately and cost more. Competitor AMD said in a CES keynote that a new mainstream Radeon product will be released within the first half of this year. Meanwhile, Nvidia will soon sell as much of the RTX 3060 as possible, hoping that there will be enough supply.
Nvidia rarely discloses the technical details of the GeForce RTX 3060, but says it offers up to twice the rasterization performance of the GeForce RTX 1060 and 10 times the ray tracing performance. (The GTX 1060 doesn’t have a dedicated real-time ray tracing core like the RTX-class GeForce cards). Interestingly, the GeForce RTX 3060 comes with 12GB of GDDR6 memory, which is more than the 8GB offered by the more powerful RTX 3060 Ti and 3070. However, it uses a smaller 192-bit bus. The AMD Radeon RX 6000 series, which has a larger VRAM capacity, seems to have affected’Team Green’ as well.
But on the GeForce RTX 3060 product page, you can find more details than the $400 RTX 3060 Ti. It has 3,584 CUDA cores running at a 1.78GHz boost. The table doesn’t include physical details such as size and connector loadout, but it should be around 170W to match it with 550W power. The RTX 3060 Ti is a 200W card and requires 600W of power.
Of course, as an RTX 30 series card, the GeForce RTX 3060 supports all of NVIDIA’s advanced features, such as real-time ray tracing, DLSS, NVIDIA Reflex, NVENC encoding, Shadowplay, and NVIDIA Broadcast.
Nvidia Reflex Latency Analyzer
Nvidia recently introduced the Reflex Latency Analyzer G-Sync display, which was unveiled in a super-fast 360Hz 1080p panel with the esports specialty in mind. The five new models announced at CES include the new 1440p and ultra-wide products. It’s a really cool premium feature, and we’re excited to see this technology spread its wings.
The final announcement was a performance upgrade for all PC gamers using the latest hardware. NVIDIA announced that the new GeForce RTX 30 series notebook GPUs will support the PCIe resizable BAR feature, which MD first announced as SAM. The desktop GeForce RTX 3060 and all future products will be supported at the time of release, and the existing RTX 30 series GPU will be supported through a later VBIOS update.
Resizeable BAR allows a computer’s processor to access the full memory capacity of a graphics card beyond the usual 256MB limit. As we saw when testing the SAM with AMD’s Radeon RX 6900 XT, the performance improvement through resizable BAR varies from game to game, GPU to GPU, and resolution. Sometimes performance doesn’t improve. In most cases, the frame is improved by 4-5%. In some cases (e.g. Assassin’s Creed Valhalla) performance can be improved by a factor of two.
A compatible CPU is required to activate the resizable BAR. Intel has said its new Tiger Lake H35 processor will support this feature, and will probably also support AMD’s new Ryzen 5000 mobile chip. This technology is introduced in the AMD desktop Ryzen 5000 series. AMD did not give a definite answer, but Nvidia said the notebook GPU is compatible with AMD notebooks in addition to Intel.
On the desktop, even if the graphics card supports it, a relatively modern processor and motherboard are required to activate the resizable BAR. AMD unveiled the SAM along with the Ryzen 5000 desktop processor and X570 motherboard. The technology is also included in Tiger Lake, so hopefully it will be applied to Intel’s new 11th-gen Core Rocket Lake CPUs, which will be released later this quarter. Some motherboard manufacturers have already enabled this feature on some Z490 boards, so it could be applied to Rocket Lake’s Z590 products as well.
The resizable BAR doesn’t provide a huge performance boost in most cases, but it seems to quickly become a general purpose feature. Each frame is a precious time. firstname.lastname@example.org
*The article has been translated based on the content of ITWorld Korea by www.itworld.co.kr. If there is any problem regarding the content, copyright, please leave a report below the article. We will try to process as quickly as possible to protect the rights of the author. Thank you very much!
*We just want readers to access information more quickly and easily with other multilingual content, instead of information only available in a certain language.
*We always respect the copyright of the content of the author and always include the original link of the source article.If the author disagrees, just leave the report below the article, the article will be edited or deleted at the request of the author. Thanks very much! Best regards!