Network Security Internet Technology Development Database Servers Mobile Phone Android Software Apple Software Computer Software News IT Information

In addition to Weibo, there is also WeChat

Please pay attention

WeChat public account

Shulou

AMD: it's not that you can't make a RTX 4090 graphics card, just because the high price and high power consumption are not cost-effective for gamers.

2025-02-27 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > IT Information >

Share

Shulou(Shulou.com)11/24 Report--

Thanks to CTOnews.com netizens OC_Formula and Wu Yanzu in South China for their clue delivery! CTOnews.com, March 12, AMD executives said in an interview that it was technically possible to develop a RDNA 3 GPU for Standard Invidia RTX 4090, but decided not to go this way because of power consumption and cost.

In September last year, Nvidia officially released a new generation of RTX 4090 flagship graphics cards with 16384 CUDA + 24GB memory at a proposed price of 12999 yuan, but due to the scarcity of supply, most of them are around 14000 yuan or even 16000 yuan, and some models can even reach 18999 yuan.

Subsequently, AMD launched the Radeon RX 7000 series of graphics cards, in which the suggested price of RX 7900XTX slightly stronger than RTX 4080 (public version 9499 yuan) is 7999 yuan, and that of RX 7900XT weaker than RTX 4080 is 7399 yuan, while the latter has now dropped to the 6499 yuan price range of RTX 4070Ti.

Japanese media ITmedia interviewed Wang Qishang (David Wang), senior vice president of AMD, in charge of engineering research and development of AMD Radeon technology division, and Rick Bergman (Rick Bergman), senior vice president and general manager of product department of AMD.

The interview began by pointing out what players are concerned about why AMD did not release a RDNA 3 GPU in its Radeon RX 7000 series to rival RTX 4090, thus competing with NVIDIA in the ultra-premium market.

Technically, it is possible to develop a GPU with (NVIDIA) competing specifications. However, the graphics card developed in this way with a 600W GPU TDP and a reference price of $1600 (CTOnews.com Note: currently about 11120 yuan) is difficult to be accepted by the average PC game fans, and finally chose not to adopt this strategy after careful consideration.

It is not difficult to see from his description that this competition refers to "GeForce RTX 4090. It has also been revealed that the target TDP of GeForce RTX 4090 was 600W at the beginning of its development, but the final version was reduced to 450W.

The target price of the "Radeon RX 7900XTX" released this time is $999, which is considered to be the "upper limit" borne by high-end users of the average PC game fans. (CTOnews.com Note: the interview will take place in December 2022, and RX 7900XTX later offered an initial price of $899.)

The pricing strategy is the same as the previous RDNA 2 (Radeon RX 6000 series), with high-end "Radeon RX 6900XT" and "Radeon RX 6800XT" targeting prices of $999 and $699, respectively. However, the target price of each generation of products changes.

We use this strategy to adapt to the mainstream infrastructure (hardware environment) used by today's PC gaming enthusiasts. While requiring high performance, it should be possible to continue to use the existing "regular power" and "conventional radiator" and do not need a very large chassis to install. The Radeon RX high-end product group was designed with this in mind.

According to public information, Nvidia GeForce RTX4090 based on TSMC 4N process to create AD102 GPU, its chip bare chip area up to 608mm ², a total of 76.3 billion transistors, 16384 CUDA cores, 24 GB high-speed GDDR6X memory, the official TDP of 450W, recommended 850W power supply.

In contrast, the AMD Radeon RX 7900 XTX uses Navi 31 GPU, with only two 8Pin external power interfaces. It is AMD's first consumer-grade GPU in MCM multi-chip package, with 58 billion transistors, one GCD (graphics computing chip) and six MCD (multi-cache I / O chips). At the same time, it will use two different processes, the former is TSMC's 5nm, and the latter is 6nm.

The flagship model RX 7900 XTX has a 96CUJ 6144 stream processor, GPU frequency 1900-2500MHz, and single-precision performance up to 61TFLOPS. The video memory is 24GB 20Gbps GDDR6, and the bandwidth can reach 960GB / s. Coupled with 96MB unlimited cache, the equivalent bandwidth can reach 3.5TB / s. The TBP power consumption of the video card is 355W, and the recommended power supply is 800W. The length of RX 7900 XTX is 287mm,2.5 slot thickness and provides interfaces such as DP 2.1and USB-C.

The sub-flagship model RX 7900 XT has an 84CUPower5376 stream processor, GPU frequency 1500-2400MHz, and single-precision performance up to 52TFLOPS. The video memory is 20GB 20Gbps GDDR6, and the bandwidth can reach 800GB / s. Coupled with 80MB unlimited cache, the equivalent bandwidth can reach 2.9TB / s. The TBP power consumption of the video card is 300W, and the recommended power supply is 750W. The length of RX 7900 XT is 276mm,2.5 slot thickness, and the interfaces include DP 2.1and USB-C.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.

Views: 0

*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.

Share To

IT Information

Wechat

© 2024 shulou.com SLNews company. All rights reserved.

12
Report