According to Kopite7kimi, a whistleblower with a good breaking record and more than 12,000 followers, the latest news on the 15th, the much-anticipated new generation of NVIDIA GeForce RTX 40 series graphics cards is expected to be launched ahead of schedule. The month is slower than August.
Q3 early
– kopite7kimi (@ kopite7kimi) May 15, 2022
Kopite7kimi even broke the news that the performance of the 40 series can be as simple as 2 times the RTX 3090:
Achieving 2x the performance of the RTX 3090 is easy and looking forward to further optimization. I found that some of my friends don’t understand what I’m saying, I mean low frequency AD102 chip without full performance.
I found that some friends didn’t understand what I said.
I mean a low-frequency AD102 without a full-fat config.– kopite7kimi (@ kopite7kimi) May 13, 2022
The foreign media “Tom’s Hardware” speculates that if the news is true, Huida may see the news of the RTX 40 series graphics card at COMPUTEX held in Taipei from May 24th to 27th, and it may be June 22nd. RX Japan (VR) event, Mobile World Congress (MWC) in Shanghai on June 29.
According to the COMPUTEX official website, 6 department heads of Huida jointly delivered a keynote speech on the COMPUTEX online platform at 11:00 on the 24th, discussing how artificial intelligence (AI) can help the network, enterprise-level data centers, and provide gamers and creators with the latest products with technology.
If Huida does launch a new generation of graphics cards ahead of schedule in the third quarter, it will ensure a first-mover advantage over rival AMD. At the same time, global miners are bound to keep an eye on the computing power of Huida, AMD, and Intel’s new graphics cards.
RTX 40 series rumored specifications
“VideoCardz” reported that the RTX 40 series graphics card uses the Ada Lovelace architecture, and many details have been leaked. For example, the RTX 4090 flagship graphics card will use the AD102 chip, and the RTX 4080 may use the AD103 chip. The new series will use the new PCIe Gen5 16pin header and higher thermal design power (TDP).
In recent weeks, Kopite7kimi has beenshare The new revelations of the RTX 40 series, such as the AD102 chip configuration, have 12 GPCs (graphics processing units), 72 TPCs (texture processing units), 144 SMs (computing units), each SM has 128 FP32 floating point cores, 64 INT32 integer cores, etc.
▲ Comparison of NVIDIA AD102 GPU and previous GPU. (Source: Kopite7kimi)
According to Kopite7kimi, the power consumption of the RTX 40 series will increase sharply compared with the previous generation. The RTX 4090 is 600W, the RTX 4080 is 400W, and the RTX 4070 is 300W; in terms of memory, the RTX 4090 is 24GB GDDR6X, the RTX 4080 is 16GB GDDR6X, and the RTX 4070 12GB GDDR6X.
▲ NVIDIA RTX 40 series rumored specifications. (Source: Kopite7kimi)
(This article is reprinted with permission from Dynamic Zone; the source of the first image: NVIDIA)
New technological knowledge, updated from time to time
<!–
–> <!–
–>
<!–
–>