Companies such as AMD, NVIDIA and Intel are exploring the future of artificial intelligence, machine learning and supercomputers. After amd demonstrated the insight mi250x OAM module based on "single dual core" Aldebaran GPU chip and NVIDIA demonstrated GH100 GPU through H100 sxm5 high-performance computing module, Intel is also entering this field now. Ponte Vecchio GPU will help the company achieve more computing algorithms in the next decade and use graphics card technology to enter data centers and machine learning systems**
Intel promises to use 100 billion transistors in the new Ponte Vecchio computing GPU, bringing the company into more mainstream artificial intelligence and high-performance computing applications.
Last week, Intel confirmed on twitter that the company has started testing and sampling Ponte Vecchio computing GPU. Their new computing GPU will enable the company to enable consumers to get more development and functions through AI and HPC applications. In Western Europe, Mikael Moreau, Intel's public relations manager, revealed modules that provide unique cooling components. For users, this tells us that the company has begun to deliver parts to partners in the United States.
Intel's next-generation Ponte Vecchio computing GPU provides up to 47 processor cores, one of which contains a staggering 100 billion transistors. A computing GPU of this scale will require a lot of power consumption, while the new OAM addition module plan requires 600W.
Since it will operate under such restrictions, the system will need sufficient cooling. Intel has chosen to study liquid cooling technology for the new computational GPU, but it occupies the smallest area compared with the liquid cooler of larger systems. What we don't know from the twitter post is whether we see the standard Ponte Vecchio computing GPU or the advanced XT variant of the next generation GPU.