Jump to content

Intel's 7nm Xe discrete graphics is reportedly codenamed "Ponte Vecchio"


zanderthunder

Recommended Posts

zanderthunder

1567613889_intelfyi2_story.jpg

 

Intel has allegedly revealed the codename of its upcoming discrete Xe graphics to the press. This update comes from folks over at Videocardz who claim to have got their hands on a presentation slide deck meant for a press briefing.

 

Dubbed "Ponte Vecchio", the 7nm GPU is reportedly being designed as a major component of the "Project Aurora", which is intended to be the first exascale supercomputer in the US. Project Aurora will reportedly be powered by:

 

  • TWO Intel Xeon Scalable Processors (Sapphire Rapids)

  • SIX Intel Xe Ponte Vecchio GPUs

  • Ultra-high cache

  • High memory bandwidth (possibly HBM)

 

Intel will reportedly share more details about the Project on November 17 with the final deployment planned for 2021.

 

"Ponte Vecchio" is an old arch bridge in Florence, Italy and the name is apparently a reference to Intel's Compute Express Link (CXL) technology that will be used for interconnection among the components in Project Aurora. Intel will also be implementing its 'One API' initiative for easy integration of the hardware in the software ecosystem. The GPUs are also said to offer a high double-precision floating-point (FP64) throughput.

 

According to one of the slides, Intel plans to use Ponte Vecchio in a wide variety of markets, like HPC/Exascale, DL/Training, workstation, and also for gaming.

 

Source: Intel's 7nm Xe discrete graphics is reportedly codenamed "Ponte Vecchio" (via Neowin)

Link to comment
Share on other sites


  • Replies 1
  • Views 800
  • Created
  • Last Reply
2 hours ago, Edward Raja said:

Project Aurora will reportedly be powered by:

 

  • TWO Intel Xeon Scalable Processors (Sapphire Rapids)

  • SIX Intel Xe Ponte Vecchio GPUs

  • Ultra-high cache

  • High memory bandwidth (possibly HBM)

 

Might take a bit more than that to make an exascale supercomputer :) :P

 

Those items would be for a single node, with thousands of nodes in the system.

Link to comment
Share on other sites


Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...