SK Hynix mentions first details of HBM3 graphics memory: does at least 665 GB/s per chip

Published by

Click here to post a comment for SK Hynix mentions first details of HBM3 graphics memory: does at least 665 GB/s per chip on our message forum
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
Hasn't HBM largely been abandoned at this point? Unless I'm mistaken, I don't think either AMD or Nvidia are using HBM in their current workstation or server GPUs.
https://forums.guru3d.com/data/avatars/m/282/282473.jpg
they are
https://forums.guru3d.com/data/avatars/m/54/54823.jpg
I like it. We get these cute little PCBs. Probably ideal for SFF.
https://forums.guru3d.com/data/avatars/m/198/198862.jpg
Fury/Vega was the last HBM consumer cards. I dont think we'll see it again.
data/avatar/default/avatar22.webp
Intel's Sapphire Rapids platform is going to have HBM support. I would not be surprised to see HBM memory on DataCenter/Server platforms from Intel/AMD in the future for specific workloads.
https://forums.guru3d.com/data/avatars/m/248/248994.jpg
schmidtbag:

Hasn't HBM largely been abandoned at this point? Unless I'm mistaken, I don't think either AMD or Nvidia are using HBM in their current workstation or server GPUs.
Nvidia A100 uses HBM2e. It's the current top server compute card from Nvidia.
https://forums.guru3d.com/data/avatars/m/270/270008.jpg
schmidtbag:

Hasn't HBM largely been abandoned at this point? Unless I'm mistaken, I don't think either AMD or Nvidia are using HBM in their current workstation or server GPUs.
No it hasn't it just isn't in consumer products now. Intel and AMD are both using HBM for HPC chips. Nvidia is still using HBM in those high end AI chips. We will see more use of HBM in consumer products if they can get the costs down but that wont be anytime soon.