Side-by-Side Comparison
Graphics Cards
HBM2 vs. HBM3
Comparing 2 standards across 6 specifications. 6 key differences identified.
Comparing:
HBM2
HBM3
HBM2, standardized by JEDEC in January 2016, doubled HBM's maximum stack height to 8 dies and its per-pin transfer rate, reaching up to 256 GB/s of bandwidth per stack and 8GB of capacity per package.
HBM3, announced by JEDEC in January 2022, doubled channel count to 16 within the same 1024-bit bus width and reaches roughly 819 GB/s of bandwidth per stack. It's used predominantly in AI and datacenter accelerators like Nvidia's H100, rather than consumer graphics cards.
Specification Comparison Table
6 Parameters| Parameter / Spec | ||
|---|---|---|
|
JEDEC Standard
|
January 2016 (JESD235a) | January 2022 |
|
Max Bandwidth / Data Rate
|
256 GB/s per stack | ~819 GB/s per stack |
|
Max Dies Per Stack
|
8 | — |
|
Available Capacity Range
|
8GB per package | — |
|
Channels
|
— | 16 x 64-bit (1024-bit total bus, same as HBM2) |
|
Primary Use
|
— | AI/datacenter accelerators |