GDDR5 vs. HBM2
Comparing 2 standards across 7 specifications. 7 key differences identified.
GDDR5 was the dominant graphics memory standard from its 2008 commercial debut through the mid-to-late 2010s, running a 32-bit bus per chip at data rates that grew from around 3.6 Gbit/s at launch to 8 Gbit/s in later products, before GDDR6 began succeeding it.
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.
Specification Comparison Table
7 Parameters| Parameter / Spec | ||
|---|---|---|
|
Commercial Introduction
|
2008 (AMD Radeon HD 4870) | — |
|
Bus Width
|
32-bit per chip | — |
|
Data Rate Range
|
~3.6–8 Gbit/s per pin | — |
|
JEDEC Standard
|
— | January 2016 (JESD235a) |
|
Max Dies Per Stack
|
— | 8 |
|
Max Bandwidth / Data Rate
|
— | 256 GB/s per stack |
|
Available Capacity Range
|
— | 8GB per package |