Side-by-Side Comparison
Graphics Cards
GDDR5 vs. HBM3
Comparing 2 standards across 7 specifications. 7 key differences identified.
Comparing:
GDDR5
HBM3
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.
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
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 2022 |
|
Channels
|
— | 16 x 64-bit (1024-bit total bus, same as HBM2) |
|
Max Bandwidth / Data Rate
|
— | ~819 GB/s per stack |
|
Primary Use
|
— | AI/datacenter accelerators |