AMD's Instinct MI455X: A New Era for Rack-Scale AI
AMD's Instinct MI455X: Aiming for the Sun
I explore AMD's new Instinct MI455X, the first GPU built specifically for rack-scale AI deployments using the CDNA5 architecture. Paired with the Helios rackscale solution, this system scales to 72 GPUs in a single pod, featuring massive memory bandwidth and a new fault-tolerant networking design. This launch marks a significant shift from the long-standing GCN microarchitecture to a design rooted in RDNA, fundamentally changing how AMD approaches datacenter acceleration.
With CDNA5 AMD has moved over to a microarchitecture that is based on the RDNA series putting a bookend to the long-lived line that was the GCN microarchitecture.
- fancyfredbot
It's mind blowing seeing these multi exaflop single rack systems.
The world's first exaflop supercomputer was Frontier. It was launched only 4 years ago in 2022.
It's not a fair comparison of course. FP4 in Helios barely qualifies as floating point. Frontier was proper fp64, 16 times the bit width and probably 256x as many transistors.
All the same just wow. Much compute.
- thyristan
> The basic GCN microarchitecture underpinned every single one of AMD’s compute accelerators for nearly 15 years [...]. With CDNA5 AMD has moved over to a microarchitecture that is based on the RDNA series putting a bookend to the long-lived line that was the GCN microarchitecture.
So everything is new again. Will this be another "buy now, have ROCM work in 3 maybe years if you are lucky"?
- dist-epoch
432 GB of RAM per chip. Tens of terabytes per rack. Clients queuing up to buy them.
RAM prices are not coming down any time soon.
- inigyou
An apt title. To get to the sun, you need to slow down from orbital speed to zero. And then you fall in. Do they run CUDA yet or is this yet more impressive but unusable hardware?
AMD never misses a chance to miss a chance.
- neogodless
Duplicate post:
https://news.ycombinator.com/item?id=49025398 (13 hours earlier)