The new AMD 400G Adaptive "Exotic" SmartNIC architecture was detailed at HC34 for the PCIe Gen5 and CXL 2.0 era
At Hot Chips 34, the Cerebras Wafer Scale Engine (WSE-2) was detailed, and the company showed how it is scaling out CS-2 deployments
At Hot Chips 34, we got a glimpse of the Tesla Dojo custom AI supercomputer, its V1 Interface NICs with HBM and how it scales data loading
In the first Hot Chips 34 Tesla talk, the company discussed the Tesla Dojo microarchitecture, the underpinnings of its AI supercomputer chips
Untether.AI Boqueria is a 1458 RISC-V core AI accelerator discussed at Hot Chips 34 that aims to scale low-power AI inference
Lightmatter Passage is a photonic wafer-scale interconnect that ties chiplet processors with silicon photonics and co-packaged optics to create something huge
The Biren BR100 is a new server GPU shown at Hot Chips 34. We discuss the new GPU and make our guess as who makes the Hearten server
At Hot Chips 34, Intel showed off features of its upcoming Ponte Vecchio GPU, but more importantly, it discussed oneAPI and D++ software tools
At Hot Chips 34, we learned more about the AMD MI250X for HPC and the MI250 OAM for mainstream GPU compute and AI/ ML workloads
At Hot Chips 34, NVIDIA went into more detail on the NVIDIA H100 Hopper generation as it waits for PCIe Gen5 CPUs and platforms