黑洞和TT-金属独立的人工智能计算机及其编程模型

Blackhole & TT-MetaliumThe Standalone AI Computer and its Programming ModelAugust 2024Jasmina Vasiljevic, Senior FellowDavor Capalija, Senior FellowAgenda•Architecture•Micro-architecture•Scale-out•Software3AI Silicon RoadmapStandalone AI ComputerWormhole•80 Tensix+ Cores•12nm•328 TOPS (FP8)•336 GB/s GDDR6•Gen4x16•16x100 Gbps Ethernet Blackhole•140 Tensix++ Cores•6nm•745TOPS (FP8)•512 GB/s GDDR6•Gen5x16•10x400 Gbps Ethernet•16 RISC-V CPU coresNetworked AI ProcessorAI ProcessorGrayskull•120 Tensix Cores•12nm•276 TOPS (FP8)•100 GB/s LPDDR4•Gen4x16202120222023GEN 1GEN 1GEN 2High Perf AI ASICScalabilityHeterogeny4Tensix coresDRAM coresETH coresPCIe coreARC coreTDEPABlackhole - A Standalone AI ComputerCRISC-V CPUs5Blackhole - A Standalone AI ComputerFeatureSpecTensix745 TFLOPs (8-bit)372 TFLOPs (16-bit)SRAM241 MBsEthernet10x 400 GbpsDRAM512 GB/s BW32 GBs capacityBaby RISC-Vs752Big RISC-Vs16PCIeGen5x16, 64 GB/sNoC2 NOCs2D Torus256 B per core6Big RISC-V & Baby RISC-VCRISC-V CPUsFeatureSpecRI