RoadBlock: Rethinking GPU Tensor Core Microarchitecture for Emerging Microscaling Format Support

arXiv:2611.08526v1 Announce Type: new Abstract: While multiple recent custom Microscaling (MX) format proposals preserve higher model quality over OCP MX, their hardware evaluations remain confined to standalone MAC-unit prototypes, ignoring the system-level integration overheads inside real GPU Tensor Cores. To this end, we present RoadBlock, an open-source framework for accurate end-to-end modeling of custom MX formats in a programmable Tensor Core, built on the RISC-V Vortex GPGPU. RoadBlock exposes data size, scale-factor size, and block size as configurable parameters and automatically derives per-thread metadata tiling for arbitrary formats, enabling format-specific Perplexity-PPA tradeoff evaluation before silicon commitment.