cs.DCNov 12, 2026

GC-MoE on SM120: Overlapping Expert-Parallel Token Dispatch with Warp-Level FP8xFP4 Block-Scaled MMA on Consumer Blackwell

Authors: Yong Li, Haoyu Wang, Kan Liu, Tao Lan

Organizations: Alibaba Inc.

Abstract

Expert-parallel Mixture-of-Experts decoding is latency bound: each layer serializes a token exchange, a grouped GEMM over routed experts, and a combine, while only a few tokens per rank are in flight. Prior GC-MoE fused these phases into one persistent, Green-Context-partitioned CUDA Graph on Hopper (SM90) and datacenter Blackwell (SM100). This paper rebuilds the operator for SM120 (compute capability 12.0, e.g. RTX PRO 5000) over PCIe. SM120 exposes block-scaled tensor cores only through warp-level mma.sync.aligned.kind::mxf8f6f4: no tensor memory, no two-CTA MMA, no cluster multicast. We rebuild the kernel around a swap-AB warp-level block-scaled MMA with FP4 weights in the register A operand and FP8 activations in B, plus SM120-specific mechanisms: scale-byte selection through MMA immediates, a compact-weights path, an adaptive token-tile geometry, an expert-granular streaming dispatch with a batched bulk-pull engine and a 64-bit deduplicating pull cache, full-width BF16 returns drained by resident relay copiers over a position-major arena, and shared experts folded into the compute role. On 8xRTX PRO 5000 over PCIe the operator is 1.522x and 1.387x faster than the FlashInfer CuTe DSL W4A8 MegaMoE and the Cake MegaMoE across 18 decode points (18/18 wins each), and 1.388x faster than Cake with shared experts fused; all 94 numerical cases and 752 per-rank checks pass, at cross-operator cosine similarity of at least 0.9999997.

Figures & tables

Explore similar work

Nov 3, 2026

GC-MoE: Overlapping Expert-Parallel Token Dispatch with FP8/FP4 Grouped GEMM via CUDA Green Contexts

Under expert parallelism, mixture-of-experts (MoE) layers pay all-to-all token dispatch, grouped GEMMs, and a combine step in every layer. In decode each rank holds only 1-256 tokens: dispatch latency and weight traffic dominate while tensor-core utilization is inherently low. In fused operators such as DeepGEMM's Mega MoE, dispatch and computation overlap only by sharing every SM: dispatch warps compete with TMA/MMA and epilogue warps for registers, shared memory, and issue slots; every phase shares one CTA resource budget, synchronization spans the whole SM array, and every replay ends with a workspace-clean pass and a cross-rank barrier. We present GC-MoE, an open-source FP8/FP4 MoE inference operator for NVIDIA Hopper and Blackwell GPUs that uses CUDA Green Contexts to split the device into a small communication partition and a large compute partition. Each role gets its own kernel, resource budget, and synchronization domain, so dispatch and GEMM tune independently. A three-node CUDA Graph (four on Hopper with fused shared experts) launches the dispatch and grouped-GEMM kernels as concurrent roots; a token-ring handshake starts computation on the first filled ring blocks, and double-buffered receive counters remove the trailing cross-rank barrier between replays. The compute partition runs operand-swapped tcgen05 block-scaled MMA on Blackwell and register-A WGMMA with in-register FP4 decoding on Hopper, with fused SwiGLU, requantization, and optional shared experts. On 8 H20 GPUs GC-MoE accelerates DeepSeek-V4.1-Flash decode by 1.188x on average on both the routed and shared-expert suites (peaks 1.296x and 1.353x); on 8 B300 GPUs it accelerates GLM-5.2 and DeepSeek-V4.1-Flash by 1.106x (1.123x with fused shared experts, peak 1.209x), beating Mega MoE at all 54 measured points at matched accuracy. GC-MoE reserves only 20 of 148 SMs on B300 for dispatch.
Jul 21, 2026

Fine-grained Computation-Communication Overlap via Tile-level Signaling and Scheduling for Mixture-of-Experts

Mixture-of-Experts (MoE) architectures increase model capacity without proportionally increasing computation cost and have become a key building block for scaling large language models (LLMs) to trillion-parameter regimes. Efficient deployment of these MoE models relies on distributed execution across multiple GPUs, where each MoE layer involves two all-to-all communications: dispatching tokens to expert ranks and returning the expert outputs to their source ranks. Conventional MoE implementations launch this return all-to-all after expert compute completes, exposing communication latency on the critical path and reducing GPU utilization. We present a fine-grained approach that overlaps expert compute with the second all-to-all via tile-level signaling and scheduling. Our producer-consumer co-design combines: (1) a persistent per-rank computation kernel (producer) that covers all local experts on the rank to eliminate repeated kernel launch overhead and prioritizes remote-critical tiles, and (2) a persistent communication kernel (consumer) on a small dedicated partition of streaming multiprocessors (SMs) that issues segment-granular transfers as tiles become ready. Our co-design avoids intrusive changes to the underlying computation operators or communication primitives, making it practical for improving distributed MoE execution efficiency on multi-GPU systems. On a 4-A100 GPU platform, evaluated on three MoE models against four state-of-the-art MoE systems, our approach achieves up to 2.64x end-to-end speedup and 2.74x MoE-layer speedup. Compared with a conventional non-overlap baseline, our approach consistently improves both operator- and MoE-layer-level performance across varying GEMM shapes, router modes, and a broad range of producer/consumer SM partitions, while preserving correctness.
Nov 13, 2026

Fuse-All MoE Inference Kernel with Load Balancing on TPUs

Mixture-of-Experts (MoE) architectures activate only a subset of experts for each token, allowing large language models (LLMs) to scale parameter capacity without a proportional increase in computation per token. However, intermediate data movement and uneven expert loads limit the efficiency of MoE inference on TPUs. To address these bottlenecks, we introduce Fuse-All MoE, a TPU kernel design that jointly optimizes expert computation, output communication, and load balancing. The kernel fuses token collection, expert projections, gated activation, and output writes, retaining expert computation intermediates on chip. Fixed token-slot destinations allow scheduling and expert placement to change while preserving routing semantics. In prefill-decode (PD) disaggregated serving, Fuse-All MoE balances EP16 prefill workloads through scheduling and expert redistribution, while symmetric EP8TP2 shares active expert computation between the two TensorCore devices within each chip during decode. Our experiments show that Fuse-All MoE reduces kernel runtime for a measured skewed EP16 prefill workload from 2.3x to 1.09x the same fixed reference time for a balanced workload. Across 32-512 decode tokens, symmetric EP8TP2 reduces kernel latency by up to 21.3% relative to fused EP16. With accompanying serving stack optimizations, the integrated system improves serving throughput by up to 34.5% over the Megablox EP16 baseline.