Posts
Nov, 16
KernelAudit: Contract-Aware Auditing of Validators for LLM-Generated GPU Kernels
The effectiveness of generating GPU kernels driven by a large language model (LLM) is highly dependent on the quality of the used validator (used to accept or reject a generated kernel candidate against the contracts of the given kernel-generation task), whose results inform reward assignment, candidate selection, and speedup claims during kernel generation. However, an […]
Nov, 16
Reconstructing Fine-grained Memory Reuse Structure in GPU Kernels
Optimizing GPU data movement requires knowing which instructions access the same data and across which execution scopes reuse occurs. Existing profilers can attribute memory behavior to individual instructions but do not expose the reuse relations among them, while aliasing, dynamic indexing, and runtime parameters prevent static analysis from establishing these relations in general. Cycle-level simulation […]
Nov, 16
GC-MoE on SM120: Overlapping Expert-Parallel Token Dispatch with Warp-Level FP8xFP4 Block-Scaled MMA on Consumer Blackwell
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 […]
Nov, 16
NVIDIA Vera: A Microarchitecture-Level Comparison with Grace, Turin, and Granite Rapids
The majority of new HPC systems are centered around NVIDIA, AMD, or Intel CPUs. Vera is NVIDIA's latest data-center CPU, including its first custom data-center core design, Olympus. This paper investigates Vera in comparison to AMD Turin and Intel Granite Rapids, using NVIDIA Grace as a baseline. We navigate up the performance ladder, from microbenchmarks […]
Nov, 16
KernelMutate: Mutation-Guided Validation of Generated GPU Kernels
Generated GPU implementations are commonly accepted after a few differential tests, although replacement obligations may remain unchecked. We present KernelMutate, an empirical study of validation reliability using controlled faults, released implementations, and replacement tests. A five-draw anchor accepts 548 of 838 CUDA candidates. The full suite finds native FP32 witnesses for 136 accepted candidates and […]
Nov, 16
Correct First, Fast Second: What a Local 7B Code Model Can and Cannot Do for GPU Kernel Optimization
We study how much the interaction strategy around a fixed, locally hosted 7B-parameter code model (Qwen2.5-Coder-7B-Instruct) affects its ability to replace PyTorch operators with custom CUDA kernels. On 30 KernelBench tasks (levels 1-3, first 10 each, evaluated on a V100), we compare four regimes of increasing structure: zero-shot prompting, profiling-guided prompting, iterative refinement on execution […]
Nov, 8
Deep Learning Compilers: Architecture, Optimization, and Empirical Characterization
As deep learning (DL) models become increasingly complex and hardware architectures grow more heterogeneous, DL compilers play a critical role in optimizing computational efficiency and enabling scalable deployment. This paper presents a systematic analysis of DL compilers, covering their architectural evolution, compilation workflows, optimization techniques, and emerging research directions. We organize DL compiler architectures into […]
Nov, 8
Characterizing CUDA Locality Domains on Blackwell GPUs
Low latency is critical to user experience. As NVIDIA Blackwell and Rubin GPUs advance multi-die scaling, data placement and cross-die communication become increasingly important. However, the latency implications and underlying costs of cross-die communication and die-aware execution remain largely unexplored. In this report, we present a comprehensive evaluation of the Locality Domain (LD) feature introduced […]
Nov, 8
Capability Gain from ISA to Shared Execution on AMD MI300A and MI250X
Each new GPU generation is showcased by its peak instruction rates, yet users hardly ever run instructions directly: they reach the hardware's performance through vendor libraries, applications and devices shared with other jobs. How much of a theoretical speedup survives these layers is rarely measured, so a team deciding whether to move to a new […]
Nov, 8
Subgrid-Scale Turbulence Closure and 44.01× Speedup on Blackwell SM120: Matrix-Free Fused Triton Kernels for High-Resolution Global Climate Modeling
High-resolution atmospheric modeling frameworks encounter severe memory bottlenecks when evaluating non-linear subgrid turbulence closures across global meshes (such as ERA5 with 103.8 million 3D points). Standard deep learning frameworks (PyTorch Eager) dispatch multiple disjoint kernels, causing effective throughput to drop below 50 GB/s. We present a register-fused, matrix-free subgrid turbulence closure operator in Triton coupling […]
Nov, 8
HPC-Bench: A Comprehensive Benchmark for High Performance Computing Codes
Large Language Models (LLMs) have shown strong capabilities in code generation and reasoning, motivating extensive benchmarks for evaluating functional correctness. However, performance optimization in High-Performance Computing (HPC) poses fundamentally different challenges that are not captured by existing evaluations: HPC performance depends on low-level code, intricate memory access patterns, and computational kernels embedded within sophisticated execution […]
Oct, 28
CHUNDER: Enabling Million-Token Qwen3 Serving on AMD ROCm
Million-token context serving is becoming a critical requirement for industry LLM systems for real deployments. However, practical 1Mcontext serving for open-source models has remained heavily CUDA-centric, largely because the required long-context runtime features are not fully available on AMD ROCm. With AMD GPUs now reported at around 7% market share (Silicon Analysts Research, 2026), this […]
