Challenges

目录

• Easy

  • 1D Convolution
  • Color Inversion
  • Gaussian Error Gated Linear Unit
  • Interleave Arrays
  • Leaky ReLU
  • Matrix Addition
  • Matrix Copy
  • Matrix Multiplication
  • Matrix Transpose
  • Rainbow Table
  • ReLU
  • Reverse Array
  • RGB to Grayscale
  • Sigmoid Activation
  • Sigmoid Linear Unit
  • Simple Inference
  • Swish-Gated Linear Unit
  • Value Clipping
  • Vector Addition

• Hard

  • All-Pairs Shortest Paths
  • BFS Shortest Path
  • Causal Self-Attention
  • Fast Fourier Transform
  • GPT-2 Transformer Block
  • K-Means Clustering
  • Linear Self-Attention
  • Llama Transformer Block
  • Multi-Agent Simulation
  • Multi-Head Attention
  • Radix Sort
  • Sliding Window Self-Attention
  • Sorting

• Medium

  • 2D Convolution
  • 2D FFT
  • 2D Jacobi Stencil
  • 2D Max Pooling
  • 2D Subarray Sum
  • 3D Convolution
  • 3D Subarray Sum
  • Adder Transformer Inference
  • Attention with Linear Biases
  • Batch Normalization
  • Batched Matrix Multiplication
  • Categorical Cross Entropy Loss
  • Causal Depthwise Conv1d
  • Count 2D Array Element
  • Count 3D Array Element
  • Count Array Element
  • Decaying Causal Attention
  • Dot Product
  • FP16 Batched Matrix Multiplication
  • FP16 Dot Product
  • Gaussian Blur
  • General Matrix Multiplication (GEMM)
  • Grouped Query Attention
  • Histogramming
  • INT4 Weight-Only Quantized MatMul
  • INT8 KV-Cache Attention
  • INT8 Quantized MatMul
  • Linear Recurrence
  • Logistic Regression
  • LoRA Linear
  • Matrix Power
  • Max Subarray Sum
  • Mean Squared Error
  • MoE Top-K Gating
  • Monte Carlo Integration
  • Nearest Neighbor
  • Ordinary Least Squares
  • Parallel Merge
  • Prefix Sum
  • Reduction
  • RMS Normalization
  • Rotary Positional Embedding
  • Segmented Exclusive Prefix Sum
  • Softmax
  • Softmax Attention
  • Sparse Matrix-Dense Matrix Multiplication
  • Sparse Matrix-Vector Multiplication
  • Speculative Decoding Verification
  • SSM Selective Scan
  • Stream Compaction
  • Subarray Sum
  • SwiGLU MLP Block
  • Top K Selection
  • Top-p Sampling
  • Weight Dequantization