Back to Study Plans
35 Problems

CUDA Kernels

Write CUDA C++ kernels and run them against correctness tests on real GPU hardware.

17 Easy15 Medium3 Hard
6 sections
Vector AdditionVector SubtractionReLUSigmoidTanhLeaky ReLUSoftmaxGELUSwish
Matrix AdditionMatrix TransposeMatrix MultiplicationOuter ProductHadamard ProductLayer NormalizationMatrix-Vector MultiplicationRMS Normalization
Dot ProductSum of ArrayL1 NormalizationL2 NormalizationMean and VarianceMax of ArrayMin of ArrayArgmaxArgmin
1D Convolution2D Convolution3D ConvolutionMax Pool 2D
Cross-Entropy LossDropoutScaled Dot-Product AttentionCosine Similarity
Tiled Matrix Multiplication