RajBhope/gpu-runtime-prediction-dataset

该数据集包含GPU工作负载性能指标数据，记录了不同GPU在不同工作负载下的性能表现，包括运行时间、浮点运算次数、内存使用量等关键指标。数据集包含51,900条训练样本，每条样本包含代码、工作负载类型、GPU名称、GPU特性、运行时间(毫秒)、浮点运算次数、内存使用量(字节)、算术强度、数据类型字节数和工作负载参数等信息。This dataset contains GPU workload performance metrics, recording the performance of different GPUs under various workloads, including key indicators such as runtime, floating-point operations, and memory usage. The dataset includes 51,900 training samples, each containing information such as code, workload type, GPU name, GPU features, runtime (ms), floating-point operations, memory usage (bytes), arithmetic intensity, data type bytes, and workload parameters.