Large-Scale TensorCircuit Contractions: Disabling XLA GPU Autotuning
Impact of disabling XLA GPU autotuning on memory savings and runtime for large TensorCircuit contractions.
When executing large-scale tensor network contractions with TensorCircuit-NG and the JAX GPU backend, disabling XLA GPU autotuning can significantly reduce memory usage. This adjustment leads to more predictable memory behavior during compilation, and in some cases, can slightly improve runtime. It is particularly recommended for workloads near memory limits.