v24.11.1 Public Minor Release
Feat
- Add stateless GEMM execution via ICPPKernel::run_op
- TensorShape class supports dynamic shapes
- Add skeletons for Dynamic GEMM operator
- Convert Double rounding to Single rounding quantization behaviour in both Cpu/Gpu backend
Fix
- Detect Advanced SIMD support on Windows®
Perf
- Implement activation heuristics for Neoverse™ V1
- Optimize PReLU on quantized datatypes
Documentation (API, build guide, contribution guide, errata, etc.) available here:
https://artificial-intelligence.sites.arm.com/computelibrary/v24.11.1/index.xhtml