QNN GPU API Revision History

QNN GPU API Version

Introduced in QNN SDK Version

Description

3.7.0

2.29.0
  • QnnGpuBackend.h: Add QnnGpuBackend_CustomConfig_t

    Add QnnGpuBackend_ConfigOption_t To QnnGpuBackend_ConfigOption_t add support for:

    QNN_GPU_BACKEND_CONFIG_OPTION_ENABLE_TUNING_MODE QNN_GPU_BACKEND_CONFIG_OPTION_PERFORMANCE_CACHE_DIR QNN_GPU_BACKEND_CONFIG_OPTION_INVALIDATE_PERFORMANCE_CACHE

  • QnnGpuOpPackage.h: Add QnnGpu_BlockEncodingInfo_t to QnnGpu_MemoryObject_t

  • QnnGpuOpPackage.h: Add kernelVariant to QnnGpuOpPackage_Node_t

    Add QnnGpu_KernelTuningOption_t Add QnnGpu_KernelLocalWorkSizeTuning_t Add QnnGpu_KernelTuningConfig_t Add tuningConfigs to QnnGpu_Kernel_t

  • QnnMem.h: Add support for the QnnMem API

3.6.0

2.27.0

3.5.0

2.25.0

3.4.0

2.21.0

3.3.0

2.14.0

  • QnnGpuOpPackage.h: Add QnnGpu_MemoryLayout_t to QnnGpu_MemoryObject_t

3.2.0

2.10.0

  • Support QNN OpPackage API Version 2.0 APIs

  • Update Internal OpPackage to utilize API Version 2.0 APIs

3.1.0

2.9.0

  • Support QNN_CONTEXT_CONFIG_OPTION_PRIORITY via QnnContext_Config_t

3.0.0

2.3.0

  • QnnGpuBackend.h: Remove deprecated QnnGpuBackend_CustomConfig_t struct

2.5.0

2.0.0

  • Branched from QNN SDK version 1.15.0