ncnn-android-mobilenetssd apk 点击按钮“识别-GPU”没有反应

[lovemory]

<<< 中央处理器 >>>

核心型号: Qualcomm Snapdragon 450 内核架构: 8x ARM Cortex-A53 @ 1804 MHz 制造工艺: 14 nm 指令集: 64位 ARMv8-A CPU 修订版本: r0p4 CPU 核心数: 8 CPU 频率范围: 652 - 1804 MHz 核心 1 频率: 1804 MHz 核心 2 频率: 1804 MHz 核心 3 频率: 1804 MHz 核心 4 频率: 1804 MHz 核心 5 频率: 1804 MHz 核心 6 频率: 1804 MHz 核心 7 频率: 1804 MHz 核心 8 频率: 1804 MHz 频率调节: schedutil 支持的 ABI: arm64-v8a, armeabi-v7a, armeabi 支持的32位 ABI: armeabi-v7a, armeabi 支持的64位 ABI: arm64-v8a AES: 支持 NEON: 支持 PMULL: 支持 SHA1: 支持 SHA2: 支持

设备信息。

<< Vulkan 设备 - Adreno (TM) 506 >> 设备名称: Adreno (TM) 506 设备类型: 集成 GPU 设备 UUID: 25-E7-14-03-43-51-00-00-00-00-00-06-00-05-00-00 设备 ID: 00005143-05000600 内存大小: 2896740 KB 最大 1D 图像大小: 16384 最大 2D 图像大小: 16384 x 16384 最大 3D 图像大小: 2048 x 2048 x 2048 最大 Cube Image Size: 16384 x 16384 最大 Image Layers: 2048 最大 Texel Buffer Elements: 65536 最大 Uniform Buffer Range: 65536 Max Storage Buffer Range: 2147483647 Max Push Constants Size: 128 字节 Max Memory Allocation Count: 4096 Max Sampler Allocation Count: 4000 Buffer Image Granularity: 1 字节 Max Bound Descriptor Sets: 4 Max Per-Stage Descriptor Samplers: 16 Max Per-Stage Descriptor Uniform Buffers: 14 Max Per-Stage Descriptor Storage Buffers: 24 Max Per-Stage Descriptor Sampled Images: 128 Max Per-Stage Descriptor Storage Images: 4 Max Per-Stage Descriptor Input Attachments: 8 Max Per-Stage Resources: 158 Max Descriptor Set Samplers: 96 Max Descriptor Set Uniform Buffers: 84 Max Descriptor Set Dynamic Uniform Buffers: 8 Max Descriptor Set Storage Buffers: 24 Max Descriptor Set Dynamic Storage Buffers: 4 Max Descriptor Set Sampled Images: 768 Max Descriptor Set Storage Images: 24 Max Descriptor Set Input Attachments: 8 Max Vertex Input Attributes: 32 Max Vertex Input Bindings: 32 MaxVertex Input Attribute Offset: 4096 Max Vertex Input Binding Stride: 2048 Max Vertex Output Components: 128 Max Fragment Input Components: 128 Max Fragment Output Attachments: 8 Max Fragment DualSrc Attachments: 1 Max Fragment Combined Output Resources: 72 Max Compute Shared Memory Size: 32 KB Max Compute Work Group Count: X: 65535, Y: 65535, Z: 65535 Max Compute Work Group Invocations: 512 Max Compute Work Group Size: X: 1024, Y: 1024, Z: 64 Subpixel Precision Bits: 4 Subtexel Precision Bits: 8 Mipmap Precision Bits: 8 Max Draw Indexed Index Value: 4294967295 Max Draw Indirect Count: 4294967295 Max Sampler LOD Bias: 15.996094 Max Sampler Anisotropy: 16.000000 Max Viewports: 1 Max Viewport Size: 16384 x 16384 Viewport Bounds Range: -32768.000000 ... 32767.000000 Min Memory Map Alignment: 64 字节 Min Texel Buffer Offset Alignment: 64 字节 Min Uniform Buffer Offset Alignment: 64 字节 Min Storage Buffer Offset Alignment: 64 字节 Min / Max Texel Offset: -8 / 7 Min / Max Texel Gather Offset: -32 / 31 Min / Max Interpolation Offset: -0.500000 / 0.437500 Subpixel Interpolation Offset Bits: 4 Max Framebuffer Size: 16384 x 16384 Max Framebuffer Layers: 2048 Framebuffer Color Sample Counts: 0x00000007 Framebuffer Depth Sample Counts: 0x00000007 Framebuffer Stencil Sample Counts: 0x00000007 Framebuffer No Attachments Sample Counts: 0x00000007 Max Color Attachments: 8 Sampled Image Color Sample Counts: 0x00000007 Sampled Image Integer Sample Counts: 0x00000007 Sampled Image Depth Sample Counts: 0x00000007 Sampled Image Stencil Sample Counts: 0x00000007 Storage Image Sample Counts: 0x00000001 Max Sample Mask Words: 1 Timestamp Period: 52.083332 ns Max Clip Distances: 8 Max Cull Distances: 8 Max Combined Clip and Cull Distances: 8 Discrete Queue Priorities: 3 Point Size Range: 1.000000 ... 1023.000000 Line Width Range: 1.000000 ... 1.000000 Point Size Granularity: 0.062500 Optimal Buffer Copy Offset Alignment: 64 字节 Optimal Buffer Copy Row Pitch Alignment: 64 字节 Non-Coherent Atom Size: 1 字节 API 版本: 1.1.66 Vulkan 库: /system/lib64/libvulkan.so Alpha To One: 支持 Anisotropic Filtering: 支持 ASTC LDR Texture Compression: 支持 BC Texture Compression: 不支持 Depth Bias Clamping: 支持 Depth Bounds Tests: 不支持 Depth Clamping: 支持 Draw Indirect First Instance: 不支持 Dual Source Blend Operations: 支持 ETC2 and EAC Texture Compression: 支持 Fragment Stores and Atomics: 支持 Full Draw Index Uint32: 支持 Geometry Shader: 不支持 Image Cube Array: 支持 Independent Blend: 支持 Inherited Queries: 支持 Large Points: 支持 Logic Operations: 不支持 Multi-Draw Indirect: 支持 Multi Viewport: 不支持 Occlusion Query Precise: 不支持 Pipeline Statistics Query: 不支持 Point and Wireframe Fill Modes: 支持 Robust Buffer Access: 支持 Sample Rate Shading: 支持 Shader Clip Distance: 支持 Shader Cull Distance: 支持 Shader Float64: 不支持 Shader Image Gather Extended: 支持 Shader Int16: 支持 Shader Int64: 不支持 Shader Resource Min LOD: 不支持 Shader Resource Residency: 不支持 Shader Sampled Image Array Dynamic Indexing: 支持 Shader Storage Buffer Array Dynamic Indexing: 支持 Shader Storage Image Array Dynamic Indexing: 支持 Shader Storage Image Extended Formats: 不支持 Shader Storage Image Multisample: 不支持 Shader Storage Image Read Without Format: 不支持 Shader Storage Image Write Without Format: 支持 Shader Tesselation and Geometry Point Size: 不支持 Shader Uniform Buffer Array Dynamic Indexing: 支持 Sparse Binding: 不支持 Sparse Residency 2 Samples: 不支持 Sparse Residency 4 Samples: 不支持 Sparse Residency 8 Samples: 不支持 Sparse Residency 16 Samples: 不支持 Sparse Residency Aliased: 不支持 Sparse Residency Aligned Mip Size: 否 Sparse Residency Buffer: 不支持 Sparse Residency Image 2D: 不支持 Sparse Residency Image 3D: 不支持 Sparse Residency Non-Resident Strict: 否 Sparse Residency Standard 2D Block Shape: 否 Sparse Residency Standard 2D Multisample Block Shape: 否 Sparse Residency Standard 3D Block Shape: 否 Standard Sample Locations: 是 Strict Line Rasterization: 是 Tesselation Shader: 不支持 Timestamps on All Graphics and Compute Queues: 支持 Variable Multisample Rate: 不支持 Vertex Pipeline Stores and Atomics: 支持 Wide Lines: 不支持 设备扩展: VK_KHR_incremental_present VK_KHR_shared_presentable_image VK_GOOGLE_display_timing VK_KHR_swapchain VK_KHR_maintenance1 VK_KHR_maintenance2 VK_KHR_maintenance3 VK_KHR_multiview VK_KHR_variable_pointers VK_KHR_storage_buffer_storage_class VK_KHR_relaxed_block_layout VK_KHR_get_memory_requirements2 VK_KHR_dedicated_allocation VK_KHR_external_memory VK_KHR_external_memory_fd VK_KHR_external_semaphore VK_KHR_external_semaphore_fd VK_KHR_external_fence VK_KHR_external_fence_fd VK_KHR_sampler_ycbcr_conversion VK_KHR_bind_memory2 VK_KHR_shader_draw_parameters VK_KHR_push_descriptor VK_KHR_descriptor_update_template VK_KHR_sampler_mirror_clamp_to_edge VK_ANDROID_external_memory_android_hardware_buffer VK_KHR_device_group VK_EXT_sampler_filter_minmax 实例扩展: VK_KHR_surface VK_KHR_android_surface VK_EXT_swapchain_colorspace VK_KHR_get_surface_capabilities2 VK_EXT_debug_report VK_KHR_get_physical_device_properties2 VK_KHR_external_memory_capabilities VK_KHR_external_semaphore_capabilities VK_KHR_external_fence_capabilities VK_KHR_device_group_creation