鲲鹏920环境Benchmark跑量化模型crash

[zhenjing]

平台(如果交叉编译请再附上交叉编译目标平台):

-- >>>>>>>>>>>>>
-- MNN BUILD INFO:
-- System: Linux
-- Processor: aarch64
-- Version: 2.8.4
-- Metal: OFF
-- OpenCL: OFF
-- OpenGL: OFF
-- Vulkan: OFF
-- ARM82: ON
-- oneDNN: OFF
-- TensorRT: OFF
-- CoreML: OFF
-- NNAPI: OFF
-- CUDA: OFF
-- OpenMP: ON
-- BF16: OFF
-- ThreadPool: OFF
-- Hidden: TRUE
-- Build Path: /home/zhenjing/MNN-2.8.4/build
-- CUDA PROFILE: OFF
-- WIN_USE_ASM:
-- Enabling AArch64 Assemblies
-- Enable INT8 SDOT
-- [*] Checking OpenMP
-- Found OpenMP_C: -fopenmp
-- Found OpenMP_CXX: -fopenmp
-- Found OpenMP: TRUE
-- Configuring done
-- Generating done

Github版本: 2.8.4

./benchmark.out models/ 50 0 0 1 2 0 1 1
MNN benchmark
Forward type: CPU thread=1 precision=2 sparsity=0 sparseBlockOC=1 testQuantizedModel=1
--------> Benchmarking... loop = 50, warmup = 0
[-INFO-]: precision=2, use fp16 inference if your device supports and open MNN_ARM82=ON.
[-INFO-]: Auto set sparsity=0 when test quantized model in benchmark...
Auto set sparsity=0 when test quantized model in benchmark...
The device support i8sdot:1, support fp16:1, support i8mm: 1
[ - ] MobileNetV2_224.mnn max = 14.354 ms min = 13.765 ms avg = 13.997 ms
Illegal instruction (core dumped)

[zhenjing]

./benchmark.out models/ 50 0 0 1 0 0 1 1
MNN benchmark
Forward type: CPU thread=1 precision=0 sparsity=0 sparseBlockOC=1 testQuantizedModel=1
--------> Benchmarking... loop = 50, warmup = 0
[-INFO-]: precision!=2, use fp32 inference.
[-INFO-]: Auto set sparsity=0 when test quantized model in benchmark...
Auto set sparsity=0 when test quantized model in benchmark...
The device support i8sdot:1, support fp16:1, support i8mm: 1
[ - ] MobileNetV2_224.mnn max = 29.129 ms min = 28.343 ms avg = 28.640 ms
Illegal instruction (core dumped)

[jxt1234]

看上去是判定 i8mm feature 出错了: support i8mm: 1
可以自行修改 cpuruntime.cpp 试试