2020-09-02-native-armv7-softfloat
ARMv7 Cortex-A9 on resy-virt 3.1.2, TUNE_FEATURES = "arm armv7a vfp thumb callconvention-hard", TARGET_FPU = "hard", native


native-armv7-neon-softfloat: 

	Processor: ARMv7 Cortex-A9 (4 Cores), Motherboard: Freescale i.MX6 Quad/DualLite (Device Tree) PHYTEC phyBOARD-Mira Quad Carrier-Board with NAND, Memory: 1024MB, Disk: 4GB SD04G

	OS: resy-virt 3.1.2, Kernel: 5.4.47-custom-ml-virt (armv7l), Compiler: GCC 9.3.0, File-System: nfs

native-armv7-neon-hardfloat: 

	Processor: ARMv7 Cortex-A9 (4 Cores), Motherboard: Freescale i.MX6 Quad/DualLite (Device Tree) PHYTEC phyBOARD-Mira Quad Carrier-Board with NAND, Memory: 1024MB, Disk: 8GB SA08G

	OS: resy-virt 3.1.2, Kernel: 5.4.47-custom-ml-virt (armv7l), Compiler: GCC 9.3.0, File-System: nfs


SciMark 2.0
Computational Test: Composite
Mflops > Higher Is Better
native-armv7-neon-softfloat . 9.63  |============
native-armv7-neon-hardfloat . 34.73 |==========================================


SciMark 2.0
Computational Test: Monte Carlo
Mflops > Higher Is Better
native-armv7-neon-softfloat . 5.92  |==================
native-armv7-neon-hardfloat . 13.70 |==========================================


SciMark 2.0
Computational Test: Fast Fourier Transform
Mflops > Higher Is Better
native-armv7-neon-softfloat . 4.20  |==============
native-armv7-neon-hardfloat . 12.68 |==========================================


SciMark 2.0
Computational Test: Sparse Matrix Multiply
Mflops > Higher Is Better
native-armv7-neon-softfloat . 11.46 |===============
native-armv7-neon-hardfloat . 32.26 |==========================================


SciMark 2.0
Computational Test: Dense LU Matrix Factorization
Mflops > Higher Is Better
native-armv7-neon-softfloat . 10.45 |==========
native-armv7-neon-hardfloat . 44.48 |==========================================


SciMark 2.0
Computational Test: Jacobi Successive Over-Relaxation
Mflops > Higher Is Better
native-armv7-neon-softfloat . 16.11 |==========
native-armv7-neon-hardfloat . 70.55 |==========================================