m4 floating-point arithmetic is very slow

[xjy198903]

On the basis of the demo, using floating-point arithmetic, I found that the M4 operation speed is very slow, ten times slower than the previous use of other programs. Does this project need to do additional configuration, such as compile options?
PS:
Compiler:gcc-arm-none-eabi-4_9-2015q1
Cflags: -g -mcpu=cortex-m4 -mfloat-abi=hard -mfpu=fpv4-sp-d16 -mthumb -Wall -fno-common -ffunction-sections -fdata-sections -ffreestanding -fno-builtin -mapcs -std =gnu99