High Current Draw With Floating Point Calculation

[Ken]

I'm attempting to optimize our application for battery life on the nrf52832.

If I execute a floating point calculation to rescale RGB to PWM, I start drawing 6.2mA, eliminating the calculation reduces current to an more reasonable 0.05mA. To be clear I am not using the calculation result to drive the PWM yet, just making the calculation & saving the result globally

volatile uint16_t blue_cmd; void rgb2pwm(uint8_t blue) { float blue_duty_cycle = (float)(blue)/2.55f; blue_cmd = (uint16_t)blue_duty_cycle; }

Is there something that needs to be done to minimize power for floating point calculations?

[Jørgen Holmefjord]

Hi,

Make sure you have implemented the workaround for PAN-87 - CPU: Unexpected wake from System ON Idle when using FPU.

Best regards,

Jørgen

[Ken]

That fixed the issue. Thank-you ken

[Marten]

Just saw that the fix is also used when calling nrf_pwr_mgmt_run() (in nrf_pwr_mgmt.c)

__STATIC_INLINE void pwr_mgmt_fpu_sleep_prepare(void)
{
    uint32_t fpscr;
    CRITICAL_REGION_ENTER();
    fpscr = __get_FPSCR();
    /*
    * Clear FPU exceptions.
    * Without this step, the FPU interrupt is marked as pending,
    * preventing system from sleeping. Exceptions cleared:
    * - IOC - Invalid Operation cumulative exception bit.
    * - DZC - Division by Zero cumulative exception bit.
    * - OFC - Overflow cumulative exception bit.
    * - UFC - Underflow cumulative exception bit.
    * - IXC - Inexact cumulative exception bit.
    * - IDC - Input Denormal cumulative exception bit.
    */
    __set_FPSCR(fpscr & ~0x9Fu);
    __DMB();
    NVIC_ClearPendingIRQ(FPU_IRQn);
    CRITICAL_REGION_EXIT();

    /*
    * Assert no critical FPU exception is signaled:
    * - IOC - Invalid Operation cumulative exception bit.
    * - DZC - Division by Zero cumulative exception bit.
    * - OFC - Overflow cumulative exception bit.
    */
    ASSERT((fpscr & 0x07) == 0);
}