Compiler/MSP432P401R: Fastest way to index

Part Number: MSP432P401R

Tool/software: TI C/C++ Compiler

I am not sure if this belongs here or the compiler forum, but here goes. I am using the red launchpad to drive a 12 bit DAC as quickly as possible. I need to know the fastest way to get data out of my array and onto the GPIO ports. Here is what I have now:

/*
 * Program to control FET switches for bridge simulation
 * NO DAC!
 *
 * Ports used:
 * Port J: Pins 2 &3 - Crystal
 * Port 1: Pin 0: Red LED
 *
 * Port4: DAC 8 bit port
 * Port3:0
 * Strobe - active low
 * When pulled low and high data is latched into the input register and then the DAC register
 */
/* DriverLib Includes */
#include <ti/devices/msp432p4xx/driverlib/driverlib.h>

/* Standard Includes */
#include <stdint.h>
#include <stdbool.h>

void strobe(void);
volatile uint32_t ii;


int main(void)
{

    int i, delay;

    uint8_t msb[41] =    { 80, 76, 73, 70, 68, 67, 66, 63, 62, 60,
                           59, 59, 58, 57, 57, 56, 56, 56, 55, 55,
                           55, 54, 54, 54, 54, 54, 53, 53, 53, 53,
                           53, 53, 52, 52, 52, 51, 51, 50, 46, 43,
                           40};

    uint8_t lsb0[41] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
                        0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
                        0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
                        0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
                        0};

    uint8_t lsb[41] = { 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
                        0x80, 0x00, 0x00, 0x80, 0x00, 0xD0, 0x70, 0x00, 0xD0, 0x70,
                        0x00, 0xD0, 0xA0, 0x70, 0x30, 0x00, 0xE0, 0xB0, 0x80, 0x40,
                        0x20, 0x00, 0xB0, 0x60, 0x00, 0x80, 0x00, 0x00, 0x00, 0x00,
                        0x00};

    /* Halting the Watchdog */
    MAP_WDT_A_holdTimer();

    //![Simple CS Config]
    /* Configuring pins for peripheral/crystal usage and LED for output */
    MAP_GPIO_setAsPeripheralModuleFunctionOutputPin(GPIO_PORT_PJ,
                                                    GPIO_PIN3 | GPIO_PIN2, GPIO_PRIMARY_MODULE_FUNCTION);
    MAP_GPIO_setAsOutputPin(GPIO_PORT_P1, GPIO_PIN0); // Main Red LED
    MAP_GPIO_setOutputLowOnPin(GPIO_PORT_P1, GPIO_PIN0);

    MAP_GPIO_setAsOutputPin(GPIO_PORT_P1, GPIO_PIN6);


    /* Just in case the user wants to use the getACLK, getMCLK, etc. functions,
     * let's set the clock frequency in the code. 
     */
    CS_setExternalClockSourceFrequency(32000,48000000);

    /* Starting HFXT in non-bypass mode without a timeout. Before we start
     * we have to change VCORE to 1 to support the 48MHz frequency */
    MAP_PCM_setCoreVoltageLevel(PCM_VCORE1);
    MAP_FlashCtl_setWaitState(FLASH_BANK0, 1);
    MAP_FlashCtl_setWaitState(FLASH_BANK1, 1);
    CS_startHFXT(false);

    /* Initializing MCLK to HFXT (effectively 48MHz) */
    MAP_CS_initClockSignal(CS_MCLK, CS_HFXTCLK_SELECT, CS_CLOCK_DIVIDER_1);
    //![Simple CS Config]

    /* Enabling MASTER interrupts */
    MAP_Interrupt_enableMaster();

    // Housekeeping
    //MSB
    P4DIR = 0xFF; // All port 4 are outputs
    P4OUT = 0;

    //LSB (top 4 bits)
    P5DIR = 0xFF;
    P5OUT = 0;

    P3DIR |= 0x01 ; // 3.0 is an out.
    P3OUT = 0x01; // turn on 3.0 (active low)

    while (1)
    {

        delay = 1;

        P4OUT = 255;
        strobe();

        // Delay to let the ADC catch up
        for (ii=0; ii<100; ii++);


        //Trigger
        MAP_GPIO_setOutputHighOnPin(GPIO_PORT_P1, GPIO_PIN0);
        MAP_GPIO_setOutputLowOnPin(GPIO_PORT_P1, GPIO_PIN0);

        P4OUT = 255;
        strobe();

// This is the part that needs to be as fast as possible!!!!

        if (delay != 0)
        {
            for (i = 0; i<41; i++)
            {
                P4OUT = msb[i];
                P5OUT = lsb[i];
                strobe();
            }

        }

        P4OUT = 0;
        strobe();

        // Delay to let the ADC catch up
        // Long delay to cool the FET
        for (ii=0; ii<200000; ii++);

    } /* while*/

}

void strobe(void)
{
    // Do a strobe on 3.0
    P3OUT = 0x00;
    __delay_cycles(2);

    P3OUT = 0x01;
}

I am not an ARM assembly expert, so if there are any better data types or indexing schemes or levels of optimization I am willing to entertain them.