Microchip - XC8 V1.40 Compiler *NEW*

[johnsondav]

Hi Everyone

If you would like the latest Microchip XC8 PIC compiler, it is now available for download. There are two option open to you.

1. New install of the latest XC8 compiler ( you would need to update your compiler options within Flowcode V7 to use it). How to link it to Flowcode, please read this:

V1.40 Compiler Settings.pdf

(334.02 KiB) Downloaded 487 times

To download installer, from:
https://www.microchip.com/mplabxc8windows


2. Update the previous version (V1.38) to V1.40, you can download the update tool here;
https://www.microchip.com/mplabxc8-ps-windows

Kind Regards
Dave

[kersing]

Keep in mind newer does not always mean better. A new compiler may not work correctly with Flowcode generated code.

I would advice anyone considering to upgrade to use the first option.

Overwriting the compiler included in the Flowcode distribution may also cause issues when applying patch updates to Flowcode, so anyone choosing option 2: make sure to use the full installer when an update becomes available!

[johnsondav]

Hi Kersing

I agree, when using the second option I was referring to anyone who had already downloaded and installed Microchip XC8 compiler V1.38 into its own default folder, NOT FC7 folder. The second option just makes the necessary changes to an already installed compiler which may have the 'FREE' Pro option activated. I should have made this clear, sorry for any inconvenience this may have caused.

Dave

[Benj]

This sounds good from the 1.40 update notes.

Improved/expanded in-built delay functions (XC8E-106, XC8E-16) The in-built delay routines (_delay(), _delaywdt(), __delay_us(), and __delay_ms(), etc.) have been improved. All devices can now generate a three-deep loop, allowing a maximum delay of 50,463,240 instruction cycles. The watchdog variant of this delay is now available for all devices. Several inaccuracies in the generated delay have been corrected. Delays for enhanced mid-range devices are more efficient and use less temporary variables. The delay times are now no longer affected by the location of temporary variables. The _delay3() in-built function is now implemented for enhanced mid-range devices.

Means we can simplify the delay routines and hopefully get closer to the actual delay amount we want.

Last time I checked 1.40 was significantly bigger (more than double) in size than 1.38 so not sure if we can build it into the next release.