Compiler in FREE mode

[PicoPuls]

I have the commercial version of Flowcode 7 and all the chip packs.

When compiling, the compiler says
Device: PIC.18F.18F4620
Generated by: Flowcode v7.0.0.4
Date: Friday, August 12, 2016 17:24:32
Users: 1
Registered to: XXX
Licence key: XXXXXXXXXX
You have compiled in FREE mode.
Using Omnicient Code Generation that is available in PRO mode,
you could have produced up to 60% smaller and 400% faster code.
See http://www.microchip.com/MPLABXCcompilers for more information.

Does this mean that Flowcode runs the Microchip compiler in the slowest mode ?
Is it necessary to by upgrades for all the compilers directly from Microchip to get top speeed ?

How flexible is Flowcode to change and use other C-compilers, Mikroe, CCS, or the Flowcode6 compilers ?

Thanks and best regards
Bengt

[QMESAR]

You should not be so concerend about the Free mode
I have been using the Free mode and the PRO modes since the release of the XC compilers a few years ago the Free mode compiler is very good (better than any other compiler for PIC)
and in my opion better than any other compiler including mikroC and CCS ,you can not even compare the BoostC compiler to it.

My advise would be if you are not doing very time critical applications such as SMPS and High frequency DSP functions you will never need a PRO compiler ,however if you have a PRO license you can run it very easily from FC by setting it up in the
FC configuration .

This is my experience working in the last 20 years with C18 ,mikroC pro for PIC and boostC in FC and as well working with XC8

You have compiled in FREE mode.
Using Omnicient Code Generation that is available in PRO mode,
you could have produced up to 60% smaller and 400% faster code.

This comment is something which is not totally true it depends how you wrote the code and with well written code you will never see this improvement and many people told Microchip that
this is a misleading comment from the compiler at the last master conference ,I am totally with people that mention this should be removed from the compiler as to judge that the optimization can result in this numbers you must evaluate thousands of coding styles and code lines

[PicoPuls]

Thanks for the valuable answer indeed.

The Flowcode7 Datasheet says

Microchip XC Compilers
For the first time, Flowcode will include
Microchip's XC compilers in version 7. For those
familiar with previous versions of Flowcode, this
will improve the speed at which 8bit PIC devices
compile by more than ten times compared to
previous versions of the software.
Here, we can see the template set for a Microchip Xpress
board being programmed in simulation mode using
Flowcode. Users can also them compile to the hardware
device easily using USB.

MPLAB® XC8
8-bit Architecture

MPLAB XC16
16-bit Architecture

MPLAB XC32
32-bit Architecture


Does that mean that the Microchip FREE mode compiles 10 times faster
than the previous compiler in Flowcode6 ?

Thanks and best regards
Bengt

[kersing]

Does that mean that the Microchip FREE mode compiles 10 times faster
than the previous compiler in Flowcode6 ?

Yes it does. The BoostC compiler previously used is slow, it takes a lot of time to compile a trivial program. Note: this statement concerns the compile speed, not the speed of the generated code. As QMESAR stated the generated code is better and faster, but not 10 times faster when compared to Flowcode 6.

[QMESAR]

Does that mean that the Microchip FREE mode compiles 10 times faster
than the previous compiler in Flowcode6 ?

Yes it does. The BoostC compiler previously used is slow, it takes a lot of time to compile a trivial program. Note: this statement concerns the compile speed, not the speed of the generated code.

Thank you for this good comment Kersing
yes I agree this spec Matrix is showing concerns the Compilation speed not the execution speed

[Bachman]

Does that mean that the Microchip FREE mode compiles 10 times faster
than the previous compiler in Flowcode6?

Maybe only the hex generation is faster. Same flowchart tested in FC v5, v6 and v7:

TestResults.jpg (128.08 KiB) Viewed 9826 times

[Benj]

Hi Bachman,

Some interesting results there!

Would it be possible for you to share your v5 flowchart with us. I would be interested to find out what is causing the increased code size. It could be some extra feature we have added but unless your specifically using it then it shouldn't be adding bloat to your program.

[Bachman]

Would it be possible for you to share your v5 flowchart with us.

Yes: DS18B20 & Memory (ROM) usage (Flowcode v6 topic).