Compilers should be free. There, I said it!

Go To Last Post
75 posts / 0 new

Pages

Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

(Don't worry - I have my flame proof undergarments on today.)

I know compilers for AVR are free, this is more aimed at our new overlords at Microchip.

 

Compilers should be free!

Come on people, like Microchip bean counters, when will you realise that you are in the business of selling chips, and great development tools will sell your chips.

Crippled and cludgy development tools make people pass over to the other manufacturers. (And you can't just buy everyone!).

 

Now, this is for the bean counters:

Say you have a team of 20 staff working in your compiler department. You need to pay them, right? So you need to charge for their products right? NO! WRONG!

By charging for their products you stem the flow of funds into your core business. By giving their product away for free you will reap rewards far greater than the cost of development.

Better yet - make the compilers fully open-source, and all your future proofing, bug squishing and development can be done by contributors who don't need paid!

 

The world has changed and moved on, it's not all about cost centres and overheads anymore - business needs to be more fluid and instead of "cost centres" you could see them as "sales enablers".

 

(Mine is the one with the furry collar, thanks).

SpiderKenny
@spiderelectron
www.spider-e.com

 

Last Edited: Wed. Feb 17, 2016 - 10:10 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

When ARM takes over the world none of this will matter - one of the best ARM compilers is arm-gcc ;-)

 

(however, back when I first started using ARM (ARM7) in the late 90's there wasn't so much choice and we used ARM SDT 2.5 - with the ICE it cost $6,500 per seat - in fact the largest capital expenditure I ever got authorised was for something like $500,000 and it was to buy a number of development seats for ARM and some other bits and pieces - how things have changed in 18..20 years!)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Free compilers can stiffle innovation - if there's no financial incentive to enter a market, then why invest the time and effort? In the early days of the IBM PC there were plenty of paid compilers all fighting for market share so the vendors had to innovate to get ahead. There's been a bit of rationalisation of that market over the years.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

clawson wrote:
... one of the best ARM compilers is arm-gcc ;-)
With the follow-on in LLVM.

ARM Compiler 6 is the paid version with the LLVM organization for the freedom-and-open version (ARM, MIPS, AVR is in-work).

IIRC Apple's version of LLVM for ARM is at a relatively low annual price.


ARM DS-5 Development Studio

DS-5 / Compare DS-5 Editions

http://ds.arm.com/ds-5/compare-ds-5-editions/

 http://ds.arm.com/ds-5/build/arm-compiler-6/

 small dragon logo

The LLVM Compiler Infrastructure Project

Features

http://llvm.org/Features.html

http://llvm.org/docs/HowToBuildOnARM.html 

 

Edit : ARM Compiler 6 URL

"Dare to be naïve." - Buckminster Fuller

Last Edited: Wed. Feb 17, 2016 - 12:04 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Microchip has done very well, despite charging for full versions of their tools. Better than Atmel, I think.

Leon Heller G1HSM

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
Free compilers can stiffle innovation - if there's no financial incentive to enter a market, then why invest the time and effort?
Dual licensing is one way to have both.

The AdaCore Blog

20 years on...

by Jamie Ayre (Marketing Director, AdaCore)

Mar 11, 2015

http://blog.adacore.com/20-years-on

Licenses : GMGPL, GPL, GPL run-time exception (AVR-Ada, etc.).

http://libre.adacore.com/tools/gnat-gpl-edition/faq/#licensing

http://libre.adacore.com/comparisonchart/

http://gnuada.sourceforge.net/pmwiki.php/Packages/HomePage

http://gnuada.sourceforge.net/pmwiki.php/Main/GMGPL

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

Microchip has done very well, despite charging for full versions of their tools.

Hard to argue with that! :-)

 

However personally I would say they have done well DESPITE their approach to compilers, not BECAUSE of it.

 

Think how much more successful they could have been if they had distributed free tools!

 

(actually take a step back and think of what are possibly the most successful phenomena in micro engineering in recent years - I think you'd be hard pressed not to agree that they are Arduino and RPi - both are a success in part because they have free dev tools available. The teams might have made $20..$25 development boards "for the masses" but suppose it had then cost the user $100 for some tools to develop code - exactly how successful would those things have been then?).

Last Edited: Wed. Feb 17, 2016 - 12:29 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

gchapman wrote:

Kartman wrote:

Free compilers can stiffle innovation - if there's no financial incentive to enter a market, then why invest the time and effort?

 

 

That'll be why gcc was such a failure, and linux failed to take off... ;-)

SpiderKenny
@spiderelectron
www.spider-e.com

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Another way with freedom-and-open compilers is to add value (regression testing, porting, training, etc.)

Embecosm Logo

Embecosm

Open Source Support

http://www.embecosm.com/services/support/

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I totally agree with this. I've come across multiple instances where some IC offer a C SDK but only for IAR workbench for example. I believe that when I was looking into AVR VS PIC, the pic C compiler would not allow for any optimizations if you didn't pay for it. I'm not saying I always need to be spoiled with a gcc compiler that can be used for free (almost) without restrictions, just a full free version for non-commercial use would be nice like what Eagle does for example. This really lowers the barrier to entry. (Here's a rant by Dave Jones that touches on some of this on this: https://www.youtube.com/watch?v=9GjBFQgVWv0 )

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
Free compilers can stiffle innovation - if there's no financial incentive to enter a market, then why invest the time and effort? In the early days of the IBM PC there were plenty of paid compilers all fighting for market share so the vendors had to innovate to get ahead. There's been a bit of rationalisation of that market over the years.

 

Agreed.  I was also there during the rise of the x86--I wonder how old Kenny is?

 

SpiderKenny wrote:
I know compilers for AVR are free, ...

???  Ridiculous.  Yes, there is one flavour that incurs no direct cost.

 

SpiderKenny wrote:
Compilers should be free!

Ridiculous, again.  Who feeds those that develop all these "free" tools?

 

Back when I was your age, Intel made a decent set of dev tools, including very decent compilers.  Free?  Heck, no.  And at the prices, perhaps not even subsidized.

 

If a chip manufacturer chooses to subsidize a dev tool package, I don't have any problems with that.  Might it be the best that the developer has to offer?  Probably not.

 

 

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I can also remember the PC compiler wars.
First off there was competent but very expensive.
Then there was cheap but cr@p.
Then along came Borland with cheap but good.
I was using M68000. DigitalResearch was expensive but cr@p.
Then Mark Williams Company came along. The compiler was cr@p but all the Unix tools worked perfectly.
Finally Borland produced TurboC for 68k that was pretty good.

As far as I can remember, GCC did not arrive until after these developments.
GCC was massively better than the Amsterdam Compiler Kit that came with Minix.

Linux killed Minix and Coherent. But it was the vast number of enthusiasts that made GNU work.
I doubt if the world would be like it currently is without Torvalds and Stallman.

No, I do not mind paying a modest price for Tools. You need some form of commercialism to get innovation and progress.

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

"In the early days of the IBM PC there were plenty of paid compilers"

 

Really? I always thought those machines were so slow they must be running interpreted C... The only way to get useable speed was to handcraft assembler.

 

 

Quebracho seems to be the hardest wood.

Last Edited: Wed. Feb 17, 2016 - 05:32 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Oh you young ones and your PCs with MS-DOS!

 

I remember the glory days of CP/M where the OS came with ASM.COM - a fully functional 8080 assembler that made it quite possible for you to write your own CP/M applications. Sadly the editor they gave you was ED.COM :-(

 

(back then I wrote several magazine articles showing folks how to write applications knowing that each and every one of them already had the necessary tools on hand to do it)

 

I suppose the original MS-DOS (sorry "PC-DOS" of course) did come with GWBASIC - so it was on the right lines - but you could hardly use that to actually write "applications" like you could with an assembler.

Last Edited: Wed. Feb 17, 2016 - 05:49 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John_A_Brown wrote:
"In the early days of the IBM PC there were plenty of paid compilers" Really? I always thought those machines were so slow they must be running interpreted C... The only way to get useable speed was to handcraft assembler.

 

I tend to disagree.  Yes, an x86 ASM guru could beat any compiler.  Still true today, and many threads here on that.

 

As with early versions of say C compilers for AVR, code generation for straightforward source was decent.  And got better; just like with AVR compilers.

 

At the time I was an active SIGPLAN member, wrote some cross-compilers, evaluated tools for embedded x86 apps, and similar.  So I dug into the code generation of many C compilers.  So I have a bit of a resume in the area, FWIW.

 

Borland was mentioned, and indeed that put pricing pressure in the x86 C compiler world.  Everyone benefited from those compiler wars even though the products were not free.

 

 

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I remember the glory days of CP/M where the OS came with ASM.COM

laugh

 

But Mchip  does a free compiler NOW. wink it may end up being as popular as the Altium "free" PCB package.

http://embedded-control-europe.c...

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John_A_Brown wrote:

"In the early days of the IBM PC there were plenty of paid compilers"

 

Really? I always thought those machines were so slow they must be running interpreted C... The only way to get useable speed was to handcraft assembler.

I used David Conroy's K&R C compiler on an 0.15 MIPs, 56 kbyte PDP-11/03 running RT-11 off of dual 8-inch DSDD floppy disks. That predates the PC. It wasn't bad.

- John

Last Edited: Wed. Feb 17, 2016 - 08:36 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

theusch wrote:
I wonder how old Kenny is?

I am 47 :-)

 

As to who feeds the compiler authors - That's my whole point! Compilers which are great, and free, sell chips. The chip company makes more money, they pay the sales-augmentation department formerly known as the compiler team.

 

There is also a market for high quality, paid for, after-market compilers and dev tools too, but manufacturer supplied compilers should be free.

SpiderKenny
@spiderelectron
www.spider-e.com

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

With free tools, you get a lot of tirekickers that add to the support cost. So by putting a price hurdle in, you ensure you get people that are serious about using the tools.
When Borland released turbo pascal for cp/m at usd $99, people were queueing up. They had the best show in town. It was one of the cheapest compilers around and one of the best.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

When I took a C programming course in college in 1989, we used Microsoft QuickC. I ran it from a 5-1/4 floppy disk (couldn't afford a $300, 10 MB hard drive) on a 8088 XT running at 4.77 MHz with 64K of RAM. It was slow, but worked.

 

but you could hardly use that to actually write "applications" like you could with an assembler.

Not true. I wrote the billing program for my Dad's company in GWBASIC. Changes to code were slow and painful. But he could send invoice output to a Radio Shack printer (using pre-printed invoice forms) instead of having to have them typed on a typewriter.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

bphillips wrote:
XT running at 4.77 MHz with 64K of RAM. It was slow, but worked.

In 1989 Your 4.77 MHz XT should have had a "turbo" button to take it to a whopping 8 MHz!

 

I remember going to a site visit to trouble shoot a tape drive that had stopped working. Turned out it only worked with the "turbo" off, the SCSI interface card couldn't handle the 8 MHz option!

SpiderKenny
@spiderelectron
www.spider-e.com

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
With free tools, you get a lot of tirekickers

Well they could give the compilers for free and offer paid up support for those that need the priority access to support.

How many people have actually paid the £689+VAT for an XC32 licence, and then actually used the priority support access?

 

Microchip's website says they only have 4 copies of the XC32 in stock!

SpiderKenny
@spiderelectron
www.spider-e.com

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

"Compilers should be free. There, I said it!"

 

It's all well and good to say that here, but do you have the b@lls to actually say that on the Microchip forum?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

RickB wrote:
It's all well and good to say that here, but do you have the b@lls to actually say that on the Microchip forum?

Been there, done that. Also on twitter to Microchip.

The official response was something along the lines of "thank you for your valuable feedback"

 

SpiderKenny
@spiderelectron
www.spider-e.com

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

SpiderKenny wrote:
How many people have actually paid the £689+VAT for an XC32 licence, and then actually used the priority support access?
 

We did for the XC32 and the XC16 or whatever it is called for the pic24.

And we paid over $5k/seat for IAR and hassled the support guys as the thing was barely fit for use. After denying there was a problem, a new release magically fixed some of the problems we were experiencing then they wanted us to pay for a support agreement! Altium do a similar thing. You pay $$$ for a product that is riddled with defects and pay for support so if you're lucky they might fix them.

 

I dare say Microchip knows their customers and where their income comes from better than you do. On a slightly different tangent - the 8051 'free' compiler (SDCC) is far from optimal (note: I haven't touched it for some time - it may have got better). IAR and Keil have ruled the roost for the 8051 for many years and you pay dearly for them. Hasn't stopped the proliferation of 8051s in just about everything. Its probably the #1 8 bit core in terms of volume. Having said that, for my hobby projects I use a paid compiler for the AVR ('cos I bought it years ago before avr-gcc existed) and various free compilers for things like MIPS, ARM. X86. Professionally, mainly paid compilers.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
... and Keil have ruled the roost for the 8051 for many years and you pay dearly for them. Hasn't stopped the proliferation of 8051s in just about everything. Its probably the #1 8 bit core in terms of volume.
Did Silicon Labs purchase a wide usage license for 8051 Keil?

Silicon Labs

Silicon Labs

Top 8 Silicon Labs 8-bit Microcontroller Technology Features

...

Simplicity Studio™ Software

http://www.silabs.com/products/mcu/8-bit/Pages/8-bit-microcontroller-technology.aspx#studio

...

... free unlimited code size Keil Compiler, ...

...

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Tell me about it! I've got a dev board turning up tomorrow. As a hobbyist, free dev tools! I'll probably get harassed with emails from Keil.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It seems pretty straightforward to me. If you have an 8051, you can either use the "serviceable" free SDCC compiler or you pay a fortune for the "excellent" Keil. I suspect that the Silcon Labs arrangement is limited to their chips.

It is much the same for P*C. You either use the serviceable compiler or you pay for the "professional" version.
With the AVR, the free GCC compiler is almost as good as the IAR compiler.
I suspect that your decision is based on whether you want to be able to work every day of the week. Or whether you can live with AS7 and tools crashing without notice.
Yes, it would be nice if manufacturers supplied better tools than commercial companies. In practice there is little difference in serviceabilty nowadays. What incentive is available to IAR etc if they do not get sales?
David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:

Tell me about it! I've got a dev board turning up tomorrow. As a hobbyist, free dev tools! I'll probably get harassed with emails from Keil.

 

I've been playing with the EFM8 and have had just two emails from Keil.

 

david.prentice wrote:

I suspect that the Silcon Labs arrangement is limited to their chips.

 

It is, although given the generic nature of the '51 I wonder how locked-in you really are.

 

 

david.prentice wrote:

With the AVR, the free GCC compiler is almost as good as the IAR compiler.

 

And that is my argument with people who say all Microchip compiler should be free. There are free compilers that are almost as good as the paid-for version.

 

OF course, you now have further options as Microchip now offer a monthly subscription version. And a cloud-based version.

#1 This forum helps those that help themselves

#2 All grounds are not created equal

#3 How have you proved that your chip is running at xxMHz?

#4 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand." - Heater's ex-boss

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

clawson wrote:

Oh you young ones and your PCs with MS-DOS!

 

I remember the glory days of CP/M where the OS came with ASM.COM - a fully functional 8080 assembler that made it quite possible for you to write your own CP/M applications. Sadly the editor they gave you was ED.COM :-(

 

(back then I wrote several magazine articles showing folks how to write applications knowing that each and every one of them already had the necessary tools on hand to do it)

 

I suppose the original MS-DOS (sorry "PC-DOS" of course) did come with GWBASIC - so it was on the right lines - but you could hardly use that to actually write "applications" like you could with an assembler.

I think I'm actually a bit older than you, Cliff!

 

Quebracho seems to be the hardest wood.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

theusch wrote:

John_A_Brown wrote:
"In the early days of the IBM PC there were plenty of paid compilers" Really? I always thought those machines were so slow they must be running interpreted C... The only way to get useable speed was to handcraft assembler.

 

I tend to disagree.  Yes, an x86 ASM guru could beat any compiler.  Still true today, and many threads here on that.

 

As with early versions of say C compilers for AVR, code generation for straightforward source was decent.  And got better; just like with AVR compilers.

 

At the time I was an active SIGPLAN member, wrote some cross-compilers, evaluated tools for embedded x86 apps, and similar.  So I dug into the code generation of many C compilers.  So I have a bit of a resume in the area, FWIW.

 

Borland was mentioned, and indeed that put pricing pressure in the x86 C compiler world.  Everyone benefited from those compiler wars even though the products were not free.

 

 

I was being slightly provocative.

However, when I first started using what was, back then, still called the IBM PC, I was also running Flex on a homebrew 6809 system, and I couldn't believe how long the PC took to boot up. I genuinely thought it must be broken.

 

Quebracho seems to be the hardest wood.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Cliff wrote:

"When ARM takes over the world none of this will matter - one of the best ARM compilers is arm-gcc ;-)"

 

Agreed. I've recently done some stuff for the nRF51822, and currently looking at using the Kinetis KL26Z. It is possible to develop for both using free tools(GCC and Eclipse or similar). I have experienced some pain in getting it all to work, but not enough to make me want to pay IAR several thousand pounds.

Quebracho seems to be the hardest wood.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

As it turns out the compiler of choice for Microchip's newly acquired ARM business just happens to be arm-gcc ;-)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

By charging for their products you stem the flow of funds into your core business.

 You seem to be assuming that compilers are written by chip vendors.  Historically, and in general, that is NOT the case.  Atmel was not a significant contributor to "their" C compiler.  Microchip, Freescale/Motorola, ARM, and NXP, all bought formerly-independent compiler vendors (and a lot of people will tell you that those are not good examples of How Things Should Be Done.)  Intel has a C compiler these days, but it took a long time!), and AMD still doesn't, AFAIK.  I don't know of any vendor-provided MIPS compilers (other than Microchip XC32.

 

ARM and x86 are particularly interesting cases: I could see various chip companies sitting around hoping that one of their competitors does a good job on a compiler, so they won't have to.

 

In the old days, a lot of the (affordable) compilers came out of the Universities.  (Shucks, that was where the LANGUAGES got invented!)

 

(Now, it IS a bit disconcerting when you find an interesting chip that doesn't even have a free assembler!  Which also used to be pretty common.   I had to buy "Microsoft Assembler" for my original IBM PC.)

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In the early days of the IBM PC there were plenty of paid compilers"  ... The only way to get useable speed was to handcraft assembler.

 And the IBM PC assembler was something you had to buy, don't forget.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

westfw wrote:
I had to buy "Microsoft Assembler" for my original IBM PC.)

Aye, but MSDOS came with "debug.com" and that crappy blind editor - what was it called, edlin or something.

You couldn't actually see what you were editing, you told it a line number and typed some stuff, and saved it then you had to use the "type" command to see what the file looked like!

Ah, the joys.

SpiderKenny
@spiderelectron
www.spider-e.com

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Westfw - in the early days (80's) you paid for an assembler. I did a lot of 8051 and i had a AD2500 macro asm. It was only until the 90's when the newly formed Microchip ( from the ashes of General Instrument Corp) brought out the revitalised PIC with free assembler. About the only 'open source' and 'free' assembler i can recall was Palasm and it was written in fortran! There was probably plenty of free stuff if you had access to a unix machine and the newsgroups.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have paid for an AVR compiler since 2001.  The purchased compiler has evolved with all the chip family additions.  It is feature  rich, and saves me time.  It has been reliable, and upward compatible.  For me money well spent, no regret.   My only complaint, wish the compiler would extend beyond AVR chips.  Also thank you Theusch, you helped make the compiler better.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In any sensible business, paying for quality tools should not be a problem. Analyse the offerings, present a business case, then get finance to sign it off. That may be how some small businesses can work.

 

In most of the places I've been at though, the managers are not interested in evaluating tools - they are busy fighting fires rather than planning ahead. Meanwhile the engineers are often independently doing their own research, but can't get POs signed without project approval, so can only download and evaluate free trial editions. I know engineers who do this sort of research at home.

 

It's then those engineers that have tried the tools who push management in a particular direction when the inevitable "OMG - we need to buy a tool to finish the project" panic moment comes. So in my case, when it was realised that the project had been created by an outside contractor with a free trial version and we didn't actually have a license in house, my manager says "go order one for < $10k and I will sign it off". Since I had already tried the free versions of Keil, IAR etc, it was easy decision for me.

 

So I think there is a lot of sense for vendors to offer some free/evaluation tool, that might be limited in some way, but doesn't have a time limit or node locked license scheme. Some shops may use those free licenses for ever, but many get converted into paid for licenses.

Bob.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

donotdespisethesnake wrote:
So I think there is a lot of sense for vendors to offer some free/evaluation tool, that might be limited in some way, but doesn't have a time limit or node locked license scheme. Some shops may use those free licenses for ever, but many get converted into paid for licenses.

I still maintain that the compilers should be fully featured and free. This will maximise chip sales, and support should be paid for.

SpiderKenny
@spiderelectron
www.spider-e.com

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

SpiderKenny wrote:
This will maximise chip sales

To who? Hobbyists? It probably does. Something has to pay for the development of the free tools the vendor supplies - most likely chip sales. So that means we pay extra per chip to subsidise this. It's a competitive world and the market is crowded. So unless your silicon is significantly better than the others, then you're in a fight with the rest so margins are tight. So does free tools = chip sales or chip price = chip sales? Someone. somewhere has to pay the bills, so nothing is 'free' in a commercial environment.

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Brian Fairchild wrote:
OF course, you now have further options as Microchip now offer a monthly subscription version.

EETimes

Microchip opens PIC compiler licensing to low-cost rolling monthly subscription

January 13, 2016

by Graham Prophet

http://www.electronics-eetimes.com/en/microchip-opens-pic-compiler-licensing-to-low-cost-rolling-monthly-subscription.html?cmp_id=7&news_id=222927143

...

Subscribers will also have the ability to receive updated versions of the MPLAB XC compiler type to which they subscribe, without the need for an active Microchip MPLAB XC High Priority Access (HPA) maintenance subscription.

...

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

westfw wrote:
I don't know of any vendor-provided MIPS compilers (other than Microchip XC32.
Imagination Technologies has the Codescape MIPS SDK (GCC, C libraries, GDB, QEMU); usage of it for FreeRTOS on the forth coming Imagination Technologies Creator Ci40 board (in lieu of the OpenWrt already on the board).

FreeRTOS Interactive

FreeRTOS 8.2.3 for MIPS32 processor cores

http://interactive.freertos.org/entries/107729283-FreeRTOS-8-2-3-for-MIPS32-processor-cores

...

This port is dependent on features provided by Codescape SDK.

...

Building the Demo(for a ci40 target):

...

Imagination Technlogies

Creator Ci40 - The ultimate IoT-in-a-box dev kit

https://community.imgtec.com/platforms/creator-ci40/

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

MIPs is prominent in wireless routers and dvd players. Don't ask me why! There's a mature gcc for this architecture. How the pic32 differs, i don't know.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
... About the only 'open source' and 'free' assembler i can recall was Palasm and it was written in fortran! There was probably plenty of free stuff if you had access to a unix machine and the newsgroups....
There was quite a bit of free, and what we would now call open "open source" software, even before Unix became popular. Most of it was tied to a particular manufacture and operating system. Unless you had one of those machines, you probably would not know about all its free software. I was into small DEC minicomputers, of the sort that powered the embedded systems of their era, and greatly benefited from the Digital Equipment Computer Users Society (DECUS). Digital supported DECUS, who would send you software out of the DECUS library for little more than the cost of media and duplication. They also held national symposia where you could walk home with swap tapes of the new and updated software, created and collected by your fellow members in the special interest groups you belonged to.

- John

Last Edited: Fri. Feb 19, 2016 - 08:37 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In 1989 I don't believe that any PC's only had 64K RAM! (yes the output program .com could only be one page (64K)) 

 

My first PC (1986) had full 640K RAM, and if it didn't had a bootdisk it complained that no BASIC rom found :) (it was the time of MS2.11)

And at that time I don't think there was any real C compilers (it was the time for either turbo pascal or basic).

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

sparrow2 wrote:

And at that time I don't think there was any real C compilers (it was the time for either turbo pascal or basic).

 

Small C was certainly available as it had been ported from CP/M.

#1 This forum helps those that help themselves

#2 All grounds are not created equal

#3 How have you proved that your chip is running at xxMHz?

#4 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand." - Heater's ex-boss

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

sparrow2 wrote:
In 1989 I don't believe that any PC's only had 64K RAM! (yes the output program .com could only be one page (64K))

Not as late as 1989 perhaps. But when IBM first sold the IBM  PC 5150 in August 1981 you got a choice of two configurations. It came with either 16KB or 64KB of RAM. (that was either 9 or 36 16kbit RAM chips - not 8/32 because of the parity bit!).

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes but then you probably used the BASIC in ROM and not an external compiler. 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

But as I remember it was slowwww.

Because we had turbo pascal (compas pascal, poly pascal same thing, based on BLS pascal) on CP/M (Had to be a Z80 as I remember) I guess that everything that didn't make native code was slow, so the first program I bought for the PC was a turbo pascal. 

 

And I should add that 4.77MHz PC XT was not that much faster than a z80 at that time. (x2-x5 or so), but the 640K RAM made a difference.

 

 

Last Edited: Fri. Feb 19, 2016 - 11:57 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

westfw wrote:
Intel has a C compiler these days, but it took a long time!), ...

 

???  Now, agreed that the first high-level-language compiler offered by Intel for x86 wasn't C.  Probably PLM.

 

I'm going back to "blue box" days circa 1980 for 8086/80186/8088 work.  Now, perhaps the ASM86/LINK86/LOC86/PLM86 were purchased or some similar outside contract as others have mentioned.  But that wouldn't be my guess.  These were fully Intel packages, and one called Intel for support.  I'd guess they had their own software division.  Blue box would have first support for a new processor family.  Then the VAX-hosted versions.

 

About the time PC/AT came out, there were MSDOS versions.

 

I cannot remember exactly when C86 came out, but it would have been in the '80s as I remember it was included in an evaluation of C compilers that I did for a CNC controller manufacturer.  (80186-based in the '80s.)

 

===================================

The above is one of my lifetimes.  In an earlier lifetime, I worked for Burroughs Corp; mainframe stuff.  No mot microprocessors but I'd say a fair analogy to y'all that are saying "manufacturers should provide free tools".  Burroughs certainly didn't; whether operating system or compilers or application program packages or system and utility programs.  That's how I fed myself the first decade after college:  Working for Burroughs developing software that was >>sold<< to the customers.  Should it all have been free?  How would I have gotten paid?  What incentive would there be for Burroughs to do innovative software if they had to give it away?

 

==============================

;)  I was going to stay out of this after my early response.  But I had to question Intel offering a C compiler...

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

Pages