HI-TECH C for PSoC going bye-bye!

[wrightpc]

I had the following email from HI-TECH Software. It implies that PSoC will no longer be supported

Dear HI-TECH Customer,

We are pleased to inform you that HI-TECH Software has joined Microchip Technology as a fully owned subsidiary! This is the natural progression of our long-standing, friendly relationship.

In order to provide the best possible products for Microchip microcontrollers and digital signal controllers, we will focus our energies exclusively on Microchip-related products. Of course, support agreements for other products will be honored for the duration of those agreements.

If you already own a HI-TECH C compiler for any of the Microchip products, you can look forward to an even higher level of performance and responsiveness. Our teams are joining forces to provide you with the very best tools for your 8-, 16- and 32-bit PIC(R) microcontroller and dsPIC(R) digital signal controller development.

From now until September 30, 2009, Microchip is offering a 20% discount on Microchip development tools for HI-TECH customers through http://www.microchipdirect.com, and a 50% exchange discount to owners of HI-TECH compilers for non-Microchip products. Customers should contact HI-TECH for details.

Please also see this document for an at-a-glance overview of the complete Microchip C compiler portfolio and other information:
http://www.htsoft.com/HI-TECH Customers Q&A_Final.pdf

Read the press release here:
http://www.htsoft.com/news/Microchip Acquires HI-TECH Release_Final.pdf

Sincere regards,

The HI-TECH and Microchip Team

This is not good news if you don't currently own a HI-TECH compiler, but maybe looking to use it with current PSoC range.

I hope we can continue to activate the compiler to support our product into the future if my computer ever needs to be rebuilt!


Paul.

[rod]

It's lucky I stuck with PD4.4 and ICC.
At least ICC works, maybe not great, but significantly better than Hi-Tech's L-sh-ITE compiler.

I wonder how Cypress feel about this news, after going to all the trouble to p*ss Imagecraft off?

---

This message is offensive only to certain corporations. I make no apologies for that.
It is the individuals that are, once again, the real losers in all of this.

[wrightpc]

I have talked to Clyde Stubbs about support for existing customers. He says that existing customers can continue to use the tools and licences remain valid. I have no concerns for existing users, but of course future users miss out on a great product.

Hi-Tech are a great company (I was one of the early adopters and helped with some bug fixes) and I think it is a shame that the support for PSoC is going. I think the PSoC community will miss out.

As to future PSoC's, hopefully the new PSoC's will arrive SOON! They don't use M8C, so will use new tools anyway.

Paul.

[GHamblin]

Yeah, I was pretty bummed when I got the letter. I was very pleased with the Hi Tech compiler. I'll certainly continue using it as long as I can.

[richard@imagecraft.com]

Speaking for ImageCraft, we will see if we can draft a new Agreement with Cypress regarding M8C compiler usage. As far as we know, our base compiler is competitive with the Hi-Tech LITE compiler and this opportunity will allow us to re-examine the prospect of provided an optimized version of our compiler. The PSoC1 has a sizable customer base and we hope to be able to serve them again. We know there are things that we can do better and we will be addressing those issues.

Stay tuned. Feel free to email me if you wish richard @imagecraft.com

// richard

[mcody]

Maybe it is time to consider an open source solution. I have been considering, off and on, porting the sdcc C compiler to the M8C platform. There is already a code base for the 8051, PIC14, and PIC16 microcontrollers. Since the PIC and M8C controllers are Harvard architectures, I think that a lot of what has been learned there could be used for an M8C sdcc. The m8cutils package could be used for assembly code generation.

Doing this would put control of a M8C C compiler into the hands of the PSoC community, rather than the hands of a company that may or may not be dedicated to the future of C development on the M8C. If one is not happy about the optimization of some part of the C compiler, contribute to the code base and improve it. All this said, I have been doing my development exclusively in M8C assembly. Assembly language programming produces optimal code and, if clearly written and documented, can be as easily understood. Still, a freely available M8C C compiler would be a nice thing to have.

Comments?

[toalan2001]

I am going to cry myself to sleep tonight.

[BitBangerB]

It certainly seems that we are being dropped like a hot potato. This really s*cks. Expletive, expletive, expletive! I've been using the Hi-Tech Pro version practically since it came out. It would be next to impossible for me to go back to ImageCraft at this point. Even though Richard said he'd look into it, there is no guarantee that IC and Cypress will get back together. Even if they do, it will probably be weeks or months before we see a new version of ICC.

At least Hi-Tech said they would honor their support agreements. Does that mean that they will stil fix bugs, or will it be tech support only? At least they've gotten most of the bugs out by now and it seems to be compiling my project correctly. I guess I'll just be stuck with yet another abandoned compiler. This is one unhappy PSoC camper.

- B

[richard@imagecraft.com]

Keep the faith

[toalan2001]

Holy crap you can not even find mention of a PSOC compiler on the HTsoft website anymore, you can not download an evaluation nor can you purchase it.

[sriramvs]

ICC is better than Hi-TECH sometimes. The 256byte limit for variables and no support for paging by HI-TECH makes me dislike it eventhough I use it everytime for new projects.

[nan358]

Emitted code density is one thing, and compliance to the language standard is totally different a matter. HI-TECH C has always kept ahead of the compliance over ICC. That was exactly where I wanted ICC to be improved (or "fixed" in my feeling) but what wasn't realized...

--
Nobu

[GHamblin]

8 bit pointers to arrays of an 8 bit type when you only have 256 bytes.

[wrightpc]

The HI-TECH compiler saved our project as we were running out of space big-time! It took 90% FLASH usage down to about 60%. Our completed project sits at about 92% usage. I was hoping that later compiler releases would have better optimisations, but that hope is gone now.

I do look to you, Richard, to keep us informed of improvements to the ICC product for future possible products.

Paul.

[wrightpc]

Holy crap you can not even find mention of a PSOC compiler on the HTsoft website anymore, you can not download an evaluation nor can you purchase it.

The forums still exist, but I don't think any improvements will ever be made. Hopefully any major bugs will get fixed though.

http://forum.htsoft.com

Paul.

[richard@imagecraft.com]

Hi Paul,

Any work will of course depends on how our negotiation go with Cypress, but I can tell you that if we were to enter a new Agreement, Cypress will be our number one priority.

In terms of Standard conformance, we have added a new __flash keyword. A switch controls whether "const" is treated the same as __flash for backward compatibility.

In terms of a potential optimized compiler beyond what we were doing, we have some ideas. Over the base compiler that has been available, the big ones are:
- Function level global optimizations. This should give us 5%-10% better code.
- Resolve any issues with the Code Compressor. This should give 5%-15%.
- Remove unnecessary SRAM paging IO register writes. This should give 2-4% in LMM code.
- Better "flow" backend to remove unnecessary temp use. This should give 5% better code.
- Implement 8 bits data pointers. This should give 10-15% better code.

I have some ideas beyond the above, but the above list has the advantage of quicker Time to Market. We will see how things go. I will share when I can.

Thanks.

[GHamblin]

Hi Richard,
Those seem like right goals. Lets hope negotiations go well. I wonder when we will hear something from Cypress on this, or whether someone will chime in here with rumors.

[wrightpc]

Richard,

I also hope the negotiations go well. I guess one hard thing for you is what plans Cypress have for the new core's compilers (are you in there anywhere???)

Optimisations in the areas you have identified are definitely going to improve the compiler, both in code size and speed. The RAM paging architecture definitely kills code density - optimising RAM usage to minimise this will help. Similarly, for functions that are not recursive or re-entrant, having fixed auto variables rather than on the stack can save a lot of overhead.

I look forward to any news on this front in the near future.

Good luck!

Paul.

[IceCube]

This is a 1:0 for Microchip.

For Microchip, the knife cuts on both sides, first Microchip is no longer depending on a third party to make good compilers and second, non-Microchip compilers will no longer be developed and also soon cut on support.

I am pretty sure the support for the M8C compiler will go to a minimum (as far as the agreement lets it).

Time for Cypress to see third parties should be treated with respect. Cypress shouldn't be relying on just a single M8C compiler developer. And good tools are of major importance for micros! I hope Imagecraft can make a good deal to develop a 'pro' version, fortunate the HI-TECH compiler leaves us some very nice features to look at.

Besides this, I hope Cypress will encourage people to start making a sort of GNU compiler by providing code-samples and documentation on how to integrate with the latest PD.

-Rolf

[orob]

What about a gcc or miniGW compiler? I was just introduced to codeblocks that uses a number of flavors of C under the hood for a proprietary radio programming project...

It sucks that Hi-tech is getting snatched from PSoC. That compiler saved my first project, however this one I'm on now doesn't work unless imagecraft compiles it... It is good to have options.