RL-ARM TCP/IP source code.

Source code is not provided for the stack.

Dave

Source code is not provided for the stack.

Dave

What a crazy! I paid ~4000 USD for RL-ARM and what I got is just a library without source code. Keil, you made me leave your product!!!

"I paid ~4000 USD ... and what I got is just a library"

Plus a free development tool suite, of course!

No, RealView-MDK: ~4000USD, RL-ARM:~4000USD, Total: ~8000 USD. From now, I'll begin to learn GCC and Linux.

"No, RealView-MDK: ~4000USD, RL-ARM:~4000USD, Total: ~8000 USD"

Oh yes - my mistake. Sorry.

"From now, I'll begin to learn GCC and Linux."

Keil sells you a ready-made, built, tested & supported library for ~USD4k (and, it says, you do get source code for some of that);

Even if the GCC/Linux source code is free of charge, you are going to have to do some work to get it configured & built to your requirements - and ~USD4k doesn't buy very much development time...

Of course, I don't know where your break-even point lies - but do be sure to consider all the costs...

Don't forget to add in the cost of them not supporting their product!

I still have open support issues from January 2006.

I have also paid 4000+ for development, 4000+ for source code and 2000+ to renew "maintenance"

Just out of interest, how much did the old Keil (pre-ARM) CARM toolset cost?

No difference. But the RealView Compiler is top notch / Great addition. Keil's products just need to catch up with ARM.

I also paid ~4K for the RV-MDK toolset. What I really wanted is to be able to get additional seats for a fraction of the full price. We need 4 developers to use it, and $16K is hard to push up to management.

So far, it showed to be worth it. I don't know CARM, but RV is really shiny, with a very well planned implementation, very aggressive optimizer, and well-documented lib, even for bare-machine C. If you want to integrate the system in a multithreading environment, you have all the lowlevel calls, so you can use the clib standard calls in your software. It supports DLL libs and code generation for position-independent code to run on RAM.

As for the RL lib, I didn't feel compelled to buy it. It is not uncommon for RT libs to come without full source code, but the user must have freedom to control aspects of the implementation.

"We need 4 developers to use it, and $16K is hard to push up to management."

But what is the cost of having those developers sit idle...?

If they were field support guys, would they be expected to share a single car...?

ARM's stated aim in acquiring Keil was to ease the path for 8-bit users migrating to 32-bit.

While ~$4k might be common enough in the 32-bit marketplace, a lot of 8-bit users would consider it excessive...

An another option for you can be having Floating License.
Check if that suits your requirement.

Suvidh

"While ~$4k might be common enough in the 32-bit marketplace, a lot of 8-bit users would consider it excessive..."

In a not too distant past, when product prices were confortably high, engineering budgets were never a concern. But the economics of current embedded development are much more complex. My case is of a company that manufactures some 20K units/yr and saw a significant product market price reduction on every new family. We have invested over $50K in development systems in the 8032 era. Today, engineering costs must be taken into account for the product pricing, because a product lifetime is way shorter than the one decade of those times. In that scenario, a $25K investment in new tools have to be negotiated, instead of taken for granted.

It is so much so that it makes no sense to have many different platforms for different products anymore. With ARM uC at $8, we are migrating all our product lines to the same chip, even when it is obscenely overkill to do so, to leverage on production, tools and development time.