Xilinx ISE drops support for more parts

If you're having trouble getting software for 3K devices, let me know.

The issue is not the software, it's the testing. With thousands of device/package/speed grade combinations, plus the platforms and OSs, plus all the EDA interfaces we support, there are millions of possibilities that need to be tested. Since the Spartan software was high quality, and we didn't want to destabalize it with enhancements needed for newer architectures, we decided to freeze the Spartan software. This allows us to focus our testing efforts where they are needed most.

Spartan software is available for free at:

By the way, I don't expect us to be dropping any more architectures from our FPGA tools.

Steve

Thanks, I appreciate the lack of HTML in the news group.

At least for XC4000E, EX, XL, XLA, and Spartan, the SW is available as Xilinx "Classic" software (free, but need normal xilinx user id/password).

As for maintaining multiple versions of the software on the same computer, I have found that VMWARE

is a very good solution. I basically set up a virtual computer for each version of the software, and each is kept separate and there are no conflicts. Costs a few GBytes of disk, which these days is only a few $.

Philip Freidin

Philip Freidin Fliptronics

I am a bit surprised that with software this complex that there was no way to de-couple the various parts. I am sure Xilinx does a lot of testing on their software, but I am always amazed when I find a bug that is known about and it takes them several major releases to fix it. My favorite was VHP__0854 The expression can not be converted to type std_logic_vector. That bug was there forever and finally fixed in

4.x.

We currently hold 7 Xilinx licenses that cost us about $10K a year to maintain. With Xilinx claiming to be best in class I expect more.

I find it interesting when I had called the Xilinx hotline that they said there was no way to get the older software. It's not an issue for us as we maintain them. For others this may be a problem.

The bigger problem is that we have some parts that we use that requires us to keep an old Window 95 PC setup just to support the old software. Plus, we don't get to leverage the newer interfaces as they are developed. Instead we are forced to remember how to use the old tools. This costs us time and money.

Thanks for picking up Synplicity. We use their full tools. Fast, don't drop support, and best of all, their tools work.

Good advice, but it addresses they symptom, not the root problem. I don't want to have to work around what Xilinx is doing. I want them to do a better job for us.

I feel for you, and every other company who is now in this trap thank to Xilinx. I think with Xilinx dropping Spartan, you really have to wonder what their long term plans are. Will we have one "stable" version of software for each series? Each version with it's own interface, bugs and PC requirements? Is this really what we expect from a company who claims to be the best?

I had called Xilinx marketing to ask some of these questions, and like the person I spoke to on the Hotline, they just don't have a clue what the long term plans are. Who is driving the ship?

It is real simple. The FPGA marketplace is mainly between two players. So a maker can focus on a few aspects of the product to compete well in the marketplace. I don't expect that long term support in the tools is a major issue with the majority of users at selection time and even if it is, who else can you choose? Is one really that much better than the other? My experience with the A vendor had one very bad example of tool support for an older product. It was still in the tool, but they wouldn't consider a bug fix even when there was no viable work around. I don't mean to keep harping on this problem, but it was very signficant to us and I now realize that there were a lot of ramifications other than just the technical issue.

--

Rick "rickman" Collins

rick.collins@XYarius.com
Ignore the reply address. To email me use the above address with the XY
removed.

Arius - A Signal Processing Solutions Company
Specializing in DSP and FPGA design      URL http://www.arius.com
4 King Ave                               301-682-7772 Voice
Frederick, MD 21701-3110                 301-682-7666 FAX

Can we get a quick summary of what's removed, and what legacy versions of SW are needed to support which family ?

Peter A mentioned SpartanXL as being still active, (and lowest power?) but not supported in the latest SW ?

I also noted in another thread that Altera ADDED support for an older 10Kxx family to Quartus 3.0.

-jg

Steve Lass wrote: [snip] :I'm driving the ship and like I said, we have no plans to drop any other :architectures from our :software. All the FPGAs we have in the software now are derivatives of :the Virtex arcitecture :so keeping them in the release is not difficult.

My apologies if I've missed something: I join this thread late. But now I'm worried: I was about to start a new design using Spartan: exactly what has been dropped? (It wasn't evident from the Xilinx website).

All the tools co-exist from what I've seen. You just need to set your PATH and XILINX environment variables appropriately for each version. That probably precludes selecting any version of any tool from the Start menu, but you can easily write a batch file to choose a version then run the tool from the command line. (floorplanner, fpga_editor, par, pace, etc.)

I don't think dropping support for old devices is too unreasonable. Otherwise the QA effort for each new software version (major, minor, even service pack) just grows bigger and bigger, and the design of the software gets more complex and messier etc. We just did something similar at my work - dropped support for old hardware in a new software version. The complexity and QA effort was killing us.

At least Xilinx keep all the support notes for the old tool versions on their web site.

Hamish

--
Hamish Moffatt VK3SB

In article , snipped-for-privacy@cloud.net.au writes: |> I don't think dropping support for old devices is too unreasonable. |> Otherwise the QA effort for each new software version (major, minor, |> even service pack) just grows bigger and bigger, and the design of |> the software gets more complex and messier etc.

Why would this be so? If the software is modularized, e.g. the fitter (placer/router) is its very own piece of command-line software there is no need to touch that code again (plus, doing so eases portability).

If integration of the necessary calls into the GUI is an issue, well, then just leave it out. From what I read in this and other "tech" groups, developers seem to prefer to write their own batch scripts anyway. (If the shared development machines sit "two networks further", you don't want to wait for GUI updates anyway.)

Just my $0.02, Rainer

Our software is modular, but each module is not self contained. There are common routines we use for constraint parsing, command line handling, IO interfaces, database management, base level mappers, etc. Making changes to these common utilities can have unexpected effects on older software. This was a bigger issue for us with the pre-Virtex architectures. Like I said, since all FPGAs in the software are based on the Virtex architecture, there is no reason for us to drop any more families.

Right now, our software is over 20 million lines of code. If each architecture's software was self contained, my guess is, that would double. I would think that most of you would rather we put our testing effort into the more recent architectures.

Again, the 4K and Spartan/XL software is available, fully supported, and free.

Steve

Steve, you have a tough job making these calls, and I am sure you are guiding Xilinx in a maner that is best suited for them. I believe you when you made the statement "..we have no plans to drop ...". I am sure there were no plans to drop any parts from the tools at a given time.

Fool me once, shame on me. (3.x drops 3xxx parts) Fool me twice (5.x drops Spartan +), .....

Our company has been using Xilinx exclusively. This latest move from Xilinx has caused us to re-think our stratagy. All of the tools we use to interface with Xilinx will support both companies and we mostly deal with strick VHDL. Supporting both Altera and Xilinx is not a problem.

Xilinx is sending someone next week to talk with us. Should be interesting.

Here are some tutorial notes, to calm the waves:

The physical life-length of typical FPGAs is >>10 years continuously at the max allowed stress. In practical terms this means they work and are reliable for 50 years or even longer, effectively forever. "Old FPGAs never die, they just get obsolete".

The commercial availability of popular parts is well beyond 10 years after introduction. Unpopular sleepers may be discontinued earlier. Any responsible manufacturer gives you 1 to 2 years of warning and "last buy" opportunities, and sometimes transfers the die inventory to an "afterlife" supplier.

Software support in the most current releases can be a problem, and has been discussed here. For Xilinx, the architectural commonality of all Virtex and Sparten-II ( and later) devices makes it easier to maintain support for all these families. Old software is available for free, but may require an old computer OS.

Design-in life is much shorter, since newer devices make the older devices appear slow and expensive. "One year in the life of an FPGA equals 15 years of human life" means that a device introduced 2 years ago is in its prime, a device introduced 4 years ago is now a senior citizen, fully capable but not to be raced against the newer parts. (There are special cases: Spartan-XL, with its roots in XC4000, is today the last really low idle-power FPGA family, and is 5-V tolerant and -compatible.)

Freshly announced parts are exciting, but the user must ascertain their availability. The manufacturer should not be criticized for releasing product details ahead of availability. That gives the user a chance to plan and evaluate.

And finally, we all know and understand that the frentic pace of FPGA innovation is at odds with the long design cycle time of miltary and aerospace projects.

Just some common thoughts...

Peter Alfke ==========================

The current software (version 5.1i, 5.2i, 6.1i) supports Virtex, Virtex-E, Virtex-II, Virtex-II Pro, Spartan II, Spartan IIE, and Spartan-3. ISE Classics (version 4.2i) supports XC4000E, XC4000L, XC4000EX, XC4000XL, XC4000XLA, Spartan, and SpartanXL.

Contact the hotline if you need software for:

3.1i supporting XC3000A, XC3000L and XC5200. XACT 6 supporting XC2000, XC3000, and XC4000, XC4000A and XC5200.

Steve

:> Old software is available for free, but may require an old computer :> OS.

: What about releasing the source of unsupported tools into the public : domain, or as Free Software? Unmaintained binary only software is not : a big help, but old source code that one can maintain oneself is a : winner.

: You might also learn a thing or two about the 'Free Software / Open : Source community' that way and how it might help you make Xilinx more : popular.

To cite the Bible:

"Eher geht ein Kamel durch's Nadeloehr, als dass ..."

Probably the software is hopelessly interweaved with externelly licensed software and other parts that are still considered crucial. I doubt that there is a big difference between an FPGA suite and Netscape. And look what effort it took to get from Netscape to Mozilla. The audience for a FPGA suite is oders of magnitude smaller.

Anyways, such a move would be great...

Bye

B.t.w.: For mantaining legacy version of windows software, the windows emulator wine

might be an option for the not-so-faint-hearted.

--
Uwe Bonnes                bon@elektron.ikp.physik.tu-darmstadt.de

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

:> Old software is available for free, but may require an old computer :> OS.

: What about releasing the source of unsupported tools into the public : domain, or as Free Software? Unmaintained binary only software is not : a big help, but old source code that one can maintain oneself is a : winner.

: You might also learn a thing or two about the 'Free Software / Open : Source community' that way and how it might help you make Xilinx more : popular.

To cite the Bible:

"Eher geht ein Kamel durch's Nadeloehr, als dass ..."

Probably the software is hopelessly interweaved with externally licensed software and other parts that are still considered crucial. I doubt that there is a big difference between an FPGA suite and Netscape. And look what effort it took to get from Netscape to Mozilla. The audience for a FPGA suite is orders of magnitude smaller.

Anyways, such a move would be great...

Bye

B.t.w.: For maintaining legacy version of windows software, the windows emulator wine

might be an option for the not-so-faint-hearted.

--
Uwe Bonnes                bon@elektron.ikp.physik.tu-darmstadt.de

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

This sounds like I better terminate my Xilinx software subscription, and use the free versions instead.

I recently changed my PC to a new one, and tried to install the licensed version of ISE 4.2, because I need Spartan and Spartan XL. It was a pain to make it work again because one design uses FPGA express, and it always complained about the license being invalid.

The only solution was, as a Xilinx FAE told me, to use use tool to change the volume serial number of my hard disk to the one that the old PC had. Fortunately I didn' have to reregister Windows XP again.

Thomas

Which tool did you find that would let you change the disk serial number?

--

Rick "rickman" Collins

rick.collins@XYarius.com
Ignore the reply address. To email me use the above address with the XY
removed.

Arius - A Signal Processing Solutions Company
Specializing in DSP and FPGA design      URL http://www.arius.com
4 King Ave                               301-682-7772 Voice
Frederick, MD 21701-3110                 301-682-7666 FAX