Quartus 13.1

--- Quote Start ---

Nope, Altera is not planning to add the CII in the future versions

--- Quote End ---

Hi,

I use Altera Boards in the lab here at university and we just ordered about 60 Altera DE1 boards:

http://www.altera.com/education/univ/materials/boards/de1/unv-de1-board.html

which are equipped with Cyclone II FPGAs. Quartus v13.0 supports Cyclone II. Some students had problems

installing 13.0 on their windows systems so they upgraded to 13.1 which was now useless. I see a big problem

with our lab investments if there is no roadmap for Cyclone II in Quartus. Maybe someone from Altera can comment on the plans for the university program and educational boards.

Regards

Friedrich

--- Quote Start ---

Hi,

I use Altera Boards in the lab here at university and we just ordered about 60 Altera DE1 boards:

http://www.altera.com/education/univ/materials/boards/de1/unv-de1-board.html

which are equipped with Cyclone II FPGAs. Quartus v13.0 supports Cyclone II. Some students had problems

installing 13.0 on their windows systems so they upgraded to 13.1 which was now useless. I see a big problem

with our lab investments if there is no roadmap for Cyclone II in Quartus. Maybe someone from Altera can comment on the plans for the university program and educational boards.

Regards

Friedrich

--- Quote End ---

I second that. We have a lot of DE2-70 boards with a Cyclone II for lab work and face the same problem.

Jesse

--- Quote Start ---

I second that. We have a lot of DE2-70 boards with a Cyclone II for lab work and face the same problem.

Jesse

--- Quote End ---

Both Cyclone II and Quartus II 13.0 are still supported by Altera. If someone's having trouble installing v13.0, try opening a mySupport request with Altera for us to fix it (after checking the normal readme files, etc).

Cyclone II has not had any significant improvements in Quartus II in years, and most (if not all) of the new features in the latest versions would not support or enhance CII in any way. What kinds of "roadmap" are you looking for?

--- Quote Start ---

Both Cyclone II and Quartus II 13.0 are still supported by Altera. If someone's having trouble installing v13.0, try opening a mySupport request with Altera for us to fix it (after checking the normal readme files, etc).

Cyclone II has not had any significant improvements in Quartus II in years, and most (if not all) of the new features in the latest versions would not support or enhance CII in any way. What kinds of "roadmap" are you looking for?

--- Quote End ---

Let me explain our situation. We have installed Quartus 11.1 on our machines. The install was done by our wonderful IT-service some time ago. This is the subscription edition. We had tested the install over and over again until it finally worked (there are over 350 software packages our school maintains, we use package-maintaining software and distribute only the PCs which should have Quartus installed). It took us three months or so. So we don't want to install every time there's a new version.

Our students have new laptop of their own with windows 8. They install always the latest version, as it should contain the least of bugs. Versions before v13 give some

trouble with installing the device driver (signed etc). Now we could tell them to install 12.1 or 13.0, but they claim that the latest is latest is always the better one and some things might not work on win8 or future versions of windows (win9, win10, ...)

Our roadmap is to use DE2-70 and DE0 boards for some years on. We cannot buy new boards every year. A school needs 20,30,40 boards and not just one.

Anyway, why is it a problem to include the CII in v13.1? And why skip support in the middle of major version? Oh, and please revert the RTL viewer to a previous version. The schematic it creates in v13.x look horrible. The one of v11 look good, I can use the schematics in my slides. And my students ask over and over if OSx will be supported in the future. It's a Unix after all ;-)

Greets,

Jesse op den Brouw

Hi Jesse,

The reason we're removing families is that family support is far more than just a database file - there can be hundreds of thousands of new lines of code written for most new families, because they often contain at least some features that were entirely new. For example, Cyclone II was the first family without a 6-LUT that supported independent placement of LUTs and FFs during detailed placement. In general, there can be far more variability between different FPGA fabrics than there are between different processor architectures, which means that *all* aspects of the compiler (including supporting tools like timing analysis) can be affected, not just a backend machine instruction generator (as you might find in GCC or LLVM).

However, over time the focus of development shifts to the newer families, which can have very little in common with older families despite a common name (Cyclone V looks very little like a Cyclone II). At this point, all that code must still be maintained and tested, but are not being improved. On the contrary - sometimes an algorithm will be rewritten in a way that helps the newer families but hurts the older ones. Or perhaps a piece of infrastructure is rewritten to handle massive new memory requirements, but back-porting that piece of infrastructure to an old family would have a negligible impact on memory yet cost a lot of time and probably generate a lot of bugs.

Even without backporting, features are constantly upgraded with the new infrastructure in mind. Unfortunately, developers (who are only human) might not fully understand the older version and break features that used to work. And because our tests coverage isn't perfect and many of our customers don't use the old families anymore, these bugs can escape testing and make it into the wild. So the claim that your students are making - that the latest version is always the best - is not always true.

In summary, maintaining old families indefinitely results in minimal improvements to those old families and the possibility of new bugs that increases over time. At some point, therefore, it makes sense to drop support for older families. This is just like any software package dropping support for older hardware and OS (the latest Photoshop requires Windows 7, which was only released in 2009).

Other notes: 13.1 is not a "minor" release, it's simply the second release in 2013. We generally release twice a year; the first one is called [20]xx.0 and the second is [20]yy.1. I can't comment on the RTL viewer but if you file a mysupport with screenshots of the old and the new, perhaps the relevant team could make some improvements. And while no Altera software supports Windows 8 (yet), I'd be surprised if you had trouble running 13.0 on W8 since most OSes are pretty good about backwards compatibility. File a bug through mySupport if you run into trouble with that; I'll also check to see if we have any formal plans to validate 13.0 on Win8 and other newer OSes.

--- Quote Start ---

...

Other notes: 13.1 is not a "minor" release, it's simply the second release in 2013. We generally release twice a year; the first one is called [20]xx.0 and the second is [20]yy.1. I can't comment on the RTL viewer but if you file a mysupport with screenshots of the old and the new, perhaps the relevant team could make some improvements. And while no Altera software supports Windows 8 (yet), I'd be surprised if you had trouble running 13.0 on W8 since most OSes are pretty good about backwards compatibility. File a bug through mySupport if you run into trouble with that; I'll also check to see if we have any formal plans to validate 13.0 on Win8 and other newer OSes.

--- Quote End ---

Hi,

this may sound silly from the pure "straight ahead numbering" point of view but maybe just shifting from 13.0 to 13.1 seems to indicate this is a minor change / improved release. Most companies increase the version number index to indicate those minor stuff and to indicate greater changes the (main) version number is increased. Maybe for these changes in the QII Software between 13.0 and 13.1 using the main Version increase Option (calling 13.1 to be 14.0) would have been the better indication (I think for similar changes ALTERA also increased the main version number in the past.) While there is no "common rule for Version numbering" my Feeling is that most of us would think of "Major changes = Version, minor changes = Version index"...

This makes no difference on the Software itself, but - to compare it with Windows - the 13.0 to 13.1 naming would be similar to calling Windows 8 just Windows 7.1, wouldn't it?

--- Quote Start ---

Most companies increase the version number index to indicate those minor stuff and to indicate greater changes the (main) version number is increased.

--- Quote End ---

Hi Carl, I didn't come up with this numbering scheme :) However, I looked around and it seems to be fairly common in the EDA business, although some other companies use the full year (eg "2013.2") which might be a bit more clear. My sense though is that there aren't many customers who make the decision based on this and there's probably enough infrastructure in place that assumes the current scheme that it would be a bit annoying to change!

Hi,

maybe this Version - Subindex Interpretation I gave is influenced by the requirement management tools I have to deal with. Comparing this with baselines (which "freeze" the actual state as a recovery Point or Version history), there is normally the Option to either increase the Version number or just the subindex to diverse between Major changes or just formal corrections / updates :-)

Maybe I'm wrong but I think there were studies which version number is rated as "most trustworthy" by customers - the outcome was - IIRC - that low numbers are less trustworthy as they indicate the Software to be in an early development stage and maybe prone to bugs while high numbers indicate that the Software was modified many times already and is either still buggy (even being modified many times) or quite old.. But this really leaves the technical sector and goes into psychology and/or mystery :-D

BTW: To prevent a lot of tools refusing to be installed due to false main Version number, win8 reports to be 6.3 (Win7 was 6.2) ...