synthesis tools

Yes. They have better quality of results and less bugs. Well, Synplify does anyway.

No doubt they will be gutted.

Depends on the design.

Synplify Pro.

Cheers, Jon

I would not buy special synthesis tools until I found a problem with the vendor tools. I find that good simulation tools get the most use and provide the most value during design. I need synthesis to check fit and fmax, but edit, sim, edit, sim,... is the tight loop.

I know of one case where a Mentor tool fit a design that wouldn't fit with ise. I know of no such cases with quartus, but I don't doubt that they exist.

-- Mike Treseler

Hi,

For FPGA's I'd rate Synopsys DC -FPGA/ Synopsys FPGA compiler and Synplify Pro. These tools are really goor and provide excellent synthesis results. The Vendor provided tools are also okay, in terms of the IP cores supplied and other P&R options. Synopsys DC has built- in Design Ware products that are good when implementing complex designs. In terms of compilation time and speed, both these tools are equally good.

I've used the Mentor Graphics tools for FPGA, namely, Precision Synthesis, Leonardo Spectrum. They're okay, but I didnt get good synthesis results from them. Regarding Xilinx and Altera, I prefer Xilinx coz the ISE pack is good in terms of features.

Regards, Abraham

If you are looking for a synthesis tool to enable ASIC prototyping on an FPGA then you probably have to go for the more expensive tools like Precision and Synplify. Precision (and I assume Synplify) support all sorts of gated clocks/clock muxing/ripple counter, SDC constraints, Synopsys DesignWare and some SystemVerilog.

That all depends on the design but in general you can say that for large designs the more expensive tools have the upper hand. They also provide more control, better inferring capability (memory/DSP) and better debug capability (e.g. analysing critical path).

Impossible to say, simply get as many designs as you can find, evaluation licenses for Precision and Synplicity and try them out. You should also investigate what each tool can do in terms of constraining, language support, inferring capability, ASIC prototyping (if you need it), debugging, level of support (simply log a bug/enhancement and see how quick they responds), number of updates etc etc.

Due to the level of R&D invested in these product (including XST/QIS) I suspect that you have to repeat this process every year! I would also suggest you ignore any posting saying that tool A is better than tool B unless the person can back it up with an extensive up2date benchmarks :-)

Hans

Unless you are prototyping an ASIC (and want portability of the code to ASIC tools), I would not recommend Synopsys for FPGA synthesis. Their support of VHDL is hideous (they still have bugs handling direct entity instantiations, even in their simulator, and that's a '93 feature, 14 years old!). They cannot infer memory from arrays. So things like records, arrays, integers, booleans, and enumerated state types must be manually converted to SLV before being stored in instantiated RAMs. And just try debugging your design in the simulator when your data is spread out bitwise among a host of memory primitives, instead of neatly stored in an array.

Synplify Pro and Precision are very good products that generally give better QOR, and have better VHDL language coverage than the FPGA vendors' synthesis tools.

Andy

Thanks for the insight everybody. I will definitly do a more thorough analysis of all the tools available if needed. For now we use XST where I work and it seems to work fine. But in the future as designs get bigger, you never know, we might need to migrate to a tool that provide better debuggin capabilities. Thanks a lot, Amish

Which tool are you talking about? DC, FPGA Compiler, FPGA Compiler II, or DC-FPGA? The latest Synopsys tool in the FPGA-camp was DC-FPGA, but they terminated the product long time ago. A bit pity I think because it was great to have an ASIC flow acompatible FPGA synthesis tool.

Petter

The last time I tried (i.e. was told) to use it was a couple of years ago (dc-fpga). Good riddance, as far as I'm concerned.

Thanks,

Andy