| ??? 12/01/09 17:16 Read: times |
#171353 - re: absolutely. Responding to: ???'s previous message |
Richard Erlacher said:
Andy Peters said:
Barry Demers said:
Since I am currently pursuing a logic design, Would ModelSim be a good option for working with and, or, not gates and being able to watch timing issues, etc? ModelSim is an HDL simulator. As such, you will need HDL models of whatever you want to simulate. And it's digital only -- no analog. It will do timing simulations, assuming you include the worst-case timing information in your models, or you back-annotate with an SDF file. But it won't do timing ANALYSIS, which is a different problem. So, if your logic design is going to live in an FPGA or CPLD, (and it probably will, it ends up being less expensive and more reliable than old-skool 74xxx-series TTL), While I have little doubt, myself, that this is the case, have you any authoritative source that can be cited to support this notion? Sheesh -- I can buy a Spartan 3AN device that holds as much logic as three or four VME-size boards full of 74-xxx DIP logic devices. The benefits here should be obvious. I'm not persuaded that multiple monitors are ABSOLUTELY necessary, having done what you describe via (A) multiple monitors on multiple computers, and (B) large single monitor on multiple computers, and (C) multiple windows on a single computer. Now, I understand it's a matter of preference choosing what to use. Perhaps monitor size makes a difference, too. It may be easier or more convenient for you, but, for example, my own preferences spring from the fact that I still find the non-windows-based app's I use every day far more productive than the Windows-based ones that do the same things, yet the multi-computer-single-monitor model allows me to display mixes of Windows and non-Windows app's without having to run a multi-OS scheme on my computer. That, of course, provides yet another alternative, but I haven't yet gone there. One of my colleagues has run DOS, Windows, and Linux on the same box with a single monitor, and with multiple monitors, and found that to be a bit "sticky." Now he uses the multi-computer model with a single large monitor, and finds that quite satisfactory. I got the notion from him after seeing how seamlessly it allows him to do what's necessary. After all, PC's cost a small fraction of what a decent-size monitor costs. Well, since I have no need to run obsolete DOS software, that aspect of your workflow is uninteresting to me. And if I needed to run such, I'd use VMWare and create a virtual DOS machine. I don't need junky computers cluttering up my work area. As for the two displays vs one -- it's all about screen real-estate. I can buy two 1920-column displays for less than half of what a single 2560-column display costs. As for using a second cheap machine (one that costs what a display costs), that simply doesn't work for what I do. Example: when I run ModelSim, I often have the waveform display window maximized in one monitor, and the rest of the app (tcl command line, signal list, etc) as well as editor windows open in the second display. Obviously ModelSim runs on only one machine, so there's no way to do what you do (use a KVM switch to toggle between different computers running different apps). And I like to see all of this stuff at the same time. Our layout guy keeps schematics open in one monitor and the PCB design open in the other (they are tightly coupled and you can click on a part in the layout and find it on the schematic, etc). And this only uses one license token; doing what you do (schematic on one computer, layout on the other) would require the use of two license tokens. Again, just because YOU don't see the utility of working in this manner, it doesn't mean that others think like you. -a |
