I think I can confirm that V4 silicon is breaking the 1GHz barrier in the main FPGA fabric, as I did measure clock actual speeds of 952MHz in lowest speed grade V4-LX25. Faster speed grades would defenetly run over 1GHz. Sure there isnt much that can be done at those internal clock rates, but at least some flip-flops can toggle at that rate. What is already pretty amazing.
FAST! but be aware not all internal circuitry speeds are scaled similarly in V4 - LUT4 propagation delay as example is MUCH faster than in S3, but the minimum clock low/high times are almost the same as in S3.
Well, the secret is out. Yes, V4 is darn fast. We have seen the DSP48 filters running at > 1 GHz clock rate (and working fine). Even some amount of fabric runs this fast. Don't expect everything to work together at those rates, however (e.g. DCM, PPC, EMAC...).
I would suppose that for some real hotshots out there, who wanted to get the fastest possible performance, they could design, characterize, and test an application running in these stratospheric ranges.
But, they would be on their own, as they would have to closely regulate the temperature and voltage to keep these working (we just do not verify, characterize, nor specify it there).
I prefer to use my Intel Pentium IV where is was designed to work, and not overclock it. In the same way, you can get really impressive performance from the V4, but we don't test it there.
Similar, are folks who want really low power, and run V4 at 1.0 volts (+/- 5%). They get 44% less dynamic power, and less than 1/2 the static power, but it runs probably at 1/3 the max clock rates.
How many folks out there would like to buy a super low power version of V4? We could test to a special screen program for 1 V operation. Let me know...if no one wants it, we are unlikey to do anything. But if there is business out there, we'd like to be your supplier of choice!
In my past lives, I have designed product that was used where the manufacturer didn't specify, and yes, I was on my own, but with careful engineering, the products worked great.
I don't see that much anymore, as engineers have become "risk adverse" in many areas (as they have become very "jobless adverse").
Of course, no risk, no reward. So, those who never take any risks are most likely to also never really get anywhere.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.