I've recently visited a Microsemi/Actel tutorial about the SmartFusion2.
Tenor about security is: "ProASIC3 is about 10 years old and hacking capabilities led to some vulnerability. Thus we've improved security - making the SmartFusion2 (and its - hardcore-less - descendants) fit for the next decade." In terms of security: nothing lasts forever.
About the omission of analog interfaces: Integrating fast digital logic and flash on the same chip is sufficiently difficult. Adding analog circuitry doesn't exactly simplify the task. "Universally" applicable analog circuitry is even more difficult. My best guess is that - alltogether - the analog front-end proved more cumbersome than advantageous. Thus the omission. IMHO no big loss but maybe a plus on the bill side.
@Tom: ...the real fruit of this union may be yet to come.
One of the big things here is that Microsemi supply a lot of discrete analog parts -- one of the things they can do now is to create reference designs that combine the Actel FPGAs with the Microsemi discrete parts -- when they are all in a referehnce design with part numbers etc., there's a much stringer incentive for the end user to simply say "what the heck -- I'll take the lot" :-)
Yes there were a flurry of reports and denials about backdoors into devices.
I could see that for the vendor it would be useful to have a secret way to access and test devices, see http://www.eetimes.com/document.asp?doc_id=1266567
This includes the statement: "Microsemi acknowledged that there is a privileged internal test facility that is typically used for initial factory testing and failure analysis but said that this feature is disabled on all shipped devices".
There may be arguments for enhanced security on FLASH devices, if it weren't for the back doors that researchers recently found.. Using sophisticated correlation techniques, the researchers found that Actel/Microsemi had fabricated backdoor registers and passwords for their FPGAs, allowing access to "secured" devices in the field. One can only imagine the reasons to do this - but in any case - can we really trust ANY silicon manufacturer with security?
I remember that it raised quite a few eyebrows when this acqusition happened. I also recall that the original SmartFusion was a big deal and made a lot of waves. I guess it was indeed inevitable that Microsemi would refocus SmartFusion on security, but I think it's kind of a bummer. I remember how excited the guys (and gals) at Actel were about it when they launched it.
Yes the acquisition was perhaps surprising for us in the programmable logic arena, but logical for Microsemi. Their quarterly revenue reported in April was $235M total, so it looks to be working for them.
My slight regret is that the initiative of pioneering what I consider to be the first "SoC" device that combined FPGA, processor and analog, has not been extended.
Microsemi also got access to the Rad Hard FPGAs from Actel to support their military and aerospace customers, of course. Xilinx has offered Rad Hard competition for many years, but has moved significantly ahead with a Virtex-5 device with impressive capabilities.
Acquisitions happen in the chip market in large part so that Company A can capitalize on the advances of Company B. But as noted in this very nice blog, it's not easy to merge cultures and processes. I think it may take a while to get revenue to jump, but surely that is a longer term goal. Don't think that the process is complete at this point; the real fruit of this union may be yet to come.
@Paul: I was really impressed with the SmartFusion devices with their mix of hard ARM Cortex-M3 MUC, programmable analog, and programmable digital.
As you say, the SmartFusion2 devices have a lot of enhancements, especially in the area of secure designs and such, but I have to say that I was surprised and sorry to see them lose their analog component (sad face)
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...