Actel dropped ARM7, when comes Xilinx ARM enabled silicon?

Started by Antti September 19, 2009
Actel has silently dropped all M7(ARM7 enabled) silicon.
They have not announced this as big news, but true it is, there are no
more M7 devices in product tables.

Years ago when Actel announced M7 i wanted to get some of them, but
Actel lied about the licensing, saying that there is no license fee,
actual story was that the M7 silicon (what is plain FPGA + AES key)
did cost maybe 100$ more then the same silicon without the key. And
even so it was not possible to obtain it. So well now it is dropped
for ever, no chances to get hold on it any more (unless some disti has
overleft stock?).

Another thing Actel did not advertize is that M7 silicon is NOT
SECURE, the AES key that is programmed to the FPGA silicon is also
used by Actel software tools, and hence can be easily extracted from
the Actel software. This is possible, and verified, a remote site did
this research for me on my request and they did deliver the "secret"
M7 AES key withing 23 hours. The key was verified to really be the M7
key from Actel ARM7 enabled devices by me.

So 3 FPGA companies have dropped ARM7:
1) Triscend (was killed by Xilinx)
2) Altera silicon discontinued (now offering Cortex-M3 soft-core)
3) Actel dropping M7

so the only FPGA company actually offering ARM could be xilinx with
their plans for Spartan-6 extended family?

Antti






On Sep 19, 5:26=A0pm, Antti <antti.luk...@googlemail.com> wrote:
> Actel has silently dropped all M7(ARM7 enabled) silicon. > They have not announced this as big news, but true it is, there are no > more M7 devices in product tables. > > Years ago when Actel announced M7 i wanted to get some of them, but > Actel lied about the licensing, saying that there is no license fee, > actual story was that the M7 silicon (what is plain FPGA + AES key) > did cost maybe 100$ more then the same silicon without the key. And > even so it was not possible to obtain it. So well now it is dropped > for ever, no chances to get hold on it any more (unless some disti has > overleft stock?).
An interesting move, and it suggests the License costs were exceeding the Silicon revenue ?!
> Another thing Actel did not advertize is that M7 silicon is NOT > SECURE, the AES key that is programmed to the FPGA silicon is also > used by Actel software tools, and hence can be easily extracted from > the Actel software. This is possible, and verified, a remote site did > this research for me on my request and they did deliver the "secret" > M7 AES key withing 23 hours. The key was verified to really be the M7 > key from Actel ARM7 enabled devices by me.
So that meant product creepage was easy? - maybe another reason to pull the plug ?
> > So 3 FPGA companies have dropped ARM7: > 1) Triscend (was killed by Xilinx) > 2) Altera silicon discontinued (now offering Cortex-M3 soft-core) > 3) Actel dropping M7
You forgot Atmel's CAP series, which is probably the smartest way to add ARM. For small volumes, having ARM is unimportant, so Nios or Microblaze or Power PC is fine, for larger volumes FPGAs are too expensive/ powerhogs so Atmel's pathway is better there.
> so the only FPGA company actually offering ARM could be xilinx with > their plans for Spartan-6 extended family?
Plans are cheap.. ;) Cypress have plans for a Cortex with a moderate CPLD - does that count ? :) Another important point is not the core, but the memory: An ARM core in a FPGA Still needs memory, so it's only a mirage-solution, like stone-soup. Plenty of CHEAP Arm+Flash offerings out there, called Microntollers, so a designer could find a much smaller FPGA can go alongside a decent microcontroller. High Speed USB is also appearing on these devices. -jg
On Sep 19, 9:02=A0am, malcolm <malcolm...@gmail.com> wrote:
> On Sep 19, 5:26=A0pm, Antti <antti.luk...@googlemail.com> wrote: > > > Actel has silently dropped all M7(ARM7 enabled) silicon. > > They have not announced this as big news, but true it is, there are no > > more M7 devices in product tables. > > > Years ago when Actel announced M7 i wanted to get some of them, but > > Actel lied about the licensing, saying that there is no license fee, > > actual story was that the M7 silicon (what is plain FPGA + AES key) > > did cost maybe 100$ more then the same silicon without the key. And > > even so it was not possible to obtain it. So well now it is dropped > > for ever, no chances to get hold on it any more (unless some disti has > > overleft stock?). > > =A0An interesting move, and it suggests the License costs were exceeding > the Silicon revenue ?! > > > Another thing Actel did not advertize is that M7 silicon is NOT > > SECURE, the AES key that is programmed to the FPGA silicon is also > > used by Actel software tools, and hence can be easily extracted from > > the Actel software. This is possible, and verified, a remote site did > > this research for me on my request and they did deliver the "secret" > > M7 AES key withing 23 hours. The key was verified to really be the M7 > > key from Actel ARM7 enabled devices by me. > > So that meant product creepage was easy? - maybe another reason to > pull the plug ? > > > > > So 3 FPGA companies have dropped ARM7: > > 1) Triscend (was killed by Xilinx) > > 2) Altera silicon discontinued (now offering Cortex-M3 soft-core) > > 3) Actel dropping M7 > > You forgot Atmel's CAP series, which is probably the smartest way to > add ARM. For small volumes, having ARM is unimportant, so Nios or > Microblaze
no i did not, its structured ASIC, not FPGA so it does NOT count
> or Power PC is fine, for larger volumes FPGAs are too expensive/ > powerhogs so Atmel's pathway is better there.
PowerPC is dead no longer offered in V-6 so its dead end path already
> > > so the only FPGA company actually offering ARM could be xilinx with > > their plans for Spartan-6 extended family? > > =A0Plans are cheap.. ;) > > =A0Cypress have plans for a Cortex with a moderate CPLD - does that > count ? :) >
no, its Cortex-M3 not ARM7
> =A0Another important point is not the core, but the memory: An ARM core > in a FPGA Still needs memory, so it's only a mirage-solution, like > stone-soup. > > Plenty of CHEAP Arm+Flash offerings out there, called Microntollers, > so a designer could find a much smaller FPGA can go alongside a decent > microcontroller. > High Speed USB is also appearing on these devices. > > -jg
where can i find info about "Microntollers" ? is it something like transputers or XMOS? and who are you, you signed as -jg what would be Jim Granville, but your account says Malcolm? Antti Lukats aka Antti Lukats
On Sep 20, 12:34=A0am, "Antti.Luk...@googlemail.com"
<antti.luk...@googlemail.com> wrote:
> and who are you, you signed as > -jg > > what would be Jim Granville, but your account says Malcolm?
oops, yes Antti, - did not spot that my son had re-logged in... -jg
On Sep 20, 12:34=A0am, "Antti.Luk...@googlemail.com"
<antti.luk...@googlemail.com> wrote:
> > =A0Cypress have plans for a Cortex with a moderate CPLD - does that > > count ? :) > > no, its Cortex-M3 not ARM7
Oh, If you are making that distinction, then ARM7 is close to NFND. So no surprise there. Cortex M3 and Cortex M0 will be the alternative FPGA-cores, on that pathway. -jg
On Sep 20, 12:31=A0am, -jg <jim.granvi...@gmail.com> wrote:
> On Sep 20, 12:34=A0am, "Antti.Luk...@googlemail.com"<antti.luk...@googlem=
ail.com> wrote:
> > > =A0Cypress have plans for a Cortex with a moderate CPLD - does that > > > count ? :) > > > no, its Cortex-M3 not ARM7 > > Oh, If you are making that distinction, then ARM7 is close to NFND. > > So no surprise there. > Cortex M3 and Cortex M0 will be the alternative FPGA-cores, on that > pathway. > > -jg
actually M1 only I think M0 is ASIC only, it would be larger than M1 M3 is much larger then M1 so as of soft core only M1 is feasible, not M0 I have compiled M3 and M1 RTL for different FPGA and M1 is small :) as of ARM7 close to NFND, you right I didnt think of that loud, but i guess it was in my mind too Antti