The agreement to put ARM cores onto Actel FPGAs is a long term
plan without significant financial impact until 2008, according to
Actel’s president and CEO John East, but then it will deliver
a big boost to Actel’s revenues and ARM’s unit
shipments.
src="../Assets/GetAsset.aspx?ItemID=397" vspace="4" /> |
|
| John East |
|
“We’ll work on ARM, and we’ll optimise it, and
make it go faster, and lower the cost and then, after a year and a
half, we’ll start winning designs with ARM. From there to
moving into production takes a year and a half. So I think
it’ll be three years before ARM shows up in the
financials.”
The choice of ARM as Actel’s 32-bit core was simple.
“ARM is the market leader in 32-bit microprocessors.
It’s got more 32-bit processors embedded than anyone else
– billions of them,” said East, “and, if they
want a programmable version they have to come to me because I have
the only programmable version of ARM.”
The reason why Actel is the only suitable FPGA for a soft ARM
core is that the normal SRAM FPGA architecture is insecure because
the chip has to be booted from a ROM and, at that point, the
netlist of the ARM is sent from the boot ROM to the FPGA. During
that transfer the netlist can be read.
Since Actel’s FPGAs, called ProASIC, are based on
non-volatile flash memory, rather than SRAM, they do not need that
boot process, and can be programmed at the factory so that the
embedded IP is secure.
“Our deal is we will give customers ARM - no charge - but
with one little caveat, the first three FPGAs with ARM, which we
give you, instead of charging $5, we’ll charge $500,”
adding, “I made that number up because we haven’t
announced that yet.”
www.actel.com