Free
Message: Re: Actel Powers Down – Again--echoes of doni

On static power, flash already has a significant advantage over its SRAM-based rivals. For an SRAM FPGA to stay alive, it needs to power a lot of SRAM cells just to hold its configuration – the state of the routing and look-up-table (LUT) truth tables. True flash-based FPGAs keep this same configuration information in flash cells which, of course, require no power to retain their state. The majority of static power in a programmable logic device is consumed by keeping the configuration intact, so flash-based FPGAs can have a couple orders of magnitude advantage in static power over even the thriftiest of SRAM-based devices.

If you read the above statement a few times, slowly, you will see, that the SRAM is not being utilized in the manner that it was previous.

"Static power in the microwatts is certainly better for batteries than the tens of milliwatts required for a typical SRAM-based device, so they’ve got a good point."

They have simply reduced the foot print of the SRAM, "Static power" is SRAM or static RAM. The code that would normally held there, is not being held there.

Prior to using SRAM to hold its configuration, they utilized SDRAM

To this point we have waited for the DRAM, SDRAM and now SRAM configurations normally utilized to hold the apps, to slide to the side.

The question is, what logic within the flash is allowing them to do this, utilizing a tiny foot print of SRAM?

""Another nice feature that Actel has provided in the new family is the “Flash*Freeze” capability, which allows you to cycle the power on and off almost instantly, maintaining state and keeping circuit-friendly I/O behavior. This makes putting your device in standby mode a snap and skips the time- and power-consuming reconfiguration required in an SRAM device when the power pipeline is shut down. ""

The above gets back to a post of mine some time back of organizing the flash/cache in a manner where you just put the cache in a dead state. All it's use for is a pass through to the processor.

It gets tiring seeing where the industry is going, and where we have already been with published patents.

gernb1, it's nice to see that you have become familiar with it all, thanks for your efforts.

doni


Jan 24, 2008 10:38AM

Jan 24, 2008 10:45AM

Jan 24, 2008 10:55AM

Jan 25, 2008 07:58AM

Jan 25, 2008 08:17AM

Jan 25, 2008 01:48PM
1
Jan 25, 2008 01:53PM

Jan 25, 2008 01:56PM

Jan 25, 2008 03:37PM

Jan 25, 2008 03:49PM

Jan 25, 2008 03:52PM

Jan 25, 2008 04:01PM

Jan 25, 2008 04:12PM
Share
New Message
Please login to post a reply