News & Analysis
Comment
Steve Knapp
Steve Knapp
Actually, SiliconBlue FPGAs also use static memory configuration cells, just ...
SiliconBlue samples 40-nm FPGAs for tablets
Dylan McGrath
7/11/2011 1:02 AM EDT
SAN FRANCISCO—Programmable logic startup SiliconBlue Technologies Ltd. Monday (July 11) announced it is sampling its next-generation 40-nm FPGAs for smartphones and media tablets.
SiliconBlue (Santa Clara, Calif.) said its iCE40 family of mobile FPGAs, codenamed "Los Angeles," includes a low-power series (LP) targeting smartphones and a high-speed series (HX) targeting tablets. The devices, built on Taiwan Semiconductor Manufacturing Co.'s 40-nm standard CMOS process, provide twice the logic capacity of SiliconBlue's previous generation 65-nm devices, the company said.
Kapil Shankar, CEO of SiliconBlue, said the company's devices are typically implemented in smartphones and tablets as companion chips to the applications processors provided by the likes of Qualcomm Inc., Nvidia Corp, Texas Instruments Inc. and others to provide additional functionality. With so many smartphones and tablets using the Android operating system and relying on silicon from the small group of companies marketing applications processors for the devices, adding a companion chip is essential for any vendor that wants to differentiate its offerings, he said.
"The software is becoming identical for everyone," Shankar said. "To differentiate, you have to add hardware features to the devices. We are the instrument to add differentiation."
FPGAs have traditionally been a hard sell for handheld consumer applications because of their size, cost and power consumption. But SiliconBlue claims to have overcome this perception with its single-chip custom mobile devices, which offer ultra low power consumption, small die size and low cost (iCE40 devices start at $1.99 per unit in high volumes).
SiliconBlue claims to have shipped 7 million devices and says it is on track to ship 10 million this year. The company also claims more than 250 active end customers, including Tier One customers like Samsung Electronics Co. Ltd. and Huawei Technologies Co. Ltd.
According to Shankar, SiliconBlue's custom mobile device FPGAs are found in 30 to 40 products already available on the market, including smartphones, cameras, personal media devices, e-books and even a watch made by Citizen. The company has to date focused on end customers in Asia, but is currently building out its sales force to target more customers in the U.S. and Europe, according to Shankar.
Last month, SiliconBlue closed $18 million in a fourth round of venture capital funding, bringing the company's total to nearly $74 million since it was founded in 2006 by Shankar, a programmable logic veteran and former high-ranking executive at market leader Xilinx Inc.
SiliconBlue (Santa Clara, Calif.) said its iCE40 family of mobile FPGAs, codenamed "Los Angeles," includes a low-power series (LP) targeting smartphones and a high-speed series (HX) targeting tablets. The devices, built on Taiwan Semiconductor Manufacturing Co.'s 40-nm standard CMOS process, provide twice the logic capacity of SiliconBlue's previous generation 65-nm devices, the company said.
Kapil Shankar, CEO of SiliconBlue, said the company's devices are typically implemented in smartphones and tablets as companion chips to the applications processors provided by the likes of Qualcomm Inc., Nvidia Corp, Texas Instruments Inc. and others to provide additional functionality. With so many smartphones and tablets using the Android operating system and relying on silicon from the small group of companies marketing applications processors for the devices, adding a companion chip is essential for any vendor that wants to differentiate its offerings, he said.
"The software is becoming identical for everyone," Shankar said. "To differentiate, you have to add hardware features to the devices. We are the instrument to add differentiation."
FPGAs have traditionally been a hard sell for handheld consumer applications because of their size, cost and power consumption. But SiliconBlue claims to have overcome this perception with its single-chip custom mobile devices, which offer ultra low power consumption, small die size and low cost (iCE40 devices start at $1.99 per unit in high volumes).
SiliconBlue claims to have shipped 7 million devices and says it is on track to ship 10 million this year. The company also claims more than 250 active end customers, including Tier One customers like Samsung Electronics Co. Ltd. and Huawei Technologies Co. Ltd.
According to Shankar, SiliconBlue's custom mobile device FPGAs are found in 30 to 40 products already available on the market, including smartphones, cameras, personal media devices, e-books and even a watch made by Citizen. The company has to date focused on end customers in Asia, but is currently building out its sales force to target more customers in the U.S. and Europe, according to Shankar.
Last month, SiliconBlue closed $18 million in a fourth round of venture capital funding, bringing the company's total to nearly $74 million since it was founded in 2006 by Shankar, a programmable logic veteran and former high-ranking executive at market leader Xilinx Inc.
|
|
|
|
|
Navigate to related information
GREAT-Terry
7/11/2011 9:49 PM EDT
I agree the phase "To differentiate, you have to add hardware features to the devices", but how this FPGA be different from the two big giants - Altera and Xilinx? What actually the FPGA help in improving the hardware features? I guess adding more sensors may also sound.
Sign in to Reply
Code Monkey
7/12/2011 2:16 PM EDT
SiBlue's FPGAs are essentially antifuse parts, while Altera and Xilinx are mostly SRAM parts. You can only program an antifuse once, but it's a very efficient way to configure FPGA cells.
Sign in to Reply
KB3001
7/14/2011 11:14 AM EDT
Are you sure their FPGA technology is one-time programmable?
Sign in to Reply
Steve Knapp
7/19/2011 3:09 PM EDT
Actually, SiliconBlue FPGAs also use static memory configuration cells, just like Xilinx and Altera. Consequently, SiliconBlue FPGAs are reprogrammable.
SiliconBlue FPGAs also have on-chip and *optional* Nonvolatile Configuration Memory (NVCM). The NVCM provides a low-cost and secure configuration option, essentially providing single-chip integration.
You can configure a SiliconBlue FPGA from an SPI Flash, download it from a processor, or use the on-chip NVCM. It's your choice.
Sign in to Reply
Patk0317
7/13/2011 4:30 PM EDT
Try to find an Altera or Xilinx device for $1.99. They are all about the high end. This is about the low end. You can sell a goodly number of chips into servers with 60% margins, or a gazillion into cell phones with much lower margins.
Sign in to Reply
Steve Knapp
7/19/2011 3:14 PM EDT
Whats the difference from the two FPGA giants?
How about standby and low-frequence active power consumption measured in MICROamps not tens of milliamps?
How about packaging measuring just a few millimeters square?
How about production prices of a few dollars or less?
Sign in to Reply
genovalente
7/15/2011 2:37 PM EDT
SiBlue is not antifuse; it is a CMOS FPGA with a non-volatile configuration memory section that allow the single package to boot itself. http://www.siliconbluetech.com/technology/NVCM.aspx
Nice packaging for mobile OEMs, very low power.. as long as the software works well, they have a really nice niche, I think.
Sign in to Reply