Startup Slashes SRAM Power With Standard Logic Process
Nick Flaherty, EETimes
3/4/2014 03:30 PM EST
A UK startup has developed an entirely new way of building static memory on a standard logic process that cuts the power in half without paying a penalty in area or speed.
"Fundamentally memory has some real problems everyone does it the same way," said Mark de Souza, chief executive at Silicon Basis in Bristol and formerly at the memory IP supplier Virage Logic. "They all take the TSMC arrays and put them together."
Silicon Basis sees two key advantages for its technology: It can save up to 50% of the power consumption, and it can go below the bit cell voltage of the foundry memories. This allows the memories to be powered by the same voltage source as the logic and so eliminates the need for a second DC-DC converter. It also makes the technology foundry independent and scalable to new technologies such as FinFet, says de Souza. The company, based in the SETsquared center in Brunel's EngineShed in Bristol, has produced all the models needed and is now working on 28nm silicon to prove the implementation.
E-mail This Article | Printer-Friendly Page |
|
Related News
- MoSys' 1T-SRAM Memory Silicon-Verified on DongbuAnam's 0.18-Micron Standard Logic Process; 0.13-Micron Verifications Initiated
- MoSys' 1T-SRAM-Q Memory Silicon-Verified on Chartered's 0.13-Micron Industry Standard Logic Process
- MoSys' 1T-SRAM Memory Silicon-Verified on SMIC's 0.18-Micron Standard Logic Process
- MoSys' 1T-SRAM-R Memory is Silicon-Proven On UMC's 0.13 Micron Logic Process
- Spectral releases Silicon proven High Speed Low Power SRAM compilers in the 40/45nm CMOS/RFSOI process nodes targeted for a wide range of IOT & 5G Applications
Breaking News
- Jury is out in the Arm vs Qualcomm trial
- Ceva Seeks To Exploit Synergies in Portfolio with Nano NPU
- Synopsys Responds to U.K. Competition and Markets Authority's Phase 1 Announcement Regarding Ansys Acquisition
- Alphawave Semi Scales UCIe™ to 64 Gbps Enabling >20 Tbps/mm Bandwidth Density for Die-to-Die Chiplet Connectivity
- RaiderChip Hardware NPU adds Falcon-3 LLM to its supported AI models