A 500 MHz 4 Mb CMOS pipeline-burst cache SRAM with point-to-point noise reduction coding I/O

A secondary cache SRAM is an indispensable CPU partner in a high-performance system. The main objectives are: 1) pipeline burst operation; 2) 32b 500MHz (2GB/s) I/Os, and 3) point-to-point communication with a CPU, as well as shortened latency and reduced noise and power caused by high-speed, high-bandwidth I/O operation. A pre-fetched pipeline scheme enables the cycle time for an internal memory core (I-cycle) to be extended by N times that of an external bus cycle (E-cycle). This is modified to an SRAM to achieve both 4b pipeline-burst cache operation and 500MHz I/O frequency. In this case, I-cycle time of 8ns is four times E-cycle time (2ns).

[1]  Kazuyuki Nakamura,et al.  PLL timing design techniques for large-scale, high-speed, low-power, and low-cost SRAMs , 1994, Proceedings of IEEE Custom Integrated Circuits Conference - CICC '94.

[2]  K. Hose,et al.  A 200 MHz 256 kB second level cache with 1.6 GB/s data bandwidth , 1996, 1996 IEEE International Solid-State Circuits Conference. Digest of TEchnical Papers, ISSCC.

[3]  Frederick R. Eirich,et al.  High speed testing , 1960 .

[4]  F. Matsuoka,et al.  A 400 MHz 4.5 Mb synchronous BiCMOS SRAM with alternating bit-line loads , 1996, 1996 IEEE International Solid-State Circuits Conference. Digest of TEchnical Papers, ISSCC.

[5]  Jeffrey J. Tabor.  Noise Reduction Using Low Weight and Constant Weight Coding Techniques , 1990 .

[6]  H. Pilo,et al.  A 300 MHz, 3.3 V 1 Mb SRAM fabricated in a 0.5 /spl mu/m CMOS process , 1996, 1996 IEEE International Solid-State Circuits Conference. Digest of TEchnical Papers, ISSCC.

[7]  Mircea R. Stan,et al.  Bus-invert coding for low-power I/O , 1995, IEEE Trans. Very Large Scale Integr. Syst..

[8]  H. B. Bakoglu,et al.  Circuits, interconnections, and packaging for VLSI , 1990 .

[9]  M.A. Horowitz,et al.  A 50% noise reduction interface using low-weight coding , 1996, 1996 Symposium on VLSI Circuits. Digest of Technical Papers.