Intel Corporation

R Kenneth Hose

David P DiMarco

Hong Yun Tan

Harry Muljono

Anant S Deval

Javed Sabir Barkatullah

Jeffrey L Miller

Chris B Freeman

Anan S Deval

Described is an apparatus which comprises: a first node to provide an un-gated power supply; a second node to provide a threshold dependent supply; an inverter with an input and an output, the inverter coupled to the first and second nodes, the inverter to receive the un-gated power supply at its power supply node, and to receive the threshold dependent supply for supplying ground supply at its ground node; and a transistor with its gate terminal coupled to the output of the inverter, the transistor to provide gated power supply to one or more logic units.

Threshold voltage dependent power-gate driver

Implementing termination on a bus. According to one embodiment of the present invention a driver drives a default signal on to a line, then drives an information signal on to the line, and then drives the default signal on to the line after driving the information signal on to the line.

Implementing termination with a default signal on a bus line

A circuit contains a microprocessor die, containing a microprocessor, and a cache memory die, containing a cache memory, for operation in conjunction with the microprocessor. A microprocessor clock and a cache memory clock are generated for operation of the microprocessor and the cache memory, respectively. The microprocessor and cache memory clocks are generated on the microprocessor die, and the cache memory clock is transmitted to the cache memory die. In order to transmit data between the microprocessor die and the cache memory die, clock cycles are designated. The microprocessor clock and the cache memory clock are synchronized to the clock cycles including compensation for the propagation delay between the two dies. The microprocessor includes a stop clock function which halts the cache memory clock and the microprocessor clock on the same clock cycle so that data integrity, in both the microprocessor and cache memory, are maintained. In order to provide functional operation over a range of clock cycle frequencies, the data, from cache memory die, becomes valid on the falling edge of the cache clock signal, and is subsequently sampled, in the same clock cycle, on the rising edge of the microprocessor clock.

Method and apparatus for synchronizing clock signals in a multiple die

A CPU is coupled to the cache memory over a system bus having a width of 64 data bits. The cache memory is organized into a left array and a right array, with data bits stored as lines of data wherein each line is comprised of 256 data bits defined into four data 'chunks' of 64 bits each. Each memory read access by the CPU to the cache results in a complete line of data to be read in the cache. The chunks comprising the line of data are coupled over an internal cache bus to a 'chunk' multiplexor. The chunk multiplexor stages the data chunks in an order defined by the CPU, and sequentially send the data chunks over the system bus to the CPU. The chunks are organized as high and low order chunks. The multiplexor includes a first multiplexor for receiving the high order chunks and a second multiplexor for receiving the low order chunks. Latches are provided which are coupled to the first and second multiplexors to receive the high order chunks upon the receipt of a first clock signal, and the low order chunks upon the receipt of a second clock signal. Enabling signals are provided to the latches such that the high order and low order chunks are coupled to the system bus in a sequential order which is determined by the CPU.

Cache memory having a multiplexor assembly for ordering output on a data

An improved cache memory architecture is disclosed, having particular application in a cache having static random access memory (RAM). In a typical static RAM memory utilized as a cache, the cache has the requirement that it must access many more bits than is required for selection. A single wordline of the RAM may span an entire memory array, and the activation of the entire wordline results in many more bitlines activated than will actually be selected by the Y decoder. As a result, power is wasted. The present invention provides a cache memory in which even and odd columns are segregated, wherein the even addressed columns may be placed in a first set (0) and the odd addressed columns in a second set (1). The wordline decode includes two wordlines per row rather than the typical single wordline in prior art systems. The first wordline corresponds to the 'even' wordline, and the second wordline corresponds to the 'odd' wordline (set 1). Only one wordline is activated at any time to save power. The wordline decoder of the present invention utilizes an address bit (for example, the low order bit) to select either the driver for the columns corresponding to the even wordline or to the odd wordline. Although the present invention requires additional drivers, only one driver is activated at any one time. It has been found that the architecture of the present invention provides a total power savings in a read operation approaching fifty percent.

Growing community of inventors

Aloha, OR, United States of America

R Kenneth Hose