Cache memory utilizing pseudo static four transistor memory cell
AssigneeIntel Corporation
InventorMoty Mehalel
Life of this patent
Estimating the term…
Abstract
A four transistor memory cell having a pair of cross coupled transistors and a pair of pass gates is disclosed. The four transistor memory cell is refreshed by charge transfer between the bit lines the internal nodes during bit line precharge.
Claims9
A cache memory comprising:
- an array of memory cells arranged in rows and columns, each row having a word line and each column having a first bit line and a second bit line, each memory cell having a first transistor coupled between a first node and a first operating potential;
- a second transistor coupled between a second node and the first operating potential, the gates of the first and second transistors being coupled to the second and first nodes, respectively, such that the first node stores a data bit and the second node stores a complement of the data bit;
- a third transistor coupled between the first bit line and the first node, the gate of the third transistor being coupled to the word line to selectively couple the first node to the first bit line;
- and a fourth transistor coupled between the second bit line and the second node, the gate of the third transistor being coupled to the word line to selectively couple the second node to the second bit line;
- and
- row decoder means for overlapping a refresh cycle on a memory cell in a selected column and row of the array during a precharge operation on the memory cell.
A method of operating a cache memory which comprises an array of memory cells arranged in rows and columns, each row having a word line and each column having a first and second bit lines, each memory cell having a first transistor coupled between a first node and a first operating potential, and a second transistor coupled between a second node and the first operating potential, the gates of the first and second transistors being coupled to the second and first nodes, respectively, such that the first node stores a data bit and the second node stores a complement of the data bit, the method comprising the steps of:
- (a) deactivating a word line of a selected row in the array;
- (b) precharging first and second bit lines of a selected column in the array to a precharge voltage level, the selected row and column identifying a memory cell of the array;
- (c) activating the word line of the selected row so that charge flows to the first node of the memory cell from the first bit line of the selected column, thereby refreshing the first node;
- (d) deactivating the word line of the selected row.
A method of overlapping refresh and precharge operations between read/write accesses to a cache memory which comprises an array of memory cells arranged in rows and columns, each row having a word line and each column having a first and second bit lines, each memory cell having a first transistor coupled between a first node and a first operating potential, and a second transistor coupled between a second node and the first operating potential, the gates of the first and second transistors being coupled to the second and first nodes, respectively, such that the first node stores a data bit and the second node stores a complement of the data bit, the method comprising the steps of:
- (a) accessing a memory cell to read/write data on first and second bit lines of a selected column by activating a word line of a selected row in the array;
- (b) deactivating the word line;
- (c) providing a control pulse having a duration to a circuit which couples a high voltage to the first and second bit lines in response thereto;
- (d) activating the word line so that charge flows to the first node of the memory cell from the first bit line of the selected column, thereby refreshing the first node;
- (e) deactivating the word line prior to the end of the duration of the control pulse;
- (f) re-accessing the memory cell to read/write data.
Dependencies are inferred from the claim text.