Programming Non Volatile Memories

Information

  • Patent Application
  • 20100146189
  • Publication Number
    20100146189
  • Date Filed
    December 09, 2008
    15 years ago
  • Date Published
    June 10, 2010
    14 years ago
Abstract
Non volatile memories and methods of programming thereof are disclosed. In one embodiment, the method of programming a memory array includes receiving a series of data blocks, each data block having a number of bits that are to be programmed, determining the number of bits that are to be programmed in a first data block, determining the number of bits that are to be programmed in a second data block, and writing the first and the second data blocks into a memory array in parallel if the sum of the number of bits that are to be programmed in the first data block and the second data block is not greater than a maximum value.
Description
TECHNICAL FIELD

The present invention relates generally to semiconductor memories, and in particular embodiments to programming non volatile memories.


BACKGROUND

A non volatile memory array includes non volatile devices such as floating gate transistors. The floating gate transistors are programmed by storing a charge on the floating gate. The charge on the floating gate shifts the conductivity of the channel upon application of an activation voltage. The channel conductivity of the memory transistors (hence, the memory state) is detected by sensing the current flowing through the device.


The programming of memory cells is typically accomplished a word at a time but conventionally requires that the drain of selected cells be placed at e.g., six or seven volts, the gate at eleven or twelve volts, and the source at ground. This programming operation draws substantial source to drain current because the gate terminal is raised above the level of the drain and source terminals while a significant potential difference is placed between the drain and source terminals.


One way of improving the speed of the programming process is to program more devices in parallel. However, programming more transistors draws a larger current. The current to the transistors is provided by charge pump circuits due to the high voltage required for programming. Further, many elements of non volatile memories (such as gate dielectric) do not scale easily and limit any decrease in programming voltage.


Charge pump circuits provide a high voltage output from a lower voltage source. However, the maximum current that can be drawn from a charge pump is limited. An increase in total current output from the charge pump requires a corresponding increase in die area that is counter productive to the economics driving scaling.


Hence, what are needed in the art are improved circuits and ways of programming semiconductor memories.


SUMMARY OF THE INVENTION

These and other problems are generally solved or circumvented, and technical advantages are generally achieved, by preferred embodiments of the present invention.


Embodiments of the invention include programming non volatile memories. In accordance with a preferred embodiment of the present invention, the method of programming a memory array comprises receiving a series of data blocks, each data block having a number of bits that are to be programmed, determining the number of bits that are to be programmed in a first data block, determining the number of bits that are to be programmed in a second data block, and writing the first and the second data blocks into a memory array in parallel if the sum of the number of bits that are to be programmed in the first data block and the second data block is not greater than a maximum value.


The foregoing has outlined rather broadly the features of an embodiment of the present invention in order that the detailed description of the invention that follows may be better understood. Additional features and advantages of embodiments of the invention will be described hereinafter, which form the subject of the claims of the invention. It should be appreciated by those skilled in the art that the conception and specific embodiments disclosed may be readily utilized as a basis for modifying or designing other structures or processes for carrying out the same purposes of the present invention. It should also be realized by those skilled in the art that such equivalent constructions do not depart from the spirit and scope of the invention as set forth in the appended claims.





BRIEF DESCRIPTION OF THE DRAWINGS

For a more complete understanding of the present invention, and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawing, in which:



FIG. 1, which includes FIGS. 1a and 1b, illustrates a write algorithm wherein the data blocks are summed over two adjacent data blocks, in accordance with an embodiment of the invention;



FIG. 2, which includes FIGS. 2a and 2b, illustrates a write algorithm wherein the data blocks are summed over any two adjacent data blocks to minimize the number of write operations, in accordance with an embodiment of the invention;



FIG. 3, which includes FIGS. 3a and 3b, illustrates a write algorithm wherein the data blocks are summed over two or three adjacent data blocks to minimize the number of write operations, in accordance with an embodiment of the invention;



FIG. 4 illustrates a write algorithm wherein the data blocks are summed over two adjacent data blocks and inverse data blocks, in accordance with an embodiment of the invention;



FIG. 5, which includes FIGS. 5a and 5b, illustrates a write algorithm wherein the data blocks are summed over two or three adjacent data blocks and inverse data blocks, in accordance with an embodiment of the invention;



FIG. 6 illustrates a write algorithm wherein the data blocks are summed over any two or any three adjacent data blocks and inverse data blocks, in accordance with an embodiment of the invention;



FIG. 7 illustrates a write algorithm wherein the data blocks are summed over adjacent data blocks, wherein the total programming current is reduced without degrading the writing efficiency, in accordance with an embodiment of the invention;



FIG. 8 illustrates a write algorithm wherein the data blocks are summed over any two or any three data blocks, in accordance with an embodiment of the invention; and



FIG. 9 describes an embodiment of the invention illustrating a block diagram of a memory circuit.





Corresponding numerals and symbols in the different figures generally refer to corresponding parts unless otherwise indicated. The figures are drawn to clearly illustrate the relevant aspects of the embodiments and are not necessarily drawn to scale.


DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS

The making and using of the presently preferred embodiments are discussed in detail below. It should be appreciated, however, that the present invention provides many applicable inventive concepts that can be embodied in a wide variety of specific contexts. The specific embodiments discussed are merely illustrative of specific ways to make and use the invention, and do not limit the scope of the invention.


The present invention will be described with respect to preferred embodiments in a specific context, namely programming non volatile memories. The invention may also be applied, however, to other operations as well as to other devices such as volatile memories.


In many non volatile memories, the number of bits (e.g., the number of “1”s) that may be programmed simultaneously is limited by the maximum available current. In particular, non volatile memories that are programmed with hot electron injection or hot hole injection require relatively high currents to generate these hot carriers. Examples of non volatile memories that require high currents include conventional flash devices as well as emerging memories such as phase change memories, MRAM, and CBRAM, where in some cases the number of memory cells with a toggling content defines the consumed current. The maximum available current is limited by the application, for example, hand held devices with limited battery have very limited available current. Similarly, size of the charge pump limits the available current. An increase in the maximum number of simultaneously programmable bits (nmax) requires an appropriate increase in the area of the charge pump.


When the total available current is Imax, and the current needed for a single write operation is Is, the maximum number of bits that can be programmed in parallel comprises Imax/Is. For example, if every written bit requires 100 μA and an overall current of 1 mA is available, not more than 10 bits can be written into the non volatile memories in parallel. Typically, when a certain amount of data bits (x) is written, blocks of nmax bit data are written into the non volatile memories in parallel until all the x bits are written independent of the number of bits that are to be programmed (e.g. number of “1”s) within these bits. Hence, the write sequence is selected for the worst case so that when all the bits in the data bits (x) comprise bits that are to be programmed, then the writing is performed at 100% efficiency. However, in most applications in average, only half the data bits comprise bits that are to be programmed, and hence only half the available current is used. In various embodiments, the present invention overcomes these limitations and writes at a higher efficiency and effectively uses the available programming current.


In a different approach, a variable number of data bits are written simultaneously so that the available current is used effectively. If the number of data bits written is selected to completely use all the programming current, an efficiency of 100% is possible. However, this approach has many drawbacks. For example, the write time is not constant but rather depends on the data, for example, the number of bits that are to be programmed in the data. Both the software and the hardware required to implement such an algorithm is quite complex and prohibitive because the number of cases to be handled is equal to the number of possible widths. In various embodiments, these limitations are overcome by preselecting data widths (or transistors) that are programmed in parallel. For example, only a certain numbers of transistors can be programmed in parallel. Examples include programming data widths (or transistors) of q, 2q, 3q, 4q, in one embodiment, wherein q is the smallest preselected data width that is programmed. Alternately, data widths of q, 1.5q, 2q are chosen in another embodiment. The larger data width is selected from the preselected data widths within the constraint that the current drawn by all the transistors is less than or equal to the maximum current achievable from the charge pump.


An embodiment of the invention of improving programming efficiency will be described with respect to FIG. 1, wherein the data is written using only two sizes, the larger one being double the smaller one. Alternate embodiments will then be described using the table shown in FIGS. 2-7. A block diagram of a memory circuit used in various embodiments will be described using FIG. 8.



FIG. 1, which includes FIGS. 1a and 1b, illustrates an embodiment of the invention improving the write performance efficiency without a significant increase in hardware complexity. FIG. 1a illustrates a table, whereas FIG. 1b illustrates a flow chart of the programming method.


As illustrated in the table of FIG. 1a, an arbitrary array of data block sequence (data to be written) comprising 16 data blocks is shown. Each data block is comprised of, for example, 8 data bits. Each data bit can be either a “1” or a “0”. A data bit that is to be programmed comprises a value “1.” Hence as described above, the number of “1”s limits the number of data bits that can be programmed in parallel. The first row is a data block sequence to be programmed, and the second row illustrates the number of data bits that are to be programmed (nr) in the data sequence. Hence, the second column of the second row shows a value of 4 for the first data block corresponding to the number of data bits that are to be programmed. The third row illustrates the number of bits that are to be programmed in adjacent data blocks, for example, first and second data blocks, and then third and fourth data blocks (n1+n2, n3+n4, . . . nr+nr+1). Hence, the third row comprises half the number of data blocks as the data to be written (8 data blocks in this case). Hence, if the maximum programmable data bits nmax (data bits that can be programmed in parallel) is 10, the third row illustrates that the first two data blocks can be written simultaneously. If the number of data bits in each element of the third row is not greater than the maximum programmable data bits nmax, then a 100% increase in efficiency is achieved. However, if an element in the third row is more than the maximum programmable data bits nmax, the data bits are written as separate blocks.


The data blocks written together are illustrated in the fourth row for the data bit sequence illustrated in the first row of FIG. 1a. Hence, the first and second data blocks are written together whereas the third and fourth data blocks are written separately. The data sequence is hence written as eleven data blocks using this approach. Hence, a 45% increase in efficiency is obtained for this particular data block sequence. When on an average only half the data bits are “1s,” the average efficiency increase obtained for random data using this algorithm is about 81%. Hence, about 20% of the time, the data blocks can not be combined.


Although illustrated with summing up two adjacent blocks, in various embodiments more adjacent data blocks may be summed together. For example, in one embodiment, both adjacent two adjacent blocks (e.g., nr+nr+1) as well three adjacent blocks (e.g., nr+nr+1+nr+2) can be summed. If the number of bits that are to be programmed in the three adjacent blocks is not greater than the maximum programmable bits nmax, then the three adjacent blocks are written together. For example, in FIG. 1, the data blocks in columns thirteen, fourteen and fifteen can be programmed together, resulting in an overall increase in efficiency of 60% for this example. Similarly, in another embodiment, upto four blocks of data can be combined. Hence, for example, the data blocks in columns thirteen, fourteen, fifteen, and sixteen can be programmed together, resulting in an overall increase in efficiency of 78% for this example.



FIG. 2, which includes FIGS. 2a and 2b, illustrates an embodiment of the invention for further optimization of the write sequence. FIG. 2a illustrates a table as in FIG. 1a, whereas FIG. 2b illustrates a flow chart. As in the prior embodiment, the first row lists the data sequence, the second row lists the number of bits that are to be programmed, and the third row lists the number of bits that are to be programmed summed over adjacent rows. As described above, the third row comprises bits that are to be programmed in adjacent data blocks, (for example, n1+n2, n3+n4, . . . nr+nr+1). Unlike the prior embodiment, a further row (fourth row) is added that sums over adjacent blocks in the second and third data blocks, (for example, n2+n3, n4+n5, . . . nr+1+nr+2).


The fifth row, which corresponds to the fourth row in FIG. 1a, illustrates the data blocks used to write the data sequence. Similar to FIG. 1a, the first column includes the first and the second data block. The second column also similar to FIG. 1a uses the third data block. However, the third column unlike the prior embodiment includes both the fourth and fifth data blocks. This is because as illustrated in the fourth row, the fourth and the fifth data blocks can be combined to less than the maximum programmable bits nmax. Hence, using this embodiment further efficiency gain is obtained. For example, for the data sequence shown in the first row, the data can be programmed as ten data blocks instead of the eleven in the prior embodiment. Hence, an improvement of 60% is obtained. When on an average only half the data bits are “1s,” the average efficiency increase for random data obtained using this algorithm is about 83%. This embodiment may also include addition over multiple data blocks (e.g., nr+1+nr+2+nr+3) to improve efficiency further as described above.



FIG. 3
a illustrates a write algorithm wherein the data blocks are summed over two or three adjacent data blocks to minimize the number of write operations, in accordance with an embodiment of the invention. As in the prior embodiment of FIG. 2a, the first row lists the data sequence, the second row lists the number of bits that are to be programmed, and the third row lists the number of bits that are to be programmed summed over adjacent data blocks. The fourth row also lists the summation over two adjacent data blocks as described in FIG. 2a. Unlike the prior embodiment, the fifth, sixth and seventh rows list the summation over three adjacent blocks, for example, the fifth row lists summation over nr+nr+1+nr+2, the sixth row lists summation over nr+1+nr+2+nr+3, and the seventh row lists summation over nr−1+nr+nr+1. Hence, as listed in the eighth row, the two or three adjacent data blocks are selected to minimize the number of write operations. For example, the twelfth, thirteenth and fourteenth data blocks are selected together and written in parallel. As shown in the table, the data sequence shown is processed in nine steps with an overall efficiency gain of about 78% for this example.



FIG. 4 illustrates an embodiment that writes data as either as “1” as described above, or as inverse data, wherein “1”s are stored as “0”s and “0”s are stored as “1”s. Storing data in inverse can save write operations if the number of “0”s are lower than the number of “1”s. However, the data block must include an additional bit to identify the data type as either inverse or not. Although there is a loss in storage efficiency to include this additional bit, this can be offset by the gains obtained by using such an algorithm, especially for particular types of data.


Referring to FIG. 4, additional rows are added to FIG. 1a to show the number of inverse data bits that need to be programmed in a single data block (ur) (fourth row) and the number of inverse data bits in adjacent columns (fifth row). For example, in the fifth row, both adjacent two adjacent blocks (e.g., ur+ur+1) are summed. If the inverse bits that are to be programmed in the adjacent blocks is less than the maximum programmable bits nmax, then the adjacent blocks are written together as inverse data, while indicating the inversion by writing an additional “1” in the additional bit described above.


Hence, the first and the second data blocks are programmed together in the normal mode as described in FIG. 1. However, the third and fourth blocks are programmed together using inverse mode, thus reducing the number of write operations needed. Hence, all the data blocks are programmed in eight data blocks while the normal mode embodiment requires eleven data blocks. While there is a loss in storage efficiency that arises from the need to include the information regarding the data bit mode, this can be offset by the gains in efficiency in some embodiments. The efficiency gain of this process is 100. Advantageously, and unlike the prior embodiments, this gain in efficiency for this example with 8 bits per data block and nmax=10 is independent of the data.



FIG. 5, which includes FIGS. 5a and 5b, illustrates an embodiment of the invention using normal and inverse data. While the first, second, and third row are similar to FIG. 4, further rows are added as described above with respect to FIG. 3 to add more than one adjacent data block. While the third row sums over two adjacent blocks (e.g., nr+nr+1), the fourth row sums over three adjacent blocks (e.g., nr+nr+1+nr+2). If the bits that are to be programmed in the three adjacent blocks is not greater than the maximum programmable bits nmax, then the three adjacent blocks are written together. Similarly, an additional row is added that sums over three adjacent blocks in the inverse data block (e.g., ur+ur+1+ur+2). If the number of inverse bits that are to be programmed in the three adjacent blocks is less than the maximum programmable bits nmax, then the three adjacent blocks are written together as inverse data, again adding an “1” for indication of inverse data. The efficiency gain is about 128% for this example.



FIG. 6 illustrates an embodiment in which three adjacent blocks are summed in both normal and inverse data mode. Additional rows are added to show additional summations (e.g. fifth row shows summation over nr+1+nr+2+nr+3 and sixth row shows summation over nr−1+nr+nr+1). Similarly, additional rows are added in the inverse data space. As shown in the table, the data sequence shown is processed in six steps with an overall efficiency gain of about 167% for this example. Hence, at any given data sequence, the selection proceeds by selecting the minimum number of bits summed over two adjacent blocks or three adjacent blocks.


In various embodiments, the write optimization algorithms described above can be configured on a chip. As the summation is performed over a finite number of data blocks (e.g. 2 for the embodiments of FIG. 1), the level of complexity of the hardware is minimal. In various embodiments, the hardware can either use embodiments of the invention or use a fixed number of data blocks, thus using writing widths independent of the data.



FIG. 7 illustrates an embodiment wherein the maximum number of programmable bits nmax is reduced without a significant decrease in write efficiency. In FIG. 7, the maximum number of programmable bits nmax is reduced to 8 compared to FIG. 1a wherein the maximum number of programmable bits nmax is 10. In this embodiment, the total number of write operations is 13, a 23% increase in efficiency despite the reduction in the total current available for programming.


As illustrated in FIG. 7, the improved writing operation can be used to reduce circuit area of the device, for example, by reducing the area of the charge pumps. This is possible by reducing the total current available for the write operation. For example, the maximum number of programmable bit nmax can be reduced using embodiments of the invention without having an impact on the total write time. With a reduction in the total current from the charge pump, the area of the charge pump can be scaled down as the size of the capacitors can be reduced. As the capacitors of the charge pumps consume considerable area, a significant saving in die size is possible. This embodiment may be further preferred in power sensitive devices such as hand held devices.



FIG. 8, illustrates an embodiment wherein data blocks are selected to maximize the number of bits written by selecting data block combinations as near to the maximum programmable bits nmax bits to program as possible. In this embodiment, the data is first buffered and the number of data bits in each data block is calculated as illustrated in FIG. 1. However, in this embodiment, data blocks are selected to maximize the number of bits that are written together. For example, FIG. 8 shows the data sequence of FIG. 1 written using this embodiment. In the first cycle, the data blocks in the first, and the third column are selected to form a single block of ten bits that is programmed in parallel. In the second cycle, the second and the fourth data block are selected. Similarly, as illustrated in FIG. 8, the rest of the data sequence is written in just 7 cycles.



FIG. 9 describes an embodiment of the invention illustrating a block diagram of a memory circuit. The memory circuit includes a charge pump 11 coupled to a power source node 12. The charge pump is designed to supply voltage that is higher than the supply voltage. For example, the charge pump supplies an output voltage of 12V used in the write or erase operations of the memory array 15. Voltage from the charge pump is supplied to the memory array, for example, through switches 16. For example, word line and bit line switches are coupled to respective word lines and bit lines of the memory array and to the charge pump 11 in one embodiment. The bit lines and the word lines provide the voltage on each memory device of the memory array 15. The controller 20 processes the data sequence 10 to be written onto the memory array 15 and provides information regarding the number and position of bits and hence the number of memory devices that are simultaneously written. In one embodiment the controller 20 comprises a processor. The switches 16 use the information from the controller 20 and couples a set of memory devices in parallel. In various embodiments, the controller 20 uses any one of the algorithms as described in the embodiments of FIGS. 1-8 in selecting the number of bits written together. In various embodiments, the memory array 15 comprises non volatile memory such as Flash memory devices.


In describing embodiments of the invention, tables are used only as a way to explain the algorithms, the actual implementation may not require generation of such tables.


Although the present invention and its advantages have been described in detail, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims. For example, many of the features and functions discussed above can be implemented in software, hardware, or firmware, or a combination thereof. As another example, it will be readily understood by those skilled in the art that many of the features, functions described herein may be varied while remaining within the scope of the present invention.


Moreover, the scope of the present application is not intended to be limited to the particular embodiments of the process, machine, manufacture, composition of matter, means, methods and steps described in the specification. As one of ordinary skill in the art will readily appreciate from the disclosure of the present invention, processes, machines, manufacture, compositions of matter, means, methods, or steps, presently existing or later to be developed, that perform substantially the same function or achieve substantially the same result as the corresponding embodiments described herein may be utilized according to the present invention. Accordingly, the appended claims are intended to include within their scope such processes, machines, manufacture, compositions of matter, means, methods, or steps.

Claims
  • 1. A method of programming a memory array, the method comprising: dividing data to be written into a plurality of first data blocks of a first length;calculating a first value, the first value being the number of bits that are to be programmed in each first data block;dividing data to be written into a plurality of second data blocks of a second length, the second length being larger than the first length;calculating a second value, the second value being the number of bits that are to be programmed in each second data block; andwriting the bits that are to be programmed in the second data block in parallel if the second value is not greater than a maximum allowable number of bits programmed in parallel.
  • 2. The method of claim 1, further comprising writing the bits that are to be programmed in the first data block in parallel if the second value is greater than the maximum allowable number of bits programmed in parallel.
  • 3. The method of claim 1, wherein the second length is an integer multiple of the first length.
  • 4. The method of claim 1, wherein the second length is less than four times the first length.
  • 5. The method of claim 1 further comprising dividing data to be written into a plurality of third data blocks of a third length, the third length being larger than the second length, calculating a third value, the third value being the number of bits that are to be programmed in each third data block; and writing the bits that are to be programmed in the third data block in parallel if the third value is not greater than the maximum allowable number of bits programmed in parallel.
  • 6. The method of claim 5, wherein the second length and the third lengths are integer multiples of the first length.
  • 7. The method of claim 1 further comprising: dividing data to be written into a plurality of third data blocks of the second length, the third data blocks and the second data blocks being translated by the first length;calculating a third value, the third value being the number of bits that are to be programmed in each third data block; andwriting the bits that are to be programmed in the third data block in parallel if the third value is not greater than the maximum allowable number of bits programmed in parallel.
  • 8. The method of claim 7, wherein the second length is an integer multiple of the first length.
  • 9. The method of claim 8 further comprising: dividing data to be written into a plurality of fourth data blocks of a third length, the third length being larger than the second length;calculating a fourth value, the fourth value being the number of bits that are to be programmed in each fourth data block; andwriting the bits that are to be programmed in the fourth data block in parallel if the fourth value is not greater than the maximum allowable number of bits programmed in parallel.
  • 10. The method of claim 9, wherein the second length and the third length are integer multiples of the first length.
  • 11. The method of claim 9 further comprising: dividing data to be written into a plurality of fifth data blocks of the third length, the fifth data blocks and the fourth data blocks being translated by the first length;calculating a fifth value, the fifth value being the number of bits that are to be programmed in each fifth data block; andwriting the bits that are to be programmed in the fifth data block in parallel if the fifth value is not greater than the maximum allowable number of bits programmed in parallel.
  • 12. The method of claim 7, wherein the second length and third lengths are integer multiples of the first length.
  • 13. The method of claim 1 further comprising: dividing data to be written into a plurality of third data blocks of a second length, the plurality of third blocks comprising the inverse of the data to be written;calculating a third value, the third value being the number of bits that are to be programmed in each third data block; andwriting the bits that are to be programmed in the third data block in parallel if the third value is not greater than the maximum allowable number of bits programmed in parallel.
  • 14. A method of programming a memory array, the method comprising: receiving a series of data blocks, each data block having a number of bits that are to be programmed;determining the number of bits that are to be programmed in a first data block;determining the number of bits that are to be programmed in a second data block; andwriting the first data block into the memory array in parallel if the sum of the number of bits that are to be programmed in the first data block and the second data block is greater than a maximum value.
  • 15. The method of claim 14, wherein the first and the second data blocks are adjacent data blocks.
  • 16. The method of claim 14 further comprising writing the first and the second data blocks into the memory array in parallel if the sum of the number of bits that are to be programmed in the first data block and the second data block is not greater than the maximum value.
  • 17. The method of claim 14 further comprising determining the number of bits that are to be programmed in a third data block.
  • 18. The method of claim 17 further comprising writing the first, the second, and the third data blocks into the memory array in parallel if the sum of the number of bits that are to be programmed in the first data block, the second data block, and the third data block is not greater than the maximum value.
  • 19. The method of claim 17 further comprising writing the first, and the second data blocks into the memory array in parallel if the sum of the number of bits that are to be programmed in the first data block, the second data block, and the third data block is greater than the maximum value.
  • 20. The method of claim 19, wherein the first, the second, and the third data blocks are adjacent data blocks.
  • 21. The method of claim 17 further comprising writing the first and the third data blocks into a memory array in parallel if the sum of the number of bits that are to be programmed in the first data block and the third data block is not greater than the maximum value.
  • 22. The method of claim 21, wherein the first and the third data blocks are not adjacent data blocks.
  • 23. The method of claim 17 further comprising determining the number of bits that are to be programmed in a fourth data block.
  • 24. The method of claim 23 further comprising writing the first, the second, the third, and the fourth data blocks into a memory array in parallel if the sum of the number of bits that are to be programmed in the first data block, the second data block, the third data block, and the fourth data block is not greater than the maximum value.
  • 25. The method of claim 24, wherein the first, the second, the third, and the fourth data blocks are adjacent data blocks.
  • 26. The method of claim 23 further comprising writing the first, the second, and the third data blocks into a memory array in parallel if the sum of the number of bits that are to be programmed in the first data block, the second data block, the third data block, and the fourth data block is greater than the maximum value.
  • 27. The method of claim 14 further comprising: determining the inverse of the data in the first and the second data blocks;writing the inverse of the first and the second data blocks into the memory array in parallel if the sum of the number of bits that are to be programmed in the first data block and the second data block is greater than the maximum value.
  • 28. The method of claim 27 further comprising writing the first and the second data blocks into the memory array in parallel if the sum of the number of bits that are to be programmed in the first data block and the second data block is not greater than a maximum value.
  • 29. The method of claim 27, wherein the first and the second data blocks are adjacent data blocks.
  • 30. A method of programming a memory array, the method comprising: dividing data to be written into a plurality of first data blocks of a first length;calculating a first value, the first value being the number of bits that are to be programmed in each data block of the plurality of first data blocks;calculating a second value, the second value being the number of bits that are to be programmed in an immediately subsequent data block of the plurality of first data blocks;calculating a third value, the third value being the sum of the first and the second values; andwriting the bits that are to be programmed in the first data block and the immediately subsequent data block in parallel if the third value is not greater than a maximum allowable number of bits programmed in parallel.
  • 31. The method of claim 30, further comprising writing the bits that are to be programmed in the first data block in parallel if the third value is greater than the maximum allowable number of bits programmed in parallel.
  • 32. The method of claim 31, further comprising: converting the first data block and the immediately subsequent data block into an inverse data;writing the bits that are to be programmed in the inverse data in parallel if the third value is greater than the maximum allowable number of bits programmed in parallel.
  • 33. A memory circuit comprising: a non volatile memory array comprising non volatile memory devices;a charge pump coupled to the plurality of non volatile memory array, the charge pump powering the programming of the non volatile memory devices;a controller coupled to the plurality of non volatile memory array, the controller comprising a means to count the number of bits that are to be programmed in a series of data blocks of data to be written, wherein the controller generates a first value, the first value being the number of bits to be written in parallel, wherein the first value is an integer multiple of a second value, wherein the second value is a predefined value and comprises the smallest number of bits written in parallel;a switch coupled between the charge pump and the non volatile memory array, wherein the switch comprises a means to couple a portion of the non volatile memory devices totaling the first value with the charge pump.
  • 34. The memory circuit of claim 33, wherein the controller generates the first value by: receiving the series of data blocks, each data block having a number of bits that are to be programmed;determining the number of bits that are to be programmed in a first data block;determining the number of bits that are to be programmed in a second data block; andassigning the number of bits that are to be programmed in the first data block to be the first value if the sum of the number of bits that are to be programmed in the first data block and the second data block is greater than a maximum value.
  • 35. The memory circuit of claim 34, wherein the first and the second data blocks are adjacent data blocks.
  • 36. The memory circuit of claim 34, wherein the maximum value is determined by a capacity of the charge pump.
  • 37. The memory circuit of claim 34 further comprises assigning the number of bits that are to be programmed in the first and the second data blocks to be the first value if the sum of the number of bits that are to be programmed in the first data block and the second data block is not greater than the maximum value.