US20080178021A1 - Power efficiency in microprocessors - Google Patents
Power efficiency in microprocessors Download PDFInfo
- Publication number
- US20080178021A1 US20080178021A1 US12/011,891 US1189108A US2008178021A1 US 20080178021 A1 US20080178021 A1 US 20080178021A1 US 1189108 A US1189108 A US 1189108A US 2008178021 A1 US2008178021 A1 US 2008178021A1
- Authority
- US
- United States
- Prior art keywords
- instruction
- bit
- instructions
- unused bits
- bits
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
- 238000000034 method Methods 0.000 claims abstract description 31
- 230000005540 biological transmission Effects 0.000 claims abstract 4
- 230000008859 change Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010899 nucleation Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/30—Arrangements for executing machine instructions, e.g. instruction decode
- G06F9/38—Concurrent instruction execution, e.g. pipeline, look ahead
- G06F9/3802—Instruction prefetching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/30—Arrangements for executing machine instructions, e.g. instruction decode
- G06F9/30145—Instruction analysis, e.g. decoding, instruction word fields
Abstract
A method of reducing the power consumption of microprocessor system is provided, wherein: said microprocessor system comprises a microprocessor (2) and a memory (4) connected by a bus (6); said memory (4) contains a plurality of data values, each represented by a number of bits, for transmission to said microprocessor (2) via the bus (6); and at least some of said data values contain unused bits; and wherein said method includes assigning values to said unused bits in such a way as to reduce the Hamming distance between successive data values by a greater extent than setting all of said unused bits to an arbitrary predetermined value.
Description
- This application is a continuation of U.S. patent application Ser. No. 10/478,168, filed Aug. 25, 2004, which claims the benefit of International Patent Application No. PCT/GB02/02034, filed May 2, 2002, which claims priority to Great Britain Patent Application No. 0112276.1, filed May 19, 2001.
- The invention relates to improved power efficiency in microprocessors.
- The concept of Hamming distance will first be described. The Hamming distance between two binary numbers is the count of the number of bits that differ between them. For example:
-
Numbers in binary (inc. Numbers in decimal leading zeros) Hamming distance 4 and 5 0100 and 0101 1 7 and 10 0111 and 1010 3 0 and 15 0000 and 1111 4 - Hamming distance is related to power efficiency because of the way that binary numbers are represented by electrical signals. Typically a steady low voltage on a wire represents a binary 0 bit and a steady high voltage represents a binary 1 bit. A number will be represented using these voltage levels on a group of wires, with one wire per bit. Such a group of wires is called a bus. Energy is used when the voltage on a wire is changed. The amount of energy depends on the magnitude of the voltage change and the capacitance of the wire. The capacitance depends to a large extent on the physical dimensions of the wire. So when the number represented by a bus changes, the energy consumed depends on the number of bits that have changed—the Hamming distance—between the old and the new value, and on the capacitance of the wires.
- If one can reduce the average Hamming distance between successive values on a high-capacitance bus, keeping all other aspects of the system the same, the system's power efficiency will have been increased.
- The capacitance of wires internal to an integrated circuit is small compared to the capacitance of wires fabricated on a printed circuit board due to the larger physical dimensions of the latter.
- EP 0,926,596 describes a method of optimizing assembly code of a VLIW processor or other processor that uses multiple-instruction words, each of which comprise instructions to be executed on different functional units of the processor.
- According to the invention there is provided a method of reducing the power consumption of a microprocessor system, a memory, a computer readable medium, computer programs, and a microprocessor system, as set out in the accompanying claims.
- Embodiments of the invention will now be described, by way of example only, with reference to the accompanying FIGURE, which shows the interconnections between a microprocessor and its memory.
- In the accompanying FIGURE, a
microprocessor 2 is connected to memory 4 by a number of buses which are implemented on a printed circuit board (not shown). - The embodiments described here aim to reduce the average Hamming distance between successive values on the microprocessor-memory interface buses, shown as the lighter colored
nets FIG. 1 , as this will have a significant influence on power efficiency. - Even in systems where microprocessor and memory are incorporated into the same integrated circuit the capacitance of the wires connecting them will be larger than average, so even in this case reduction of average Hamming distance on the microprocessor-memory interface is worthwhile.
- Processor-memory communications perform two tasks. Firstly, the processor fetches its program from the memory, one instruction at a time. Secondly, the data that the program is operating on is transferred back and forth. The embodiments described here focus on instruction fetch, which makes up the majority of the processor-memory communications, but the invention is by no means limited to instruction fetch.
- The
instruction fetch bus 6 is the bus on which instructions are communicated from the memory 4 to theprocessor 2. Embodiments described here aim to reduce the average Hamming distance on this bus, i.e. to reduce the average Hamming distance from one instruction to the next. - It is common for instruction sets to be redundant, i.e. for instructions to contain information that is ignored by the processor. In particular some instructions contain unused bits.
- For example in the instruction set considered here all instructions are 32 bits long. There are three instruction formats:
- Bits marked ‘X’ are unused bits. The other bits convey useful information to the
processor 2. - In two of these formats, 31 of the 32 bits are used. In the third format, 26 of the 32 bits are used. The remaining six bits are completely ignored by the
processor 2. Some further features of our illustrative instruction set mean that other bits will also sometimes be unused, but the exact details of this have been omitted for clarity. - Any or all of the unused bits can be assigned to a combination of ‘0’s or ‘1’s without changing the meaning of the program.
- Many other
common processor 2 instruction sets also have unused bits. Typically, existing compiler tool chains will set all of the unused bits to ‘0’. - Although the
processor 2 ignores these bits, they still contribute to the average Hamming distance for instruction fetches. The embodiments described here assign values to unused bits in a way that reduces the Hamming distance and hence maximizes the power efficiency. - For example, consider the following sequence of three instructions:
- A: X0111000011010001101000110100111
- B: X000000010011110XXXXX01101100011
- C: X0001101111000111100110011011100
- Note the group of unused bits, marked ‘XXXXX’ in instruction B. A conventional system might set all unused bits to ‘0’ giving the following code:
- A: 00111000011010001101000110100111
- B: 00000000100111100000001101100011
- C: 00001101111000111100110011011100
- The Hamming distances between these instructions are:
- From A to B: 16
- From B to C: 22
- One embodiment described here instead gives the unused bits the following values:
- A: 00111000011010001101000110100111
- B: 00000000100111101101001101100011
- C: 00001101111000111100110011011100
- In this case the Hamming distances have been reduced to:
- From A to B: 13
- From B to C: 21
- This technique does not require any modifications to the
microprocessor 2. Power is saved through only changing the program bit pattern. - A first embodiment of the invention uses the following method for assigning unused bits in a sequence of instructions:
- For the first instruction in the program, set any unused bits to 0.
- Considering each subsequent instruction in the program sequentially:
-
- Considering each bit in this instruction:
- If this bit is unused, assign the value of the corresponding bit from the previous instruction to it.
- Considering each bit in this instruction:
- An example of this method will now be given using the following sequence of 8-bit instructions:
- Bit number: 76543210
-
- 01X01X01
- X001X010
- XX000X1X
- 00X1X100
- 1X001X00
- 1X001X00
- 00001X00
- 01X0XX10
- 10000XX1
- 01X100X0
- 1X1X00X0
- 01X01000
- XX001X0X
- 110X01X0
- The changes to the bits of the first three instructions will now be given in detail.
- The first instruction is 01X01X01.
-
- Bit 7 of instruction is 0 Do nothing
-
Bit 6 of instruction is 1 Do nothing - Bit 5 of instruction is X Set it to 0
- Bit 4 of instruction is 0 Do nothing
- Bit 3 of instruction is 1 Do nothing
-
Bit 2 of instruction is X Set it to 0 - Bit 1 of instruction is 0 Do nothing
- Bit 0 of instruction is 1 Do nothing
- After processing the first instruction it has been changed to 01001001.
- The second instruction is X001X010.
-
- Bit 7 of instruction is X Set it to 0
- (Copied from bit 7 of first instruction)
-
Bit 6 of instruction is 0 Do nothing - Bit 5 of instruction is 0 Do nothing
- Bit 4 of instruction is 1 Do nothing
- Bit 3 of instruction is X Set it to 1
- (Copied from bit 3 of first instruction)
-
Bit 2 of instruction is 0 Do nothing - Bit 1 of instruction is 1 Do nothing
- Bit 0 of instruction is 0 Do nothing
- Bit 7 of instruction is X Set it to 0
- After processing the second instruction it has been changed to 00011010.
- The third instruction is XX000X1X.
-
- Bit 7 of instruction is X Set it to 0
- (Copied from bit 7 of second instruction)
-
Bit 6 of instruction is X Set it to 0- (Copied from
bit 6 of second instruction)
- (Copied from
- Bit 5 of instruction is 0 Do nothing
- Bit 4 of instruction is 0 Do nothing
- Bit 3 of instruction is 0 Do nothing
-
Bit 2 of instruction is X Set it to 0- (Copied from
bit 2 of second instruction)
- (Copied from
- Bit 1 of instruction is 1 Do nothing
- Bit 0 of instruction is X Set it to 0
- (Copied from bit 0 of second instruction)
- Bit 7 of instruction is X Set it to 0
- After processing the third instruction it has been changed to 00000010.
- The complete sequence of output instructions, after processing according to the method of the first embodiment, is given in the following table.
-
Input instruction Output instruction 00X1X100 00010100 1X001X00 10001100 1X001X00 10001100 00001X00 00001100 01X0XX10 01001110 10000XX1 10000111 01X100X0 01010010 1X1X00X0 11110010 01X01000 01101000 XX001X0X 01001000 110X01X0 11000100 - In this example, the mean inter-instruction Hamming distance after the unused bits have been assigned is 2.61. If all unused bits had been assigned to zero, then the mean inter-instruction Hamming distance would be 2.92, indicating a power saving of around 5%.
- This method produces optimal results for straight-line code, i.e. code that has no branches in the flow-of-control. To take into account non-linear flow-of-control a more sophisticated method is required, as will be described below.
- The following table shows an example of a fragment of pseudo high-level code and corresponding pseudo assembly language instructions:
-
if a=1 then Compare a with 1 (1) Branch, if not equal, to L1 (2) b:=0; set b to 0 (3) jump to L2 (4) else L1: set c to 1 (5) c:=1; end if; L2: add 1 to d (6) d:=d+1; e:=a; set e to a (7) -
-
- A second embodiment of the invention incorporates an unused bit assignment method that takes into account flow of control.
- When an unused bit is both preceded and followed by used bits in the adjacent instructions, setting the unused bit to the value of either the preceding bit or the following bit will optimize Hamming distance. For example:
-
Preceding bit Unused bit Following bit A: 0 X 1 B: 1 X 1 - In example A, the value of the preceding bit can be copied into the unused bit giving the bit-sequence 001, or the value of the following bit can be copied into the unused bit giving the bit-sequence 011. In both cases there is exactly one transition overall. In example B, whichever bit's value is copied into the unused bit it will be a 1, giving no transitions in either case.
- The first embodiment described above always copies from the preceding instruction. A modification of the first embodiment could run in reverse and always copy from the following instruction. The method of the second embodiment may copy from either.
- In the example,
instruction 2 is a point of divergence, because the following instruction can be either instruction 3 or instruction 5.Instruction 6 is a point of convergence, because the preceding instruction can be either instruction 4 or instruction 5. - Instructions at points of convergence have more than one possible preceding instruction. If we used only preceding instructions to determine how to assign unused bits we would have to make a decision about which of the two possible preceding instruction to use. Instead in these cases we can use the following instruction as the basis for unused bit assignment. So in the example, unused bits in
instruction 6 are filled in from instruction 7. - The method of the second embodiment is based on the following three rules.
- 1) If an instruction has more than one preceding instruction, i.e. it is at a point of convergence, assign unused bits based on the following instruction.
- 2) If an instruction has more than one following instruction, i.e. it is at a point of divergence, assign unused bits based on the preceding instruction.
- 3) If an instruction has exactly one preceding and exactly one following instruction, i.e. it is neither a point of convergence nor of divergence, then assign unused bits based on either the preceding or the following instruction.
- The following table shows how this can be applied to the example shown above:
-
Possible Possible Assign preceding following unused bits instructions instructions based on compare ‘a’ with 1 (1) (None) 2 2 branch if not equal, to L1 (2) 1 3, 5 1 set ‘b’ to 0 (3) 2 4 2 or 4 jump to L2 (4) 3 6 3 or 6 L1: set ‘c’ to 1 (5) 2 6 2 or 6 L2: add 1 to ‘d’ (6) 4, 5 7 7 set ‘e’ to ‘a’ (7) 6 (None) 6 - In a variant embodiment an instruction's unused bits are based on a non-adjacent instruction. For example, instruction (5) could be based on instruction (2). This will typically occur at the target of an unconditional branch. In an implementation of this algorithm, such an assignment may be less practical than an assignment from an adjacent instruction because it requires that the implementation compute the target of the branch. In some cases, this may even be impossible, particularly when the branch target is computed at run time.
- The following are hypothetical machine code bit patterns corresponding to these instructions, including some unused bits:
-
Compare a with 1 00001X00 (1) Branch if not equal to L1 01X0XX10 (2) Set b to 0 10000XX1 (3) Jump L2 01X100X0 (4) Set c to 1 1X1X00X0 (5) Add 1 to d 01X01000 (6) Set e to a XX001X0X (7) - In accordance with the second embodiment, combining the above two tables gives the following assignments for unused bits, in which the arrows show how unused bits are assigned from adjacent instructions.
- Where an arrow joins a used bit to an unused bit the unused bit can be assigned from that used bit. For example, bit 3 of instruction (2) can be assigned from bit 3 of instruction (1). When-a used bit is connected to an unused bit, and that unused bit is connected in turn to other unused bits, the value can be propagated to all of them. For example, bit 1 of instruction (2) can be propagated to unused bit 1 in instructions (3) and (4).
- Here is the complete list of the assignments in accordance with the second embodiment:
- Some unused bits can remain unassigned after the method of the second embodiment has been carried out. This will occur when a bit is unused in all instructions between a point of convergence and a point of divergence. In the example,
bit 2 of instructions (1) and (2) are unassigned for this reason. A third embodiment, given later, will “seed” the group of unused bits using a used bit in an adjacent instruction. In this example,bit 2 of instruction (2) could be seeded with a ‘0’ from instruction (3). - The mean inter-instruction Hamming distances for this example are:
- Difficulties arise when an instruction is both a point of convergence and a point of divergence. This will occur when a branch instruction leads to another branch instruction. The example in the following table illustrates this.
-
... ... ... ... (1) jump L1 ... ... ... ... (2) jump L1 (3) ... ... ... ... (4) L1: branch if negative, to L2 (5) ... ... ... ... (6) L2: - Possible paths through this code are:
-
- (1)(4)(5) (1)(4)(6)
- (2)(4)(5) (2)(4)(6)
- (3)(4)(5) (3)(4)(6)
- Instruction (4) is a point of both convergence and divergence, and the embodiments described so far offer no instruction on which it should base any unused bits that it may contain.
- In a further variant embodiment, unused bit assignment is based on the values of bits in the multiple preceding and following instructions. In general an optimal assignment requires knowledge of the relative probabilities of each path. However if all of the preceding instructions, or all of the following instructions, are known and have the same bit value an optimal assignment is still possible. The third embodiment given later simply copies a value from an adjacent instruction.
- Many processors incorporate an architectural feature called pipelining that affects the sequence in which instructions are fetched. In a pipelined processor, one or more instructions sequentially after a branch instruction will be fetched and then discarded if the branch is taken. For the purposes of this analysis these instruction fetches are as important as any other and need to be taken into account.
- Considering the same example as before:
-
Compare a with 1 (1) Branch if not equal to L1 (2) set b to 0 (3) Jump to L2 (4) L1: set c to 1 (5) L2: add 1 to d (6) set e to a (7) -
- The second and third embodiments can function correctly for pipelined processors with only a minor change. “Points of divergence”, which were previously considered to be branch instructions, arm now the instructions an appropriate distance after the branch instructions.
- The method of the third embodiment can be summarized by the following set of instructions
- Let C be the set of instructions that are at points of convergence. This means all instructions that are labeled as branch targets.
- Let D be the set of instructions that are at points of divergence. This means all branch instructions, or if pipelining is being taken into account, all instructions that are the appropriate distance after a branch instruction.
- Let E be the set of pairs (I, J) where I and J are instructions that satisfy either or both of the following conditions:
-
- J is not an element of C, and I immediately precedes J
- J is not an element of D, and I immediately follows J
(Note: Each element of E corresponds to an arrow in FIGURE, with I being the instruction at the tail of the arrow and J being the instruction at the head of the arrow).
- For each bit in turn:
-
Let B be the bit-number of the current bit While there are any instructions where bit B is unused While there is an element (I, J) of E, such that bit B of instruction J is unused and bit B of I is used Set bit B of instruction J to the value of bit B of instruction I End-while If there are still any instructions where bit B is unused (Note: This step implements the “seeding” process mentioned earlier) Find any two instructions I and J, such that bit B of instruction I is used, bit B of J is unused, and I and J are adjacent instructions [for the purposes of this step, the first instruction in the program should be considered to be preceded by and the last instruction followed by an Instruction containing all zeros] Set bit B of instruction J to the value of bit B of instruction I End-if End-while End-for
Claims (12)
1. A method of reducing the power consumption of a microprocessor system which comprises a microprocessor and a memory connected by at least one bus, said memory containing a plurality of data values, each represented by a plurality of bits, for transmission to said microprocessor via said at least one bus, and at least some of said data values containing unused bits, the method comprising the steps of:
reading an initial data value from memory;
setting unused bits in the initial data value to a predetermined bit value;
separately considering each remaining data value in sequence, assigning a bit value to each unused bit in the considered data value, the assigned bit value being equal to a bit value of a corresponding bit in an adjacent data value.
2. A method as claimed in claim 1 , wherein the data values are instructions for execution by the microprocessor,
3. A method as claimed in claim 1 , wherein said adjacent data value is a data value to be transmitted to the microprocessor via the at least one bus immediately before or after transmission of the considered data value.
4. A method as claimed in claim 1 , wherein said adjacent data value is stored immediately before the considered data value in said memory.
5. A method as claimed in claim 1 , wherein said adjacent data value is stored immediately after the considered data value in said memory.
6. A method of reducing the power consumption of a microprocessor system which comprises a microprocessor and a memory connected by at least one bus, said memory containing a plurality of instructions, each represented by a number of bits, for transmission in a sequence for execution to said microprocessor via said at least one bus, and at least some of said instructions containing unused bits, the method comprising the steps of:
for an initial instruction, assigning unused bits to a predetermined bit value;
for remaining instructions, taken in the sequence:
if the instruction has more than one possible following instruction, assigning unused bits in the instruction to be equal to a bit value of a corresponding bit in a preceding instruction;
if the instruction has more than one possible preceding instruction, assigning unused bits in the instruction to be equal to a bit value of a corresponding bit in a following instruction; and
in other instructions, assigning unused bits in the instruction to be equal to a bit value of a corresponding bit in an adjacent instruction; and
storing the instructions in the memory.
7. A method as claimed in claim 6 , wherein, in the other instructions, the unused bits in the instruction are assigned to be equal to a bit value of a corresponding bit in a following instruction.
8. A method as claimed in claim 6 , wherein, in the other instructions, the unused bits in the instruction are assigned to be equal to a bit value of a corresponding bit in a preceding instruction.
9. A method as claimed in claim 6 , wherein assignment of unused bits in the instruction is based on a consideration of the probabilities of different paths to and from said instruction.
10. A method as claimed in claim 6 , wherein in the case of an instruction having more than one possible preceding instruction and more than one possible following instruction, assignment of unused bits in the instruction is based on a consideration of bits in multiple preceding and following instructions.
11. A method as claimed in claim 6 , wherein in the case of an instruction having more than one possible preceding instruction and more than one possible following instruction, assignment of unused bits in the instruction is based on relative probabilities of each path.
12. A method of reducing the power consumption of a microprocessor system which comprises a microprocessor and a memory connected by at least one bus, said memory containing a plurality of instructions, each represented by a number of bits, for fetching via said at least one bus in a sequence for execution by said microprocessor, and at least some of said instructions containing unused bits, said microprocessor being a pipelined microprocessor that fetches n extra instructions ahead in the sequence, the method comprising the steps of:
for an initial instruction, assigning unused bits to a predetermined bit value;
for remaining instructions, taken in the sequence:
if an instruction has more than one possible following instruction, assigning unused bits in a subsequent instruction that is n instructions after the instruction, to be equal to a bit value of a corresponding bit in an instruction preceding the subsequent instruction;
if the instruction has more than one possible preceding instruction, assigning unused bits in the instruction to be equal to a bit value of a corresponding bit in a following instruction; and
in remaining instructions with unused bits, assigning unused bits in the instruction to be equal to a bit value of a corresponding bit in an adjacent instruction; and
storing the instructions in the memory.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US12/011,891 US20080178021A1 (en) | 2001-05-19 | 2008-01-30 | Power efficiency in microprocessors |
Applications Claiming Priority (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
GB0112276A GB2375695B (en) | 2001-05-19 | 2001-05-19 | Improved power efficency in microprocessors |
GB0112276.1 | 2001-05-19 | ||
US10/478,168 US7353411B2 (en) | 2001-05-19 | 2002-05-02 | Power efficiency in microprocessors using reduction of hamming distance between data value |
PCT/GB2002/002034 WO2002095574A1 (en) | 2001-05-19 | 2002-05-02 | Improved power efficiency in microprocessors |
US12/011,891 US20080178021A1 (en) | 2001-05-19 | 2008-01-30 | Power efficiency in microprocessors |
Related Parent Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/GB2002/002034 Continuation WO2002095574A1 (en) | 2001-05-19 | 2002-05-02 | Improved power efficiency in microprocessors |
US10/478,168 Continuation US7353411B2 (en) | 2001-05-19 | 2002-05-02 | Power efficiency in microprocessors using reduction of hamming distance between data value |
Publications (1)
Publication Number | Publication Date |
---|---|
US20080178021A1 true US20080178021A1 (en) | 2008-07-24 |
Family
ID=9914956
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/478,168 Expired - Fee Related US7353411B2 (en) | 2001-05-19 | 2002-05-02 | Power efficiency in microprocessors using reduction of hamming distance between data value |
US12/011,891 Abandoned US20080178021A1 (en) | 2001-05-19 | 2008-01-30 | Power efficiency in microprocessors |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/478,168 Expired - Fee Related US7353411B2 (en) | 2001-05-19 | 2002-05-02 | Power efficiency in microprocessors using reduction of hamming distance between data value |
Country Status (3)
Country | Link |
---|---|
US (2) | US7353411B2 (en) |
GB (1) | GB2375695B (en) |
WO (1) | WO2002095574A1 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140173296A1 (en) * | 2012-12-14 | 2014-06-19 | International Business Machines Corporation | Chip level power reduction using encoded communications |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2378269B (en) * | 2001-07-28 | 2005-08-10 | At & T Lab Cambridge Ltd | Improved power efficiency in microprocessor systems |
JP3896087B2 (en) * | 2003-01-28 | 2007-03-22 | 松下電器産業株式会社 | Compiler device and compiling method |
WO2005048115A2 (en) * | 2003-11-13 | 2005-05-26 | Koninklijke Philips Electronics N.V. | Electronic data processing circuit that transmits packed words via a bus |
CN100592711C (en) * | 2004-03-08 | 2010-02-24 | 皇家飞利浦电子股份有限公司 | Integrated circuit and method for packet switching control |
US20080043757A1 (en) * | 2004-08-12 | 2008-02-21 | Koninklijke Philips Electronics, N.V. | Integrated Circuit And Method For Packet Switching Control |
US9710277B2 (en) * | 2010-09-24 | 2017-07-18 | Intel Corporation | Processor power management based on class and content of instructions |
US20120079249A1 (en) * | 2010-09-28 | 2012-03-29 | Wei-Han Lien | Training Decode Unit for Previously-Detected Instruction Type |
TWI442317B (en) * | 2011-11-07 | 2014-06-21 | Ind Tech Res Inst | Reconfigurable instruction encoding method and processor architecture |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5572736A (en) * | 1995-03-31 | 1996-11-05 | International Business Machines Corporation | Method and apparatus for reducing bus noise and power consumption |
US5790874A (en) * | 1994-09-30 | 1998-08-04 | Kabushiki Kaisha Toshiba | Information processing apparatus for reducing power consumption by minimizing hamming distance between consecutive instruction |
US5854935A (en) * | 1995-10-18 | 1998-12-29 | Nec Corporation | Program transformation system for microcomputer and microcomputer employing transformed program |
US6535984B1 (en) * | 1998-11-25 | 2003-03-18 | Texas Instruments Incorporated | Power reduction for multiple-instruction-word processors with proxy NOP instructions |
US6587939B1 (en) * | 1999-01-13 | 2003-07-01 | Kabushiki Kaisha Toshiba | Information processing apparatus provided with an optimized executable instruction extracting unit for extending compressed instructions |
US20030212914A1 (en) * | 2001-07-28 | 2003-11-13 | Paul Webster | Power efficiency in microprocessor systems |
US6725450B1 (en) * | 1999-06-21 | 2004-04-20 | Matsushita Electric Industrial Co., Ltd. | Program conversion apparatus, processor, and record medium |
US6807528B1 (en) * | 2001-05-08 | 2004-10-19 | Dolby Laboratories Licensing Corporation | Adding data to a compressed data frame |
US20050010830A1 (en) * | 2001-08-10 | 2005-01-13 | Paul Webster | Power reduction in microprocessor systems |
US20050229017A1 (en) * | 2001-08-10 | 2005-10-13 | At&T Corp. | Power reduction in microprocessor systems |
US7386844B2 (en) * | 2003-01-28 | 2008-06-10 | Matsushita Electric Industrial Co., Ltd. | Compiler apparatus and method of optimizing a source program by reducing a hamming distance between two instructions |
US7386777B2 (en) * | 2004-04-05 | 2008-06-10 | Verigy (Singapore) Pte. Ltd. | Systems and methods for processing automatically generated test patterns |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6002878A (en) * | 1997-07-01 | 1999-12-14 | Motorola, Inc. | Processor power consumption estimator that using instruction and action formulas which having average static and dynamic power coefficients |
DE69838374T2 (en) * | 1997-12-23 | 2008-05-29 | Texas Instruments Inc., Dallas | Processor and method for reducing its power consumption |
-
2001
- 2001-05-19 GB GB0112276A patent/GB2375695B/en not_active Expired - Fee Related
-
2002
- 2002-05-02 WO PCT/GB2002/002034 patent/WO2002095574A1/en not_active Application Discontinuation
- 2002-05-02 US US10/478,168 patent/US7353411B2/en not_active Expired - Fee Related
-
2008
- 2008-01-30 US US12/011,891 patent/US20080178021A1/en not_active Abandoned
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5790874A (en) * | 1994-09-30 | 1998-08-04 | Kabushiki Kaisha Toshiba | Information processing apparatus for reducing power consumption by minimizing hamming distance between consecutive instruction |
US5572736A (en) * | 1995-03-31 | 1996-11-05 | International Business Machines Corporation | Method and apparatus for reducing bus noise and power consumption |
US5854935A (en) * | 1995-10-18 | 1998-12-29 | Nec Corporation | Program transformation system for microcomputer and microcomputer employing transformed program |
US6535984B1 (en) * | 1998-11-25 | 2003-03-18 | Texas Instruments Incorporated | Power reduction for multiple-instruction-word processors with proxy NOP instructions |
US6587939B1 (en) * | 1999-01-13 | 2003-07-01 | Kabushiki Kaisha Toshiba | Information processing apparatus provided with an optimized executable instruction extracting unit for extending compressed instructions |
US6725450B1 (en) * | 1999-06-21 | 2004-04-20 | Matsushita Electric Industrial Co., Ltd. | Program conversion apparatus, processor, and record medium |
US6807528B1 (en) * | 2001-05-08 | 2004-10-19 | Dolby Laboratories Licensing Corporation | Adding data to a compressed data frame |
US20030212914A1 (en) * | 2001-07-28 | 2003-11-13 | Paul Webster | Power efficiency in microprocessor systems |
US20050010830A1 (en) * | 2001-08-10 | 2005-01-13 | Paul Webster | Power reduction in microprocessor systems |
US20050229017A1 (en) * | 2001-08-10 | 2005-10-13 | At&T Corp. | Power reduction in microprocessor systems |
US7386844B2 (en) * | 2003-01-28 | 2008-06-10 | Matsushita Electric Industrial Co., Ltd. | Compiler apparatus and method of optimizing a source program by reducing a hamming distance between two instructions |
US7386777B2 (en) * | 2004-04-05 | 2008-06-10 | Verigy (Singapore) Pte. Ltd. | Systems and methods for processing automatically generated test patterns |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140173296A1 (en) * | 2012-12-14 | 2014-06-19 | International Business Machines Corporation | Chip level power reduction using encoded communications |
US9189051B2 (en) * | 2012-12-14 | 2015-11-17 | International Business Machines Corporation | Power reduction by minimizing bit transitions in the hamming distances of encoded communications |
Also Published As
Publication number | Publication date |
---|---|
US20050022041A1 (en) | 2005-01-27 |
GB0112276D0 (en) | 2001-07-11 |
GB2375695A (en) | 2002-11-20 |
US7353411B2 (en) | 2008-04-01 |
GB2375695B (en) | 2004-08-25 |
WO2002095574A1 (en) | 2002-11-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20080178021A1 (en) | Power efficiency in microprocessors | |
EP0994413B1 (en) | Data processing system with conditional execution of extended compound instructions | |
US7386844B2 (en) | Compiler apparatus and method of optimizing a source program by reducing a hamming distance between two instructions | |
US7546476B2 (en) | Power reduction in microprocessor systems | |
CN102609311B (en) | Device, method and system for selecting reception interrupted processor | |
US7107434B2 (en) | System, method and apparatus for allocating hardware resources using pseudorandom sequences | |
CN101676865A (en) | Processor and computer system | |
US20100138676A1 (en) | Microprocessors with improved power efficiency | |
US20020087833A1 (en) | Method and apparatus for distributed processor dispersal logic | |
US7076775B2 (en) | Power efficiency in microprocessor systems | |
JPH09114676A (en) | Program conversion method for microcomputer and microcomputer using the program | |
US20050010830A1 (en) | Power reduction in microprocessor systems | |
WO1990010267A1 (en) | Distributed pipeline control for a computer | |
CN100375012C (en) | Address offset generation within a data processing system | |
Sangireddy | Reducing rename logic complexity for high-speed and low-power front-end architectures | |
US20030101326A1 (en) | Reducing transitions on address buses using instruction-set-aware system and method | |
US20100005456A1 (en) | Compiling method, compiling apparatus and computer system for a loop in a program | |
CN115599445B (en) | Method for executing out-of-order instructions | |
US7275149B1 (en) | System and method for evaluating and efficiently executing conditional instructions | |
US7802077B1 (en) | Trace indexing via trace end addresses | |
CN114528021A (en) | Time-sharing multiplexing quantum measurement and control system and low-power-consumption high-efficiency quantum measurement and control compiling method | |
CN116860329A (en) | Compiling optimization method based on logical operation linear reconstruction | |
CN1410884A (en) | Guess execution command cancel device combined with exception treatment | |
Uht et al. | Levo: IPC in the 10‘s via Resource Flow Computing | |
Sima | 2 System Design |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION |