Clock generator based on adjustable replication circuit
By designing an oscillator circuit with a configurable delay path in an integrated circuit and adjusting the clock signal frequency, the problem of minimizing the supply voltage and coping with process speed, voltage and temperature changes is solved, and more efficient circuit performance and stability is achieved.
Patent Information
- Application Number
- CN202280094029.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-30
- Publication Date
- 2024-11-08
AI Technical Summary
The prior art is difficult to effectively minimize the power supply voltage in integrated circuits, and at the same time deal with changes in process speed, voltage and temperature, resulting in complex adjustment of clock signal frequency, affecting circuit performance and stability.
Design an oscillator circuit with configurable delay paths, adjust the frequency of the clock pulse signal, respond to process, voltage and temperature changes, and decouple at reset and set inputs of the latch , to achieve frequency adjustment for each cycle.
It minimizes the power supply voltage of the integrated circuit, reduces the complexity of the clock signal frequency, improves the flexibility and stability of the circuit, and can respond to power disturbances instantly, reduces margin, and improves power and performance.
Smart Images

Figure CN118923041A_ABST
Abstract
Description
Background Art
[0001] In some embodiments of the present invention, the present invention relates to an adjustable clock generator circuit for providing a clock signal and minimizing the power supply voltage of an integrated circuit (IC) or a CPU core, and more specifically, to an oscillator circuit with a configurable delay path, but not exclusively, to a specific oscillator circuit. Summary of the invention
[0002] The object of the present invention is to provide an apparatus, a system, a computer program product and a method, which relate to an adjustable clock generator for minimizing the supply voltage of an integrated circuit or a CPU core, and more particularly, to an oscillator circuit with a configurable delay path, but not exclusively, to a specific oscillator circuit.
[0003] A clock generator circuit comprises: a gate having a first input electrically connected to the output of a first delay path circuit of another clock generator circuit and a second input electrically connected to the output of a second delay path circuit; an enable circuit having at least one input electrically connected to the output of the gate; a latch circuit comprising: a reset (R) having an input connected to the output of a third delay path circuit, a set (S) having an input connected to the output of the enable circuit, an inverted Q output connected to the input of an inverter, the output of the inverter being connected to the input of the third delay path circuit, the output of the inverter providing a clock pulse output (cp) signal of the clock generator circuit, and a Q output connected to the inverting input of the second delay path circuit; the gate merges the first input with the second input to adjust the frequency of the clock pulse output (cp) signal in response to the slowest delay path of the first path circuit or the second path circuit.
[0004] The enabling circuit may include: an inverter including an enable input (en); an OR gate including an input connected to the reset (R) and another input from the output of the inverter; and a NOR gate including an input from the output of the OR gate and the at least one input of the enabling circuit.
[0005] The frequency of the clock signal output (cp) may be adjustable in response to power supply disturbances in different locations of the integrated circuit. The gate may be a NAND gate. The first delay path circuit may be located in another area of the integrated circuit relative to the location of the second delay path circuit and the third delay path circuit. The outputs of the second delay path circuit and the third delay path circuit are decoupled at the reset and set inputs of the latch to enable the clock pulse output (cp) signal to be adjusted at each cycle of the frequency. The frequency of the clock pulse output (cp) signal may be reduced by a throttle operation unit in the integrated circuit. The frequency of the clock pulse output (cp) signal may be adjusted in response to a user machine interface signal. The frequency of the clock pulse output (cp) signal may be reduced by a macro request of the throttle operation unit in the integrated circuit.
[0006] A method for establishing a frequency of a clock generator circuit includes: combining a first output of a first delay path circuit of another clock generator circuit with a second output of a second delay path circuit of the clock generator circuit. The combining determines which of the first delay path circuit or the second delay path circuit is slower. The frequency of the clock pulse output (cp) signal is adjusted in response to the combining.
[0007] The first delay path circuit may be located in another area of the integrated circuit relative to the location of the second delay path circuit and the third delay path circuit of the clock generator circuit. The merging may be responsive to power supply disturbances in different locations of the integrated circuit. The second delay path circuit and the third delay path circuit of the clock generator circuit may be decoupled in response to the adjustment of the frequency of the clock pulse output (cp) signal. In response to the adjustment of the frequency of the clock pulse output (cp) signal, circuit protection bands may be enabled at different locations of the integrated circuit. The adjustment of the frequency of the clock pulse output (cp) signal may vary from one clock cycle of the clock pulse output (cp) signal to the next clock cycle to facilitate frequency hopping and efficient speed debugging of the integrated circuit. When the deceleration operation unit in the integrated circuit is turned on, the frequency of the clock pulse output (cp) signal may be reduced.
[0008] The merging of the output of the first delay path circuit of another clock generator circuit with the second output of the second delay path circuit of the clock generator circuit can eliminate the sensing of multiple voltage-frequency (VF) operating points in the integrated circuit by the clock generator circuit. This avoids the clock generator circuit from switching between the voltage-frequency (VF) operating points in the integrated circuit. The merging of the first delay path circuit of another clock generator circuit with the second output of the second delay path circuit of the clock generator circuit can be achieved by performing a logical "NAND" operation on the first output and the second output. The merging can obtain the slowest delay path from the first path circuit or the second path circuit.
[0009] According to one aspect, a process speed, voltage and temperature (PVT) responsive clock generator is implemented at multiple locations across an IC macro to generate an instantaneous variable cycle time clock signal. The average frequency of the instantaneous variable cycle time clock signal is controlled to meet a target. The target is provided so that the target does not exceed a previously calibrated value of the PVT clock generator in response to the supply voltage, temperature and process-related maximum operating frequency of the timing macro of the IC. If the provided target exceeds the error-free operation cycle time, the PVT responsive clock generator will issue a signal to indicate this to the power supply controller of the IC. The power supply system should increase the power supply to allow the PVT responsive clock generator to achieve the target and never exceed the error-free operation limit.
[0010] The above and other objects are achieved by the features claimed in the independent claims. Other implementations are apparent from the dependent claims, the description and the drawings.
[0011] Unless otherwise defined, the meaning of all technical and / or scientific terms used herein is the same as that known to those of ordinary skill in the art to which the present invention belongs. Although methods and materials similar or equivalent to the methods and materials described herein can be used in the practice or testing of embodiments of the present invention, exemplary methods and / or materials are described below. In the event of a conflict, this patent specification (including definitions) shall prevail. In addition, these materials, methods and examples are illustrative only and are not necessarily restrictive. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] Some embodiments of the present invention are described herein with reference to the accompanying drawings as examples only. Now specifically with reference to the accompanying drawings, it should be emphasized that the details shown are only for illustration and for the purpose of illustrative discussion of the embodiments of the present invention. In this regard, it is easy for those skilled in the art to understand how to practice the embodiments of the present invention based on the accompanying drawings.
[0013] In the attached picture:
[0014] Figure 1 A combination circuit and block diagram of an oscillator 10 in an integrated circuit (IC) provided by some embodiments are shown;
[0015] Figure 1A shows further details of a delay circuit provided by some embodiments;
[0016] Figure 2 A block diagram of an oscillator included in an integrated circuit (IC) provided by some embodiments is shown;
[0017] Figure 2A shows further details of a half-rate splitter circuit provided by some embodiments;
[0018] Figure 2B shows further details of the unit model provided by some embodiments;
[0019] Figure 2C shows a graph of frequency versus voltage for three configurable delay path cells according to some embodiments;
[0020] Figure 2D Shown is the method for implementing Figure 2C More details of the delay path circuitry of the delay path cells in each cell model are shown;
[0021] Figure 2E shows more details of the combiner circuit provided by some embodiments;
[0022] Figure 3 A flow chart of a method provided by some embodiments is shown. DETAILED DESCRIPTION
[0023] In some embodiments of the present invention, the present invention relates to an adjustable tunable clock generator for minimizing the supply voltage of an integrated circuit or a CPU core, and more particularly, but not exclusively, to an oscillator circuit with a configurable delay path.
[0024] Before explaining at least one embodiment of the present invention in detail, it should be understood that the present invention is not necessarily limited in its application to the details of the structure and arrangement of the components and / or methods set forth in the following description and / or illustrated in the drawings and / or examples. The present invention may have other embodiments or may be practiced or implemented in various ways.
[0025] The present invention is described in detail in the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products provided in conjunction with the embodiments of the present invention. It should be understood that each block of the flowchart and / or block diagram and the combination of blocks in the flowchart and / or block diagram can be implemented by computer-readable program instructions.
[0026] The present invention may be a system, a method and / or a computer program product. The computer program product may include one (or more) computer-readable storage media having computer-readable program instructions, the instructions being used to cause a processor to execute various aspects of the present invention.
[0027] The computer readable storage medium may be a tangible device capable of retaining and storing instructions for use by an instruction execution device. The computer readable storage medium may be, for example but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the above devices.
[0028] The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a corresponding computing / processing device, or downloaded to an external computer or external storage device via a network such as the Internet, a local area network, a wide area network, and / or a wireless network.
[0029] The computer-readable program instructions can be executed completely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or completely on a remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), and can also be connected to an external computer (e.g., by using the Internet of an Internet service provider). In certain embodiments, the electronic circuit including a programmable logic circuit, a field programmable gate array (FPGA) or a programmable logic array (PLA) etc. can execute the computer-readable program instructions by utilizing the state information of the computer-readable program instructions to customize the electronic circuit, thereby performing various aspects of the present invention.
[0030] The flow chart and block diagram in the figure show the architecture, function and operation of the possible implementation of the system, method and computer program product provided by various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a section or a partial instruction, which includes one or more executable instructions for implementing one or more specified logical functions. In some alternative implementations, the functions described in the box may not be in the order described in the figure. For example, in fact, the two boxes shown in succession can be executed almost at the same time, or sometimes can be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart and the box combination in the block diagram and / or flow chart can be implemented by a system based on special hardware, which performs specific functions or actions, or performs a combination of special hardware and computer instructions.
[0031] By the following introductory aspects of the present invention, an oscillator is described that utilizes multiple configurable distributed delay models located at various parts in an integrated circuit (IC). Distributed delay models can be provided from the same oscillator circuit 10 located at different parts of the IC. Other delay models can be used under the control of a controller to sense the process, voltage and temperature (PVT) of the entire integrated circuit (IC). In addition, an immediate (next cycle) response to power disturbances can be provided, thereby potentially reducing the required margin and improving the power and performance of the IC. The decoupling of the set S and reset R paths of the latch included in the oscillator supports tuning at each cycle of the clock pulse (in high phase or low phase) to provide agile operation in response to tuning from one cycle to the next cycle, which further provides speed debugging opportunities. Therefore, the delay path of the oscillator can be used as a process monitor and an aging monitor. The distribution of delay models at different locations of the IC enables power detection of the entire IC to capture the voltage frequency (voltage frequency, VF) curve and to calibrate delay paths of multiple voltage threshold (Vt) types. The distribution of delay models and operation of oscillators of the above features supports modeling of wire delays in integrated circuits, to support cycle-by-cycle compensation of voltage drops at distant sites in macros of an IC, and to support modeling of clock network effects and arbitrary logic behavior.
[0032] Reference now Figure 1, shows a combinational circuit and block diagram of an oscillator 10 in an integrated circuit (IC) provided by some embodiments. The oscillator 10 is used to generate a clock pulse cp, which can be used by other components in the integrated circuit (IC). The clock pulse cp provides the output of the inverter gate I2 at the node n2. The input of the inverter I2 is provided from the output of the latch 16 and the output of the NOR gate G4. The output of the delay circuit 12 is connected to the input of the OR gate G5 of the enable circuit 18 and the reset R of the latch 16. The input enable signal En is connected to the input of the inverter gate I1, and the output of the inverter gate I1 is connected to the second input of the OR gate G5. The output of the OR gate G5 is connected to the set S input of the latch 16. The latch 16 is shown as being implemented using two NOR gates G3 and G4. The latch 16 can be implemented using other types of gates. Reset R is connected to the input of NOR gate G3, the other input of which is connected to node n2, which is connected to the output of NOR gate G4 and the input of inverter gate I2. The output of NOR gate G3 is connected to the input of NOR gate G4 at node n1, the other input of which is the set S input of latch 16. Node n1 is connected to the input of inverter I3. The output of inverter I3 is connected to the input of delay circuit 14. The output of delay circuit 14 is connected to the input labeled I p2 The output of other delay circuits including other tunable or configurable oscillators (not shown) located in different areas of the integrated circuit (IC) may be applied to the input I of the NAND gate G1. p1 and I p3 The output of the NAND gate G1 is connected to the input of the NOR gate G2 included in the enable circuit 18. The input X of the NOR gate G2 can realize another plurality of inputs I from the output of other delay circuits of other tunable or configurable oscillators 10 via the NAND gates Gn. p4 to I p7 Among other things, other tunable or configurable oscillators 10 are included and located in different areas of an integrated circuit (IC).
[0033] Reference now Figure 1A, further details of the delay circuit 12 provided by some embodiments are shown. The delay circuit 12 implemented as shown can also be used to implement the delay circuit 14. As a non-limiting example and for the purpose of simplification, the delay circuit 12 includes a combinational logic portion of four two-input NAND gates and a trigger portion including four triggers. For the required delay range, any odd or even number of combinational logic gates and triggers can be used. In the combinational logic portion, input I is connected to one input of the NAND gate xdl3, and the other pin is connected to the clock CK. The output n3 of the NAND gate xdl3 is connected to one input of the NAND gate xdl2. The other input of the NAND gate xdl2 is connected to the configuration input cfg1. The output n2 of the NAND gate xdl2 is connected to one input of the NAND gate xdl1. The other input of the NAND gate xdl1 is connected to the clock CK. The output n1 of the NAND gate xdl1 is connected to one input of the NAND gate xdl0. The other input of the NAND gate xd0 is connected to the configuration input cfg0. The output O of the NAND gate xdl0 provides the output of the delay circuit 12. Table T1 shows possible inputs on each of the two inputs cfg0 and cfg1 of each of the four flip-flops. The fourth flip-flop has inputs a and b on two corresponding inputs, similarly the third flip-flop has inputs c and d, similarly the second flip-flop has inputs e and f, and similarly the first flip-flop has inputs g and h.
[0034] Included in the operation of the oscillator 10, the output of the latch 16 is inverted by the inverter I2, and the output of the inverter I2 is the generated clock output cp provided at the node n2. The phase delay circuit 12 delays the rise of the generated clock output cp to the rising edge of the "reset" R input applied to the latch 16. The rise of the generated clock output cp is adjustable or configurable. Together with the latch 16, the phase delay circuit 12 is actually a "one-time" circuit, which may only be in a high state during a period of time. The period is determined by the process speed, voltage and temperature (PVT) characteristics and the configuration data of the phase delay circuit 12.
[0035] Similarly, if phase delay circuit 14 is identical to phase delay circuit 12 and is identically controlled, then phase delay circuit 14 delays the falling edge of the output of latch 16 to the rising edge of the set S input of latch 16 in response to the reset R assertion of phase delay circuit 12. Therefore, phase delay circuit 14 is another "one-shot" circuit of equal delay. However, the output of phase delay circuit 14 will only remain low for a long time. The period during which the output of latch 16 remains low is determined by the delay of phase delay circuit 14. Although phase delay circuits 12 and 14 are identically controlled by a controller (not shown), the delays of these phase delay circuits may not be exactly equal due to power supply variations and other factors throughout the operating cycle of oscillator 10 and IC.
[0036] The set S input of latch 16 is supplied by the output of enable circuit 18. The set S input of latch 16 comes from NAND gate G2 which performs a NAND operation. The phase delay circuits of other delay elements in other delay circuit paths may also be adjustable or configurable before the operation time of the IC and may be controllable at the operation time of the IC. The phase delay circuits of other delay elements may be similar or identical to delay element 12. The outputs of other adjustable or configurable delay paths of delay circuits may be applied to input I p1 and I p3 The output of the latch circuit 16 at the node n2 is applied to the input Ip2 via the inverter I3 and the delay circuit 14. The NAND gate G2 is connected to the input I p1 ,I p2 and I p3 Perform a NAND operation so that the output of the enabling circuit 18 is only at the input I p1 ,I p2 and I p3 The "set" S signal is asserted to the latch 16 only when all paths supplied by the NAND gate have switched to 1. Therefore, the characteristic of the NAND gate G2 is that it merges the input I p1 ,I p2 and I p3 Input I p1 ,I p2 and I p3 The combination allows the frequency of the clock pulse output (cp) signal to be adjusted by responding to the input I p1 ,I p3 ,I p4 to I p7 and I p2 To configure the slowest delay path of the other delay elements provided, the input I p2 Supplied from the delay circuit 14 of the oscillator 10 .
[0037] Under the control of a controller included in the IC, the IC can be a CPU or can be included in the CPU. The frequency of the clock pulse output (cp) signal of the oscillator 10 can be reduced to cause a delay when the slow-motion operation unit in the integrated circuit IC is turned on. If the voltage drops, the delay will increase without the need to tune the phase delay circuits 12 and 14 or any phase delay paths at different locations in the IC. However, the user and / or the operating system may request a different frequency. On the other hand, when the slow-motion unit is turned on, the CPU macro itself may request a momentary slow speed. The user can change the clock pulse output (cp) signal in response to the user's request. When the macro of the IC or CPU requests to momentarily turn on the slow-motion operation unit in the integrated circuit, the frequency of the clock pulse output (cp) signal may be reduced.
[0038] Generally, according to the following description, the operation changes of the oscillator 10 and other delay paths at various other locations in the IC enable the other delay paths to be used as sensors or monitors in the IC. A feature of the other delay paths at various other locations in the IC is that they can be used to perform specific sensing or monitoring functions under the control of a controller included in the IC.
[0039] Reference now Figure 2 , a block diagram of an oscillator 10 included in an integrated circuit (IC) 20 provided by some embodiments is shown. For the purpose of the following discussion, the integrated circuit (IC) 20 may be further included in a CPU core (not shown). All components in the IC 20 share a controller 24, which includes control lines (indicated by dashed lines) connected to the oscillator 10, the half-rate splitter 26, and the unit models 22a and 22b. The outputs of the unit models 22a and 22b are connected to the input I of the oscillator 10 via the combiner 28. p1 Typically, multiple inputs I are provided on the oscillator 10. p1 and I pn The outputs of the cell models 22a and 22b are provided from phase delay circuits of other delay paths included in the cell models 22a and 22b. The other delay paths are located at different locations of the IC 20. The other delay paths may be similar or identical to the oscillator 10. The inputs of the cell models 22a and 22b are from the outputs of the half rate splitter 26. The inputs of the half rate splitter 26 are from the clock pulse output (cp) signal of the oscillator 10 via the buffer B2.
[0040] Reference now Figure 2A, further details of the half-rate separator 26 provided by some embodiments are shown. The half-rate separator 26 includes a clock pulse cp applied to one input of the NAND gate G6 and a rising edge enable (En) input of the D-type latch F1 at its input. The D input of the latch F1 receives the enable signal E. The Q output of the latch F1 is applied to the other input of the NAND gate G6, the input of the inverter gate I5, and the enable input of the flip-flop F2. The output (loc_cp) of the NAND gate G6 is applied to the rising edge input (cp) of the flip-flop F2. The Q output (loc_q) of the flip-flop F2 is fed back to the D input of the flip-flop F2 via the inverter I4, and fed back to one input of the AND gate G7 and the input of the NOR gate G8. The other input of the AND gate G7 is connected to the Q output of the latch F3. The D input of the latch F3 is connected to the input of the inverter I5 and the Q output of the latch F1. The level enable (E) input of the D-type latch F3 is connected to the clock pulse cp. The output of inverter I5 is connected to the other input of NOR gate G8. Half rate separator 26 provides two enable outputs en0 and en1 at the corresponding outputs of AND gate G7 and NOR gate G8. Clock pulse cp is a pulse signal with a PVT fixed width of about 50% duty cycle. The output of G6 enters the falling edge triggered flip-flop F2. The bubble at the cp input of flip-flop F2 indicates "inversion" or falling edge sensitivity, and the bubble at the Q output of flip-flop F2 indicates the same.
[0041] Reference now Figure 2B, further details of the cell model 22 provided by some embodiments are shown. The cell model 22a or the cell model 22b can be implemented as a cell model 22 including a pulse generator circuit 30. When the cell model 22a or the cell model 22b is implemented as the cell model 22, respectively, the enable outputs en0 and en1 of the half-separator circuit 26 can be applied to the corresponding inputs of the cell model 22a or the cell model 22b. Specifically, for the cell model 22, the pulse generator 30 includes an input connected to the clock signal cp. The clock signal cp is connected to the rising edge enable (En) input of the latches L1 and L2 and one input of the NAND gate G9. Enable en0 can be applied to the input of the latch L1. The output of the latch L1 is connected to the input of the latch L2 and the second input of the NAND gate G9. The output of the latch L2 is connected to the input of the inverter gate I6, and the output of the inverter gate I6 is connected to the third input of the NAND gate G9. The output of the NAND gate G9 is connected to the input of the buffer I7, and the output of the buffer gate I7 provides a pulse (pls) output, which is connected in parallel to the clock (clk) input of the delay path units 34U, 34L and 34S. The tuning control 32 from the controller 214 can be provided on a plurality of latches L3. The plurality of outputs of the latch L3 are connected to the configuration (cfg) input of each of the delay path units 34U, 34L and 34S. The three outputs of each delay path unit 34U, 34L and 34S are connected to the three corresponding inputs of the NAND gate G10.
[0042] Reference now Figure 2C, a frequency versus voltage graph 200 of three adjustable or configurable delay path units 34U, 34L, and 34S is shown according to some embodiments. The three adjustable or configurable delay path units 34U, 34L, and 34S may be included in the unit models 22a and 22b. Three scaling curves 341, 342, and 343 are shown for the corresponding delay path units 34S, 34L, and 34U for the corresponding standard voltage threshold (SVT), low voltage threshold (LVT), and ultralow voltage threshold (ULVT) regions. Arrows 341b, 342b, and 343b indicate low voltage fault points, medium voltage fault points, and high voltage fault points, respectively. The voltage axes marked at points 341a, 342a, and 343a indicate the supply voltage levels of the low performance power supply, the normal power supply, and the high performance power supply, respectively, at different locations of the IC. Prior to run time of the IC, a test time enables adjustment or configuration of the three delay path units 34U, 34L, and 34S. Three scaling curves 341, 342, and 343 are used to adjust the three adjustable or configurable delay path units 34U, 34L, and 34S. Included in the adjustment and subsequent run time, the IC delay path units 34U, 34L, and 34S are used to provide monitoring or sensing of frequency and power supply levels for the delay path units 34U, 34L, and 34S located at different locations of the IC.
[0043] The unit models 22a and 22b, which are respectively implemented as unit models 22, are used as additional loop paths. The additional loop path will delay the rising edge of the clock pulse cp to the rising edge (R2R). The unit models 22a and 22b delay the rising edge by generating short negative pulses (pulses pls) respectively. The pulses pls come from each pulse generator circuit 30 included in the unit model 22a and the unit model 22b. The pulses pls quickly reset the output of the loop path to zero, and then each pulse pls passes through adjustable or configurable delay path units 34U, 34L and 34S. The trailing edge (rising) of the pulse pls is delayed by the delay path units 34U, 34L and 34S and a rising edge is generated. The rising edge is tuned at the test time of the delay at a certain position representing the integrated circuit (IC).
[0044] At the run time of the IC, after tuning and the voltage threshold (Vt) gate constants derived from the test time are loaded into the controller 24, the average frequency of the clock pulses cp is monitored. The average frequency can be monitored by the controller 24 counting the clock pulses of the clock pulses cp in the window and comparing them with the external slow fixed frequency reference clock. Only one set of controls of the controller 24 can be modified at a time to reduce the clock frequency of the oscillator that is too high, because the oscillator 10 waits for the last path of the cell model 22a and 22b to rise. The criterion for reducing the clock frequency is that at any time, the clock period of the clock signal may not be shorter than the clock period defined by all other paths set at the test time. This criterion will prevent the clock of the macro of the IC 20 from running too fast, so that the clock of the macro of the IC 20 or the macro of the CPU fails functionally. In summary, in the cell model 22, the duration of each clock signal (cp) cycle is adjustable or configurable, and responds to process, voltage and temperature (PVT) through the adjustable or configurable delay path cells 34U, 34L and 34S in each cell model 22. The clock cycle duration is determined by the maximum delay (longest) of the delay path cells 34U, 34L and 34S located at different positions on the timing macro of the IC, wherein each of the delay path cells 34U, 34L and 34S is affected by the average supply voltage near it during the clock cycle.
[0045] Reference now Figure 2D , shows more details of the delay path circuit 34 provided by some embodiments for implementing the delay path units 34U, 34L and 34S in each unit model 22. The clock pulse cp input of the delay path circuit 34 is provided from the pulse pls. For example, the pulse pls comes from each pulse generator circuit 30 included in the unit model 22a and the unit model 22b. The input cfg00 of the NAND gates G19 and G17 is a plurality of outputs of the latch L3, and the plurality of outputs are connected to the configuration (cfg) input of each delay path unit in the delay path units 34U, 34L and 34S. The output Z of the multiplexer MX4 is the output of each delay path unit 34U, 34L and 34S.
[0046] The clock pulse cp input is connected to a series connection chain of inverters I10 to I17. The outputs of inverters I10 to I13 provide four inputs to multiplexer MX1, and the outputs of inverters I14 to I17 provide four inputs to multiplexer MX2. Multiplexers MX1 and MX2 are provided with two corresponding selection inputs S0 and S1, respectively. The outputs of multiplexers MX1 and MX2 provide two inputs for multiplexer MX3 that can be selected by selection input S2. Selection inputs S0, S1 and S2 are provided by controller 24. Selection inputs S0, S1 and S2 set the expected delay of delay path 34 under the control of controller 24. Selection inputs S0, S1 and S2 set the number of inversion delays added to the path. In binary code, S0 is the least significant bit (LSB), and binary 000 of S2, S1 and S0 represent zero, respectively, with only minimal overhead except for the contribution from inverter I1. Thus, binary 000 sets the fastest delay of delay path 34. Binary 111 of S2, S1, and S0 represents the addition of an additional seven inversions from inverters I11 through I17, thus representing the slowest delay of delay path 34.
[0047] The output of the multiplexer MX3 provides a selection input for the multiplexer MX4. The clock pulse cp input is connected to the input of the inverter I18, and the output (ckb) of the inverter I18 is connected to one input of the OR gate G18 and the OR gate G16. The other input of the OR gate G16 is connected to the selection input S0, and the other input of the OR gate G18 is connected to the output of the NAND gate G17. The output of the OR gate G16 is connected to the input of the NAND gate G17. The output of the OR gate G18 is connected to the input of the NAND gate G19. The input cfg00 to the other corresponding inputs of the NAND gates G19 and G17 is a plurality of outputs of the latch L3, and the plurality of outputs are connected to the configuration (cfg) input of each of the delay path units 34U, 34L and 34S. The two inputs s0b and s0bb to the multiplexer MX4 are provided from the NAND gate G17 and the NAND gate G19, respectively. The output (z) of multiplexer MX4 is the output of each delay path unit 34U, 34L and 34S.
[0048] In the operation delay path circuit 34, when the clock pulse cp switches from high (1) to low (0), both inputs of the output multiplexer MX4 are caused to be zero quickly (one gate delay). Therefore, regardless of the select input of the multiplexer MX4 provided from the output of the multiplexer MX3, the output Z switches to zero (0) (it is originally 1). After the clock pulse cp is zero (0) for a period of time, the value of the select node of the multiplexer MX4 (where S0 is its value) will cause the output Z of the multiplexer MX4 to be zero (0) anyway. When the clock pulse cp switches from high (1) to low (0), nothing happens except that the transition begins to propagate in the input chain of inverters I10 to I17 (8 gates). One of the nodes provided to the inputs of multiplexers MX1 and MX2 at the output of inverters I10 to I17 is selected by the select inputs S0, S1 and S2 as three configuration bits. The values of the configuration bits do not change during operation and only change during calibration of the delay path circuit 34. Thus, when the transition reaches a certain point, the transition propagates from multiplexer MX3 to the select input of multiplexer MX4 and switches output Z high. The exact delay depends on how many gates the signal must traverse before reaching the selected node in the series connection of inverters I10 to I17. Thus, the three selectable configuration bits provide eight delay possibilities. Since the selected inputs of multiplexers MX1 and MX2 may be rising (even number of gates) or falling (odd number of gates), the XOR gates will invert the falling edge to make the falling edge rising and not change the polarity of the even selection.
[0049] The goal is to keep the falling edge from rising because the goal is to get a configurable delayed rising edge of Cfg00, which if 0 forces both s0b and s0bb to be high (Z is high), making this delay path circuit 34 inactive. The XOR gate is implemented by multiplexer MX4 (when clock pulse cp=1, inputs s0b and s0bb of multiplexer MX4) with inputs s0b and s0bb being either 0-1 or 1-0, depending on the select input S0. The reliance on the select input S0 means that the output of MX3 in response to the rise of clock pulse cp will result in output Z rising. If the select input S0 is assigned, the opposite value of the same clock pulse cp rising will have the other polarity. However, the inputs s0b and s0bb of multiplexer MX4 are also inverted, producing the same result of Z rising in response to the rise of clock pulse cp. Only the delay depending on how many gates are traversed until the signal reaches the selected node in the series connection of inverters I10 to I17 will be different.
[0050] Reference now Figure 2E, showing more details of the combiner 28 provided by some embodiments. The clock route cp_rt typically supplies the clock pulse (cp) input of the flip-flops F4, F5 and F8. The clock route cp_rt is provided from the clock pulse output (cp) signal of the oscillator 10. However, the clock signal from the clock tree of the IC can be provided as a local clock signal to the unit model 22 located in various parts of the IC. The multiple unit models 22 located in various parts of the IC are shown as multiple rising edge to rising edge (R2R) paths 230, shown in dashed boxes. About Figure 2 , the inputs 220a and 220b of flip-flops F7 and F6 may be provided from the outputs of cell models 22a and 22b, respectively. An enable signal E is applied from controller 24 to the D input of flip-flop F4.
[0051] The output of G10 enables (E) combiner 28. The enable signal E is inverted by inverter I9, and the output of inverter I9 is connected to the clock data (cd) input of flip-flops F4 and F5. The enable signal E is also connected to one input of AND gate G11. The other input of AND gate G11 is connected to the Q output of flip-flop F4. The output of AND gate G11 is connected to the other multiple rising edge (R2R) paths 230 and the D input of flip-flop F5. The Q output (uEbf) is inverted by inverter I8 to obtain the inverted output uenb. The inverted output uenb is connected to the input of NOR gate G12. The other input of NOR gate G12 is connected to the Q output rdtrst of flip-flop F8. The output of NOR gate G12 is connected to the D input of flip-flop F8 and the input rdypls0 of AND gate G13. The Q output rdyrst of flip-flop F8 is connected to the input of AND gate G14. The other inputs of AND gates G13 and G14 are connected to the clock routing cp_rt. The output rdypls0 of the AND gate G13 is connected to the clock data (cd) input of the flip-flop F7. The output rdypls1 of the AND gate G14 is connected to the clock data (cd) input of the flip-flop F6. The outputs rdy0 and rdy1 of the Q outputs of the corresponding flip-flops F7 and F6 are connected to the two inputs of the OR gate G15. Figure 1 and Figure 2 In the context of OR gate G15, the output of the input I p1 Provided to the NAND gate G1 of the oscillator 10.
[0052] The D inputs of flip-flops F7 and F6 are connected to "1", so when clocked (falling edge, note the bubble on the cp input of flip-flops F7 and F6), output Q asserts "1". Output Q asserting "1" causes the next cycle on the oscillator located in the center (oscillator 10) to rise, which in turn resets the active flip-flop in flip-flop F6 or F7 via a pulse derived from the clock route cp_rt.
[0053] In other words, when the flip-flop F6 or F7 is clocked by a falling edge (the output of the NAND gate from the multiple rising edge (R2R) path 230), the output pin Q of the flip-flop F6 or F7 becomes 1. The output pin Q of the flip-flop F6 or F7 will remain at 1 until FF is reset by the reset pulse rdyrst0 or rdypls1. After the reset, the Q output will be 0 again, waiting for the next falling of the NAND gate of the multiple rising edge (R2R) path 230. In essence, the flip-flop F6 or F7 holds the information of the slowest path of the NAND gate from one of the three inputs of the multiple rising edge (R2R) path 230, which depends on the voltage at the location of the path. It may not be possible to tell in advance which path has completed its programmed delay and turned to 1. The unit center oscillator 10 now "agrees" that the master clock cp rises again. However, the clock route cp_rt will rise after all other oscillators 10 and their corresponding root paths have "agreed". Therefore, the flip-flop in each combiner 28 will hold this data until the clock cp actually rises. This rise will generate a reset pulse, which will clear the flip-flops F6 or F7 in all other combiners 28. The clearing of the flip-flops in all other combiners 28 is for even and odd cycles (rdy0, rdy1). Clearing is performed every two cycles, and the previous cycle and the following cycle are processed by another flip-flop F6 or F7 in each combiner 28. cp_rt is a pulse (positive) from the rising edge of the main clock cp. The E input comes from the controller 24, which selects to enable each combiner 28 or not to enable each combiner 28.
[0054] Reference now Figure 3 , a flow chart of a method 300 provided by some embodiments is shown. Prior to the run time step 301, as part of the manufacturing and testing process, a comprehensive test is run on a macro clocked by the oscillator 10 while changing the control of the adjustable or configurable delay paths of the cell models 22a and 22b. The input I connected to the oscillator 10 p1 and I p3 The control of the outputs of the cell models 22a and 22b is tuned so that the functional test of the macro will successfully pass at the fastest passable speed. The outputs of the cell models 22a and 22b are provided from phase delay circuits of other delay paths included in the cell models 22a and 22b. The other delay paths are located and connected at various locations in the IC 20. Only when the test passes will the delay be reduced, which means that the macro clocked by the oscillator 10 is working properly. An adjustable or configurable delay path that is too fast causes an unacceptable functional failure, and the tuning is restored to the transmission setting of the delay control provided by the controller 24. The details of how the delay path is adjustable or configurable are described in detail. Figure 1A The details of the delay circuit 12 shown are determined by the binary values of the selectable inputs a to h and the Figure 2BThe details of the cell model 22 shown in FIG. 1 are shown in FIG. The delay controlled transmission setting is performed at multiple voltage points in the IC to tune the delay path constructed by gates of different voltage thresholds (Vt). Among them, the threshold voltage is related to the gate speed and leakage classification. The delay controlled transmission setting effectively captures and protects the function of the oscillator 10 over the entire operating supply voltage range of the IC.
[0055] During the manufacturing and testing of IC20, the input I p1 ,I p2 and I p3 The combination with the control constants obtained for the controller 24 prior to step 301 enables the frequency of the clock signal output (cp) of the oscillator 10 to be configured. The configuration of the clock pulse output (cp) signal of the oscillator 10 is responsive to the input I p1 ,I p2 and I p3 The slowest delay path of the oscillator and oscillator 10 supplied on the clock. The duration of each clock pulse output cp cycle is adjustable or configurable and responds to process, voltage and temperature (PVT). The clock cycle duration is determined by the maximum delay (longest) of the paths located at various locations on the timing macro, and each path is affected by the average supply voltage near it during the clock cycle.
[0056] In step 301, the clock output cp of the oscillator 10 is combined with the outputs of the cell models 22a and 22b during the run-time of the oscillator 10 that has been previously tested, manufactured, and calibrated in the IC 20. The clock output cp of the oscillator 10 is further combined with the outputs of the delay circuits 12 and 14 alone or with the outputs of the delay circuits 12 and 14. p1 ,I p2 and I p3 The selection of the merge can be done by the controller 24 or the macro of the IC 20. In terms of circuit, the NAND gates G1 and Gn perform the merge step 301. Since the oscillator 10 uses multiple configurable distributed delay models to sense process, voltage, temperature (PVT), the merge step 301 eliminates the need to sense multiple voltage-frequency (VF) operating points and switch between voltage-frequency (VF) operating points in the IC 20.
[0057] Prior to run-time step 301 and during run-time, the average frequency is monitored and controlled by control circuit (24). The average frequency counts how many clock (cp) cycles are generated in a time period defined by a configurable number of reference clock cycles. The reference clock is a slow, external, fixed frequency clock. For example, a 2.5 GHz target cp frequency with a reference clock of 25 MHz means that there are 100 cp cycles per reference clock cycle. Therefore, in a typical average period of 100 reference cycles, 10,000 cp cycles are required. Numbers greater than this value will cause control (24) to increase the adjustable or configurable delay of delay circuits 12, 14, while smaller numbers will cause control (24) to decrease the delay of delay circuits 12, 14. Reasonable average periods may also be 1,000 or 10,000 reference clock cycles, with the expected cp cycle count being scaled accordingly.
[0058] In step 303, the frequency of the clock pulse output (cp) signal is adjusted in response to the merging step 301, so that the unit models 22a and 22b are used as additional loop paths because they delay the rising edge of the clock pulse cp to the rising edge (R2R). Since the unit models 22a and 22b delay the rising edge by generating a short negative pulse, the short negative pulse quickly resets the output of the loop path to zero, and then the short negative pulse passes through a set of adjustable or configurable delay elements. This set of configurable delay elements is provided by the delay circuits 12 and 14 and the delay circuits of other delay paths included in the unit models 22a and 22b. The trailing edge (rising) of the short negative pulse is delayed by a set of configurable delay elements and a rising edge is generated. The rising edge is configured in time to represent the delay of a certain position of the integrated circuit (IC). Therefore, the adjustment of the frequency of the clock output cp at step 303 is responsive to the configuration of the frequency of the clock pulse output (cp) signal of the oscillator 10 before step 301 and the power supply disturbances in different positions of the IC 20 during operation time.
[0059] The configuration of the clock pulse output before step 301 can be further enabled by decoupling the delay circuits 12 and 14 at the reset R input and the set S input of the latch 16 to enable the configuration of the clock pulse output (cp) signal on each cycle of the frequency. During runtime, when the slow-motion operation unit in the integrated circuit (IC) 20 is turned on to cause a delay, the frequency of the clock pulse output (cp) signal of the oscillator 10 may decrease. If the voltage drops, the delay will also increase without reconfiguring the clock pulse output. However, the user or the operating system may request a different frequency. On the other hand, when the slow-motion unit is turned on, the CPU macro itself may request a momentary slow speed. The user can change the clock pulse output (cp) signal in response to the user's request. When the macro of IC 20 requests to momentarily turn on the slow-motion operation unit in the integrated circuit 20, the frequency of the clock pulse output (cp) signal may decrease.
[0060] During run time, the local voltage drop of the logic at each oscillator at different locations on the IC 20 will cause the delay at the output of the cell models 22a and 22b to increase the rise and protect the logic. In addition, by not modeling the delay path of the logic of the integrated circuit (IC) 20, additional margin or protection band will be forced so that the logic will not fail even when such local voltage drops occur.
[0061] In summary, the oscillator 10 utilizes multiple configurable distributed delay models in the IC 20 to sense the process, voltage and temperature (PVT) of the entire IC 20, and provides an immediate (next cycle) response to power disturbances, thereby potentially reducing the required margin and improving the power and performance of the IC 20. Before step 301, the decoupling of the set S and reset R paths of the latch 16 supports configuration at each cycle of the clock pulse cp (at the high phase or low phase) to provide agile operation in response to the tuning from one cycle to the next cycle, which further provides speed debugging opportunities. In other words, the falling time of the clock pulse output signal cp is configured by different paths other than the period path and the rising path. In each cycle, when the path is not activated, configuration through different paths enables the controller 24 to change control. Therefore, before step 301, there is a configuration of the falling path after the fall and tuning, and the rising path after the rise but before the fall of the falling path. Therefore, the delay path of the oscillator 10 can act as a process monitor and an aging monitor during the operation time of the IC 20. The oscillators are distributed at different locations of the IC 20 so that the power supply detection of the entire IC 20 can capture the voltage frequency (VF) curve and can calibrate the delay paths of various voltage threshold (Vt) types.
[0062] Other systems, methods, features and advantages of the present invention will be or become apparent to those skilled in the art after studying the following drawings and detailed description. It is intended that all such other systems, methods, features and advantages be included in this description, within the scope of the present invention, and protected by the accompanying claims.
[0063] The description of the various embodiments of the present invention is for illustrative purposes only, and these descriptions are not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terms used herein are selected to best explain the principles, practical applications, or technological advances of the embodiments, or to enable other persons skilled in the art to understand the embodiments disclosed herein, compared to the technologies available on the market.
[0064] As used herein, the term "about" refers to ± 10%.
[0065] The terms "including," "having," and variations thereof mean "including but not limited to." This term encompasses the terms "consisting of" and "consisting essentially of.
[0066] The phrase "consisting essentially of" means that the composition or method may include additional ingredients and / or steps, but only if the additional ingredients and / or steps do not materially alter the basic and novel characteristics of the claimed composition or method.
[0067] As used herein, the singular forms "a", "an" and "the" include plural references unless the context clearly indicates otherwise. For example, the term "a compound" or "at least one compound" may include a plurality of compounds, including mixtures thereof.
[0068] The word “exemplary” is used herein to mean “serving as an example, instance, or illustration.” Any embodiment described as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments, and / or does not exclude the incorporation of features of other embodiments.
[0069] As used herein, the word “optionally” means “provided in some embodiments and not provided in other embodiments.” Any particular embodiment of the present invention may include a number of “optional” features, unless such features are mutually incompatible.
[0070] Throughout this application, embodiments of the present invention may be presented in a range format. It should be understood that the description of the range format is only for convenience and brevity, and should not be interpreted as a fixed limitation on the scope of the present invention. Therefore, the description of the range should be considered to have specifically disclosed all possible sub-ranges and individual numerical values within the range. For example, the description of a range such as from 1 to 6 should be considered to have specifically disclosed sub-ranges such as 1 to 3, 1 to 4, 1 to 5, 2 to 4, 2 to 6, 3 to 6, and individual numbers such as 1, 2, 3, 4, 5, and 6 within the range. This applies regardless of how wide the range is.
[0071] When a numerical range is indicated herein, any number (fractional or integer) listed within the indicated range is included. The phrases "a range between a first indicated number and a second indicated number" and "a range from a first indicated number to a second indicated number" are used interchangeably herein to include the first indicated number and the second indicated number and all fractions and integers between the two.
[0072] It should be understood that certain features of the present invention described in the context of a separate embodiment for the sake of clarity may also be provided in a single embodiment by combination. Conversely, for the sake of brevity of description, the various features of the present invention described in the context of a single embodiment may also be provided individually or in any suitable sub-combination or as any suitable other embodiment of the present invention. Some features described in the context of various embodiments are not considered to be essential features of these embodiments, unless the embodiment is inoperable without these elements.
[0073] The purpose of the applicant is that all publications, patents and patent applications mentioned in this specification are incorporated into this specification by reference in their entirety, just as each individual publication, patent or patent application is specifically and individually indicated when it is mentioned to be incorporated by reference herein. In addition, the citation or identification of any reference in this application should not be construed as an admission that the reference is available as a prior art of the present invention. With regard to the use of section headings, section headings should not be construed as necessary limitations. In addition, any one or more priority documents of the present application are incorporated into this application by reference in their entirety.
Claims
1. A clock generator circuit, characterized in that: include: a gate having a first input electrically connected to an output of a first delay path circuit of another clock generator circuit and a second input electrically connected to an output of a second delay path circuit; an enable circuit having at least one input electrically connected to an output of the gate; A latch circuit, the latch circuit comprising: reset (R) having an input connected to the output of the third delay path circuit, Set (S), having an input connected to the output of the enable circuit, an inverting Q output connected to an input of an inverter, wherein the output of the inverter is connected to an input of the third delay path circuit, wherein the output of the inverter provides a clock pulse output (cp) signal of the clock generator circuit, a Q output connected to an inverting input of the second delay path circuit; The gate combines the first input and the second input to adjust the frequency of the clock pulse output (cp) signal in response to the slowest delay path of the first path circuit or the second path circuit.
2. The clock generator circuit according to claim 1, characterized in that: The enabling circuit comprises: An inverter including an enable input (en); an OR gate including an input connected to said reset (R) and another input from said output of said inverter; A NOR gate includes an input from the output of the OR gate and the at least one input of the enable circuit.
3. The clock generator circuit according to any one of the preceding claims, characterized in that The frequency of the clock signal output (cp) may be adjusted in response to power supply disturbances in different locations of an integrated circuit.
4. The clock generator circuit according to any one of the preceding claims, characterized in that The gate is a NAND gate.
5. The clock generator circuit according to any one of the preceding claims, characterized in that The first delay path circuit is located in another area of the integrated circuit relative to the second delay path circuit and the third delay path circuit.
6. The clock generator circuit according to any one of the preceding claims, characterized in that The outputs of the second and third delay path circuits are decoupled at the reset and set inputs of a latch to enable adjustment of the clock pulse output (cp) signal at each cycle of the frequency.
7. The clock generator circuit according to any one of the preceding claims, characterized in that The frequency of the clock pulse output (cp) signal is reduced by a slow operation unit in the integrated circuit.
8. The clock generator circuit according to any one of the preceding claims, characterized in that The frequency of the clock pulse output (cp) signal is adjusted in response to a user machine interface signal.
9. The clock generator circuit according to any one of the preceding claims, characterized in that The frequency of the clock pulse output (cp) signal is reduced by a macro request of the slow operation unit in the integrated circuit.
10. A method for establishing the frequency of a clock generator circuit, characterized in that include: combining a first output of a first delay path circuit of another clock generator circuit with a second output of a second delay path circuit of the clock generator circuit, wherein the combining determines which of the first delay path circuit or the second delay path circuit is the slowest delay path; The frequency of the clock pulse output (cp) signal is adjusted in response to the combining.
11. The method according to claim 10, characterized in that The first delay path circuit is located in another area of the integrated circuit relative to the second and third delay path circuits of the clock generator circuit, wherein the merging is responsive to a power supply disturbance in a different location of the integrated circuit.
12. The method according to any one of the above method claims, characterized in that Also includes: In response to the adjusting, the second delay path circuit and the third delay path circuit of the clock generator circuit are decoupled.
13. The method according to any one of the above method claims, characterized in that Also includes: In response to the adjusting, circuit protection bands are enabled at different locations of the integrated circuit.
14. The method according to any one of the above method claims, characterized in that The adjustment varies from one clock cycle to the next clock cycle of the clock pulse output (cp) signal, facilitating frequency hopping and efficient speed debugging of the integrated circuit.
15. The method according to any one of the above method claims, characterized in that When a slow operation unit in the integrated circuit is turned on, the frequency of the clock pulse output (cp) signal is reduced.
16. The method according to any one of the above method claims, characterized in that The merger also includes: Eliminating sensing of a plurality of voltage-frequency (VF) operating points in the integrated circuit by the clock generator circuit; The clock generator circuit is thereby prevented from switching between the voltage-frequency (VF) operating points in the integrated circuit.
17. The method according to any one of the above method claims, characterized in that The merging is performed by performing a logic “NAND” operation on the first output and the second output to obtain the slowest delay path from the first path circuit or the second path circuit.
