Method, system and medium for DMSA leakage optimization and time sequence maintenance
The PrimeTime tool sets the dont_touch attribute and optimized setup_margin value, which solves the problem of degradation in the timing performance of leakage current power consumption optimization in the existing technology, and achieves the balance of low power consumption and high performance, reducing timing violations and iterations.
Patent Information
- Application Number
- CN202411051140.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-01
- Publication Date
- 2024-11-26
AI Technical Summary
While optimizing leakage current power consumption, the prior art is difficult to maintain or improve the timing performance of the circuit, which easily leads to an increase in timing violations and the number of iterations.
The units of the special timing check path are extracted through the PrimeTime tool and the dont_touch attribute is set to avoid DMSA from changing these units during leakage optimization. At the same time, the value of the establishment time margin setup_margin is optimized. Use data statistics to calculate reasonable setup_margin value to control leakage power consumption .
It effectively reduces the occurrence of timing violations, quickly meets acceptance standards, meets the dual requirements of low power consumption and high performance, and reduces the number of iterations.
Smart Images

Figure CN119026541A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of chip development, and in particular to a method, system and medium for optimizing DMSA leakage and maintaining timing. Background Art
[0002] As the scale of integrated circuits continues to increase and the process technology continues to shrink, the power consumption of chips has become increasingly prominent, among which leakage power has become an important component that cannot be ignored. Traditional low-power design methods mainly focus on reducing dynamic power consumption, but as the process node enters the nanometer era, the proportion of leakage current power consumption has increased significantly, and in some cases even exceeds dynamic power consumption.
[0003] In order to reduce leakage current power consumption, people have developed technologies such as multi-threshold voltage, power gating, and body bias. However, while optimizing leakage current, these technologies often have a negative impact on the timing characteristics of the circuit, such as increasing circuit delay and reducing operating frequency. How to maintain or even improve the timing performance of the circuit while effectively reducing leakage current power consumption has become a key technical challenge that needs to be solved urgently.
[0004] At present, there are usually two ways in the industry to optimize leakage power consumption while maintaining timing: the first is to read the dont_touch attribute script spit out by the synthesis tool into the PrimeTime tool before power optimization, so that the tool does not replace these units; the second is to control the intensity of leakage power optimization by controlling the setup / hold margin variable, and replace the low threshold voltage unit with a high threshold voltage unit (HVT) on the non-timing-critical path to reduce leakage power consumption. When the reserved margin is large, the optimization intensity is small and it is difficult to reduce leakage power consumption. On the contrary, a large number of new timing violations are likely to emerge, resulting in an increase in the number of iterations.
[0005] The above-mentioned prior art has the following disadvantages: 1) The dont_touch attribute script relies on EDA tool automation, and the units that need to be dont_touch may be ignored. For example, in special timing checks (special_check), there are a large number of timing paths (timing paths usually consist of a startpoint, a wire delay, a unit delay, and an endpoint), which may cause new skew_check violations due to the replacement of timing units by DMSA leakage optimization. 2) The setup / hold margin setting depends entirely on previous project experience. When the margin setting is too large, the leakage optimization may be insufficient and the power consumption optimization may not be obvious; when the margin setting is too small, setup time timing violations may occur.
[0006] In summary, there is an urgent need for a design method that can optimize leakage current power consumption while taking into account timing performance to meet the dual requirements of modern integrated circuits for low power consumption and high performance. Summary of the invention
[0007] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method, system and medium for DMSA leakage optimization and timing maintenance. The leakage power consumption optimization solution provided by the present invention can optimize the leakage current power consumption while taking into account the timing performance, thereby meeting the dual requirements of modern integrated circuits for low power consumption and high performance.
[0008] To achieve the above objectives, the present invention provides the following technical solutions:
[0009] A method for optimizing DMSA leakage and maintaining timing, the method comprising the following steps:
[0010] Receive the special timing check list special_check list provided by the front end, extract the timing start and end information in the special_check list and write it into the preset first script file;
[0011] Use the timing analysis tool PrimeTime to obtain and execute the first script to obtain the timing report tmp_timing_path of all special timing checks special_check, wherein the timing report tmp_timing_path includes all timing path information that passes special_check and records the pin names of the standard cells in the timing path;
[0012] Extract all the standard cell names in the above timing report, set all the cells to the dont_touch attribute, and save them to the preset second script file;
[0013] PrimeTime DMSA is used to optimize leakage power consumption; wherein, the second script is read in using the remote execution command remote_execute, and the DMSA does not change the units in the second script when performing leakage optimization.
[0014] Furthermore, a tcl script is used to extract the timing start and end point information of the special_check list. The first script is a gen_special_timing_rpt script for generating a special timing report. The timing start and end point information is written into the aforementioned gen_special_timing_rpt script through the report_timing command. The content format is: report_timing-nosplit-from$startpoint-to$endpoint>>. / tmp_timing_path, where nosplit means that the timing report is displayed in a single line.
[0015] Further, the steps of extracting all standard cell names in the aforementioned timing report and setting all cells to the dont_touch attribute include:
[0016] According to the timing report tmp_timing_path obtained above, use the [get_cells-of_objects*pin_name*] command in the PrimeTime tool to get the standard cell names of all cells in all timing paths;
[0017] Configure all cells in the timing report to have the dont_touch attribute, with the following command: set_dont_touch[get_cells*cell_name*].
[0018] Furthermore, the second script reserves a special_dont_touch.tcl script for a special timing, and saves the unit configured as the dont_touch attribute into the special_dont_touch.tcl script.
[0019] Furthermore, before using PrimeTime DMSA for leakage power optimization, the following steps are also included:
[0020] Use the timing analysis tool PrimeTime to extract all timing paths whose slack is less than the preset slack threshold, and generate a timing report based on all extracted timing paths;
[0021] The slack values of all timing paths in the timing report are counted, and a reasonable slack value interval is determined according to the statistical results; the reasonable slack value interval is divided into multiple numerical intervals based on a preset time step, and a script is used to count the distribution of slack values in the reasonable slack value interval in each numerical interval;
[0022] According to a preset distribution range standard, a margin reference interval corresponding to the aforementioned distribution range standard is calculated, and a lower boundary value of the margin reference interval is obtained; the distribution range standard represents the proportion of slack in the aforementioned reasonable slack value interval falling into the aforementioned margin reference interval, and the margin reference interval is within the reasonable slack value interval;
[0023] According to the lower boundary value of the margin reference interval, the value of the setup time margin setup_margin required for leakage power optimization is calculated using the formula. The calculation formula is as follows:
[0024] $setup_margin=$lower_bound*$delta;
[0025] Among them, $lower_bound is the lower boundary value of the margin reference interval, and $delta is the set additional margin value, which is set by the system or the user.
[0026] Furthermore, the method of extracting all timing paths whose slack is less than the preset slack threshold is as follows:
[0027] Receive the margin threshold A set by the user, where A represents the time length in ns;
[0028] In the sta script of DMSA, add the additional option -slack_lesser_than A through the report_timing command. The final timing report will include all timing paths with slack less than A under all timing convergence status corners.
[0029] Furthermore, a distribution range standard collection column is configured, and the distribution range standard collection column is used to obtain the value of the distribution range standard set by the user, and send the set value to the numpy module in the python tool. The numpy module is used to count all slack data to obtain the margin reference interval corresponding to the aforementioned distribution range standard.
[0030] The present invention also provides a system for optimizing DMSA leakage and maintaining timing, the system comprising the following structure:
[0031] The timing start and end point information extraction module is used to receive the special timing check list special_check list provided by the front end, extract the timing start and end point information in the special_check list and write it into the preset first script file;
[0032] A timing report generation module, used to use the timing analysis tool PrimeTime to obtain and execute the aforementioned first script, to obtain the timing report tmp_timing_path of all special timing checks special_check, wherein the timing report tmp_timing_path includes all timing path information passing through special_check, and records the pin names of the standard cells in the timing path;
[0033] The unit attribute configuration module is used to extract all the standard unit names in the aforementioned timing report, and set all the units to the dont_touch attribute and save them into the preset second script file;
[0034] A leakage power consumption optimization module is used to optimize leakage power consumption through PrimeTime DMSA; wherein, the remote execution command remote_execute is used to read the aforementioned second script, and the DMSA does not change the units in the aforementioned second script when performing leakage optimization.
[0035] The present invention also provides a method for optimizing DMSA leakage and maintaining timing, wherein leakage power consumption is optimized by optimizing the value of setup time margin setup_margin, and the method comprises the following steps:
[0036] Use the timing analysis tool PrimeTime to extract all timing paths whose slack is less than the preset slack threshold, and generate a timing report based on all extracted timing paths;
[0037] The slack values of all timing paths in the timing report are counted, and a reasonable slack value interval is determined according to the statistical results; the reasonable slack value interval is divided into multiple numerical intervals based on a preset time step, and a script is used to count the distribution of slack values in the reasonable slack value interval in each numerical interval;
[0038] According to a preset distribution range standard, a margin reference interval corresponding to the aforementioned distribution range standard is calculated, and a lower boundary value of the margin reference interval is obtained; the distribution range standard represents the proportion of slack in the aforementioned reasonable slack value interval falling into the aforementioned margin reference interval, and the margin reference interval is within the reasonable slack value interval;
[0039] According to the lower boundary value of the margin reference interval, use the formula $setup_margin = $lower_bound*
[0040] $delta, the value of the setup time margin setup_margin is calculated, which is the optimized setup_margin value; wherein, $lower_bound is the lower boundary value of the margin reference interval, and $delta is the set additional margin value, which is set by the system or the user;
[0041] Based on the previously optimized setup_margin value, use PrimeTime DMSA to perform leakage power optimization.
[0042] The present invention also provides a computer-readable storage medium storing computer program instructions, which implement the aforementioned method when executed by a processor.
[0043] Due to the adoption of the above technical solution, the present invention has the following advantages and positive effects compared with the prior art, as an example:
[0044] On the one hand, the PrimeTime tool is used to extract all cells that have passed the special timing check special_check path, and the command set_dont_touch is used to instruct the tool not to replace these cells during power optimization, thereby maintaining the timing status of the special timing check special_check.
[0045] On the other hand, when optimizing leakage power consumption, since the value of setup time margin setup_margin determines the degree of power consumption optimization, the present invention proposes to use a statistical method to analyze the slack value to calculate a reasonable setup time margin setup_margin value, that is, to provide a reasonable leakage power consumption optimization strength by optimizing the value of setup time margin setup_margin. In this way, while satisfying the maximum power consumption optimization, it is possible to make the timing state less prone to a large number of timing violations after optimization, thereby effectively reducing the number of iterations and quickly reaching the acceptance criteria. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] Figure 1 A logic flow chart of a special timing check special_check for leakage power optimization protection provided in an embodiment of the present invention.
[0047] Figure 2 A logic flow chart for setting a setup time margin setup_margin value when optimizing leakage power consumption provided by an embodiment of the present invention.
[0048] Figure 3 A partial distribution diagram of slack values in a reasonable range of slack values in a timing report provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0049] The following is a further detailed description of the method, system and medium for optimizing DMSA leakage and maintaining timing disclosed in the present invention in conjunction with the accompanying drawings and specific embodiments. It should be noted that the techniques (including methods and devices) known to ordinary technicians in the relevant field may not be discussed in detail, but in appropriate cases, the above-mentioned known techniques are considered to be part of the specification. At the same time, other examples of the exemplary embodiments may have different values. The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the invention.
[0050] In the description of the embodiments of the present application, " / " means or, and "and / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" means: A and B exist alone, B exists alone, and A and B exist at the same time. In the description of the embodiments of the present application, "multiple" means two or more.
[0051] Terminology explanation:
[0052] 1) Corner: The state corner of timing convergence.
[0053] 2) PrimeTime: A timing analysis tool.
[0054] 3)Synopsys: An IC design tool.
[0055] 4)STA: static timing analysis.
[0056] 5)DMSA: Distributed Multi-Scenario Analysis.
[0057] 6) Special_check: special timing check.
[0058] 7) Setup / hold margin: setup / hold time margin.
[0059] 8)Skew_check: timing skew check.
[0060] Example
[0061] According to the shortcomings of the two solutions for optimizing leakage power consumption while maintaining timing in the current industry, the present invention proposes improvements. On the one hand, the present invention improves the existing dont_touch script that relies on EDA tool automation. The present invention extracts all the unit information of the special timing check special_check path through the PrimeTime tool, and uses the set_dont_touch command to rewrite all the units in the timing path into dont_touch attributes and store them in the preset script; when the DMSA performs leakage power consumption optimization in the later stage, these units are not changed, so that the special timing check special_check will not have new skew_check violations, so that the timing state of special_check can be maintained. On the other hand, the setup time margin setup_margin value required for leakage power consumption optimization is optimized to obtain a reasonable setup_margin value, that is, a reasonable leakage power consumption optimization strength is provided by optimizing the value of the setup time margin setup_margin. In this way, while satisfying the maximum power consumption optimization, the timing state can be prevented from having a large number of timing violations after optimization, thereby effectively reducing the number of iterations and quickly reaching the acceptance standard.
[0062] For details, see first Figure 1 As shown, a method for optimizing leakage power consumption and maintaining timing in DMSA (distributed multi-scenario analysis) provided by the present invention comprises the following steps.
[0063] S110, extracting the timing start / end point of the special inspection and writing it into the script in a specified format.
[0064] Specifically, according to the special timing check list special_check list provided by the front end, a tcl script can be used to extract all the timing start / end point information in the special timing check list special_check list, and write it into a preset first script file.
[0065] Preferably, the first script is a gen_special_timing_rpt script for generating a special timing report, and the timing start and end information can be written into the aforementioned gen_special_timing_rpt script through the report_timing command. The report_timing command is a timing report command commonly used in timing analysis, and the required report is generated using this command, and is often used to report a certain path or paths that share a specific node.
[0066] As an example of a typical method, the content format of the report_timing command may be as follows: report_timing-nosplit-from$startpoint-to$endpoint>>. / tmp_timing_path, where nosplit indicates that the timing report is displayed in a single line.
[0067] S120, use the PrimeTime tool to execute the script to obtain the timing report of all special checks.
[0068] Then, the timing analysis tool PrimeTime is used to obtain and execute the first script to obtain a timing report tmp_timing_path file of all special timing checks special_check. The timing report tmp_timing_path file may include all timing path information that has passed special_check.
[0069] Specifically, in the timing report, what is obtained are the pin names of the standard cells in the timing path.
[0070] S130, extract the timing report using a script and set all cells to the dont_touch attribute.
[0071] Specifically, all standard cell names in the aforementioned timing report may be extracted, and all cells are set to have a dont_touch attribute and then saved in a preset second script file.
[0072] Preferably, the second script is a special timing reservation special_dont_touch.tcl script, and the unit configured as the dont_touch attribute is saved in the special_dont_touch.tcl script.
[0073] In this embodiment, the specific steps of extracting all the standard cell names in the aforementioned timing report and setting all the cells to the dont_touch attribute can be as follows: according to the aforementioned timing report tmp_timing_path, use the [get_cells-of_objects*pin_name*] command in the PrimeTime tool to obtain the standard cell names of all cells in all timing paths; configure all the cells in the timing report to the dont_touch attribute, the command is as follows: set_dont_touch[get_cells*cell_name*].
[0074] S140, PrimeTime DMSA leakage power optimization. That is, PrimeTime DMSA is used to optimize leakage power. The remote execution command remote_execute is used to read the second script. When the DMSA performs leakage optimization, the unit in the second script is not changed.
[0075] In this way, the timing status of the special timing check special_check can be protected during leakage power optimization. For the special timing checklist special_checklist that records special_check data provided by the front-end design department, the timing start point (startpoint) and end point (endpoint) can be extracted from the list through a script and written into the first script file (as a design file). Then, after the PrimeTime tool reads the design file, the design file is executed through the source command to obtain the timing report tmp_timing_path of all timing paths containing special_check. The script is used to extract all cells in all timing paths of the timing report, set them to the dont_touch attribute, and save them in the special_dont_touch.tcl script file in the format of: set_dont_touch[get_cells-of_object*cell_name / pin*]; finally, when using DMSA for power optimization, the special_dont_touch.tcl script file is sourced, and these cells in the source special_dont_touch.tcl script file are not replaced during power optimization, so that the timing status of special_check can be maintained.
[0076] In another implementation of this embodiment, before using PrimeTime DMSA to optimize leakage power consumption, the setup time margin setup_margin value may also be optimized to obtain a reasonable setup_margin value, thereby providing a reasonable leakage power consumption optimization strength. Figure 2 As shown, steps S210 to S250 are included.
[0077] S210 , PrimeTime extracts all timing paths whose time slack is less than a preset slack threshold, and generates a timing report based on all the extracted timing paths.
[0078] Preferably, the method for extracting all timing paths whose time margin slack is less than a preset margin threshold is as follows: receiving the margin threshold A set by the user, where A represents the time length in ns; adding an additional option -slack_lesser_than A through the report_timing command in the sta script of DMSA, and the timing report finally generated includes all timing paths whose slack is less than A under all timing convergence status angles corner.
[0079] As an example, Figure 2 The preset slack threshold A=5.0ns, that is, extract the timing report with a time slack less than 5.0ns, and generate a timing report based on all the extracted timing paths.
[0080] It should be noted that the specific value of the preset margin threshold A can be set by the system by default, or can be set by the user in a personalized manner. Preferably, a margin threshold setting column is provided for the user to set the margin threshold.
[0081] S220, using a script to count the distribution of slack values in various numerical ranges.
[0082] Specifically, the slack values of all timing paths in the timing report can be counted by a preset data statistics script, and then a reasonable slack value interval can be determined based on the statistical results. Then, the reasonable slack value interval is divided into multiple numerical intervals based on a preset time step, and the script is used to count the distribution of the slack values in the above-mentioned reasonable slack value interval in each numerical interval.
[0083] During specific implementation, the method for determining the reasonable range of the slack value may be any one of the following methods.
[0084] Method 1: Analyze the slack value statistical results from small to large through the aforementioned data statistics script, and based on a preset slack allowable range (which can be set by the user), remove the slack values that exceed the slack allowable range (that is, remove unreasonable extreme values, including extremely large values and extremely small values), obtain the continuous distribution interval of the remaining slack values, and determine the continuous distribution interval as a reasonable slack value interval. As an example and not a limitation, in a certain application scenario, all slacks with slack values less than 5.0ns are analyzed through a data statistics script. After removing several extreme values that exceed the slack allowable range, the obtained reasonable slack value interval is -0.300ns to 0.300ns, that is, greater than or equal to -0.300ns and less than or equal to 0.300ns.
[0085] Method 2: Output the slack value statistics to the user, collect the user's interval selection operation (including the interval upper limit and the interval lower limit) for the above statistics, and use the interval selected by the user as the reasonable slack value interval. As an example but not a limitation, for example, according to the output slack value statistics chart, the user selects -0.300ns and 0.300ns as the interval upper limit and the interval lower limit, respectively, then the user-configured reasonable slack value interval can be obtained as -0.300ns to 0.300ns, that is, greater than or equal to -0.300ns and less than or equal to 0.300ns.
[0086] After determining the reasonable range of slack values, this reasonable range of slack values can be divided into multiple numerical ranges based on the preset time step t. As an example but not a limitation, for example, t = 0.01ns is taken as the step size. Then, a script is used to count the distribution of all slack values in the above-mentioned reasonable range of slack values -0.300ns to 0.300ns in each numerical range, and a distribution bar chart of slack values in each range is obtained. For example, Figure 3 shown.
[0087] exist Figure 3 In the figure, the horizontal axis represents the slack value in ns, and the vertical axis represents the frequency of the normalized slack value. The normalization method is: obtain the numerical interval with the highest frequency (number of times), and use the frequency (number of times) corresponding to the slack value in the numerical interval as the normalization base, as an example but not a limitation, Figure 3 The slack values that appear in the numerical range of 0.17 to 0.18 are the most, that is, the frequency is the highest. The frequency corresponding to the slack value in the numerical range of 0.17 to 0.18 is taken as the normalization base. After dividing the frequency (number of times) corresponding to the slack values in all numerical ranges by the aforementioned base, the slack value distribution information of each numerical range after normalization is obtained.
[0088] S230, using the numpy module in the python tool, most of the slack value data are located at the upper and lower boundaries of the interval.
[0089] Specifically, the margin reference interval corresponding to the aforementioned distribution range standard can be calculated according to the distribution range standard preset by the user, and the lower boundary value of the margin reference interval can be obtained. The distribution range standard represents the proportion of slack in the aforementioned reasonable slack value range that falls into the aforementioned margin reference interval, and the margin reference interval is within the reasonable slack value range. In this embodiment, the distribution range standard is preferably a ratio parameter greater than 90%.
[0090] Preferably, a distribution range standard collection column may be configured, and the distribution range standard collection column is used to obtain the value of the distribution range standard set by the user, and send the set value to the numpy module in the python tool. Then, all slack data are counted by the numpy module to obtain the margin reference interval corresponding to the aforementioned distribution range standard.
[0091] As an example but not limitation, taking the distribution range standard set by the user as 99.75%, according to the distribution range standard set by the user, it is necessary to find the interval range in which 99.75% of the slack values in the reasonable range of -0.300ns to 0.300ns fall, and this interval range is the margin reference interval. In specific implementation, the margin reference interval can be counted and calculated by using the numpy module in the python tool, as an example but not limitation, such as Figure 3 The margin reference interval corresponding to 99.75% is (0.0134~0.3), that is, the interval (0.0134~0.3) contains 99.75% of the slack values.
[0092] Combination Figure 3 It can also be seen from the histogram that 99.75% of the slack is located on the right side of the interval data (0.01-0.02), and the lower boundary value of the margin reference interval is within the numerical interval (0.01-0.02). The specific lower boundary value can be found to be 0.0134 through the numpy module.
[0093] It should be noted that when the margin reference interval calculated according to the distribution range standard set by the user is not unique (that is, it corresponds to multiple numerical intervals), the system can output multiple numerical intervals to the user, and collect one of the numerical intervals selected by the user as the target margin reference interval, and obtain the lower boundary value of the target margin reference interval for subsequent calculations.
[0094] S240, according to the lower boundary value of the interval, use a formula to calculate the value of setup_margin.
[0095] Specifically, according to the lower boundary value of the margin reference interval, the value of the setup time margin setup_margin required for leakage power optimization can be calculated using the following formula:
[0096] $setup_margin=$lower_bound*$delta
[0097] Among them, $lower_bound is the lower boundary value of the margin reference interval, and $delta is the set additional margin value, which is set by the system or the user.
[0098] As an example but not a limitation, taking $lower_bound=0.0134 and the user-preset additional margin value $delta=1.1 as an example, when using the PrimeTime tool for power optimization, the required setup_margin is $lower_bound*$delta, that is, 0.0134*1.1=0.0147ns≈0.015ns, and the 0.015ns is the reasonable setup_margin value after optimization.
[0099] S250, PrimeTime DMSA leakage power optimization.
[0100] Based on the setup_margin value obtained above, PrimeTime DMSA is used to optimize leakage power consumption, that is, the tool leaves a margin of 0.015ns during optimization.
[0101] The above solution provided by the present invention uses a script to count the distribution of slack values in a reasonable range (-0.3 to 0.3ns), uses the numpy module of python to count the range where 99.75% of the slack values are located, and uses the formula $setup_margin = $lower_bound*$delta to obtain the optimized value of setup_margin. While meeting the maximum power consumption optimization, it makes it difficult for the timing state to have a large number of timing violations after optimization, which is very useful in the later stages of the project, and can effectively reduce the number of iterations and quickly meet the acceptance criteria.
[0102] Another embodiment of the present invention also provides a system for optimizing DMSA leakage and maintaining timing.
[0103] The system includes a timing start point and end point information extraction module, a timing report generation module, a unit attribute configuration module and a leakage power consumption optimization module.
[0104] The timing start and end point information extraction module is used to receive the special timing check list special_check list provided by the front end, extract the timing start and end point information in the special_check list, and write it into the preset first script file.
[0105] The timing report generation module is used to use the timing analysis tool PrimeTime to obtain and execute the above-mentioned first script to obtain the timing report tmp_timing_path of all special timing checks special_check, and the timing report tmp_timing_path includes all timing path information passing special_check and records the pin names of standard cells in the timing path.
[0106] The unit attribute configuration module is used to extract all standard unit names in the aforementioned timing report, and set all units to dont_touch attributes and save them in a preset second script file.
[0107] The leakage power consumption optimization module is used to optimize leakage power consumption through PrimeTime DMSA; wherein, the remote execution command remote_execute is used to read the aforementioned second script, and the DMSA does not change the units in the aforementioned second script when performing leakage optimization.
[0108] In specific implementation, the timing start and end point information extraction module is configured as follows: using a tcl script to extract the timing start and end point information of the special_check list, the first script is a gen_special_timing_rpt script for generating a special timing report, and the timing start and end point information is written into the aforementioned gen_special_timing_rpt script through the report_timing command, and the content format is: report_timing-nosplit-from$startpoint-to$endpoint>>. / tmp_timing_path, where nosplit means that the timing report is displayed in a single line.
[0109] The cell attribute configuration module is configured as follows: according to the timing report tmp_timing_path obtained above, use the [get_cells-of_objects*pin_name*] command in the PrimeTime tool to obtain the standard cell names of all cells in all timing paths; configure all cells in the timing report as dont_touch attributes, the command is as follows: set_dont_touch[get_cells*cell_name*].
[0110] For other technical features, please refer to the description of the previous embodiment and will not be repeated here.
[0111] Another embodiment of the present invention further provides a method for optimizing DMSA leakage and maintaining timing, wherein the method optimizes leakage power consumption by optimizing the value of setup time margin setup_margin.
[0112] The method comprises the following steps: using a timing analysis tool PrimeTime to extract all timing paths whose time margin slack is less than a preset margin threshold, and generating a timing report based on all the extracted timing paths;
[0113] The slack values of all timing paths in the timing report are counted, and a reasonable slack value interval is determined based on the statistical results; the reasonable slack value interval is divided into multiple numerical intervals based on a preset time step, and a script is used to count the distribution of slack values in the aforementioned reasonable slack value interval in each numerical interval; according to a preset distribution range standard, a margin reference interval corresponding to the aforementioned distribution range standard is calculated, and the lower boundary value of the margin reference interval is obtained; the distribution range standard indicates the proportion of slack in the aforementioned reasonable slack value interval that falls into the aforementioned margin reference interval, and the margin reference interval is within the reasonable slack value interval; according to the lower boundary value of the margin reference interval, the formula $setup_margin=$lower_bound*$delta is used to calculate the value of the setup time margin setup_margin, which is the optimized setup_margin value; wherein, $lower_bound is the lower boundary value of the margin reference interval, and $delta is the set additional margin value, which is set by the system or the user; according to the aforementioned optimized setup_margin value, PrimeTime is used DMSA performs leakage power optimization.
[0114] Compared with the method in the previous embodiment, this embodiment optimizes leakage power consumption only by optimizing the value of setup time margin setup_margin. The principle of this optimization method is that when optimizing power consumption, scientific statistics are used to obtain a reasonable setup_margin value, which not only meets the maximum power consumption optimization, but also makes it less likely for the timing state to have a large number of timing violations after optimization. It is very useful in the later stages of the project, and can effectively reduce the number of iterations and quickly meet the acceptance criteria.
[0115] In this embodiment, the method for extracting all timing paths whose time margin slack is less than the preset margin threshold is as follows: receiving the margin threshold A set by the user, A represents the time length, and the unit is ns; in the sta script of DMSA, an additional option -slack_lesser_than A is added through the report_timing command, and the final timing report includes all timing paths whose slack is less than A under the state angle corner of all timing convergence.
[0116] In addition, a distribution range standard collection column is configured, and the distribution range standard collection column is used to obtain the value of the distribution range standard set by the user, and send the set value to the numpy module in the python tool. The numpy module is used to count all slack data to obtain the margin reference interval corresponding to the aforementioned distribution range standard.
[0117] For other technical features, please refer to the description of the previous embodiment and will not be repeated here.
[0118] Another embodiment of the present invention further provides a computer-readable storage medium storing computer program instructions, wherein the computer program instructions implement the aforementioned method when executed by a processor.
[0119] The storage medium may include any medium that can store program codes, such as a USB flash drive, a mobile hard disk, a read only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.
[0120] For other technical features, please refer to the description of the previous embodiment and will not be repeated here.
[0121] In the above description, the disclosure of the present invention is not intended to limit itself to these aspects. Rather, within the scope of the target protection of the present disclosure, the components can be selectively and operatively combined in any number. In addition, terms such as "include", "encompass" and "have" should be interpreted as inclusive or open by default, rather than exclusive or closed, unless they are explicitly defined to the contrary. All technical, scientific or other terms are consistent with the meanings understood by those skilled in the art, unless they are defined to the contrary. Common terms found in dictionaries should not be interpreted too idealistically or too impractically in the context of relevant technical documents, unless the present disclosure explicitly defines them as such. Any changes and modifications made by a person of ordinary skill in the field of the present invention based on the above disclosure belong to the scope of protection of the claims.
Claims
1. A method for optimizing DMSA leakage and maintaining timing, characterized in that Includes steps: Receive the special timing checklist special_checklist provided by the front end, extract the timing start and end information in the special_checklist and write it into the preset first script file; Use the timing analysis tool PrimeTime to obtain and execute the first script to obtain the timing report tmp_timing_path of all special timing checks special_check, wherein the timing report tmp_timing_path includes all timing path information that passes special_check and records the pin names of the standard cells in the timing path; Extract all the standard cell names in the above timing report, set all the cells to the dont_touch attribute, and save them to the preset second script file; PrimeTime DMSA is used to optimize leakage power consumption; wherein, the second script is read in using the remote execution command remote_execute, and the DMSA does not change the units in the second script when performing leakage optimization.
2. The method according to claim 1, characterized in that: A tcl script is used to extract the timing start and end point information of the special_checklist. The first script is a gen_special_timing_rpt script for generating a special timing report. The timing start and end point information is written into the aforementioned gen_special_timing_rpt script through the report_timing command. The content format is: report_timing-nosplit-from$startpoint-to$endpoint>>. / tmp_timing_path, where nosplit means that the timing report is displayed in a single line.
3. The method according to claim 1, characterized in that: The steps of extracting all the standard cell names in the aforementioned timing report and setting all the cells to the dont_touch attribute include: According to the timing report tmp_timing_path obtained above, use the [get_cells-of_objects*pin_name*] command in the PrimeTime tool to get the standard cell names of all cells in all timing paths; Configure all cells in the timing report to have the dont_touch attribute, with the following command: set_dont_touch[get_cells*cell_name*].
4. The method according to claim 3, characterized in that: The second script is a special_dont_touch.tcl script reserved for special timing, and the unit configured as the dont_touch attribute is saved in the special_dont_touch.tcl script.
5. The method according to claim 1, characterized in that: Before using PrimeTime DMSA for leakage power optimization, the following steps are also included: Use the timing analysis tool PrimeTime to extract all timing paths whose slack is less than the preset slack threshold, and generate a timing report based on all extracted timing paths; The slack values of all timing paths in the timing report are counted, and a reasonable slack value interval is determined according to the statistical results; the reasonable slack value interval is divided into multiple data intervals based on a preset time step, and a script is used to count the distribution of slack values in the above reasonable slack value interval in each data interval; According to a preset distribution range standard, a margin reference interval corresponding to the aforementioned distribution range standard is calculated, and a lower boundary value of the margin reference interval is obtained; the distribution range standard represents the proportion of slack in the aforementioned reasonable slack value interval falling into the aforementioned margin reference interval, and the margin reference interval is within the reasonable slack value interval; According to the lower boundary value of the margin reference interval, the value of the setup time margin setup_margin required for leakage power optimization is calculated using the formula. The calculation formula is as follows: $setup_margin=$lower_bound*$delta; Among them, $lower_bound is the lower boundary value of the margin reference interval, and $delta is the set additional margin value, which is set by the system or the user.
6. The method according to claim 5, characterized in that: The method of extracting all timing paths whose slack is less than the preset slack threshold is as follows: Receive the margin threshold A set by the user, where A represents the time length in ns; In the sta script of DMSA, add the additional option -slack_lesser_than A through the report_timing command. The final timing report will include all timing paths with slack less than A under all timing convergence status corners.
7. The method according to claim 5, characterized in that: A distribution range standard collection bar is configured, which is used to obtain the value of the distribution range standard set by the user and send the set value to the numpy module in the python tool. The numpy module is used to count all slack data to obtain the margin reference interval corresponding to the aforementioned distribution range standard.
8. A system for optimizing DMSA leakage and maintaining timing, characterized in that include: The timing start and end point information extraction module is used to receive the special timing check list special_checklist provided by the front end, extract the timing start and end point information in the special_check list and write it into the preset first script file; A timing report generation module, used to use the timing analysis tool PrimeTime to obtain and execute the aforementioned first script, to obtain the timing report tmp_timing_path of all special timing checks special_check, wherein the timing report tmp_timing_path includes all timing path information passing through special_check, and records the pin names of the standard cells in the timing path; The unit attribute configuration module is used to extract all the standard unit names in the aforementioned timing report, and save all the units into a preset second script file after setting the dont_touch attribute; A leakage power consumption optimization module is used to optimize leakage power consumption through PrimeTime DMSA; wherein, the remote execution command remote_execute is used to read the aforementioned second script, and the DMSA does not change the units in the aforementioned second script when performing leakage optimization.
9. A method for optimizing DMSA leakage and maintaining timing, wherein the leakage power consumption is optimized by optimizing the value of the setup time margin setup_margin, characterized in that Includes steps: Use the timing analysis tool PrimeTime to extract all timing paths whose slack is less than the preset slack threshold, and generate a timing report based on all extracted timing paths; The slack values of all timing paths in the timing report are counted, and a reasonable slack value interval is determined according to the statistical results; the reasonable slack value interval is divided into multiple data intervals based on a preset time step, and a script is used to count the distribution of slack values in the above reasonable slack value interval in each data interval; According to a preset distribution range standard, a margin reference interval corresponding to the aforementioned distribution range standard is calculated, and a lower boundary value of the margin reference interval is obtained; the distribution range standard represents the proportion of slack in the aforementioned reasonable slack value interval falling into the aforementioned margin reference interval, and the margin reference interval is within the reasonable slack value interval; According to the lower boundary value of the margin reference interval, the value of the setup time margin setup_margin is calculated using the formula $setup_margin=$lower_bound*$delta, which is the optimized setup_margin value; wherein $lower_bound is the lower boundary value of the margin reference interval, and $delta is the set additional margin value, which is set by the system or the user; Based on the previously optimized setup_margin value, use PrimeTime DMSA to perform leakage power optimization.
10. A computer-readable storage medium storing computer program instructions, characterized in that: The computer program instructions implement the method according to any one of claims 1 to 7 when executed by a processor, or the computer program instructions implement the method according to claim 9 when executed by a processor.
