KR102634255B1 - Dynamic adaptive buffer for distributed simulation and method controlling thereof - Google Patents

Abstract

The dynamic simulation buffer according to an embodiment of the present invention is a memory space in which simulation samples generated by executing a simulation for a submodel associated with a subsystem are stored. When a read operation is performed and simulation samples are stored in the memory space, the dynamic simulation buffer is a memory space in which simulation samples are stored. A first indicator (front) indicating a memory location changed according to a read operation, a second indicator (rear) indicating a memory location changed according to the write operation when a simulation sample is output from the memory space as a write operation is executed, and As the write operation and the read operation are performed, the position of the first indicator (rear) and the second indicator (front) corresponding to the write operation are moved to determine whether overflow or underflow has occurred, and the determination is made. and a memory space control unit that reallocates and rearranges the memory space of the simulation buffer or changes the end position of the simulation buffer according to the results.

Description

Dynamic simulation buffer for distributed simulation and its control method {DYNAMIC ADAPTIVE BUFFER FOR DISTRIBUTED SIMULATION AND METHOD CONTROLLING THEREOF}

The present invention relates to a dynamic simulation buffer for distributed simulation and its control method. More specifically, it improves reusability by minimizing model coupling in a dynamic simulation system, and improves packet delay variation and batch processing performance in a network environment. It relates to a dynamic simulation buffer for distributed simulation and its control method.

An autonomous vehicle refers to a vehicle that recognizes the surrounding environment and driving conditions while driving on its own without driver intervention, controls the vehicle's operation according to the recognition results, and drives itself to a preset destination.

Recently, much research has been conducted on the necessity and effectiveness of simulation for generating artificial intelligence learning data for autonomous driving. In particular, research is actively underway on simulation methods that distribute large computational throughput when reproducing large amounts of data in autonomous driving simulators.

Autonomous driving requires real cars to communicate with each other, but autonomous driving simulators process each module by distributing it to a separate simulator for performance. At this time, since a large amount of packets are moved between simulators and the simulation control synchronization time is important, it is required that very low communication delay be stably provided.

Additionally, existing object-oriented simulation approaches make optimizations such as multithreading difficult. Because objects are encapsulated by possessing property values, the property values of each object must be simulated by moving to different, discontinuous locations in memory. This is recognized as a factor that reduces overall simulation processing performance by spending a lot of time reading and writing attribute values from memory in actual simulation.

Recently, much research has been conducted on data-centric simulation methods, but there is a lack of research on simulation performance optimization methods in configuring variable simulation models that take into account packet delay variations in the network environment.

Registered Patent No. 10-2126270 (June 18, 2020) Registered Patent No. 10-2142876 (August 4, 2020) Registered Patent No. 10-2103430 (April 16, 2020)

The purpose of the present invention is to increase reusability by minimizing model coupling in a dynamic simulation system, and to provide a dynamic simulation buffer and its control method for distributed simulation to improve packet delay variation and batch processing performance in a network environment. do.

In addition, the present invention proposes a model and system dependency injection method through the input and output of a dynamic simulation buffer, and a dynamic simulation buffer and its The purpose is to provide a control method.

Additionally, the purpose of the present invention is to provide a dynamic simulation buffer for distributed simulation and a control method thereof that can increase the effect of improving service quality in a network environment through a dynamic simulation buffer algorithm.

Additionally, the purpose of the present invention is to provide a dynamic simulation buffer for distributed simulation and a method for controlling the same, which enables the configuration of a highly reusable subsystem and a distributed simulation system.

In addition, the purpose of the present invention is to provide a dynamic simulation buffer for distributed simulation and a control method thereof that enables management of experimental data based on a dynamic simulation buffer that can be applied and expanded in various fields.

The objects of the present invention are not limited to the objects mentioned above, and other objects and advantages of the present invention that are not mentioned can be understood by the following description and will be more clearly understood by the examples of the present invention. Additionally, it will be readily apparent that the objects and advantages of the present invention can be realized by the means and combinations thereof indicated in the patent claims.

A method of controlling a dynamic simulation buffer for distributed simulation to achieve this purpose includes the steps of running a simulation on a submodel associated with a subsystem to generate experimental data and storing it in the memory space of the dynamic simulation buffer. As a write operation and a read operation are performed, determining whether overflow or underflow has occurred by moving the position of the first indicator (rear) and the second indicator (front) corresponding to the write operation; Depending on the determination result, reallocating and rearranging the memory space of the dynamic simulation buffer or changing the end position of the dynamic simulation buffer.

In addition, the dynamic simulation buffer to achieve this purpose is a memory space in which simulation samples generated by executing simulations for submodels associated with the subsystem are stored. When the simulation samples are stored in the memory space as a read operation is executed, the dynamic simulation buffer is a memory space in which simulation samples are stored. A first indicator (front) indicating a memory location changed according to a read operation, a second indicator (rear) indicating a memory location changed according to the write operation when a simulation sample is output from the memory space as a write operation is executed, and As the write operation and the read operation are performed, the position of the first indicator (rear) and the second indicator (front) corresponding to the write operation are moved to determine whether overflow or underflow has occurred, and the determination is made. and a memory space control unit that reallocates and rearranges the memory space of the dynamic simulation buffer or changes the end position of the dynamic simulation buffer according to the results.

As described above, the present invention has the advantage of increasing reusability by minimizing model coupling in a dynamic simulation system and improving packet delay variation and batch processing performance in a network environment.

In addition, the present invention has the advantage of providing a model and system dependency injection method through the input and output of a dynamic simulation buffer, thereby improving CPU cache efficiency and improving simulation performance by batch executing continuous experimental data.

Additionally, according to the present invention, there is an advantage in that the effect of improving service quality in a network environment can be increased through a dynamic simulation buffer algorithm.

Additionally, according to the present invention, there is an advantage of being able to configure a highly reusable subsystem and a distributed simulation system.

Additionally, according to the present invention, there is an advantage in that experimental data can be managed based on a dynamic simulation buffer that can be applied and expanded in various fields.

Figure 1 is a block diagram for explaining a general dynamic simulation system.
Figures 2 and 3 are exemplary diagrams to explain execution assumptions of a general simulation buffer.
Figure 4 is a diagram for explaining a dynamic simulation system according to an embodiment of the present invention.
Figure 5 is a diagram for explaining a dynamic simulation buffer for distributed simulation according to an embodiment of the present invention.
Figures 5 and 6 are exemplary diagrams to explain the execution process of a dynamic simulation buffer for distributed simulation according to an embodiment of the present invention.
FIG. 7 is a diagram illustrating a process of sharing a dynamic simulation buffer for distributed simulation according to an embodiment of the present invention.

The above-mentioned objects, features, and advantages will be described in detail later with reference to the attached drawings, so that those skilled in the art will be able to easily implement the technical idea of the present invention. In describing the present invention, if it is determined that a detailed description of known technologies related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description will be omitted. Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the attached drawings. In the drawings, identical reference numerals are used to indicate identical or similar components.

Among the terms used in this specification, “simulation buffer” is a device for reading or writing between different systems or modules. A read operation to a buffer refers to an input operation, and a write operation to a buffer refers to an output operation.

Among the terms used in this specification, if no simulation samples are stored in the memory of the “simulation buffer,” the first indicator (rear) and the second indicator (front) identically indicate the initial location of the memory. In this state, when a simulation sample is input to the memory as a write operation occurs, the position of the first indicator (rear) is moved by the size of the input simulation sample, and as a read operation occurs, the simulation sample is output from the memory. The position of the second indicator (front) is moved by the size of the simulated sample.

Figure 1 is a block diagram for explaining a general dynamic simulation system. Figures 2 and 3 are exemplary diagrams to explain execution assumptions of a general simulation buffer.

Referring to Figure 1, the dynamic simulation system includes a simulation unit (1), a central model unit (2), and a simulation buffer (8).

The simulation unit (1) includes subsystems (i.e., artificial intelligence, dynamics, scenarios, graphics), and the submodels (AI model, vehicle model, traffic model, object model) associated with the subsystem among the central model unit (2) Generate experimental data by running a simulation.

At this time, the simulation unit 1 repeatedly performs simulations on submodels associated with the subsystem at the same cycle. That is, each subsystem of the simulation unit 1 generates experimental data by executing a simulation targeting the submodel, and stores the execution results of the experimental data again.

As described above, the experimental data for the sub-model can be reused, and the upper logic is determined according to the input/output flow 6 of the experimental data generated between the central model unit 2 and the simulation buffer 8. For example, experimental data may be needed for multiple objects in the driving environment, and considering simulation input/output time differences, it is advantageous to maintain it in a continuous memory space.

Simulation samples are stored in the simulation buffer 8. That is, the input/output experimental data at a specific time is defined as the simulation sample (7), and the component that manages the set of continuous time units of simulation samples is defined as the simulation buffer (8).

In the present invention, experimental data generated according to input and output between the central model unit 2 and the simulation buffer 8 are defined as simulation samples 7, and a module that manages a set of continuous time units of the simulation samples 7 is provided. Defined as simulation buffer (8).

The memory size of the simulation buffer 8 is determined by the size of the simulation sample and the number of simulation samples (buffer size), and the simulation sample may include all experimental data for each object.

In one embodiment, the simulation buffer 8 configures simulation samples in a memory that is continuous with experimental data of all objects to be processed at the same time (7), and sequentially configures simulation samples for the next time in the simulation buffer (8). ).

For example in Figure 2, simulation samples 9 plotting the position of an object and simulation samples 9 plotting the velocity of the object may be placed in a contiguous memory space of the simulation buffer 8. .

In the structure of the simulation buffer 8, simulation samples are placed in a continuous memory space, and the write operation 12 and the read operation 13 are configured to be performed at different times.

As the write operation 12 for the simulation buffer 8 is executed and a simulation sample is input from the memory space, the memory location indicated by the second indicator (rear) is changed according to the write operation 12, and a read operation is performed. As (13) is executed and the simulation sample is output to the memory space, the memory location indicated by the first indicator (front) is changed according to the read operation (13).

As described above, when the position of the first indicator (rear) changes due to the write operation 12 for the simulation buffer 8, the memory is instructed to start the read operation 13 up to the position of the first indicator (rear). The position (front) of the second indicator must move.

As such, an overflow may occur in the write operation 12 that exceeds the buffer size during the process of moving the position (front) of the second indicator to indicate the memory starting the read operation 13.

The memory size of the simulation buffer 8 is determined according to the size and number of simulation samples. At this time, the simulation sample may include experimental data for each object. Hereinafter, the operation process of the simulation buffer 8 will be described with reference to FIGS. 2 and 3.

As shown in Figure 2(a), simulation samples configuring the position of the object and simulation samples configuring the velocity of the object are arranged in a continuous memory space of the simulation buffer.

Thereafter, as shown in FIG. 2(b), the write operation 12 and the read operation 13 are configured to be performed at different times. As the write operation 12 is executed in the simulation buffer 8, when a simulation sample is input from the memory space, the memory location indicated by the second indicator (rear) changes according to the write operation 12, as shown in Figure 2 (c) ), the position of the first indicator (front) corresponding to the read operation 13 may be moved to the position of the second indicator (rear) corresponding to the write operation 12.

However, as shown in Figure 2(c), during the process of moving the second indicator (front) corresponding to the read operation 13 to the position of the first indicator (rear) corresponding to the write operation 12, the first indicator (rear) Since there is no memory space to move ), a write operation 12 that exceeds the buffer size may occur, resulting in an overflow 16 as shown in FIG. 2(c).

As described above, when experimental data is generated by the simulation unit 1, it is placed in the memory space as a simulation sample, and as the write operation 12 is executed in the simulation buffer 8, it is stored in the memory space. When a simulation sample is input, the memory location indicated by the first indicator (rear) changes according to the write operation 12, and the position of the first indicator (rear) corresponding to the write operation 12 is changed as shown in FIG. 2(c). The position of the second indicator (front) corresponding to the read operation 13 may be moved.

In this way, during the process of moving the position of the second indicator (front) corresponding to the read operation (13) to the position of the first indicator (rear) moved by the write operation (12), the write operation (12) exceeds the buffer size. ), overflow may occur.

To solve this problem, in a situation where overflow 16 occurs in a write operation that exceeds the simulation buffer size as described above, the second indicator (front) is available by the amount to be moved due to the read operation 13. A write operation may be performed in the available space 25. At this time, the available space 25 may correspond to the size to which the second indicator (front) must be moved due to the read operation.

To this end, the read operation is performed until the memory location indicated by the second indicator (front) reaches the end 26 of the simulation buffer, and then the insufficient simulation samples start again from the first location of the simulation buffer and cycle. (27).

However, if the write operation (12) and the read operation (13) are performed at regular intervals, they can cycle indefinitely, but the position of the first indicator (rear) corresponding to the write operation (12) is the position of the first indicator (rear) corresponding to the read operation (13). There may be an overflow situation that overtakes the position of the second indicator (front) again. Conversely, because the read operation is executed faster than the write operation, a situation in which the sample for which the read operation is to be performed may be empty (28) or underflow (29) may occur.

Figure 4 is a diagram for explaining a dynamic simulation system according to an embodiment of the present invention.

The dynamic simulation system of FIG. 4 can be applied to each of the transmitting simulator and the receiving simulator. Even if the dynamic simulation system is applied to each of the transmission simulator and the reception simulator, the clocks between the transmission simulator and the reception simulator are different, resulting in an error in data transmission and reception delay time. This error in data transmission and reception delay time can produce poor simulation results.

Therefore, the transmitting simulator records the time stamp (Ts) of the local time generated in one frame in the transmission packet and transmits it to the receiving simulator. Therefore, the reception simulator updates the delay time (D) using the time stamp (Ts) and reception time (Tr) recorded in the transmission packet. That is, the transmission simulator can calculate the delay time (D) based on [Equation 1] below.

[Equation 1]

D(n) = Tr(n) - Ts(n)

D: delay time,

Ts: Timestamp recorded on transmission packet

Tr: reception time

The reason for this is that the transmission simulator and the reception simulator are initialized at different times or include variable time for transmission preparation, network transmission time, etc., so the timeline of the transmission simulator and the timeline of the reception simulator need to be synchronized.

Therefore, the transmission simulator continuously tracks the delay time between the two clocks for each of the transmission simulator and the reception simulator, updates the relative offset (Offst) to the minimum value as in [Equation 2], and then uses [Equation 3] and Likewise, the time (T) of the receiving simulator's timeline is derived by adding an offset to the transmitter time stamp.

[Equation 2]

Offset(n) = min(D(n), Offset(n-1))

Offset: relative offset,

min(): Function to determine the minimum value among relative offsets,

D: delay time,

[Equation 3]

T(n) = Ts(n) + Offset(n)

T: Time of the receiving simulator's timeline

Offset: offset,

Ts: Timestamp recorded on transmission packet

Typically, the clocks between the transmitting and receiving simulators may run at slightly different rates, which can result in overflow or underflow of the simulator buffer.

Figure 5 is a diagram for explaining a dynamic simulation buffer for distributed simulation according to an embodiment of the present invention.

Referring to FIG. 5, the dynamic simulation buffer for distributed simulation includes a memory space 110, a first indicator 120, a second indicator 130, and a memory space control unit 140.

The memory space 110 stores simulation samples generated by executing simulations on submodels associated with subsystems.

When a simulation sample is output to the memory space as a read operation is performed, the first indicator 120 indicates a memory location changed according to the read operation.

The second indicator 130 indicates a memory location changed according to the write operation when the simulation sample is stored in the memory space as the write operation is performed.

As a write operation and a read operation are performed on the memory space 110, the memory space control unit 140 moves the positions of the first indicator (rear) and the second indicator (front) corresponding to the write operation to prevent overflow. Or, determine whether underflow has occurred.

In one embodiment, when a write operation on the dynamic simulation buffer is performed, the memory space control unit 140 determines the position of the second indicator (front) corresponding to the read operation up to the position of the first indicator (rear) corresponding to the write operation. is moved, and if the position of the second indicator (front) exceeds the size of the simulation buffer during the movement, it may be determined that an overflow has occurred.

In the above embodiment, the memory space control unit 140 reallocates the memory space of the simulation buffer and rearranges the memory state in case an overflow occurs in the simulation circular buffer. At this time, the size of the new simulation buffer is calculated by adding half of the requested size, and the buffer size is changed to the requested size. This is a reasonable choice to reduce the cost of frequent memory dynamic allocation and copying. At this time, the size of the new simulation buffer is determined based on [Equation 4] below.

[Equation 4]

In another embodiment, when a write operation on the dynamic simulation buffer is performed, the memory space control unit 140 displays a second indicator (rear) corresponding to the read operation before the position of the first indicator (rear) to be moved to the write operation. front), it can be determined that underflow has occurred.

In the above embodiment, the memory space control unit 140 may calculate a value corresponding to half the size of the dynamic simulation buffer as a delay estimate, and change the size of the dynamic simulation buffer by reflecting the delay estimate. That is, when underflow occurs in the circular simulation buffer, the memory space control unit 140 changes the end position of the simulation buffer to the requested size in order to reduce the size of the simulation buffer without reallocating memory.

At this time, due to the characteristics of the circular simulation buffer, it is known that the start point of the read operation operates more stably when the write position is at the halfway position of the buffer. In other words, the larger the simulation buffer size, the more the delay time increases, and the smaller it is, the more difficult it is to ensure stability, so the buffer size must be changed to reflect the current delay estimate. At this time, the current delay estimate is determined based on [Equation 5] below.

[Equation 5]

Figure 5 is an example diagram for explaining the execution process of a dynamic simulation buffer for distributed simulation according to an embodiment of the present invention.

Referring to Figures 5 and 6, the memory space control unit moves the positions of the first indicator (rear) and the second indicator (front) corresponding to the write operation as write and read operations are performed on the memory space. to determine whether overflow or underflow has occurred.

In one embodiment, when a write operation on the dynamic simulation buffer is performed, the memory space control unit moves the position of the second indicator (front) corresponding to the read operation to the position of the first indicator (rear) corresponding to the write operation. , If the position of the second indicator (front) exceeds the size of the simulation buffer during the process of movement, it may be determined that overflow has occurred.

In case an overflow occurs in the simulation circular buffer as shown in FIG. 5(a) (30), the memory space control unit reallocates the memory space of the simulation buffer as shown in FIG. 5(b) (31) and changes the memory state. Reorder (33). At this time, the size of the new simulation buffer is calculated by adding half of the requested size (31), and the buffer size is changed to the requested size (32). This is a reasonable choice to reduce the cost of frequent memory dynamic allocation and copying.

In another embodiment, when a write operation on the dynamic simulation buffer is performed, the memory space control unit sets a second indicator (front) corresponding to a read operation before the position of the first indicator (rear) moved by the write operation. If it is located, it can be determined that underflow has occurred, as shown in FIG. 6(a).

In the above embodiment, the memory space control unit may calculate a value corresponding to half the size of the dynamic simulation buffer as a delay estimate, as shown in FIG. 6(b), and change the size of the dynamic simulation buffer by reflecting the delay estimate. there is.

That is, when underflow occurs in the circular simulation buffer, the memory space control unit changes the end position of the simulation buffer to the requested size to reduce the size of the simulation buffer without reallocating memory (36).

FIG. 7 is a diagram illustrating a process of sharing a dynamic simulation buffer for distributed simulation according to an embodiment of the present invention.

Referring to Figure 7, the important purpose of simulation is to continuously simulate the model and verify the experimental model through the results. Experimenters need an experimental framework that can intuitively and repeatedly check results, such as a monitoring tool. Representative examples include a user interface that allows the experimenter to intervene in the simulation through an experiment framework, and visualization or data logging of experiment results.

These systems are used to perform simple read-only transfer of experimental data or for one-to-many flow control, and are extended to the data sharing buffer (40) and reference buffer (41) functions to enable input and output between simulation buffers.

When data in the shared buffer 40 of the embodiment is shared in the reference buffer 41, the sharing point is all managed references whenever a write operation 43 or a read operation 45 occurs in the shared buffer 40. The same data writing operation is performed in the buffer 41 (44, 46).

In addition, the write operation 48 of the reference buffer 41 is performed by the shared buffer 40 (47), but the read operation is performed by reading input data at the time of component execution (49).

Simulation logic is mainly performed between update models, and most input and output buffers are shared, so there is no difficulty in optimizing the number of buffers. However, in a distributed simulation environment, the overall system logic can become increasingly complex as the number of packet transmission/reception buffers and various output buffers for monitoring increases, so efforts are being made to reduce unnecessary input/output overhead by reducing the number of duplicate buffers and utilizing shared buffers. This is needed.

Although the present invention has been described with reference to limited embodiments and drawings, the present invention is not limited to the above embodiments, and various modifications and variations can be made by those skilled in the art from these descriptions. Accordingly, the spirit of the present invention should be understood only by the scope of the claims set forth below, and all equivalent or equivalent modifications thereof shall fall within the scope of the spirit of the present invention.

Claims (10)

In a method of controlling a dynamic simulation buffer for distributed simulation,
Running a simulation on a submodel associated with a subsystem to generate experimental data and storing it in the memory space of a dynamic simulation buffer;
As write and read operations are performed on the dynamic simulation buffer, the position of the first indicator (rear) corresponding to the write operation and the position of the second indicator (front) corresponding to the read operation are moved to cause overflow or underflow. determining whether has occurred; and
Characterized in that it includes the step of reallocating and rearranging the memory space of the dynamic simulation buffer or changing the end position of the dynamic simulation buffer according to the determination result.
Control method of dynamic simulation buffer for distributed simulation.
According to paragraph 1,
The step of determining whether overflow or underflow has occurred is
When a write operation on the dynamic simulation buffer is performed, moving the position of a second indicator (front) corresponding to a read operation to the position of a first indicator (rear) corresponding to the write operation; and
Characterized in that it includes the step of determining that an overflow has occurred if the size of the dynamic simulation buffer is exceeded during the process of moving the position of the second indicator (front).
Control method of dynamic simulation buffer for distributed simulation.
According to paragraph 1,
The step of reallocating and rearranging the memory space of the dynamic simulation buffer or changing the end position of the dynamic simulation buffer according to the determination result is
Characterized in that it includes the step of reallocating and rearranging the memory space of the dynamic simulation buffer using half of the request size due to the overflow and the request size.
Control method of dynamic simulation buffer for distributed simulation.
According to paragraph 1,
The step of determining whether overflow or underflow has occurred is
When a write operation to the dynamic simulation buffer is performed, if the second indicator (front) corresponding to the read operation is located before the position of the first indicator (rear) moved by the write operation, it is determined that an underflow has occurred. Characterized by comprising the steps
Control method of dynamic simulation buffer for distributed simulation.
According to paragraph 1,
The step of reallocating and rearranging the memory space of the dynamic simulation buffer or changing the end position of the dynamic simulation buffer according to the determination result is
Calculating a value corresponding to half the size of the dynamic simulation buffer as a delay estimate, and changing the size of the dynamic simulation buffer by reflecting the delay estimate.
Control method of dynamic simulation buffer for distributed simulation.
In the dynamic simulation buffer,
A memory space where simulation samples generated by running a simulation for a submodel associated with a subsystem are stored;
When a simulation sample is stored in the memory space as a read operation is performed, a second indicator (front) indicates a memory location changed according to the read operation;
When a simulation sample is output from the memory space as a write operation is performed, a first indicator (rear) indicates a memory location changed according to the write operation;
As the write and read operations are performed, the position of the first indicator (rear) corresponding to the write operation and the position of the second indicator (front) corresponding to the read operation are moved to determine whether overflow or underflow has occurred. Characterized by comprising a memory space control unit that determines, and reallocates and rearranges the memory space of the dynamic simulation buffer according to the determination result, or changes the end position of the dynamic simulation buffer.
Dynamic simulation buffer.
According to clause 6,
The memory space control unit
When a write operation to the dynamic simulation buffer is performed, the position of the second indicator (front) corresponding to the read operation is moved to the position of the first indicator (rear) corresponding to the write operation, and the second indicator (front) is moved to the position of the first indicator (rear) corresponding to the write operation. Characterized in that it is determined that an overflow has occurred if the size of the dynamic simulation buffer is exceeded during the process of moving the position.
Dynamic simulation buffer.
According to clause 6,
The memory space control unit
Characterized by reallocating and reordering the memory space of the dynamic simulation buffer using half of the request size due to the overflow and the request size.
Dynamic simulation buffer.
According to clause 6,
The memory space control unit
When a write operation to the dynamic simulation buffer is performed, if the second indicator (front) corresponding to the read operation is located before the position of the first indicator (rear) moved by the write operation, it is determined that an underflow has occurred. characterized by
Dynamic simulation buffer.
According to clause 6,
The memory space control unit
Calculating a value corresponding to half the size of the dynamic simulation buffer as a delay estimate, and changing the size of the dynamic simulation buffer by reflecting the delay estimate.
Dynamic simulation buffer.