阅读说明：本技术 矢量加速器及其控制方法、装置 (Vector accelerator and control method and device thereof ) 是由 刘玉佳 陈钦树 朱伏生 朱晓明 卢华 于 2021-11-08 设计创作，主要内容包括：本申请公开了一种矢量加速器及其控制方法、装置,矢量加速器包括矢量超前运行单元和矢量执行单元,所述矢量超前运行单元预取所述矢量执行单元进行矢量运算的矢量数据,所述矢量加速器的控制方法包括：在矢量运算的过程中,从矢量超前运行单元获取目标矢量数据；将所述目标矢量数据发送至矢量执行单元,以供所述矢量执行单元接收到所述目标矢量数据后执行矢量运算。解决了矢量加速器进行矢量运算时的耦合度较高的技术问题,提高了矢量加速器进行矢量运算的速度。(The application discloses a vector accelerator and a control method and device thereof, the vector accelerator comprises a vector advance operation unit and a vector execution unit, the vector advance operation unit prefetches vector data of vector operation performed by the vector execution unit, and the control method of the vector accelerator comprises the following steps: in the process of vector operation, target vector data are obtained from a vector advance operation unit; and sending the target vector data to a vector execution unit so that the vector execution unit can execute vector operation after receiving the target vector data. The technical problem of high coupling degree when the vector accelerator performs vector operation is solved, and the speed of performing vector operation by the vector accelerator is improved.)

1. A control method of a vector accelerator, characterized in that the vector accelerator comprises a vector look-ahead execution unit and a vector execution unit, the vector look-ahead execution unit prefetches vector data for vector operations by the vector execution unit, the control method of the vector accelerator comprising:

in the process of vector operation, target vector data are obtained from a vector advance operation unit;

and sending the target vector data to a vector execution unit so that the vector execution unit can execute vector operation after receiving the target vector data.

2. The method of controlling a vector accelerator according to claim 1, wherein the step of obtaining target vector data from the vector look-ahead unit is preceded by the step of:

receiving a vector accelerator configuration instruction, and configuring a vector length and a register number based on the vector accelerator configuration instruction;

receiving a starting vector operation instruction, and acquiring a data acquisition instruction cached in a cache region based on the starting vector operation instruction;

acquiring vector data according to the acquired data acquisition instruction cached in the cache region;

and storing the vector data in the vector look-ahead operation unit.

3. The method for controlling a vector accelerator according to claim 2, wherein the step of fetching vector data according to the fetched data fetch instruction cached in the cache region comprises:

acquiring address information of the vector data carried by the data acquisition instruction cached in the cache region, register information for storing the vector data and data length information;

and acquiring the vector data from the cache region according to the address information, the register information and the data length information.

4. The method for controlling a vector accelerator according to claim 3, wherein the step of obtaining address information of the vector data carried by the data fetch instruction cached in the cache area, register information for storing the vector data, and data length information comprises:

and analyzing the data acquisition instruction cached in the cache region to obtain address information of the vector data, register information for storing the vector data and data length information.

5. The control method of the vector accelerator according to claim 1, wherein the step of acquiring target vector data from the vector advance operation unit during the vector operation comprises:

during the vector operation, determining to acquire target vector data for executing the process from the vector data cached by the vector advance operation unit according to the process of vector operation.

6. The method of controlling a vector accelerator according to claim 1, wherein before the step of sending the target vector data to a vector execution unit for the vector execution unit to execute a vector operation after receiving the target vector data, the method comprises:

receiving a configuration instruction of a central processing unit;

and setting the number of registers of the vector accelerator according to the configuration instruction.

7. The method of controlling a vector accelerator of claim 6, wherein the step of sending the target vector data to a vector execution unit is followed by further comprising:

and acquiring an execution process of the vector operation executed by the vector execution unit, and determining the data volume of the target vector data sent to the vector execution unit according to the execution process.

8. A control apparatus of a vector accelerator, characterized by comprising:

the acquisition module is used for acquiring target vector data from the vector advance operation unit in the vector operation process;

and the sending module is used for sending the target vector data to a vector execution unit so that the vector execution unit can execute vector operation after receiving the target vector data.

9. A vector accelerator, characterized in that it comprises a vector accelerator, a memory and a control program of the vector accelerator stored on the memory and operable on the vector accelerator, which control program, when executed by the vector accelerator, implements the steps of the control method of the vector accelerator according to any one of claims 1 to 7.

10. The vector accelerator of claim 9, wherein the vector accelerator is connected to a central processing unit, receives an instruction from the central processing unit, and executes a control operation corresponding to the instruction.

Technical Field

The present application relates to the field of computer technologies, and in particular, to a vector accelerator and a control method and apparatus thereof.

Background

With the rapid development of artificial intelligence, many application programs include a relatively high percentage of vector operations, and in the process of running applications, an arithmetic unit needs to execute a large number of vector operations. The existing vector operator, such as a Cray-1 type vector accelerator, has low decoupling degree between vector data access and vector execution, and reduces the operation speed of the vector accelerator.

Disclosure of Invention

The embodiment of the application aims to reduce the coupling degree of a vector accelerator during vector operation by providing the vector accelerator and a control method and device thereof.

To achieve the above object, an aspect of the present application provides a control method of a vector accelerator, the vector accelerator including a vector advance execution unit and a vector execution unit, the vector advance execution unit prefetching vector data for vector operations by the vector execution unit, the control method comprising:

in the process of vector operation, target vector data are obtained from a vector advance operation unit;

and sending the target vector data to a vector execution unit so that the vector execution unit can execute vector operation after receiving the target vector data.

Optionally, before the step of obtaining the target vector data from the vector advance operation unit, the method further includes:

receiving a vector accelerator configuration instruction, and configuring a vector length and a register number based on the vector accelerator configuration instruction;

receiving a starting vector operation instruction, and acquiring a data acquisition instruction cached in a cache region based on the starting vector operation instruction;

acquiring vector data according to the acquired data acquisition instruction cached in the cache region;

and storing the vector data in the vector look-ahead operation unit.

Optionally, the step of obtaining vector data according to the obtained data obtaining instruction cached in the cache region includes:

acquiring address information of the vector data carried by the data acquisition instruction cached in the cache region, register information for storing the vector data and data length information;

and acquiring the vector data from the cache region according to the address information, the register information and the data length information.

Optionally, the step of obtaining address information of the vector data carried by the data obtaining instruction cached in the cache region, register information for storing the vector data, and data length information includes:

and analyzing the data acquisition instruction cached in the cache region to obtain address information of the vector data, register information for storing the vector data and data length information.

Optionally, the step of acquiring target vector data from the vector advance operation unit during the vector operation includes:

during the vector operation, determining to acquire target vector data for executing the process from the vector data cached by the vector advance operation unit according to the process of vector operation.

Optionally, before the step of sending the target vector data to a vector execution unit for the vector execution unit to execute vector operation after receiving the target vector data, the method includes:

receiving a configuration instruction of a central processing unit;

and setting the number of registers of the vector accelerator according to the configuration instruction.

Optionally, after the step of sending the target vector data to a vector execution unit, the method further includes:

and acquiring an execution process of the vector operation executed by the vector execution unit, and determining the data volume of the target vector data sent to the vector execution unit according to the execution process.

In order to achieve the above object, the present invention also provides a control device for a vector accelerator, including:

the acquisition module is used for acquiring target vector data from the vector advance operation unit in the vector operation process;

and the sending module is used for sending the target vector data to a vector execution unit so that the vector execution unit can execute vector operation after receiving the target vector data.

In addition, in order to achieve the above object, the present invention further provides a vector accelerator, which includes a vector accelerator, a memory, and a control program of the vector accelerator stored on the memory and operable on the vector accelerator, wherein when the control program of the vector accelerator is executed by the vector accelerator, the steps of the control method of the vector accelerator are implemented as described above.

Optionally, the vector accelerator establishes a connection with a central processing unit, receives an instruction of the central processing unit, and executes a control operation corresponding to the instruction.

In the embodiment, when the vector processor executes the vector operation, in the process of the operation of the vector accelerator, the vector processor acquires the target vector data from the vector advance operation unit and sends the target vector data to the vector execution unit, so that the vector execution unit executes the vector operation after receiving the target vector data. When the vector accelerator performs vector operation, the vector advance operation unit can pre-fetch target vector data of the vector operation performed by the vector execution unit from the vector advance operation unit in advance and send the target vector data to the vector execution unit. The vector execution unit directly receives the target vector data sent by the vector look-ahead operation unit, namely, the vector data acquisition and the vector operation are executed separately, and the efficiency of the vector operation of the vector accelerator is improved by reducing the coupling of the vector operation of the vector accelerator.

Drawings

Fig. 1 is a schematic terminal structure diagram of a hardware operating environment according to an embodiment of the present application;

FIG. 2 is a flowchart illustrating an embodiment of a control method for a vector accelerator according to the present application;

FIG. 3 is a schematic structural diagram of a vector accelerator according to a control method of the vector accelerator;

FIG. 4 is a schematic structural diagram of a vector advance operation unit according to the control method of the vector accelerator of the present application;

fig. 5 is a functional block diagram of a control method of a vector accelerator according to the present application.

The implementation, functional features and advantages of the objectives of the present application will be further explained with reference to the accompanying drawings.

Detailed Description

It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

As an implementation solution, referring to fig. 1, fig. 1 may be a schematic diagram of a hardware architecture of a terminal device according to an embodiment of the present invention, as shown in fig. 1, the terminal device may include a vector accelerator 101, for example, a GPU, a memory 102, and a communication bus 103, where the communication bus 103 is used for implementing connection communication between these modules.

The memory 102 may be a high-speed RAM memory or a non-volatile memory (e.g., a disk memory). As shown in fig. 1, a control program of a vector accelerator may be included in a memory 102 as a readable storage medium; and the vector accelerator 101 may be configured to call a control program of the vector accelerator stored in the memory 102 and perform the following operations:

in the process of vector operation, target vector data are obtained from a vector advance operation unit;

and sending the target vector data to a vector execution unit so that the vector execution unit can execute vector operation after receiving the target vector data.

Further, the vector accelerator 101 may be configured to call a control program of the vector accelerator stored in the memory 102, and perform the following operations:

receiving a vector accelerator configuration instruction, and configuring a vector length and a register number based on the vector accelerator configuration instruction;

receiving a starting vector operation instruction, and acquiring a data acquisition instruction cached in a cache region based on the starting vector operation instruction;

acquiring vector data according to the acquired data acquisition instruction cached in the cache region;

and storing the vector data in the vector look-ahead operation unit.

Further, the vector accelerator 101 may be configured to call a control program of the vector accelerator stored in the memory 102, and perform the following operations:

acquiring address information of the vector data carried by the data acquisition instruction cached in the cache region, register information for storing the vector data and data length information;

and acquiring the vector data from the cache region according to the address information, the register information and the data length information.

Further, the vector accelerator 101 may be configured to call a control program of the vector accelerator stored in the memory 102, and perform the following operations:

and analyzing the data acquisition instruction cached in the cache region to obtain address information of the vector data, register information for storing the vector data and data length information.

Further, the vector accelerator 101 may be configured to call a control program of the vector accelerator stored in the memory 102, and perform the following operations:

during the vector operation, determining to acquire target vector data for executing the process from the vector data cached by the vector advance operation unit according to the process of vector operation.

Further, the vector accelerator 101 may be configured to call a control program of the vector accelerator stored in the memory 102, and perform the following operations:

receiving a configuration instruction of a central processing unit;

and setting the number of registers of the vector accelerator according to the configuration instruction.

Further, the vector accelerator 101 may be configured to call a control program of the vector accelerator stored in the memory 102, and perform the following operations:

and acquiring an execution process of the vector operation executed by the vector execution unit, and determining the data volume of the target vector data sent to the vector execution unit according to the execution process.

Referring to fig. 2, fig. 2 is a flowchart illustrating an embodiment of a control method of a vector accelerator according to the present application.

While the embodiments of the present application provide an embodiment of a control method for a vector accelerator, it should be noted that, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in an order different from that shown or described herein.

The control method of the vector accelerator comprises the following steps:

step S10, in the process of vector operation, obtaining target vector data from the vector advance operation unit;

step S20, sending the target vector data to a vector execution unit, so that the vector execution unit executes vector operation after receiving the target vector data.

Because the existing vector operator, such as the Cray-1 type vector accelerator, the decoupling degree between the vector data access and the vector execution is not high, and the operating speed of the vector accelerator is reduced.

In the present embodiment, the vector accelerator receives an instruction transmitted from a Central Processing Unit (CPU), parses the received instruction, and executes an operation corresponding to the instruction.

Referring to fig. 3, fig. 3 is a schematic structural diagram of a vector processor according to the present application. In the present application, the vector processor includes a buffer, wherein the buffer is an external buffer of the vector accelerator (external buffer L1, external buffer L2, wherein the external buffer L1 and the external buffer L2 may perform data interaction), a vector lookahead unit, and a vector execution unit. The vector advance operation unit obtains data which needs to be subjected to vector operation by the vector execution unit from the external buffer area L1 or the external buffer area L2 in advance, and obtains target vector data in the vector data when the vector operation is performed by the vector execution unit, and the target vector data is sent to the vector execution unit from the vector advance operation unit.

Optionally, the vector run-ahead unit sends the target vector data to a register of the vector execution unit by receiving the load instruction, so that the vector execution unit performs the vector operation by acquiring the target data stored in the register when performing the vector operation.

And sending the target vector data to a vector execution unit, so that before the step of executing vector operation after the vector execution unit receives the target vector data, a vector accelerator receives a configuration instruction of a CPU (Central processing Unit), and the number of registers of the vector accelerator is set according to the configuration instruction.

Further, referring to fig. 4, fig. 4 is a schematic structural diagram of a vector advance operation unit according to the present application, and in the present embodiment, the vector advance operation unit includes a throttle manager. The throttle manager controls the data amount of the target vector data sent from the vector advance operation unit to the vector execution unit, manages the data amount sent to the vector execution unit, and can prevent vector data accumulation in the vector execution unit from influencing the calculation speed of the vector execution unit. That is, when the vector accelerator performs a vector operation, the throttle manager of the vector look-ahead unit acquires an execution process of the vector operation performed by the vector execution unit, and determines the data amount of the target vector data to be sent to the vector execution unit based on the execution process.

In the embodiment, when the vector processor executes the vector operation, the vector operation unit acquires the vector data from the vector advance operation unit and sends the target vector data to the vector execution unit, so that the vector execution unit executes the vector operation after receiving the target vector data. When the vector accelerator performs vector operation, the vector advance operation unit can pre-fetch target vector data of the vector operation performed by the vector execution unit from the vector advance operation unit in advance and send the target vector data to the vector execution unit. The vector execution unit does not need to acquire target vector data for vector calculation from the vector advance operation unit, and the target vector data sent by the vector advance operation unit can be directly received, namely, the acquisition of the vector data and the vector operation are executed separately, and the efficiency of the vector operation of the vector accelerator is improved by reducing the coupling of the vector operation of the vector accelerator.

Further, another embodiment of the control method of the vector accelerator is provided. Before the step of obtaining the target vector data from the vector advance operation unit, the method further comprises:

step S1, receiving a vector accelerator configuration instruction, and configuring the length of a vector and the number of registers based on the vector accelerator configuration instruction;

step S2, receiving a starting vector operation instruction, and acquiring a data acquisition instruction cached in a cache region based on the starting vector operation instruction;

step S3, vector data are obtained according to the obtained data obtaining instruction cached in the cache region;

step S4, storing the vector data in the vector advance operation unit.

In this embodiment, the vector lookahead unit receives a vector accelerator configuration instruction, configures a vector length and a register number based on the vector accelerator configuration instruction, and when a launch vector operation instruction (e.g., VectorFetch instruction) is received, acquires a data acquisition instruction from a buffer (an external buffer L1 or an external buffer L2) based on the received launch vector operation instruction, acquires vector data cached in the buffer according to the acquired data acquisition instruction, and stores the vector data in the buffer of the vector lookahead unit, so that when the vector execution unit executes a vector operation, it determines to acquire target vector data for executing the process from the vector data cached by the vector lookahead unit according to a process of the vector operation.

In the embodiment, the vector advance operation unit can acquire the vector data needing to perform vector calculation from an external cache in advance, and the vector data is cached in the cache region of the vector advance operation unit, so that the speed of sending the target vector data to the vector execution unit by the vector advance operation unit is increased.

The step of obtaining vector data of the cache region according to the data obtaining instruction comprises:

step S31, acquiring address information of the vector data carried by the data acquisition instruction cached in the cache region, register information for storing the vector data, and data length information;

step S32, obtaining the vector data from the buffer according to the address information, the register information, and the data length information.

In this embodiment, the vector accelerator acquires address information of the vector data, register information for storing the vector data, and data length information carried by the data acquisition instruction, and acquires the vector data from the cache region according to the register information and the data length information.

Optionally, in this embodiment, the command decoding module of the vector look-ahead unit parses the data obtaining instruction cached in the cache region to obtain address information of the vector data, register information for storing the vector data, and data length information.

In this embodiment, the vector data is obtained by analyzing the data obtaining instruction cached in the cache region by the command decoding module of the vector advance operation unit, so that the accuracy of obtaining the vector data is improved.

In this embodiment, the vector lookahead unit fetches a target instruction based on an address in the launch vector operation instruction. And the target instruction is the first execution for starting to acquire the vector data. For example, when vector a and vector b are added, the target instruction is the instruction that fetches vector a. And acquiring address information of the target vector data and the data length of the target vector data based on the target instruction, and acquiring the target vector data from the cache region according to the address information and the data length.

In this embodiment, the vector run-ahead unit, upon receiving a target instruction (e.g., a VLW instruction), parses the vector allocate instruction and the vector length allocate instruction to obtain address information of target vector data and a data length of the target vector data, and obtains the vector data from the external cache in advance according to the parsed address information and data length.

In the embodiment, the vector data can be acquired from the external cache in advance and stored in the vector advance operation unit, so that the speed of acquiring the vector data is improved.

In order to more clearly describe the technical solution of the present invention, the technical solution of the present invention is described below by a specific example. For example, vector a = [ a0, a1, a2, a3, … …, a254, a255 ]; b = [ b0, b1, b2, b3, … …, b254, b256], and the operation result of the vector a + b is obtained.

The execution process comprises the following steps: a vector processor is configured first, and data required for operation is taken out and operated.

First, the cpu issues a vsectg (vector configuration command) command to configure the vector register vcfg to set the number of used registers. Wherein the VSETCFG (vector configuration command) command is passed through the vector/vector look ahead queue to the vector processor.

The CPU issues a VSETVL (vector length configuration instruction) instruction. This instruction sets the value of the vector length register vlen and this information is passed to the vector processor via the vector/vector advance queue for recording. In this example, the length is set to 256, i.e., the value contained in the address rs1 is 256, and the cpu writes the value in r1 to the rd register.

31 25	24 20	19 15	14 12	11 7	6 0
						0000000	00000	rs1	110	rd	0001011

A VMCA (data move to vector address register instruction) instruction moves the address information of the first element of the a-vector and the address information of the first element of the b-vector into two vector address registers, respectively (address registers mentioned at the end of the text).

Such as: vmca va0, t 0;

vmca va1，t1。

the control execution is passed through a command queue to a scalar unit and a vector look ahead unit.

The cpu issues a vf instruction, and the address of the instruction fetch is obtained by adding the value in r1 in the vf instruction and the immediate imm.

31 25	24 20	19 15	14 12	11 7	6 0
						imm[11:5]	10000	rs1	010	imm[4:0]	0101011

Such as: imm [11:0] = 1001_0010_0000 (0 x920 in hexadecimal), rs1= 01010.

31 25	24 20	19 15	14 12	11 7	6 0
						1001001	10000	01010	010	00000	0101011

Where rs1=01010 stores data at 0x80001000, the scalar processor fetches the first instruction from address 0x80001000 + 0x920 = 0x80001920 via the external cache L1. The instruction contained in the address 0x80001920 is a VLW (unit stride vector sign extended load word) instruction that fetches a vector data, which is decoded by a scalar unit (which does not execute the instruction) and provided to a vector execution unit for execution. The vector lookahead unit obtains instructions from the external cache L1, decodes the load instruction, and issues a prefetch instruction directly to the external cache L2 or the external cache L1 to predict the fetch data load issued by the vector channel (i.e., to prefetch the operation data required for the vector operation) using the address information gathered by the VMCA (data move to vector address register instruction), the vector length information gathered by the VSETVL (vector length configure instruction) instruction. Wherein as1 is va 1.

63 48	47 45	44 42	41	40 33	32	31 24	23 16	15 12	11 0
										000000000000000	000	010	0	00000000	1	as1	vd	p	101100111111

Then a second instruction is fetched, which is also a VLW (Single stride vector sign extended load word) instruction, to load the b vector data. And taking down a vector addition instruction, transmitting the prefetched target vector data to a vector execution unit for calculation, and writing the result back to the memory through the memory controller.

In the embodiment, when the vector processor executes the vector operation and the vector accelerator executes the operation, the vector operation unit acquires the vector data from the vector advance operation unit and sends the target vector data to the vector execution unit, so that the vector execution unit executes the vector operation after receiving the target vector data. When the vector accelerator performs vector operation, the vector advance operation unit can pre-fetch target vector data of the vector operation performed by the vector execution unit from the vector advance operation unit in advance and send the target vector data to the vector execution unit. The vector execution unit does not need to acquire target vector data for vector calculation from the vector advance operation unit, and the target vector data sent by the vector advance operation unit can be directly received, namely, the acquisition of the vector data and the vector operation are executed separately, and the efficiency of the vector operation of the vector accelerator is improved by reducing the coupling of the vector operation of the vector accelerator.

Referring to fig. 5, fig. 5 is a schematic structural diagram of a control device of the vector accelerator according to the present invention. In order to achieve the above object, the present invention further provides a control device of a vector accelerator, including:

an obtaining module 10, configured to obtain target vector data from a vector advance operation unit in a vector operation process;

a sending module 20, configured to send the target vector data to a vector execution unit, so that the vector execution unit executes vector operation after receiving the target vector data.

In addition, in order to achieve the above object, the present invention further provides a vector accelerator, which includes a vector accelerator, a memory, and a control program of the vector accelerator stored on the memory and operable on the vector accelerator, wherein when the control program of the vector accelerator is executed by the vector accelerator, the steps of the control method of the vector accelerator are implemented as described above.

In addition, in order to achieve the above object, the present invention further provides a vector accelerator, where the vector accelerator establishes a connection with a central processing unit, receives an instruction from the central processing unit, and executes a control operation corresponding to the instruction.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

It should be noted that in the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The application can be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The usage of the words first, second and third, etcetera do not indicate any ordering. These words may be interpreted as names.

While alternative embodiments of the present application have been described, additional variations and modifications of these embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following appended claims be interpreted as including alternative embodiments and all such alterations and modifications as fall within the scope of the application.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.