The present invention relates generally to data communications systems, and more specifically to synchronizing a data communications system to ensure data signals transferred by the system are successfully transferred.
A conventional computer system 100 includes a processor 102 coupled through a system bus 104 to a system memory 106 as illustrated in
Conventional processors 102 generally operate at a relatively high speed compared to the memory modules 110A-C. Because access to system memory 106 is a frequent operation of the processor 102, the slower operating speed of the memory 110A-C greatly slows the overall operation of the computer system 100, as will be appreciated by those skilled in the art. As a result, much effort has been put into increasing the data transfer rate or bandwidth of the data bus DQ to enable system memory 106 to transfer data more quickly. To increase the bandwidth of the data bus DQ, the width of the data bus has been increased, and new types of DRAM technology having much higher transfer speeds, such as RAMBUS DRAMs (“RDRAMs”) and synchronous link DRAMs (“SLDRAMs”), have also been developed. As the operating speed of the data bus DQ increases, however, noise, signal skew, a smaller data eye—which defines the duration for which the data signals are valid—and other factors make it more difficult to reliably transfer data over the data bus.
One approach that has been utilized in transferring data more reliably at high data transfer rates is an adaptive process of adjusting the delay between data signals and a clock signal transmitted along with the data signals. A receiving device captures the data signals in response to the clock signal. The process involves applying the clock signal and a pseudorandom bit pattern having known values on the data bus DQ. The timing relationship or delay between the data signals forming the pseudorandom bit pattern and the clock signal is thereafter adjusted through a series of values, and the bit pattern captured at each value of the delay. Because the pseudorandom bit pattern has known values, the captured data can be compared to the expected values to determine whether the bit pattern was successfully captured at each value of the delay between the bit pattern and the clock signal. The values for the delay between the bit pattern and the clock signal where the bit pattern was unsuccessfully captured are designated failing values, and the values where the bit pattern was successfully captured are designated passing values. The range of the passing values define the data eye of the applied bit pattern, and a final value of the delay between the bit pattern and the clock signal may be selected in the middle of the data eye to optimize the delay for successful capture of the bit pattern. This approach is described in more detail in U.S. Pat. No. 6,338,127 to Manning entitled METHOD AND APPARATUS FOR RESYNCHRONIZING A PLURALITY OF CLOCK SIGNALS USED TO LATCH RESPECTIVE DIGITAL SIGNALS, AND MEMORY DEVICE USING SAME, and in U.S. Pat. No. 6,374,360 to Keeth et al. entitled METHOD AND APPARATUS FOR BIT-TO-BIT TIMING CORRECTION OF A HIGH SPEED MEMORY BUS, both of which are incorporated herein by reference.
In addition to systems communicating via electrical signals, such as the RDRAM and SLDRAM technologies mentioned above, optically-based memory systems including an optical communications link between the memory controller 108 and memory modules 110A-C have also been developed to increase the bandwidth of system memory 106. In such optically-based systems, however, problems of transmitting and receiving optical signals between the memory controller 108 and the memory modules 110A-C result in unacceptably high bit error rates (BER) and continue to hamper the commercialization of such systems, particularly in system memories 106 having parallel, closely-spaced memory modules of the type found in many existing personal computer systems. For example, in an optically-based system the optical transmitter and receiver must be designed to have sufficient dynamic range to accommodate all possible variations in system memory parameters such as the total number of memory modules. Dynamic range defines the required operating range of a parameter of the receiver or transmitter, such as required power of a received optical signal for a receiver, as will be understood by those skilled in the art. Sufficient dynamic range is required to ensure optical signals are reliably transmitted and received, and transmitters and receivers having larger dynamic range are more costly, increasing the overall cost of optically-based memory systems.
There is a need for a system memory that reliably transfers data at a high bandwidth but has a relatively low cost for use in computer systems and other cost-sensitive applications.
According to one aspect of the present invention, a method dynamically adjusts link control parameters of a communications network. The communications network includes a transmitter coupled through a first data link to a receiver, with the transmitter and receiver each have at least one associated link control parameter that affects the operation of that component. Data signals are transmitted over the first data link and the transmitted data signals are captured. The values of the captured data signals are compared to expected values for those signals, and the values of the link control parameters are adjusted to successfully capture the transmitted digital signals.
In the system memory 200, the optical communications link 202 corresponds to a data bus DQ for transferring data between the memory controller 204 and memory device 206. An address bus ADDR and control bus CONT are electrical busses through which the memory controller 204 applies address and control signals, respectively, to the memory device 206. With other embodiments of the system memory 200, the communications link 202 also includes the address and control busses ADDR, CONT for communicating address and control signals from the memory controller 204 to the memory device 206. Although only the single memory device 206 is depicted in
Before describing the overall adaptive synchronization process executed by the system memory 200, the operation of the optical transmitters 208, 214 and optical receivers 212, 216 will be described in more detail. Operation of the optical transmitters 208 and 214 is the same, as is the operation of the optical receivers 212 and 216, and thus, for the sake of brevity, only the operation of the transmitter 208 and receiver 212 will be described in more detail. The optical transmitter 208 receives a bit stream of electrical data signals from the memory controller 204, modulates an optical carrier signal with the received bit stream, and couples the modulated optical carrier into the fiber-optic cable 210. Typically, the optical transmitter 208 includes a light emitting diode (LED) or laser diode for generating the optical carrier signal. The optical transmitter 208 includes several operational parameters whose values are typically defined during design of the system containing the optical transmitter.
One such operational parameter of the optical transmitter 208 is a gain of the transmitter and determines the power of a light corresponding to the optical carrier signal that is coupled into the fiber-optic optic cable 210. The transmitter 208 must provide an optical carrier signal having sufficient power to propagate through the fiber-optic cable 210 and be received by the optical receiver 212. The gain of the optical transmitter 208 ensures the optical carrier signal a sufficient power to compensate for various losses in the communications link 202, such as attenuation in the fiber-optic optic cable 210 and losses resulting from inefficient coupling of the optical transmitter to the fiber-optic cable. Another operational parameter of the optical transmitter 208 is a pulse-shaping parameter that ensures the optical transmitter 208 generates an optical carrier signal having the required shape to allow the carrier signal to be reliably detected and demodulated by the optical receiver 212. When an LED is used as the light generation source in the optical transmitter 208, the pulse-shaping parameter compensates for differing turn ON and OFF times of the LED.
The optical receiver 212 senses or detects the received modulated optical carrier signal propagating through the fiber-optic optic cable 210, converts the modulated optical carrier signal into corresponding electrical signals, and demodulates the signal to obtain the originally transmitted data signals. Typically, the optical receiver 212 includes a photo diode for detecting the received modulated optical carrier signal. One operational parameter associated with the optical receiver 212 is an input threshold level or sensitivity that defines a minimum optical power the receiver can reliably detect at a given data rate in order to achieve a particular bit error rate. The optical receiver 212 may also include a link monitor, which is circuitry indicating when the received modulated optical carrier signal from the fiber-optic cable 210 is less than the required minimum optical power. Suitable circuitry for forming all components 202-216 in the system memory 200 will be understood by those skilled in the art, and thus, for the sake of brevity, such circuitry will not be described or depicted in more detail.
The overall process executed by the system memory 200 in adaptively adjusting the operational parameters of the optical transmitters 208, 214 and optical receivers 212, 216 will now be described in more detail with reference to the flowchart of
From step 304, the process goes to step 306 and the memory device 206 captures the synchronization data transmitted over the optical communications link 202. More specifically, the optical receiver 212 detects, converts, and demodulates the received modulated optical carrier signal to obtain electrical received data signals that ideally should correspond to the originally transmitted synchronization data. The process proceeds to step 308 in which the electrical received data signals from the optical receiver 212 are transferred to the memory device 206 which, in turn, compares the values of the received data signals to expected values for those signals. The synchronization data has known values, and thus the memory device 206 is able to generate the expected values for the received data signals. For example, when the synchronization data is a pseudorandom bit pattern having a known repeating sequence of bits, the memory device 206 can determine expected values for the received data signals.
When the determination in step 308 is negative, meaning the received data signals do not equal the expected values for those signals, the process goes to step 310 and the memory device 206 activates the error signal ES. If the determination in step 308 is positive, which means the received data signals are equal to the expected values for those signals, the process goes to step 312 and the memory device 206 deactivates the error signal ES. From either step 310 or step 312, the process of proceeds to step 314 and the current value of the error signal ES associated with the current values of a link control parameters is stored. Note, the values of the error signal ES can be stored in the memory device 206 and later transferred to the memory controller 204 to these values may be transferred as they are generated. In the embodiment of
After the value of error signal ES associated with the current link control parameters has been stored in step 314, the process goes to step 316 and the memory controller 204 and determines whether all values of the link control parameters have been tested. This determination indicates whether each link control parameter has been assigned every possible value within a possible range of values for the parameter. When the determination in step 316 is negative, the process goes to step 318 and the memory controller 204 adjusts the values of one or more link control parameters. The process then goes back and repeats steps 304-316 for the new values of the link control parameters established in step 318. In this way, the memory controller 204 repeatedly adjusts the values of link control parameters and thereafter transmits synchronization data to the memory device 206 through the optical communications link 202 which now operates according to the new values for the link control parameters. Each time the values of link control parameters are adjusted, the memory device 206 determines whether the received electrical data signals have their expected values and stores the error signal ES having the appropriate value that is associated with the current link control parameters. An array of error signals ES is thus generated, with the value of each error signal in the array being associated with particular values for the link control parameters.
The process continues executing steps 304-316 until the determination in step 316 is positive, meaning that all link control parameters have assumed all desired values. Thus, when the optical communications link 202 has been tested for all desired values of the link control parameters, the determination in step 316 is positive and the process proceeds to step 319. In step 319, if the memory controller 204 does not already contain the array of error signals ES generated during steps 304-316, the error signals are transferred to the memory controller 204. At this point, the memory controller 204 evaluates the array of error signals ES to determine optimal values for the link control parameters. The process then proceeds to step 320 and the memory controller 204 sets each of the link control parameters for the optical transmitter 208 and optical receiver 212 to the determined optimal value. From step 320, the synchronization process goes to step 322 and terminates, with the link control parameters of the optical transmitter 208 and optical receiver 212 having been set to values to be utilized during normal operation of the system memory 200.
Once the link control parameters of the optical transmitter 208 and optical receiver 212 have been set, the system memory 200 executes substantially the same process just described to adjust the link control parameters of the optical transmitter 214 and optical receiver 216 to optimal values and thereby synchronize the optical communication link 202 for read data transfer operations. Because the process of setting the link control parameters of the optical transmitter 214 and optical receiver 216 is substantially the same as just described and will be understood by those skilled in the art, this process will not, for the sake of brevity, be described in more detail.
The times at which the system memory 200 (
Referring back to
The process for synchronizing the optical communications link 202 may also be combined with a synchronization process for adjusting the delay between the electronic data signals and a clock signals transmitted along with those electronic data signals. Such a synchronization process for electrical data signals and a clock signal was previously discussed with reference to the conventional system memory 106 of
As previously mentioned, the optical communications link 202 could include other suitable transmission media in place of the fiber-optic cable 210 such as free space. Moreover, although the system memory 200 is described and depicted as including the optical communications link 202, other high speed communications links such as a radio frequency or microwave link could be utilized in place of the optical communications link. The concepts described above are equally applicable for adaptively adjusting the operational parameters of such a radio frequency, microwave, or other high speed communications link, as will be appreciated by those skilled in the art. Even though various embodiments and advantages of the present invention have been set forth in the foregoing description, the above disclosure is illustrative only, and one skilled in the art will understand various equivalents and modifications of the described components and concepts that may be made in detail and yet remain within the broad principles of the invention. For example, some of the components described above may be implemented using either digital or analog circuitry, or a combination of both, and also, where appropriate, may be realized through software executing on suitable processing circuitry. Therefore, the present invention is to be limited only by the appended claims.
Number | Date | Country | |
---|---|---|---|
Parent | 10461207 | Jun 2003 | US |
Child | 11639950 | US |