Marge tester measurements using machine

Abstract

A margin tester comprises: one or more ports for allowing the margin tester to be connected to a device under test (DUT); a memory, the memory containing a margin tester signature; a transmitter; a receiver for receiving a signal from the DUT; one or more processors configured to execute code that causes the one or more processors to: receive a plurality of signals from the receiver through the one or more ports; generating a performance indicator from the plurality of signals; sending the performance indicator and the margin tester signature to one or more machine learning networks; and receiving, from the one or more machine learning networks, a result including a performance measurement prediction for the DUT.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS The present disclosure is a non-provisional application of U.S. Provisional Application No. 63 / 513,849, filed on July 14, 2023, entitled “MARGINTESTER MEASUREMENT USING MACHINE LEARNING” and U.S. Provisional Application No. 63 / 514,518, filed on July 19, 2023, entitled “MARGIN TESTER TRANSLATION FOR PASS / FAIL RESULT USING MACHINE LEARNING”, and claims the benefit of these two U.S. Provisional Applications, the disclosures of which are incorporated herein by reference in their entirety. Technical Field

[0002] The present disclosure relates to test and measurement instruments, and more particularly to using machine learning to better correlate measurements made by different test and measurement instruments. Background Art

[0003] The Tektronix Margin Tester (TMT) has both a built-in transmitter and receiver that currently supports PCIE Gen3 and Gen4 standards. See https: / / www.tek.com / en / products / pciemargintester.

[0004] The traditional method for testing a DUT transmitter uses a real-time (RT) oscilloscope to acquire the waveform, and then uses measurement software such as SigTest, Seasim, and DPOJET (see https: / / www.tek.com / en / datasheet / jitter-noise-and-eye-diagram-analysis-solution) to generate eye diagrams, BER (bit error rate) contour graphs, and measurements such as eye height, eye width, receiver DFE (decision feedback equalizer) tap values, and CTLE (continuous time linear equalizer) gain. The traditional method for testing a DUT receiver uses a BERT (bit error rate tester) to generate a signal with the desired stress.

[0005] TMT can test all PCIE (high-speed peripheral component interface) channels at the same time, up to 16 channels. This solution has a much higher throughput than the traditional method of using a RT scope that typically only has four channels. The transmitter part of TMT can create various stress signals used as a signal source for DUT receiver testing. TMT can generate signals at all channels (up to 16 channels) that are much higher than BERT at the same time. This achieves higher throughput.

[0006] However, measurement results for the same DUT may vary from one specific TMT instrument to another, and measurement results from a TMT instrument may not match measurement results using traditional methods with a real-time oscilloscope and a BERT. Embodiments of the present disclosure generally include machine learning-based methods that reduce these instrument-to-instrument mismatches. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 An embodiment of a setup for testing a device under test (DUT) using a margin tester is shown.

[0008] Figure 2 An example of a BER contour scan result on a margin tester is shown.

[0009] Figure 3 An embodiment of a setup for training a machine learning network for testing a DUT using a margin tester is shown.

[0010] Figure 4 An embodiment of training data for a neural network of a transmitter is shown.

[0011] Figure 5 An embodiment of a setup for runtime operation of a margin tester for testing a DUT using a machine learning network is shown.

[0012] Figure 6 An embodiment of runtime data for testing a DUT is shown.

[0013] Figure 7 An embodiment of a training setup for training a neural network for receiver testing is shown.

[0014] Figure 8 An embodiment of a runtime setup for receiver testing using a neural network is shown.

[0015] Fig. 9 An embodiment of a runtime setup for receiver testing using a calibrated TMT receiver with a neural network is shown. DETAILED DESCRIPTION

[0016] Embodiments herein provide solutions to the problem of mismatch between different margin testers and can speed up the testing process for transmitter and receiver devices (such as those used in PCIE devices). The use of machine learning neural networks reduces the time required to test the device under test (DUT). Providing the neural network with a "signature" of the margin tester being used for testing allows the neural network to adjust any measurements made by the margin tester to adapt to that particular margin tester.

[0017] Currently, Tektronix remains the only provider of margin testers. The following discussion refers to the margin tester as a "Tektronix margin tester" (TMT), with the understanding that other margin testers from other manufacturers may become available with similar characteristics. The claims set forth below will cover those margin testers having the same characteristics and capabilities as the margin tester discussed herein, including support for current and future standards (such as PCIE5 and PCIE6) and standards other than PCIE.

[0018] The following discussion applies to testing transmitters and receivers. It is not intended to be limited to those specific kinds of DUTs, nor should it be inferred that it is limited to those specific kinds of DUTs. DUTs may include different types of devices that TMT may be capable of testing. Further, although the transmitters and receivers used for testing by TMT may reside on TMT (the same or on two different devices), it is not intended to be limited to those transmitters or receivers, nor should it be inferred that it is limited to those transmitters or receivers in any way.

[0019] Before the advent of TMT, transmitters (Tx) and receivers (Rx) had more complex testing processes using more and typically more expensive equipment. In receiver (Rx) testing, the BERT output signal undergoes calibration using an oscilloscope (mirror). After the BERT output signal is calibrated, the BERT is connected to the DUT (Rx). The BERT receives a loopback signal from the DUT to obtain a BER count, which does not use a mirror. Rx testing generally involves target stress levels for signals from transmitters in the TMT. TMT has the ability to provide various stress levels that allow TMT transmitters to test receivers. For Tx testing, a mirror with analysis software can make pass-fail predictions to be used as labels for neural network training and measurement results as will be discussed in more detail later. The mirror plus analysis software will constitute a reference system with which the translator will view the margin tester BER data profile and make pass-fail translation predictions and / or various types of measurement results as will be discussed in more detail later. OptaML TM Analysis software includes embodiments of analysis software.

[0020] Figure 1 An overall test setup is shown for testing a DUT 10 using a TMT 12. The TMT has one or more ports, such as 14, to allow the TMT to send and receive signals from the DUT 10. As mentioned above, the TMT can test multiple channels, such as 16 channels, at a time. The connection between the DUT and the TMT may include a ribbon cable with 16 or more connections to the TMT. The TMT has a transmitter 16 that allows the TMT to transmit test signals to the DUT, and a receiver 18 that allows the TMT to receive signals. The TMT will also have one or more processors 20 and a memory 22. The one or more processors may access a machine learning system 30 on an external device, or, as will be discussed below, the machine learning network may reside on the TMT.

[0021] In one embodiment, the TMT receives a signal transmitted from the DUT, and the TMT generates a set of BER contours for various BER levels for each channel. Figure 2 An example of a collection of BER contours generated by TMT using data from one of the 16 channels on the DUT is shown. Each contour line from each channel can be represented by a single image. This will result in 16 times, however, that many contour vectors are used. Figure 2 The profiles show ten profiles for 10 BER levels for one channel, so the resulting array will be a 10×16 array of individual profile lines. The machine learning system can receive this array of images for one DUT as input, or it can receive other formats, such as images for each channel, images for each DUT profile across all channels, etc. The image format used at runtime must match the image format used when training the neural network. In addition, the input images can include other types of data.

[0022] Figure 3 An embodiment of a test setup for training a machine learning network 38 within a machine learning system 30 is shown. Figure 1 In the case of the components shown in FIG, the TMT 12 does not show its internal components. During training, the TMT will Figure 4 The DUT BER profile, parameters and measurements received from the DUT, discussed in more detail, and the manufacturing reference BER profile (MFG REF BER profile) are sent to the machine learning system, referred to herein as the instrument translator. The manufacturing reference BER profile acts as a "signature" for the TMT. As discussed above, measurements and performance vary from TMT to TMT. This allows the machine learning system 30 to take into account the measurements of the TMT.

[0023] The development of TMT signatures can take many forms. TMTs do not have a way to calibrate them to make them more consistent. In one embodiment, the process uses a "golden" reference DUT to connect to a profile set (such as, Figure 2 The TMT then saves the resulting profile as a signature of the TMT. In one embodiment, the signature profile may be placed in one of the color channels of the input image that is different from the DUT BER profile data.

[0024] The machine learning system 30 as shown includes a neural network 38 and a tensor builder 32. These components may reside in separate computing devices or in the cloud, or may reside on the TMT 12. The tensor builder 32 takes input data and constructs a tensor image for input to the neural network. The input image may have three color channels RGB, and one color channel may contain a TMT signature, as discussed above. U.S. Patent Application No. 17 / 747,954, filed 5 / 21 / 2021, “SHORT PATTERN WAVEFORM DATABASE BASED MACHINE LEARNING” may provide more information about tensors and is incorporated herein in its entirety.

[0025] The tensor input received from the tensor builder 32 may undergo normalization at 34. Although not Figure 3 The output of the neural network will undergo denormalization at 36 for use in a training environment. The neural network receives input data from the DUT, the TMT signature, and any measurements made by the TMT. The tensor builder 32 can include the measurements in the tensor image as bar graphs, pie charts, and many other graphs. These can include information about Figure 4 Many different types of measurement results are discussed, and may include other information, such as the PCIE standard under which the transmitter is operating.

[0026] During training, the neural network 38 also receives reference data from a reference test station 40. The reference test station is also connected to the DUT 10, which is shown repeatedly for drawing simplicity. A variable ISI (inter-symbol interference) board 42 resides between the DUT Tx output and an oscilloscope or "mirror" 44. The ISI 42 can typically use multiple channels. In one embodiment, the TMT 12 uses only one worst-case channel that matches the margin tester as closely as possible, so the ISI 42 will only use one channel. This minimizes the amount of training required. The mirror 44 receives the waveforms of the 16 channels and measures them using some test measurement software (such as, among many others, SigTest, DPOJET, PAM4), and then provides those measurements 46 to the neural network for correlation with the tensor data from the TMT.

[0027] Figure 4 An example of a training input 50 to a neural network is shown. The training input includes a TMT signature 52 and a profile 54 of the DUT under test as a BER profile. Additional information 56 may also be included. This may include channel, preset, temperature, measurements such as eye width (EW), eye height (EH), attenuation (ATT), settings for a continuous time linear equalizer (CTLE), gain, and any decision feedback equalizer (DFE), among many others. Following the same measurements made by the TMT, a mirror measurement reference result 62 is also input to the neural network 38. The neural network then correlates the DUT and its signature and the TMT measurements with the reference measurements.

[0028] It should be noted that in this particular embodiment, the DUT profile is used as a performance indicator of the DUT. In other margin testers or in future embodiments of the margin tester, different types of performance indicators may be used. Similarly, as will be discussed below during runtime, the neural network provides results that include predictions of one or more performance measurements for the DUT.

[0029] Figure 5 An embodiment of a testing system during runtime is shown. Figure 5 Shown with Figure 3The same components, but other settings may also exist. In this embodiment of the runtime environment, the output data from the TMT is used to provide an estimate of the measurement results for the DUT, a pass / fail indicator, or both. As mentioned above, the output of the neural network 38 can undergo denormalization at 36. Normalization converts the output of the tensor builder 32 to a range of values ​​from -1 to +1. The output of the neural network is then generally converted from this range back to the expected range for the measurement results. In one embodiment, the measurement result value or the general pass / fail indicator can be provided to the decision box 68. If the DUT fails, or if the measurement result is out of range, the user will have the option of using the reference station 40 to obtain the measurement result, enabled by the switch 66. In another embodiment, this part will not be included, and the various measurement results will only be presented to the user.

[0030] Figure 6 An embodiment of a prediction of a DUT result is shown. On the left, the data provided to the neural network 50 will be the same type of data as provided during training. However, in this environment, the neural network will provide a prediction of one or more of the measurement results. The neural network can also be configured to also generate a pass / fail result as a performance measurement result prediction.

[0031] Although common to all DUTs, the above discussion primarily discusses the configuration of a setup for transmitter testing. In the case of transmitter testing, the receiver of the TMT receives signals from the channels used by the DUT, and the received signals are used to generate profiles and other measurements. As will be discussed below, the TMT can also be used to perform receiver testing. In this instance, the TMT also sends signals to the DUT.

[0032] The characterization of the receiver involves generating a signal with a desired stress level. In traditional testing, a mirror (such as a real-time mirror) calibrates the BERT to generate an output signal with a desired stress. The BERT output is then connected to the DUT receiver. Using TMT as a signal generator improves test throughput because TMT has a much higher channel count. The TMT Tx can be configured to generate signals with various stresses. Among many others, the transmitter settings can include Tx FFE (feedforward equalizer) settings and gain settings. Tuning the Tx FFE taps results in various conditions, such as preshoot as an example. The mirror then provides a label for the training data, which is the stress level. Labels include random jitter, periodic jitter, total jitter, vertical noise, inter-symbol interference (ISI) and others.

[0033] Figure 7The training configuration is shown in abbreviated form. The mirror 42 is part of the reference test station 40 of the previous figure. The TMT Tx generates a signal with the desired pressure and sends it to the receiver 18 and the mirror 42. The receiver sends its results (such as a profile) to the neural network 38 along with the labels provided by the mirror for the training data.

[0034] During runtime, two scenarios for receiver testing can be considered. In the first scenario, a neural network external to the TMT receiver receives the profile from the TXT receiver and makes a prediction of the receiver pressure. Figure 8 An example embodiment of this is shown. TMT Tx 16 sends the test signal to TMT Rx 18 which in turn sends the profile it generates through scanning to the neural network 38. The neural network determines the stress level in the signal.

[0035] In another embodiment, if Fig. 9 As shown in and discussed above, the TMT Rx can be "calibrated" because it has a trained neural network within it. The TMT 60 may include a trained neural network, and in other embodiments may also undergo training. However, in Fig. 9 In an embodiment of the present invention, the TMT Tx 16 sends a signal to a calibrated TMT 60 having a receiver, the calibrated TMT 60 containing a trained neural network. The neural network within the TMT then detects the stress level.

[0036] As a variation, the system may use more than one neural network. For example, one neural network may be trained to predict transmitter measurements and one neural network may be trained to predict stress levels for receive tests. The training setup will be the same, the networks will only receive a subset of the information used to train a single neural network.

[0037] The above embodiments use machine learning to predict the performance of a DUT. One embodiment has a transmitter as the DUT and another embodiment has a receiver as the DUT, and the machine learning system provides a prediction of the DUT performance based on inputs to the neural network as performance indicators. The neural network then provides a prediction of the measurement results of the DUT performance. The use of a margin tester allows the DUT to be tested using multiple channels simultaneously. In one experiment, results were obtained using a margin tester in one minute or less. This speeds up the text process, and the use of the TMT signature allows the neural network to be "calibrated" based on inputs including the TMT signature to alleviate the problem of differences between TMT devices.

[0038] Aspects of the present disclosure may be operated on specially created hardware, on firmware, digital signal processors, or on a specially programmed general-purpose computer including a processor that operates according to programmed instructions. The term controller or processor as used herein is intended to include microprocessors, microcomputers, application-specific integrated circuits (ASICs), and dedicated hardware controllers. One or more aspects of the present disclosure may be embodied in computer-usable data and computer-executable instructions executed by one or more computers (including monitoring modules) or other devices, such as one or more program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types when executed by processors in computers or other devices. Computer-executable instructions may be stored on non-transient computer-readable media, such as hard disks, optical disks, removable storage media, solid-state memory, random access memory (RAM), etc. As will be appreciated by those skilled in the art, the functionality of program modules may be combined or distributed as desired in various aspects. In addition, functionality may be embodied in whole or in part in firmware or hardware equivalents (such as integrated circuits, FPGAs, etc.). Certain data structures may be used to more efficiently implement one or more aspects of the present disclosure, and such data structures are contemplated within the scope of the computer-executable instructions and computer-usable data described herein.

[0039] In some cases, the disclosed aspects may be implemented in hardware, firmware, software, or any combination thereof. The disclosed aspects may also be implemented as instructions carried by one or more non-transient computer-readable media or stored on one or more non-transient computer-readable media, which may be read and executed by one or more processors. Such instructions may be referred to as computer program products. As discussed herein, computer-readable media means any medium accessible by a computing device. As an example and not limitation, computer-readable media may include computer storage media and communication media.

[0040] The previously described versions of the disclosed subject matter have many advantages that are described or will be apparent to those skilled in the art. Even so, these advantages or features are not necessarily present in all versions of the disclosed apparatus, system or method.

[0041] Computer storage media means any medium that can be used to store computer-readable information. As an example and not limitation, computer storage media may include RAM, ROM, electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital video disk (DVD), or other optical disk storage, cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, and any other volatile or non-volatile, removable or non-removable media implemented in any technology. Computer storage media excludes the signal itself and the signal transmission in transient form.

[0042] Communication media refers to any medium that can be used for communication of computer-readable information. By way of example and not limitation, communication media can include coaxial cables, fiber optic cables, air, or any other medium suitable for communication of electrical, optical, radio frequency (RF), infrared, acoustic or other types of signals.

[0043] Example Illustrative examples of the disclosed technology are provided below. Embodiments of the technology may include one or more and any combination of the examples described below.

[0044] Example 1 is a margin tester comprising: one or more ports for allowing the margin tester to be connected to a device under test (DUT); a memory containing a margin tester signature; a transmitter; a receiver for receiving a signal from the DUT; one or more processors configured to execute code that causes the one or more processors to perform the following operations: receiving multiple signals from the receiver through the one or more ports; generating performance indicators from the multiple signals; sending the performance indicators and the margin tester signature to one or more machine learning networks; and receiving results from the one or more machine learning networks containing predictions of performance measurement results for the DUT.

[0045] Example 2 is the margin tester of Example 1, wherein the DUT includes a transmitter, and one or more of the plurality of signals from the receiver includes a performance indicator generated from reception of a transmission from the transmitter.

[0046] Example 3 is the margin tester of Example 2, wherein the code that causes the one or more processors to send the performance indicator and the margin tester signature to one or more machine learning networks includes code that causes the one or more processors to perform the following operations: sending parameters and settings for the transmitter to the one or more machine learning networks.

[0047] Example 4 is the margin tester of Example 2, wherein the code causing the one or more processors to generate the performance indicator includes code causing the one or more processors to generate a bit error rate (BER) contour.

[0048] Example 5 is the margin tester of any of Examples 1 to 4, wherein the margin tester signature comprises a BER contour unique to the margin tester.

[0049] Example 6 is the margin tester of any one of Examples 1 to 5, wherein the performance measurement result prediction includes a pass / fail prediction or one or any of one or more measurement results.

[0050] Example 7 is a margin tester of any one of Examples 1 to 6, wherein the code causing the one or more processors to send the performance indicator to the machine learning network includes code causing the one or more processors to perform the following operations: constructing a tensor image of the performance indicator; and sending the tensor image to the machine learning network.

[0051] Example 8 is the margin tester of Example 7, wherein the tensor image includes one or more of: configuration parameters, environmental parameters, and measurement results for one or more channels of the DUT.

[0052] Example 9 is the margin tester of Example 1, wherein the DUT under test includes a receiver, and the signal received from the DUT includes a signal received by the receiver.

[0053] Example 10 is the margin tester of any one of Examples 1 to X, wherein the signal received by the receiver is generated by a margin tester transmitter.

[0054] Example 11 is the margin tester of any of Examples 9 or 10, wherein the code causing the one or more processors to generate a performance metric from the signal comprises code causing the one or more processors to generate a BER contour scan of the signal.

[0055] Example 12 is the margin tester of any one of Examples 1 to 11, wherein the code causing the one or more processors to send the performance metric and the margin tester signature to the one or more machine learning networks includes code causing the one or more processors to perform the following operations: sending the performance metric and the margin tester signature to one or more external neural networks.

[0056] Example 13 is the margin tester of any one of Examples 1 to 12, wherein the code causing the one or more processors to send the performance metric and the margin tester signature to the one or more machine learning networks includes code causing the one or more processors to perform the following operations: sending the performance metric and the margin tester signature to one or more neural networks residing on the margin tester.

[0057] Example 14 is a method comprising: receiving multiple signals from a device under test (DUT) at a receiver of a margin tester, the margin tester having a margin tester signature; generating performance indicators from the multiple signals; sending the performance indicators and the margin tester signature to one or more machine learning networks; and receiving results from the one or more machine networks containing predictions of performance measurement results for the DUT.

[0058] Example 15 is the method of Example 14, wherein the DUT includes a transmitter.

[0059] Example 16 is the method of Example 15, further comprising training the one or more neural networks, the training comprising: collecting performance metrics for multiple margin testers; and labeling the performance metrics for each margin tester using the margin tester signature and one or more of the configuration of the transmitter channel, environmental parameters, and raw measurement results.

[0060] Example 17 is the method of Example 14, wherein the DUT includes a receiver.

[0061] Example 18 is the method of Example 17, further comprising training the one or more neural networks, the training comprising: collecting performance metrics for multiple margin testers; and labeling the performance metrics for each margin tester using the margin tester signature and the stress applied to the multiple signals before the multiple signals are transmitted.

[0062] Example 19 is the method of any of Examples 14 to 18, wherein the performance measurement result prediction includes one or any of a pass / fail prediction or one or more measurement result predictions for the DUT.

[0063] Example 20 is the method of any one of Examples 14 to 19, wherein sending the performance indicator and the margin tester signature to the one or more machine learning networks includes sending the performance indicator and the margin tester signature to any one of one or more neural networks external to the margin tester or one or more neural networks residing in the margin tester.

[0064] In addition, the written description mentions specific features. It should be understood that the disclosure in this specification includes all possible combinations of those specific features. Where a specific feature is disclosed in the context of a particular aspect or example, the feature may also be used in the context of other aspects and examples as much as possible.

[0065] Furthermore, when the present application refers to a method having two or more defined steps or operations, the defined steps or operations may be performed in any order or simultaneously, unless the context excludes those possibilities.

[0066] All features disclosed in the specification including claims, abstract and drawings, and all steps in any disclosed method or process may be combined in any combination except combinations in which at least some of such features and / or steps are mutually exclusive. Each feature disclosed in the specification including claims, abstract and drawings may be replaced by an alternative feature serving the same, equivalent or similar purpose, unless expressly stated otherwise.

[0067] Although specific examples of the present invention have been shown and described for purposes of illustration, it should be understood that various modifications can be made without departing from the spirit and scope of the present invention. Accordingly, the present invention should not be limited, except by the appended claims.

Claims

1. A margin tester, comprising: one or more ports for allowing the margin tester to be connected to a device under test (DUT); a memory containing a margin tester signature; transmitter; A receiver for receiving a signal from the DUT; One or more processors configured to execute code that causes the one or more processors to perform the following operations: receiving a plurality of signals from the receiver via the one or more ports; generating a performance indicator from the plurality of signals; sending the performance metric and the margin tester signature to one or more machine learning networks; as well as Results including predictions of performance measurements for the DUT are received from the one or more machine learning networks.

2. The margin tester of claim 1, wherein the DUT comprises a transmitter and one or more of the plurality of signals from the receiver comprises a performance indicator generated from reception of a transmission from the transmitter.

3. The margin tester of claim 2, wherein the code causing the one or more processors to send the performance metrics and the margin tester signature to one or more machine learning networks comprises code causing the one or more processors to send parameters and settings for the transmitter to the one or more machine learning networks.

4. The margin tester of claim 2, wherein the code causing the one or more processors to generate the performance indicator comprises code causing the one or more processors to perform the following operations: generate a bit error rate (BER) contour.

5. The margin tester of claim 1, wherein the margin tester signature comprises a BER contour unique to the margin tester.

6. The margin tester of claim 1, wherein the performance measurement result prediction comprises a pass / fail prediction or one or any one of one or more measurement results.

7. The margin tester of claim 1 , wherein the code causing the one or more processors to send the performance indicator to the machine learning network comprises code causing the one or more processors to: construct a tensor image of the performance indicator; and send the tensor image to the machine learning network.

8. The margin tester of claim 7, wherein the tensor image comprises one or more of: configuration parameters, environmental parameters, and measurement results for one or more channels of the DUT.

9. The margin tester of claim 1, wherein the DUT under test comprises a receiver, and the signal received from the DUT comprises a signal received by the receiver.

10. The margin tester of claim 9, wherein the signal received by the receiver is generated by a margin tester transmitter.

11. The margin tester of claim 9, wherein the code causing the one or more processors to generate a performance metric from the signal comprises code causing the one or more processors to generate a BER contour scan of the signal.

12. The margin tester of claim 1 , wherein the code causing the one or more processors to send the performance metric and the margin tester signature to the one or more machine learning networks comprises code causing the one or more processors to send the performance metric and the margin tester signature to one or more external neural networks.

13. The margin tester of claim 1, wherein the code causing the one or more processors to send the performance metric and the margin tester signature to the one or more machine learning networks comprises code causing the one or more processors to send the performance metric and the margin tester signature to one or more neural networks resident on the margin tester.

14. A method comprising: receiving a plurality of signals from a device under test (DUT) at a receiver of a margin tester, the margin tester having a margin tester signature; generating a performance indicator from the plurality of signals; sending the performance metric and the margin tester signature to one or more machine learning networks; as well as Results including performance measurement predictions for the DUT are received from the one or more machine networks.

15. The method of claim 14, wherein the DUT comprises a transmitter.

16. The method of claim 15, further comprising training the one or more neural networks, the training comprising: Collect performance metrics for multiple margin testers; as well as The performance indicator for each margin tester is labeled using the margin tester signature and one or more of a configuration of a transmitter channel, environmental parameters, and raw measurements.

17. The method of claim 14, wherein the DUT comprises a receiver.

18. The method of claim 17, further comprising training the one or more neural networks, the training comprising: Collect performance metrics for multiple margin testers; as well as A performance indicator for each margin tester is labeled with the margin tester signature and a stress applied to the plurality of signals before the plurality of signals are transmitted.

19. The method of claim 14, wherein the performance measurement result prediction comprises one or any of a pass / fail prediction or one or more measurement result predictions for the DUT.

20. The method of claim 14, wherein sending the performance metric and the margin tester signature to the one or more machine learning networks comprises: The performance indicator and the margin tester signature are sent to either one of one or more neural networks external to the margin tester or one or more neural networks resident in the margin tester.

Images