Abstract: In various examples, one or more components or regions of a processing unit-such as a processing core, and/or component thereof—may be tested for faults during deployment in the field. To perform testing while in deployment, the state of a component subject to test may be retrieved and/or stored during the test to maintain state integrity, the component may be clamped to communicatively isolate the component from other components of the processing unit, a test vector may be applied to the component, and the output of the component may be compared against an expected output to determine if any faults are present. The state of the component may be restored after testing, and the clamp removed, thereby returning the component to its operating state without a perceivable detriment to operation of the processing unit in deployment.

Abstract: In various examples, one or more components or regions of a processing unit—such as a processing core, and/or component thereof—may be tested for faults during deployment in the field. To perform testing while in deployment, the state of a component subject to test may be retrieved and/or stored during the test to maintain state integrity, the component may be clamped to communicatively isolate the component from other components of the processing unit, a test vector may be applied to the component, and the output of the component may be compared against an expected output to determine if any faults are present. The state of the component may be restored after testing, and the clamp removed, thereby returning the component to its operating state without a perceivable detriment to operation of the processing unit in deployment.

Abstract: Apparatuses, systems, and techniques to optimize processor resources at a user-defined level. In at least one embodiment, priority of one or more tasks are adjusted to prevent one or more other dependent tasks from entering an idle state due to lack of resources to consume.

Abstract: In training a deep neural network using reduced precision, gradient computation operates on larger values without affecting the rest of the training procedure. One technique trains the deep neural network to develop loss, scales the loss, computes gradients at a reduced precision, and reduces the magnitude of the computed gradients to compensate for scaling of the loss. In one example non-limiting arrangement, the training forward pass scales a loss value by some factor S and the weight update reduces the weight gradient contribution by 1/S. Several techniques can be used for selecting scaling factor S and adjusting the weight update.

Abstract: In training a deep neural network using reduced precision, gradient computation operates on larger values without affecting the rest of the training procedure. One technique trains the deep neural network to develop loss, scales the loss, computes gradients at a reduced precision, and reduces the magnitude of the computed gradients to compensate for scaling of the loss. In one example non-limiting arrangement, the training forward pass scales a loss value by some factor S and the weight update reduces the weight gradient contribution by 1/S. Several techniques can be used for selecting scaling factor S and adjusting the weight update.

Abstract: In various examples, one or more components or regions of a processing unit—such as a processing core, and/or component thereof—may be tested for faults during deployment in the field. To perform testing while in deployment, the state of a component subject to test may be retrieved and/or stored during the test to maintain state integrity, the component may be clamped to communicatively isolate the component from other components of the processing unit, a test vector may be applied to the component, and the output of the component may be compared against an expected output to determine if any faults are present. The state of the component may be restored after testing, and the clamp removed, thereby returning the component to its operating state without a perceivable detriment to operation of the processing unit in deployment.

Abstract: In various examples, one or more components or regions of a processing unit—such as a processing core, and/or component thereof—may be tested for faults during deployment in the field. To perform testing while in deployment, the state of a component subject to test may be retrieved and/or stored during the test to maintain state integrity, the component may be clamped to communicatively isolate the component from other components of the processing unit, a test vector may be applied to the component, and the output of the component may be compared against an expected output to determine if any faults are present. The state of the component may be restored after testing, and the clamp removed, thereby returning the component to its operating state without a perceivable detriment to operation of the processing unit in deployment.

Abstract: In various examples, one or more components or regions of a processing unit—such as a processing core, and/or component thereof—may be tested for faults during deployment in the field. To perform testing while in deployment, the state of a component subject to test may be retrieved and/or stored during the test to maintain state integrity, the component may be clamped to communicatively isolate the component from other components of the processing unit, a test vector may be applied to the component, and the output of the component may be compared against an expected output to determine if any faults are present. The state of the component may be restored after testing, and the clamp removed, thereby returning the component to its operating state without a perceivable detriment to operation of the processing unit in deployment.

Abstract: In various examples, one or more components or regions of a processing unit—such as a processing core, and/or component thereof—may be tested for faults during deployment in the field. To perform testing while in deployment, the state of a component subject to test may be retrieved and/or stored during the test to maintain state integrity, the component may be clamped to communicatively isolate the component from other components of the processing unit, a test vector may be applied to the component, and the output of the component may be compared against an expected output to determine if any faults are present. The state of the component may be restored after testing, and the clamp removed, thereby returning the component to its operating state without a perceivable detriment to operation of the processing unit in deployment.

Abstract: In various examples, one or more components or regions of a processing unit—such as a processing core, and/or component thereof—may be tested for faults during deployment in the field. To perform testing while in deployment, the state of a component subject to test may be retrieved and/or stored during the test to maintain state integrity, the component may be clamped to communicatively isolate the component from other components of the processing unit, a test vector may be applied to the component, and the output of the component may be compared against an expected output to determine if any faults are present. The state of the component may be restored after testing, and the clamp removed, thereby returning the component to its operating state without a perceivable detriment to operation of the processing unit in deployment.

Abstract: Apparatuses, systems, and techniques to optimize processor resources at a user-defined level. In at least one embodiment, priority of one or more tasks are adjusted to prevent one or more other dependent tasks from entering an idle state due to lack of resources to consume.

Abstract: A multi-pass unit interoperates with a device driver to configure a screen space pipeline to perform multiple processing passes with buffered graphics primitives. The multi-pass unit receives primitive data and state bundles from the device driver. The primitive data includes a graphics primitive and a primitive mask. The primitive mask indicates the specific passes when the graphics primitive should be processed. The state bundles include one or more state settings and a state mask. The state mask indicates the specific passes where the state settings should be applied. The primitives and state settings are interleaved. For a given pass, the multi-pass unit extracts the interleaved state settings for that pass and configures the screen space pipeline according to those state settings. The multi-pass unit also extracts the interleaved graphics primitives to be processed in that pass. Then, the multi-pass unit causes the screen space pipeline to process those graphics primitives.

Abstract: A multi-pass unit interoperates with a device driver to configure a screen space pipeline to perform multiple processing passes with buffered graphics primitives. The multi-pass unit receives primitive data and state bundles from the device driver. The primitive data includes a graphics primitive and a primitive mask. The primitive mask indicates the specific passes when the graphics primitive should be processed. The state bundles include one or more state settings and a state mask. The state mask indicates the specific passes where the state settings should be applied. The primitives and state settings are interleaved. For a given pass, the multi-pass unit extracts the interleaved state settings for that pass and configures the screen space pipeline according to those state settings. The multi-pass unit also extracts the interleaved graphics primitives to be processed in that pass. Then, the multi-pass unit causes the screen space pipeline to process those graphics primitives.

Abstract: In training a deep neural network using reduced precision, gradient computation operates on larger values without affecting the rest of the training procedure. One technique trains the deep neural network to develop loss, scales the loss, computes gradients at a reduced precision, and reduces the magnitude of the computed gradients to compensate for scaling of the loss. In one example non-limiting arrangement, the training forward pass scales a loss value by some factor S and the weight update reduces the weight gradient contribution by 1/S. Several techniques can be used for selecting scaling factor S and adjusting the weight update.

Abstract: A power subsystem is configured to manage the maximum power usage of a computer subsystem. A power detector determines when power usage approaches the maximum capability of the power supply. The power detector generates a signal that corresponds to power usage. A controller then applies the signal to the system voltage regulator as a secondary regulation function such that the output voltage is reduced in a manner that supports maximum operating voltage while limiting power usage to within the capability of the power supply. The controller may configure the signal to implement the secondary regulation function as a modification of the feedback voltage, the reference voltage, or the current feedback of the regulator. As a result the subsystem causes the computer subsystem to operate at an optimum point on the voltage-current curve of the power supply.

Abstract: A multi-pass unit interoperates with a device driver to configure a screen space pipeline to perform multiple processing passes with buffered graphics primitives. The multi-pass unit receives primitive data and state bundles from the device driver. The primitive data includes a graphics primitive and a primitive mask. The primitive mask indicates the specific passes when the graphics primitive should be processed. The state bundles include one or more state settings and a state mask. The state mask indicates the specific passes where the state settings should be applied. The primitives and state settings are interleaved. For a given pass, the multi-pass unit extracts the interleaved state settings for that pass and configures the screen space pipeline according to those state settings. The multi-pass unit also extracts the interleaved graphics primitives to be processed in that pass. Then, the multi-pass unit causes the screen space pipeline to process those graphics primitives.

Abstract: A multi-pass unit interoperates with a device driver to configure a screen space pipeline to perform multiple processing passes with buffered graphics primitives. The multi-pass unit receives primitive data and state bundles from the device driver. The primitive data includes a graphics primitive and a primitive mask. The primitive mask indicates the specific passes when the graphics primitive should be processed. The state bundles include one or more state settings and a state mask. The state mask indicates the specific passes where the state settings should be applied. The primitives and state settings are interleaved. For a given pass, the multi-pass unit extracts the interleaved state settings for that pass and configures the screen space pipeline according to those state settings. The multi-pass unit also extracts the interleaved graphics primitives to be processed in that pass. Then, the multi-pass unit causes the screen space pipeline to process those graphics primitives.

Abstract: A power subsystem is configured to manage the maximum power usage of a computer subsystem. A power detector determines when power usage approaches the maximum capability of the power supply. The power detector generates a signal that corresponds to power usage. A controller then applies the signal to the system voltage regulator as a secondary regulation function such that the output voltage is reduced in a manner that supports maximum operating voltage while limiting power usage to within the capability of the power supply. The controller may configure the signal to implement the secondary regulation function as a modification of the feedback voltage, the reference voltage, or the current feedback of the regulator. As a result the subsystem causes the computer subsystem to operate at an optimum point on the voltage-current curve of the power supply.

Abstract: A clock gating latch, a method of gating a clock signal and an integrating circuit incorporating the clock gating latch or the method. In one embodiment, the clock gating latch includes: (1) a propagation circuit having a single, first switch configured to be driven by an input clock signal, (2) a keeper circuit coupled to the propagation circuit and having a single, first switch configured to be driven by the input clock signal and (3) an AND gate coupled to the propagation circuit and the keeper circuit and having an internal node coupled to a second switch in the propagation circuit and a second switch in the keeper circuit.

Abstract: Embodiments of the present technology are directed toward circuits for gating pre-charging sense nodes within a flip-flop when an input data signal changes and a clock signal is in a given state. Embodiments of the present technology are further directed toward circuits for maintaining a state of the sense nodes.