Abstract: One embodiment provides a graphics processor comprising a memory controller and a graphics processing resource coupled with the memory controller. The graphics processing resource includes circuitry configured to execute an instruction to perform a matrix operation on first input including weight data and second input including input activation data, generate intermediate data based on a result of the matrix operation, quantize the intermediate data to a floating-point format determined based on a statistical distribution of first output data, and output, as second output data, quantized intermediate data in a determined floating-point format.

Abstract: One embodiment provides for a compute apparatus comprising a decode unit to decode a single instruction into a decoded instruction that specifies multiple operands including a multi-bit input value and a ternary weight associated with a neural network and an arithmetic logic unit including a multiplier, an adder, and an accumulator register. To execute the decoded instruction, the multiplier is to perform a multiplication operation on the multi-bit input based on the ternary weight to generate an intermediate product and the adder is to add the intermediate product to a value stored in the accumulator register and update the value stored in the accumulator register.

Abstract: Systems, methods, tangible non-transitory computer-readable media, and devices associated with the motion prediction and operation of a device including a vehicle are provided. For example, a vehicle computing system can access state data including information associated with locations and characteristics of objects over a plurality of time intervals. Trajectories of the objects at subsequent time intervals following the plurality of time intervals can be determined based on the state data and a machine-learned tracking and kinematics model. The trajectories of the objects can include predicted locations of the objects at subsequent time intervals that follow the plurality of time intervals. Further, the predicted locations of the objects can be based on physical constraints of the objects. Furthermore, indications, which can include visual indications, can be generated based on the predicted locations of the objects at the subsequent time intervals.

Abstract: In one embodiment, a method includes determining an initial cost volume associated with a plurality of potential trajectories of a vehicle in an environment based on a set of movement restrictions of the vehicle, generating a delta cost volume using the initial cost volume and environment data associated with the environment, wherein the delta cost volume is generated by determining adjustments to the initial cost volume that incorporate observed driving behavior, and scoring a trajectory of the plurality of potential trajectories for the vehicle based on the initial cost volume and the delta cost volume.

Abstract: In an example, an apparatus comprises a plurality of processing unit cores, a plurality of cache memory modules associated with the plurality of processing unit cores, and a machine learning model communicatively coupled to the plurality of processing unit cores, wherein the plurality of cache memory modules share cache coherency data with the machine learning model. Other embodiments are also disclosed and claimed.

Abstract: Embodiments provide mechanisms to facilitate compute operations for deep neural networks. One embodiment comprises a graphics processing unit comprising one or more multiprocessors, at least one of the one or more multiprocessors including a register file to store a plurality of different types of operands and a plurality of processing cores. The plurality of processing cores includes a first set of processing cores of a first type and a second set of processing cores of a second type. The first set of processing cores are associated with a first memory channel and the second set of processing cores are associated with a second memory channel.

Abstract: One embodiment provides for a graphics processing unit to accelerate machine-learning operations, the graphics processing unit comprising a multiprocessor having a single instruction, multiple thread (SIMT) architecture, the multiprocessor to execute at least one single instruction; and a first compute unit included within the multiprocessor, the at least one single instruction to cause the first compute unit to perform a two-dimensional matrix multiply and accumulate operation, wherein to perform the two-dimensional matrix multiply and accumulate operation includes to compute an intermediate product of 16-bit operands and to compute a 32-bit sum based on the intermediate product.

Abstract: Techniques to improve performance of matrix multiply operations are described in which a compute kernel can specify one or more element-wise operations to perform on output of the compute kernel before the output is transferred to higher levels of a processor memory hierarchy.

Abstract: A semiconductor device includes a substrate, a 2-D material layer, source/drain contacts, and a gate electrode. The 2-D material layer is over the substrate, the 2-D material layer includes source/drain regions and a channel region between the source/drain regions, in which the 2-D material layer is made of a transition metal dichalcogenide (TMD). The source/drain contacts are in contact with source/drain regions of the 2-D material layer, in which a binding energy of transition metal atoms at the channel region of the 2-D material layer is different from a binding energy of the transition metal atoms at the source/drain regions of the 2-D material layer. The gate electrode is over the substrate.

Abstract: An apparatus to facilitate compute optimization is disclosed. The apparatus includes one or more processing units to provide a first set of shader operations associated with a shader stage of a graphics pipeline, a scheduler to schedule shader threads for processing, and a field-programmable gate array (FPGA) dynamically configured to provide a second set of shader operations associated with the shader stage of the graphics pipeline.

Abstract: One embodiment provides a graphics processor comprising a memory controller and a graphics processing resource coupled with the memory controller. The graphics processing resource includes circuitry configured to execute an instruction to perform a matrix operation on first input including weight data and second input including input activation data, generate intermediate data based on a result of the matrix operation, quantize the intermediate data to a floating-point format determined based on a statistical distribution of first output data, and output, as second output data, quantized intermediate data in a determined floating-point format.

Abstract: A mechanism is described for facilitating intelligent thread scheduling at autonomous machines. A method of embodiments, as described herein, includes detecting dependency information relating to a plurality of threads corresponding to a plurality of workloads associated with tasks relating to a processor including a graphics processor. The method may further include generating a tree of thread groups based on the dependency information, where each thread group includes multiple threads, and scheduling one or more of the thread groups associated a similar dependency to avoid dependency conflicts.

Abstract: Embodiments provide mechanisms to facilitate compute operations for deep neural networks. One embodiment comprises a graphics processing unit comprising one or more multiprocessors, at least one of the one or more multiprocessors including a register file to store a plurality of different types of operands and a plurality of processing cores. The plurality of processing cores includes a first set of processing cores of a first type and a second set of processing cores of a second type. The first set of processing cores are associated with a first memory channel and the second set of processing cores are associated with a second memory channel.

Abstract: Techniques to improve performance of matrix multiply operations are described in which a compute kernel can specify one or more element-wise operations to perform on output of the compute kernel before the output is transferred to higher levels of a processor memory hierarchy.

Abstract: A mechanism is described for facilitating intelligent thread scheduling at autonomous machines. A method of embodiments, as described herein, includes detecting dependency information relating to a plurality of threads corresponding to a plurality of workloads associated with tasks relating to a processor including a graphics processor. The method may further include generating a tree of thread groups based on the dependency information, where each thread group includes multiple threads, and scheduling one or more of the thread groups associated a similar dependency to avoid dependency conflicts.

Abstract: A processing apparatus is provided comprising a multiprocessor having a multithreaded architecture. The multiprocessor can execute at least one single instruction to perform parallel mixed precision matrix operations. In one embodiment the apparatus includes a memory interface and an array of multiprocessors coupled to the memory interface. At least one multiprocessor in the array of multiprocessors is configured to execute a fused multiply-add instruction in parallel across multiple threads.

Abstract: An apparatus to facilitate compute optimization is disclosed. The apparatus includes a plurality of processing units each comprising a plurality of execution units (EUs), wherein the plurality of EUs comprise a first EU type and a second EU type.

Abstract: One embodiment provides for a compute apparatus to perform machine learning operations, the compute apparatus comprising a decode unit to decode a single instruction into a decoded instruction, the decoded instruction to cause the compute apparatus to perform a complex compute operation.

Abstract: An apparatus to facilitate compute optimization is disclosed. The apparatus includes one or more processing units to provide a first set of shader operations associated with a shader stage of a graphics pipeline, a scheduler to schedule shader threads for processing, and a field-programmable gate array (FPGA) dynamically configured to provide a second set of shader operations associated with the shader stage of the graphics pipeline.

Abstract: An apparatus to facilitate compute optimization is disclosed. The apparatus includes a plurality of processing units each comprising a plurality of processing cores of a first type and a second type. A first set of processing cores of a first type perform multi-dimensional matrix operations and a second set of processing cores of a second type perform general purpose graphics processing unit (GPGPU) operations.