Nvidia Corporation

Omkar Paranjape

Shirish C Gadre

Steven James Heinrich

Jack H Choquette

Michael Alan Fetterman

Xiaogang Qiu

Anjana Rajendran

John H Edmondson

Rajeshwaran Selvanesan

Eric Lyell Hill

Poornachandra Rao

Charles McCarver

Ramesh Jandhyala

Mark Alan Gebhart

Olivier Giroux

Ronny Krashinsky

Manan Patel

Stewart Glenn Carlton

Andrew Robert Kerr

Douglas J Hahn

Alan Kaatz

Kevin Mitchell

Michael Siu

Emmett Michael Kilgariff

Lacky V Shah

Rouslan Lyubomirov Dimitrov

James Robertson

Gentaro Hirota

Stefano A Pescador

Gregory Scott Palmer

Lei Yang

Rajballav Dash

Raghavan Madhavan

William Newhall

Sriharsha Niverty

Milton Lei

Jeff Schottmiller

Mitchell Hayenga

Poorna Rao

A technique for block data transfer is disclosed that reduces data transfer and memory access overheads and significantly reduces multiprocessor activity and energy consumption. Threads executing on a multiprocessor needing data stored in global memory can request and store the needed data in on-chip shared memory, which can be accessed by the threads multiple times. The data can be loaded from global memory and stored in shared memory using an instruction which directs the data into the shared memory without storing the data in registers and/or cache memory of the multiprocessor during the data transfer.

Techniques for efficiently transferring data to a processor

In a streaming cache, multiple, dynamically sized tracking queues are employed. Request tracking information is distributed among the plural tracking queues to selectively enable out-of-order memory request returns. A dynamically controlled policy assigns pending requests to tracking queues, providing for example in-order memory returns in some contexts and/or for some traffic and out of order memory returns in other contexts and/or for other traffic.

Out of order memory request tracking structure and technique

A texture processing pipeline in a graphics processing unit generates the surface appearance for objects in a computer-generated scene. This texture processing pipeline determines, at multiple stages within the texture processing pipeline, whether texture operations and texture loads may be processed at an accelerated rate. At each stage that includes a decision point, the texture processing pipeline assumes that the current texture operation or texture load can be accelerated unless specific, known information indicates that the texture operation or texture load cannot be accelerated. As a result, the texture processing pipeline increases the number of texture operations and texture loads that are accelerated relative to the number of texture operations and texture loads that are not accelerated.

Techniques for performing accelerated point sampling in a texture processing pipeline

A parallel processing unit (e.g., a GPU), in some examples, includes a hardware scheduler and hardware arbiter that launch graphics and compute work for simultaneous execution on a SIMD/SIMT processing unit. Each processing unit (e.g., a streaming multiprocessor) of the parallel processing unit operates in a graphics-greedy mode or a compute-greedy mode at respective times. The hardware arbiter, in response to a result of a comparison of at least one monitored performance or utilization metric to a user-configured threshold, can selectively cause the processing unit to run one or more compute work items from a compute queue when the processing unit is operating in the graphics-greedy mode, and cause the processing unit to run one or more graphics work items from a graphics queue when the processing unit is operating in the compute-greedy mode. Associated methods and systems are also described.

Simultaneous compute and graphics scheduling

A unified cache subsystem includes a data memory configured as both a shared memory and a local cache memory. The unified cache subsystem processes different types of memory transactions using different data pathways. To process memory transactions that target shared memory, the unified cache subsystem includes a direct pathway to the data memory. To process memory transactions that do not target shared memory, the unified cache subsystem includes a tag processing pipeline configured to identify cache hits and cache misses. When the tag processing pipeline identifies a cache hit for a given memory transaction, the transaction is rerouted to the direct pathway to data memory. When the tag processing pipeline identifies a cache miss for a given memory transaction, the transaction is pushed into a first-in first-out (FIFO) until miss data is returned from external memory. The tag processing pipeline is also configured to process texture-oriented memory transactions.

Unified cache for diverse memory traffic

One embodiment of the present disclosure sets forth an optimized way to execute pre-scheduled replay operations for divergent operations in a parallel processing subsystem. Specifically, a streaming multiprocessor (SM) includes a multi-stage pipeline configured to insert pre-scheduled replay operations into a multi-stage pipeline. A pre-scheduled replay unit detects whether the operation associated with the current instruction is accessing a common resource. If the threads are accessing data which are distributed across multiple cache lines, then the pre-scheduled replay unit inserts pre-scheduled replay operations behind the current instruction. The multi-stage pipeline executes the instruction and the associated pre-scheduled replay operations sequentially. If additional threads remain unserviced after execution of the instruction and the pre-scheduled replay operations, then additional replay operations are inserted via the replay loop, until all threads are serviced. One advantage of the disclosed technique is that divergent operations requiring one or more replay operations execute with reduced latency.

Pre-scheduled replays of divergent operations

One embodiment of the present disclosure sets forth an effective way to maintain fairness and order in the scheduling of common resource access requests related to replay operations. Specifically, a streaming multiprocessor (SM) includes a total order queue (TOQ) configured to schedule the access requests over one or more execution cycles. Access requests are allowed to make forward progress when needed common resources have been allocated to the request. Where multiple access requests require the same common resource, priority is given to the older access request. Access requests may be placed in a sleep state pending availability of certain common resources. Deadlock may be avoided by allowing an older access request to steal resources from a younger resource request. One advantage of the disclosed technique is that older common resource access requests are not repeatedly blocked from making forward progress by newer access requests.

Mechanism for waking common resource requests within a resource management subsystem

Resource management subsystem that maintains fairness and order

One embodiment of the present invention sets forth an approach for executing replay operations for divergent operations in a parallel processing subsystem. Specifically, the streaming multiprocessor (SM) includes a multistage pipeline configured to batch two or more replay operations for processing via replay loop. A logic element within the multistage pipeline detects whether the current pipeline stage is accessing a shared resource, such as loading data from a shared memory. If the threads are accessing data which are distributed across multiple cache lines, then the multistage pipeline batches two or more replay operations, where the replay operations are inserted into the pipeline back-to-back.

Batched replays of divergent operations

Mechanism for tracking age of common resource requests within a resource management subsystem

One embodiment of the present invention sets forth a technique for performing a shaped access of a register file that includes a set of N registers, wherein N is greater than or equal to two. The technique involves, for at least one thread included in a group of threads, receiving a request to access a first amount of data from each register in the set of N registers, and configuring a crossbar to allow the at least one thread to access the first amount of data from each register in the set of N registers.

Shaped register file reads

Growing community of inventors

Austin, TX, United States of America

Omkar Paranjape