Nvidia Corporation

Timothy John Purcell

Jerome Francis Duluk, Jr

Lacky V Shah

James Leroy Deming

Wishwesh Anil Gandhi

Steven E Molnar

Sean Jeffrey Treichler

Karim Abdalla

Gregory Scott Palmer

Shailendra Singh

Alan Menezes

Manas Mandal

Praveen Joginipally

Pranav Vaidya

Feiqi Su

Ziyad Sami Hakura

Emmett Michael Kilgariff

Samuel Hammond Duncan

Brian Matthew Fahs

Mark Hairgrove

Lucien Dunning

Philip Alexander Cuadra

Karan Mehra

Nick Barrow-Williams

Jonathon Stuart Ramsey Evans

Eric Rock

Franklin C Crow

Gentaro Hirota

James C Bowman

Johnny S Rhoades

Praveen Krishnamurthy

Tanmoy Mandal

Peter B Holmquist

Dale L Kirkland

Gregory Muthler

Peter B Holmqvist

George R Lynch

Luke Durant

James Robertson

Jesse David Hall

Brian S Pharris

John S Rhoades

Cynthia Ann Edgeworth Allison

Gerald F Luiz

Hemayet Hossain

Jonathon Stuart Ramsay Evans

Gernot Schaufler

Karl M Wurstner

Sonata Gale Wen

A parallel processing unit (PPU), operating in a traditional processing environment or in a virtualized processing environment, can be divided into partitions. Each partition is configured to operate similarly to how the entire PPU operates. A given partition includes a subset of the computational and memory resources associated with the entire PPU. Software that executes on a CPU partitions the PPU for an admin user. A guest user is assigned to a partition and can perform processing tasks within that partition in isolation from any other guest users assigned to any other partitions. Because the PPU can be divided into isolated partitions, multiple CPU processes can efficiently utilize PPU resources.

Techniques for configuring a processor to function as multiple, separate processors in a virtualized environment

A parallel processing unit (PPU) can be divided into partitions. Each partition is configured to operate similarly to how the entire PPU operates. A given partition includes a subset of the computational and memory resources associated with the entire PPU. Software that executes on a CPU partitions the PPU for an admin user. A guest user is assigned to a partition and can perform processing tasks within that partition in isolation from any other guest users assigned to any other partitions. Because the PPU can be divided into isolated partitions, multiple CPU processes can efficiently utilize PPU resources.

Techniques for configuring a processor to function as multiple, separate processors

Techniques for reconfiguring partitions in a parallel processing system

In one embodiment of the present invention a cache unit organizes data stored in an attached memory to optimize accesses to compressed data. In operation, the cache unit introduces a layer of indirection between a physical address associated with a memory access request and groups of blocks in the attached memory. The layer of indirection&#x2014;virtual tiles&#x2014;enables the cache unit to selectively store compressed data that would conventionally be stored in separate physical tiles included in a group of blocks in a single physical tile. Because the cache unit stores compressed data associated with multiple physical tiles in a single physical tile and, more specifically, in adjacent locations within the single physical tile, the cache unit coalesces the compressed data into contiguous blocks. Subsequently, upon performing a read operation, the cache unit may retrieve the compressed data conventionally associated with separate physical tiles in a single read operation.

Organizing memory to optimize memory accesses of compressed data

A technique for simultaneously executing multiple tasks, each having an independent virtual address space, involves assigning an address space identifier (ASID) to each task and constructing each virtual memory access request to include both a virtual address and the ASID. During virtual to physical address translation, the ASID selects a corresponding page table, which includes virtual to physical address mappings for the ASID and associated task. Entries for a translation look-aside buffer (TLB) include both the virtual address and ASID to complete each mapping to a physical address. Deep scheduling of tasks sharing a virtual address space may be implemented to improve cache affinity for both TLB and data caches.

Efficient memory virtualization in multi-threaded processing units

One embodiment of the present invention sets forth a technique for error-checking a compute task. The technique involves receiving a pointer to a compute task, storing the pointer in a scheduling queue, determining that the compute task should be executed, retrieving the pointer from the scheduling queue, determining via an error-check procedure that the compute task is eligible for execution, and executing the compute task.

Error checking in out-of-order task scheduling

One embodiment of the present invention sets forth a technique for selecting a first processor included in a plurality of processors to receive work related to a compute task. The technique involves analyzing state data of each processor in the plurality of processors to identify one or more processors that have already been assigned one compute task and are eligible to receive work related to the one compute task, receiving, from each of the one or more processors identified as eligible, an availability value that indicates the capacity of the processor to receive new work, selecting a first processor to receive work related to the one compute task based on the availability values received from the one or more processors, and issuing, to the first processor via a cooperative thread array (CTA), the work related to the one compute task.

Execution state analysis for assigning tasks to streaming multiprocessors

One embodiment of the present invention sets forth a technique for processing commands received by an intermediary cache from one or more clients. The technique involves receiving a first write command from an arbiter unit, where the first write command specifies a first memory address, determining that a first cache line related to a set of cache lines included in the intermediary cache is associated with the first memory address, causing data associated with the first write command to be written into the first cache line, and marking the first cache line as dirty. The technique further involves determining whether a total number of cache lines marked as dirty in the set of cache lines is less than, equal to, or greater than a first threshold value, and: not transmitting a dirty data notification to the frame buffer logic when the total number is less than the threshold value, or transmitting a dirty data notification to the frame buffer logic when the total number is equal to or greater than the first threshold value.

Control mechanism for fine-tuned cache to backing-store synchronization

One embodiment of the present invention sets forth a technique for managing the allocation and release of resources during multi-threaded program execution. Programmable reference counters are initialized to values that limit the amount of resources for allocation to tasks that share the same reference counter. Resource parameters are specified for each task to define the amount of resources allocated for consumption by each array of execution threads that is launched to execute the task. The resource parameters also specify the behavior of the array for acquiring and releasing resources. Finally, during execution of each thread in the array, an exit instruction may be configured to override the release of the resources that were allocated to the array. The resources may then be retained for use by a child task that is generated during execution of a thread.

Compute work distribution reference counters

One embodiment of the present invention sets forth a technique for computing line stipple using a parallel rasterizer. Stipple phases are computed in parallel for individual line segments of a line strip during the viewport scale, cull, and clipping operations. The line segments are distributed to multiple parallel rasterizers. Each line segment may be sent to only one of the parallel rasterizers. Update phase messages that include an accumulated stipple phase for a batch of line segments are broadcast to all of the multiple parallel rasterizers. The update phase messages are used by the multiple parallel rasterizers to reconstruct the stipple phases for each line segment of a line strip in order to correctly render stippled line strips and produce a continuous stippled line.

Parallel line stipple computation

One embodiment of the present invention sets forth a technique for assigning a compute task to a first processor included in a plurality of processors. The technique involves analyzing each compute task in a plurality of compute tasks to identify one or more compute tasks that are eligible for assignment to the first processor, where each compute task is listed in a first table and is associated with a priority value and an allocation order that indicates relative time at which the compute task was added to the first table. The technique further involves selecting a first task compute from the identified one or more compute tasks based on at least one of the priority value and the allocation order, and assigning the first compute task to the first processor for execution.

Scheduling and execution of compute tasks

Growing community of inventors

Provo, UT, United States of America

Timothy John Purcell