Microsoft Technology Licensing, LLC

Bradley Post

Parag Chakraborty

B Anil Kumar

Vladimir Pavlov

Rene Antonio Vega

Varadharajan Ponnappan

Pareekshit Singh

Mahesh Sharad Lotlikar

Meher P Malakapalli

Karthik Lakshminarayanan

Winston Matthew Penfold Johnston

Eric Kai-hau Han

Ed Cox

Methods, systems, and computer readable media are disclosed for creating a multi-port client server connection in a remote desktop environment. In an embodiment, the multi-client server connection can decouple client-server connection and client-server input from graphics delivery from the server to the client. Such embodiment provides higher server performance and enables dynamic server resource management. The multi-client server connection can be implemented in a virtual environment or between physical machines where one set of physical machines receives the graphics requests from the client and another set of physical machines generates and transmits display data to the client.

Delivering a single end user experience to a client from multiple servers

Memory is reserved in a virtualized computing environment for graphics processing of each child partition in the computing environment. A video memory controller can identify video settings for child partitions. The video memory controller can determine an amount of memory for graphics processing for a child partition based on the video settings for that child partition. The video memory can also request an amount of memory to be reserved for that child partition based on the calculated amount of memory. Reserving memory for graphics processing of child partitions in this way allows for a sufficient amount of memory to be reserved for a child partition without wasting memory resources by reserving significantly more memory than is needed for the child partition.

Dynamic provisioning of virtual video memory based on virtual video controller configuration

Described herein is providing GPU resources across machine boundaries for a virtual machine that migrates between servers. Data centers tend to have racks of servers that have limited access to GPUs. Accordingly, disclosed herein is providing GPU resources to computing devices that have limited access to GPUs across machine boundaries.

Delivering GPU resources to a migrating virtual machine

Described herein is providing GPU resources across machine boundaries. Data centers tend to have racks of servers that have limited access to GPUs. Accordingly, disclosed herein is providing GPU resources to computing devices that have limited access to GPUs across machine boundaries.

Delivering GPU resources across machine boundaries

Exemplary techniques for balancing 3D graphical processor unit use among virtual machines are herein disclosed. In an exemplary embodiment, a virtualization platform can load an instance of a graphics rendering module for a virtual machine; select a GPU for the graphics rendering module to run on; and configure the virtual machine to render to the selected GPU. In addition to the foregoing, other aspects are described in the detailed description, claims, and figures.

Techniques for load balancing GPU enabled virtual machines

Exemplary techniques for recovering from a graphics processor reset are herein disclosed. In an exemplary embodiment, a 3D graphics service managercan detect that a graphics processing unit reset and can restart a rendering process configured to render 3D graphics for a virtual machine and cause a graphics buffer to be established between the rendering process and the virtual machine. In addition to the foregoing, other aspects are described in the detailed description, claims, and figures.

Dynamic virtual device failure recovery

A mechanism is described for mitigating the effects of such a DoS attack by detecting a multiple TDR situation within a short duration, identifying the errant virtual machines, and suspending all rendering ability for that virtual machine or set of virtual machines. While the disclosed embodiments are described in the context of virtual machines, the principles may be extended to the general problem of DoS attacks due to TDRs and bugchecks on any physical machine that includes a GPU. For example, DoS attacks can be generated from web sites directly using GPU rendering.

Prevention of DoS attack by a rogue graphics application

A mechanism for the creation of a shared memory aperture between modes in a parent and child partition is described. The shared memory aperture can be created between any memory mode between the guest and any host. For example, a shared memory aperture can be created between the kernel mode on the child partition and the user mode on the parent partition.

Shared memory between child and parent partitions

Dynamically compiled code is stored for future use by an emulator application. The emulator application may recall the previously compiled code rather than recompile the same code again. Recycling previously compiled code prevents unneeded code compilation, and saves valuable processing time and resources. Previously compiled code can be associated with code identification information. When compiled code is needed, one or more indexes of compiled code identification information is searched. If the appropriate identification information is found, the corresponding code is loaded and executed.

Caching dynamically compiled code to storage

In a virtual computing machine, a system and method that dynamically patches the interrupt mechanism (in interrupt vector space) of a host computing architecture with guest mode software. Significant increases in performance are achieved without depending on the host code. A patching mechanism evaluates the operating system version, processor, and code to be patched. If patchable, low-level interfaces are created dynamically; a dispatcher is written into an unused location in vector space, and instructions copied from each interrupt vector to be patched to a guest interrupt vector. For an interrupt, the new, patched instructions branch to the dispatcher, which then branches to the appropriate patched interrupt guest code. If the processor is operating as a virtual machine, the guest interrupt code handles the interrupt, otherwise the original copied instructions are replayed, followed by execution at the original host instruction in vector space that exists after the copied and patched instructions.

Growing community of inventors

San Francisco, CA, United States of America

Bradley Post