Intel Corporation

Lance Cheney

Altug Koker

Varghese George

Sanjeev S Jahagirdar

Lakshminarayanan Striramassarma

Josh B Mastronarde

Iqbal Rajwani

Vikranth Vemulapalli

Naveen Matam

Melaku Teshome

Eric Finley

Binoj Xavier

Allan Wong

Brian D Rauchfuss

Anthony Babella

A disaggregated processor package can be configured to accept interchangeable chiplets. Interchangeability is enabled by specifying a standard physical interconnect for chiplets that can enable the chiplet to interface with a fabric or bridge interconnect. Chiplets from different IP designers can conform to the common interconnect, enabling such chiplets to be interchangeable during assembly. The fabric and bridge interconnects logic on the chiplet can then be configured to confirm with the actual interconnect layout of the on-board logic of the chiplet. Additionally, data from chiplets can be transmitted across an inter-chiplet fabric using encapsulation, such that the actual data being transferred is opaque to the fabric, further enable interchangeability of the individual chiplets. With such an interchangeable design, cache or DRAM memory can be inserted into memory chiplet slots, while compute or graphics chiplets with a higher or lower core count can be inserted into logic chiplet slots.

Enabling product SKUs based on chiplet configurations

Embodiments described herein provide techniques to disaggregate an architecture of a system on a chip integrated circuit into multiple distinct chiplets that can be packaged onto a common chassis. In one embodiment, a graphics processing unit or parallel processor is composed from diverse silicon chiplets that are separately manufactured. A chiplet is an at least partially and distinctly packaged integrated circuit that includes distinct units of logic that can be assembled with other chiplets into a larger package. A diverse set of chiplets with different IP core logic can be assembled into a single device.

Disaggregation of system-on-chip (SOC) architecture

Embodiments described herein provide techniques to disaggregate an architecture of a system on a chip integrated circuit into multiple distinct chiplets that can be packaged onto a common chassis. In one embodiment, a graphics processing unit or parallel processor is composed from diverse silicon chiplets that are separately manufactured. A chiplet is an at least partially packaged integrated circuit that includes distinct units of logic that can be assembled with other chiplets into a larger package. A diverse set of chiplets with different IP core logic can be assembled into a single device.

Disaggregation of System-On-Chip (SOC) architecture

A disaggregated processor package can be configured to accept interchangeable chiplets. Interchangeability is enabled by specifying a standard physical interconnect for chiplets that can enable the chiplet to interface with a fabric or bridge interconnect. Chiplets from different IP designers can conform to the common interconnect, enabling such chiplets to be interchangeable during assembly. The fabric and bridge interconnects logic on the chiplet can then be configured to confirm with the actual interconnect layout of the on-board logic of the chiplet. Additionally, data from chiplets can be transmitted across an inter-chiplet fabric using encapsulation, such that the actual data being transferred is opaque to the fabric, further enable interchangeability of the individual chiplets. With such an interchangeable design, higher or lower density memory can be inserted into memory chiplet slots, while compute or graphics chiplets with a higher or lower core count can be inserted into logic chiplet slots.

Enabling product SKUS based on chiplet configurations

Disaggregation of SOC architecture

Provided are a method and system for activating a design test mode in a graphics card having multiple execution units. A design test mode is activated in a graphics module comprising multiple execution units coupled to a cache on a bus. The bus is configured to return test instructions from the cache to the execution units in response to a request from one execution unit for the test instructions from the cache in the design test mode. The execution units execute the test instructions during the design test mode. Interrupts are prevented during the design test mode.

Activating a design test mode in a graphics card having multiple execution units to bypass a host cache and transfer test instructions directly to an instruction cache

Provided are a method and system checking output from multiple execution units. Execution units concurrently execute test instructions to generate test output, wherein test instructions are transferred to the execution units from a cache coupled to the execution units over a bus. The test output from the execution units is compared to determine whether the output from the execution units indicates the execution units are properly concurrently executing test instructions. The result of the comparing of the test output are forwarded to a design test unit.

Growing community of inventors

El Dorado Hills, CA, United States of America

Lance Cheney