US20050127490A1 - Multi-die processor - Google Patents
Multi-die processor Download PDFInfo
- Publication number
- US20050127490A1 US20050127490A1 US10/738,680 US73868003A US2005127490A1 US 20050127490 A1 US20050127490 A1 US 20050127490A1 US 73868003 A US73868003 A US 73868003A US 2005127490 A1 US2005127490 A1 US 2005127490A1
- Authority
- US
- United States
- Prior art keywords
- die
- logic
- processor
- logic portion
- execution
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F15/00—Digital computers in general; Data processing equipment in general
- G06F15/76—Architectures of general purpose stored program computers
- G06F15/78—Architectures of general purpose stored program computers comprising a single central processing unit
- G06F15/7828—Architectures of general purpose stored program computers comprising a single central processing unit without memory
- G06F15/7832—Architectures of general purpose stored program computers comprising a single central processing unit without memory on one IC chip (single chip microprocessors)
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L25/00—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof
- H01L25/03—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes
- H01L25/04—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes the devices not having separate containers
- H01L25/065—Assemblies consisting of a plurality of individual semiconductor or other solid state devices ; Multistep manufacturing processes thereof all the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/00, or in a single subclass of H10K, H10N, e.g. assemblies of rectifier diodes the devices not having separate containers the devices being of a type provided for in group H01L27/00
- H01L25/0657—Stacked arrangements of devices
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2224/00—Indexing scheme for arrangements for connecting or disconnecting semiconductor or solid-state bodies and methods related thereto as covered by H01L24/00
- H01L2224/01—Means for bonding being attached to, or being formed on, the surface to be connected, e.g. chip-to-package, die-attach, "first-level" interconnects; Manufacturing methods related thereto
- H01L2224/02—Bonding areas; Manufacturing methods related thereto
- H01L2224/04—Structure, shape, material or disposition of the bonding areas prior to the connecting process
- H01L2224/05—Structure, shape, material or disposition of the bonding areas prior to the connecting process of an individual bonding area
- H01L2224/0554—External layer
- H01L2224/05573—Single external layer
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2224/00—Indexing scheme for arrangements for connecting or disconnecting semiconductor or solid-state bodies and methods related thereto as covered by H01L24/00
- H01L2224/01—Means for bonding being attached to, or being formed on, the surface to be connected, e.g. chip-to-package, die-attach, "first-level" interconnects; Manufacturing methods related thereto
- H01L2224/02—Bonding areas; Manufacturing methods related thereto
- H01L2224/04—Structure, shape, material or disposition of the bonding areas prior to the connecting process
- H01L2224/05—Structure, shape, material or disposition of the bonding areas prior to the connecting process of an individual bonding area
- H01L2224/0554—External layer
- H01L2224/05599—Material
- H01L2224/056—Material with a principal constituent of the material being a metal or a metalloid, e.g. boron [B], silicon [Si], germanium [Ge], arsenic [As], antimony [Sb], tellurium [Te] and polonium [Po], and alloys thereof
- H01L2224/05638—Material with a principal constituent of the material being a metal or a metalloid, e.g. boron [B], silicon [Si], germanium [Ge], arsenic [As], antimony [Sb], tellurium [Te] and polonium [Po], and alloys thereof the principal constituent melting at a temperature of greater than or equal to 950°C and less than 1550°C
- H01L2224/05647—Copper [Cu] as principal constituent
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2224/00—Indexing scheme for arrangements for connecting or disconnecting semiconductor or solid-state bodies and methods related thereto as covered by H01L24/00
- H01L2224/01—Means for bonding being attached to, or being formed on, the surface to be connected, e.g. chip-to-package, die-attach, "first-level" interconnects; Manufacturing methods related thereto
- H01L2224/10—Bump connectors; Manufacturing methods related thereto
- H01L2224/12—Structure, shape, material or disposition of the bump connectors prior to the connecting process
- H01L2224/13—Structure, shape, material or disposition of the bump connectors prior to the connecting process of an individual bump connector
- H01L2224/13001—Core members of the bump connector
- H01L2224/1302—Disposition
- H01L2224/13025—Disposition the bump connector being disposed on a via connection of the semiconductor or solid-state body
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2224/00—Indexing scheme for arrangements for connecting or disconnecting semiconductor or solid-state bodies and methods related thereto as covered by H01L24/00
- H01L2224/01—Means for bonding being attached to, or being formed on, the surface to be connected, e.g. chip-to-package, die-attach, "first-level" interconnects; Manufacturing methods related thereto
- H01L2224/10—Bump connectors; Manufacturing methods related thereto
- H01L2224/12—Structure, shape, material or disposition of the bump connectors prior to the connecting process
- H01L2224/13—Structure, shape, material or disposition of the bump connectors prior to the connecting process of an individual bump connector
- H01L2224/13001—Core members of the bump connector
- H01L2224/13099—Material
- H01L2224/131—Material with a principal constituent of the material being a metal or a metalloid, e.g. boron [B], silicon [Si], germanium [Ge], arsenic [As], antimony [Sb], tellurium [Te] and polonium [Po], and alloys thereof
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2224/00—Indexing scheme for arrangements for connecting or disconnecting semiconductor or solid-state bodies and methods related thereto as covered by H01L24/00
- H01L2224/01—Means for bonding being attached to, or being formed on, the surface to be connected, e.g. chip-to-package, die-attach, "first-level" interconnects; Manufacturing methods related thereto
- H01L2224/10—Bump connectors; Manufacturing methods related thereto
- H01L2224/12—Structure, shape, material or disposition of the bump connectors prior to the connecting process
- H01L2224/14—Structure, shape, material or disposition of the bump connectors prior to the connecting process of a plurality of bump connectors
- H01L2224/141—Disposition
- H01L2224/1418—Disposition being disposed on at least two different sides of the body, e.g. dual array
- H01L2224/14181—On opposite sides of the body
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2224/00—Indexing scheme for arrangements for connecting or disconnecting semiconductor or solid-state bodies and methods related thereto as covered by H01L24/00
- H01L2224/01—Means for bonding being attached to, or being formed on, the surface to be connected, e.g. chip-to-package, die-attach, "first-level" interconnects; Manufacturing methods related thereto
- H01L2224/10—Bump connectors; Manufacturing methods related thereto
- H01L2224/15—Structure, shape, material or disposition of the bump connectors after the connecting process
- H01L2224/16—Structure, shape, material or disposition of the bump connectors after the connecting process of an individual bump connector
- H01L2224/161—Disposition
- H01L2224/16135—Disposition the bump connector connecting between different semiconductor or solid-state bodies, i.e. chip-to-chip
- H01L2224/16145—Disposition the bump connector connecting between different semiconductor or solid-state bodies, i.e. chip-to-chip the bodies being stacked
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2224/00—Indexing scheme for arrangements for connecting or disconnecting semiconductor or solid-state bodies and methods related thereto as covered by H01L24/00
- H01L2224/73—Means for bonding being of different types provided for in two or more of groups H01L2224/10, H01L2224/18, H01L2224/26, H01L2224/34, H01L2224/42, H01L2224/50, H01L2224/63, H01L2224/71
- H01L2224/732—Location after the connecting process
- H01L2224/73251—Location after the connecting process on different surfaces
- H01L2224/73253—Bump and layer connectors
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2225/00—Details relating to assemblies covered by the group H01L25/00 but not provided for in its subgroups
- H01L2225/03—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00
- H01L2225/04—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers
- H01L2225/065—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers the devices being of a type provided for in group H01L27/00
- H01L2225/06503—Stacked arrangements of devices
- H01L2225/06513—Bump or bump-like direct electrical connections between devices, e.g. flip-chip connection, solder bumps
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2225/00—Details relating to assemblies covered by the group H01L25/00 but not provided for in its subgroups
- H01L2225/03—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00
- H01L2225/04—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers
- H01L2225/065—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers the devices being of a type provided for in group H01L27/00
- H01L2225/06503—Stacked arrangements of devices
- H01L2225/06555—Geometry of the stack, e.g. form of the devices, geometry to facilitate stacking
- H01L2225/06562—Geometry of the stack, e.g. form of the devices, geometry to facilitate stacking at least one device in the stack being rotated or offset
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2225/00—Details relating to assemblies covered by the group H01L25/00 but not provided for in its subgroups
- H01L2225/03—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00
- H01L2225/04—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers
- H01L2225/065—All the devices being of a type provided for in the same subgroup of groups H01L27/00 - H01L33/648 and H10K99/00 the devices not having separate containers the devices being of a type provided for in group H01L27/00
- H01L2225/06503—Stacked arrangements of devices
- H01L2225/06555—Geometry of the stack, e.g. form of the devices, geometry to facilitate stacking
- H01L2225/06565—Geometry of the stack, e.g. form of the devices, geometry to facilitate stacking the devices having the same size and there being no auxiliary carrier between the devices
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L24/00—Arrangements for connecting or disconnecting semiconductor or solid-state bodies; Methods or apparatus related thereto
- H01L24/01—Means for bonding being attached to, or being formed on, the surface to be connected, e.g. chip-to-package, die-attach, "first-level" interconnects; Manufacturing methods related thereto
- H01L24/10—Bump connectors ; Manufacturing methods related thereto
- H01L24/15—Structure, shape, material or disposition of the bump connectors after the connecting process
- H01L24/16—Structure, shape, material or disposition of the bump connectors after the connecting process of an individual bump connector
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2924/00—Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
- H01L2924/0001—Technical content checked by a classifier
- H01L2924/00013—Fully indexed content
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2924/00—Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
- H01L2924/10—Details of semiconductor or other solid state devices to be connected
- H01L2924/11—Device type
- H01L2924/14—Integrated circuits
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Definitions
- the present disclosure relates generally to information processing systems and, more specifically, to processors whose logic is partitioned among a plurality of stacked dice.
- IC packages typically include a single IC die on a substrate or leadframe. The die and substrate are encapsulated in a material such as plastic. The encapsulated packages are then mounted to another substrate such as a PCB.
- Various packaging approaches have been employed to improve performance for such electronic devices.
- Multichip modules are IC packages that can contain two or more “bare” or unpackaged integrated circuit dice interconnected on a common substrate.
- the size of the electronic device that uses MCMs can be reduced because MCMs typically have a number of individual IC dice mounted within a single package in a laterally adjacent manner.
- SoC System on a Chip
- a system-on-a-chip for a sound-detecting device may include an audio receiver, an analog-to-digital converter, a microprocessor, memory, and input/output control logic on a single IC die.
- Stacked-CSP Stacked Chip Scale Package
- the Stacked-CSP is essentially a space-efficient MCM, where multiple die are stacked (in a face-to-back orientation) and integrated into a single package.
- Stacked-CSP packaging allows manufacturers of mobile phones and other portable devices to make their products smaller by vertically stacking heterogeneous dice, such as stacking flash and SRAM (Static Random Access Memory) dice, within a single package.
- SRAM Static Random Access Memory
- FIG. 1 is a block diagram of at least one embodiment of a multi-die processor exhibiting disclosed features.
- FIG. 2 is a block diagram illustrating further detail for at least one embodiment of a multi-die processor.
- FIGS. 3 is a data flow diagram illustrating at least one embodiment of an illustrative instruction execution pipeline.
- FIG. 4 is a data flow diagram illustrating a load access data path.
- FIG. 5 is a data flow diagram illustrating a schedule-execute data path.
- FIG. 6 is a block diagram illustrating conductive elements on the face side of multiple dice.
- FIG. 7 is a block diagram illustrating a sample scalar floor plan for logic of an illustrative scalar single-die processor.
- FIG. 8 is a block diagram illustrating at least one embodiment of partitioning and allocation of the logic illustrated in FIG. 7 between two face-to-face stacked dice.
- FIG. 9 is a plan view of a reduced-length load access data path for at least one embodiment of a multi-die processor.
- FIG. 10 is a plan view of a reduced-length schedule-execute data path for at least one embodiment of a multi-die processor.
- FIG. 11 is a block diagram illustrating a zigzag data path among blocks of processor logic partitioned between two face-to-face stacked dice according to at least one embodiment.
- FIG. 12 is a block diagram illustrating at least one embodiment of a split array partitioned between two face-to-face stacked dice.
- FIG. 13 is a block diagram of at least one embodiment of split execution logic partitioned between two face-to-face stacked dice.
- FIG. 14 is a block diagram of at least one embodiment of split scheduling logic partitioned between two face-to-face stacked dice.
- FIG. 15 is a block diagram of at least one embodiment of split floating point execution logic partitioned between two face-to-face stacked dice.
- FIG. 16 is a block diagram of at least one embodiment of split address generation logic partitioned between two face-to-face stacked dice.
- FIG. 17 is a block diagram illustrating at least one embodiment of a processing system that includes a multi-die processor.
- Described herein are selected embodiments of a multi-die processor apparatus and system.
- numerous specific details such as inter-component communication mechanisms, specified pipeline stages, overlap configurations for split logic, and the like have been set forth to provide a more thorough understanding of the present invention. It will be appreciated, however, by one skilled in the art that the invention may be practiced without such specific details. Additionally, some well-known structures, circuits, and the like have not been shown in detail to avoid unnecessarily obscuring the present invention.
- Disclosed herein is a packaging approach to stack multiple dice that, together, implement a processor device in a single package. For example, efficiencies in processor performance (as measured, for instance, by instructions per clock cycle) and heat and power management may be realized by splitting the logic of a processor core among two stacked dice that work together to cooperatively execute instructions.
- FIG. 1 is a block diagram of a multi-die processor core 100 to execute one or more instructions.
- the processor core may be a general or special purpose microprocessor, a digital signal processor, a graphics processor, a communications processor, a network processor or any other processor that executes one or more instructions.
- FIG. 1 illustrates that the processor package 100 includes a first die 102 and a second die 104 .
- the first and second dice operate cooperatively to execute instructions. That is, neither die 102 , 104 is a fully functional processor; the functionality of a processor is “split” between the two dice 102 , 104 .
- each of the first die 102 and second die 104 has a face side and a back side.
- face it is intended to refer to the side of the die with an integrated circuit formed on it. This face side may be referred to as the side of the die having active silicon.
- the “back side” of a die is the side having inactive matter (such as silicon substrate) that may be coupled to another structure, such as a heat sink, C 4 I/O bumps, a substrate, or the like.
- first die 102 and second die 104 are coupled together in a face-to-face orientation.
- the face sides of the first die 102 and second die 104 may be coupled together via a conductive interface.
- conductive elements 106 on the face side of the first die 102 are coupled to conductive elements 108 on the face side of the second die 104 .
- corresponding conductive elements 106 , 108 may be copper vias that may be bonded together in order to create a conductive inter-die interface (see 275 , FIG. 2 )
- FIG. 6 further illustrates the range of choices for placement of conductive elements 106 , 108 on a first die 602 and second die 604 .
- FIG. 6 illustrates the face sides a first die 602 and a second die 604 .
- the conductive elements 106 , 108 may be placed at any location on the face of the respective dice 602 , 604 .
- FIG. 6 illustrates a plurality of conductive elements 106 , 108 covering the entire face of each die 602 , 604 , one of skill in the art will recognize that any subset of the area of the face of each die 602 , 604 may include conductive elements 106 , 108 .
- FIG. 6 illustrates that each die 602 , 604 has a perimeter region 606 and a central region 608 .
- the conductive elements 106 for the first die 602 may be placed within the central region of the die 608 a and/or along the perimeter region of the die 606 a .
- the conductive elements 108 for the second die 604 may also be placed anywhere along the face of the die 604 , including the central region and/or the perimeter region of the die 604 .
- the first and second dice 602 , 604 may be of the same size and shape. They may thus overlap each other completely when stacked. In other embodiments, different sizes and shapes may be used and/or overlap may be more limited or partial.
- FIG. 2 is a block diagram illustrating further detail for at least one embodiment 200 of a multi-die processor made in accordance with disclosed techniques.
- the face side of first die 102 includes an active silicon layer 203 and one or more metal layers 214 .
- the face side of second die 104 also includes an active silicon layer 205 and one or more metal layers 216 .
- FIG. 2 further illustrates that the back side of each of the dice 102 , 104 includes a layer of inactive bulk silicon 220 , 230 , respectively.
- the back side of first die 102 may also include one or more backside vias 233 to provide a communication pathway from the active silicon 203 on the face of the first die 102 to metal input/output (I/O) solder bumps 212 .
- the vias 233 may be made of, for example, copper.
- the I/O 212 bumps provide a means for communicating with structures outside the multi-die processor 200 , such as an interface portion of a processing system (see 1704 , FIG. 17 ).
- the interface portion may be coupled to the I/O bumps 212 , which may reside on both dice. Alternatively, the I/O bumps 212 may reside on only one of the two dice, such as on the first die 102 as illustrated in FIG. 2 . Accordingly, the interface portion may be coupled to only one of the multiple dice.
- the interface portion, in communication with the I/O bumps 212 may provide a mechanism for the multi-die processor 200 to communicate with other components of a processing system. For example, the interface portion and the I/O bumps 212 may facilitate communication between the processor 200 and a memory system (see 1750 , FIG. 17 ) in order to process memory access requests from the processor.
- FIG. 2 illustrates that the back-side layer of inactive silicon 230 for the second die 104 may be coupled to a heat sink 210 .
- FIG. 3 is a block flow diagram illustrating a sample execution pipeline 300 for a processor.
- the illustrative execution pipeline 300 illustrated in FIG. 3 includes the following stages: instruction pointer generation 302 , instruction fetch 304 , instruction decode 306 , register rename 308 , scheduling 312 , execution 313 , and instruction retirement 314 .
- the pipeline 300 illustrated in FIG. 3 is illustrative only; the techniques described herein may be used on any processor.
- the stages of a pipeline 300 may appear in different order than that depicted in FIG. 3 or may be split into many more, different or other stages.
- a processor whose pipeline 300 may include different or additional pipeline stages to those illustrated in FIG. 3 .
- alternative embodiments of the pipeline 300 may include additional pipeline stages for rotation, expansion, exception detection, etc.
- a EPIC-type (Explicitly Parallel Instruction Computing) processor may include different pipeline stages, such as a word-line decode stage, architectural rename stage, etc., than appear in the pipeline for a processor that includes variable-length instructions in its instruction set.
- some processors may not include a decode phase 306 to decode instructions into micro-operations.
- FIGS. 4 and 5 illustrate that, during execution of an instruction, various data paths through the pipeline 300 may be followed. Some of these paths may constitute performance-critical data loops.
- FIG. 4 illustrates that, during execution of a load instruction, a load access data path 420 may be followed from a load/store execution unit (not shown) during an execution pipeline stage 313 , to a data cache 365 .
- load access data path 420 is a performance-critical loop because of data dependencies inherent in software code—some later instructions cannot be executed until data from an earlier load instruction is retrieved from memory and is made available to the later instruction.
- FIG. 5 illustrates that a schedule-execute data path 520 may also be a performance-critical loop. Again, such schedule-execute data path 520 may be a performance-critical loop due to data dependencies inherent in software code—some later instructions (consumers) cannot be executed until an earlier instruction (producer) has calculated a register value to be used by the consumer instruction.
- a multi-die processor such as, for example, the embodiments 100 , 200 illustrated in FIGS. 1 and 2 , respectively, may be partitioned among multiple dice such that critical data paths are shortened.
- the load access path 420 and the schedule-execute path 520 illustrated in FIGS. 4 and 5 may be shortened by judicious partitioning and allocation of processor logic among a first and second face-to-face die.
- FIG. 7 illustrates, for example, a sample scalar processor layout wherein logic of the processor is allocated to a single die 700 .
- FIG. 7 illustrates the data paths 420 , 520 illustrated in FIGS. 4 and 5 , respectively.
- the load access data path 420 begins at the general register file 702 , proceeds through a load/store execution unit in the execution unit block of logic 704 , and proceeds to the data cache 706 to obtain the load data.
- the time delay to traverse the execution unit logic 704 to get to the data cache 706 is indicated by reference numeral 732 .
- Once the load data is retrieved from the data cache 706 the data is returned to the front of the execution logic 704 .
- the time delay to get the desired load data from the data cache 706 to the front of the execution logic 704 is indicated by reference numeral 730 .
- FIG. 7 illustrates such path 520 on the sample scalar processor logic 700 .
- FIG. 7 illustrates that data for such path 520 starts at schedule logic 708 , proceeds through the general register file 702 , the execute logic 704 , the data cache 706 , and back to the scheduling logic 708 .
- FIG. 8 illustrates at least one embodiment of processor logic, such as the sample processor logic illustrated in FIG. 7 , which has been allocated between two separate dice 802 , 804 . While a particular allocation of the logic is illustrated in FIG. 8 , it should be understood that FIG. 8 is presented only for the general purpose of illustrating that instruction-processing logic for a processor may be split among two or more dice. The particular allocation of processor logic between the first and second dice 802 , 804 as illustrated in FIG. 8 should not be taken to be limiting.
- At least one allocation scheme for splitting processor logic between two dice 802 , 804 may be designed, for example, to ameliorate power-density concerns. That is, processors often strive to achieve a current-per-region value that is at or lower than a predetermined threshold. A relatively high power-density region requires a relatively large amount of current. By allocating a portion of the logic for the high power-density region to a first die and the remaining portion of the logic for the high power-density region to a second die, the implementation constraints for the region may be relaxed, leading to a lower power-density design. This ability to partition the logic of a high power-density region to reduce its footprint and to lower its power consumption is only one advantage of the stacking approach illustrated in FIG. 8 . As is explained immediately below, other advantages may also be realized from such an approach.
- FIG. 8 illustrates that the instruction-processing logic of the scalar processor die 700 illustrated in FIG. 7 may be partitioned such that the execution logic 704 is partitioned to a first die 802 and the data cache 706 is partitioned to a second die 804 .
- the execution logic 704 may be placed on the face side of the first die 802 and the data cache 706 may be placed on the face side of the second die 804 .
- the data cache logic 706 and the execution logic 704 may further be partitioned such that, when the first die 802 is stacked on top of the second die 804 , the data cache logic 706 overlaps the execution logic 704 .
- the logic may be partitioned such that, when the two dice 802 , 804 are stacked, the front edge 830 of the data cache 706 overlaps the front edge 820 of the execution logic 704 .
- FIG. 8 Further partitioning of scalar processor logic is also illustrated in FIG. 8 .
- the scheduling logic 708 illustrated in FIG. 7 has been partitioned into two logic portions 708 a and 708 b that have been allocated to separate dice.
- FIG. 8 illustrates that an L 3 cache may be partitioned among the two dice, as may be a memory controller (“mem”), an L 2 cache, retirement logic (“ret”), integer register file (“ireg”), floating point register file (“freg”), input/output logic (“I/O”), fetch logic (“fetch”), floating point execution logic (“fp”), and instruction queue logic (“iq”).
- FIG. 9 illustrates the new data path 420 a for the load access loop 420 illustrated in FIGS. 4 and 7 is shorter and more time-efficient under the multi-die approach illustrated in FIG. 8 .
- FIG. 9 illustrates the load access data path 420 for an embodiment wherein the logic of a processor has been split such that the data cache 706 of a second die 904 overlaps the execution logic 704 on a first die 902 such that the front edge 830 of the data cache 706 overlaps the front edge 820 of the execution logic 704 .
- the general register file 702 may also be split between the two dice 902 , 904 , as illustrated, for example, in FIG. 8 ).
- FIG. 9 illustrates that the new load access data path 420 a is more efficient in that, after load data has been retrieved from the data cache 706 , the data need not be transmitted along the full length of the execution logic 704 in order to arrive at the front edge 820 of the execution logic 704 . Instead, the data may be transmitted directly to the front edge 820 of the execution logic 704 via the conductive intra-die interface 275 . In this manner, the data need not take time to traverse the execution logic 704 and, thus, the time delay that to get the desired load data from the data cache 706 to the front of the execution logic 704 (see 730 , FIG. 7 ) is avoided or reduced. Similar time delay reduction may be realized for the portion of the data path indicated by reference numeral 732 in FIG. 7 .
- the embodiment illustrated in FIG. 9 may also provide beneficial heat-management effects, in addition to the timing efficiency discussed above. That is, for at least some processor embodiments, the execution units in execution logic 704 represent relatively “hot” logic blocks. That is, they have significant internal transistor switching and therefore generate relatively more heat than other blocks. In contrast, for at least some embodiments a data cache 706 is a relatively “cool” logic block that does not generate as much heat as a “hot” block. Accordingly, the embodiment illustrated in FIG. 9 enjoys the additional benefit of stacking a “hot” block and a “cool” block such that they at least partially overlap. Such hot/cool stacking may minimize detrimental thermal effects that otherwise might ensue from stacking “hot” blocks in a multi-die processor such that they overlap each other. That is, the cool block may effectively be used as a heat sink to dissipate heat from the hot block.
- FIG. 10 illustrates that the partitioning of the scheduling logic 708 into two portions lo 708 a , 708 b that are each allocated to different dice 1002 , 1004 may also result in efficiencies for the schedule-execute data path 520 illustrated in FIGS. 5 and 7 .
- FIG. 10 illustrates that the new schedule-execute data path 520 a also enjoys the efficiency of die-to-die communication through the intra-die interface 275 .
- Each of the intra-die communications involves only the time involved with communicating through the die-to-die interface 275 between logic elements 708 a and 708 b , 704 and 706 that overlap each other.
- FIG. 11 illustrates an embodiment of a multi-die processor 1100 that has been partitioned such that logic portions for the stages 302 , 304 , 306 , 308 , 312 , 313 , 314 of an execution pipeline 300 (see, e.g., FIG. 3 ) have been allocated between two die 1102 , 1104 .
- the logic for the pipeline stages have been allocated between the two die 1102 , 1104 such that communication signals for the pipeline follows a zigzag path 1106 between the first die 1102 and the second die 1104 via the die-to-die interface 275 to access the allocated portions of the pipeline state logic.
- This type of allocation may be advantageous in some embodiments because the die-to-die communications along the zigzag path are more efficient in that they may reduce path length and therefore improve performance.
- FIG. 11 illustrates that at least one embodiment 1100 of a multi-die processor may include I/O bumps 212 and a heat sink 210 , as is discussed above.
- FIG. 12 illustrates at least one embodiment of a multi-die processor wherein an array 1208 of a scalar floor plan 1200 is split and allocated between two dice 1202 , 1204 .
- the array 1208 may be any array, including a general register file, a cache, a floating point register file, or a microcode ROM (read-only memory).
- FIG. 12 illustrates that the array may be split into a first portion 1208 a and a second portion 1208 b that may each be allocated to a different die 1202 , 1204 .
- the two logic portions 1208 a , 1208 b may at least partially overlap.
- the array access time may be decreased by utilizing the die-to-die interface (see 275 , FIG. 2 ) to facilitate communication between the overlapping portions 1208 a , 1208 b because the total length of the array is reduced.
- This may be advantageous in some embodiments, for example, where the array is a register file array in order to reduce the total worst-case transfer time between registers.
- the two portions 1208 a , 1208 b may, but need not necessarily, completely overlap each other.
- the overlapping portions may be offset such that only part of the portions 1208 a , 1208 b overlap each other.
- FIG. 13 illustrates at least one other embodiment of a multi-die processor layout.
- a floor plan for a scalar processor 1300 includes an integer execution unit 1306 , a load/store execution unit 1308 , and a floating point execution unit 1310 .
- the execution unit logic blocks 1306 , 1308 and 1310 may be allocated between the face sides of two dice, 1302 and 1304 .
- FIG. 13 illustrates that the execution unit logic blocks 1306 , 1308 , 1310 maybe allocated such that the integer execution unit 1306 and load/store execution unit 1308 are allocated to a first die 1302 and the floating point execution unit 1310 may be allocated to a second die 1304 such that they at least partially overlap each other when the first die 1302 and second die 1304 are stacked.
- This type of allocation may be advantageous in some embodiments at least because splitting the execution units 1306 , 1308 , 1310 between the two dice 1302 , 1304 allows up to twice the transistor density for the execution units and therefore may improve performance by shortening path length
- FIG. 14 illustrates at least one other embodiment of a multi-die processor layout.
- a floor plan for a scalar processor 1400 includes scheduling logic 1408 .
- the scheduling logic 1408 may be partitioned into two logic portions, 1408 a and 1408 b .
- the scheduling logic 1408 may be partitioned, for example, into an arithmetic scheduling portion 1408 a and a memory request scheduling portion 1408 b .
- the two logic portions 1408 a and 1408 b may be partitioned between a first die 1402 and a second die 1404 , respectively, such that they at least partially overlap each other when the first die 1402 and second die 1404 are stacked.
- Such overlapping portions may operate to cooperatively accomplish scheduling operations for execution of an instruction.
- This type of allocation may be advantageous in some embodiments at least because signal bandwidth for the scheduling logic 1408 may be increased. Accordingly, it enhances performance by allowing more than one signal to be transmitted between the scheduling portions 1408 a , 1408 b concurrent
- FIG. 15 illustrates at least one other embodiment of a multi-die processor layout.
- FIG. 15 illustrates that a floor plan for a scalar processor 1500 includes floating point execution logic 1508 .
- the floating point execution logic 1508 may be partitioned into two logic portions, 1508 a and 1508 b .
- the logic 1508 may be partitioned, for example, into a non-SIMD (single-instruction-multiple-data) floating point execution logic potion 1508 a and a SIMD floating point execution logic portion 1508 b .
- SIMD single-instruction-multiple-data
- the two logic portions 1508 a and 1508 b may be partitioned between a first die 1502 and a second die 1504 , respectively, such that they at least partially overlap each other when the first die 1502 and second die 1504 are stacked. This type of allocation may be advantageous in some embodiments at least for those reasons, discussed above, related to increased transistor density and increased signal bandwidth.
- FIG. 16 illustrates at least one other embodiment of a multi-die processor layout.
- FIG. 16 illustrates that a floor plan for a scalar processor 1600 includes address generation 1608 .
- the address generation logic 1608 may be partitioned into two logic portions, 1608 a and 1608 b .
- the two logic portions 1608 a and 1608 b may be partitioned between a first die 1602 and a second die 1604 , respectively, such that they at least partially overlap each other when the first die 1602 and second die 1604 are stacked.
- this type of allocation may be advantageous in some embodiments at least for those reasons, discussed above, related to increased transistor density and increased signal bandwidth.
- FIGS. 8-16 illustrate various embodiments of a multi-die processor wherein the instruction-processing logic of the processor is allocated among two die, such embodiments are provided for illustrative purposes and should not be taken to be limiting.
- any one or more of the particular partitioning and/or allocation embodiments illustrated in FIGS. 8-16 may be practiced alone, or in conjunction with other such embodiments. That is, in a single multi-die processor the partitioning illustrated in FIG. 9 , for instance, may be practiced along with the partitioning illustrated in FIGS. 12, 13 , 14 , 15 and/or 16 .
- the logic of a processor may be partitioned among a plurality of dice.
- face-to-face die may overlap such that a portion of a first top die and a portion of a second top die overlap a third bottom die.
- the partitioned logic on the multiple dice whatever the number, cooperatively operates to execute one or more instructions.
- the logic portions allocated to respective multiple dice may be invoked to perform one or more execution operations associated with an instruction.
- the logic portions operate to cooperatively accomplish execution operations, such as those operations indicated for an execution pipeline (see, for example, sample pipeline 300 illustrated in FIG. 3 ).
- These execution operations may include, but are not limited, address generation, instruction pointer generation, fetching instructions, decoding instructions into micro-operations, renaming registers to eliminate certain types of data dependencies, dispatch, scheduling, execution, and retirement.
- the execution operations may include sub-instruction level tasks performed in response to an instruction.
- the logic portions may be allocated among the multiple dice such that certain functions are split. That is, address generation unit logic may be split into a first portion and a second portion, with the first portion being allocated to a first die and a second portion being allocated to a second die. The first and second logic portions may at least partially overlap and may act together to cooperatively perform the operations of an address generation unit.
- a scheduling unit may be split, as may an array such as a general register file, a cache, a floating point register file or a microcode memory array.
- a memory controller may also be split, as may a cache, a translation lookaside buffer, decode logic, rename logic, fetch logic, retirement logic, and floating point execution unit logic.
- logic portions may also be allocated such that, rather splitting a block of logic, the intact logic blocks for successive pipeline stages are allocated among the multiple dice of the processor. Such allocation of the logic for pipeline stages may result in a zigzag communication path 1106 through the die-to-die interface 275 as illustrated in FIG. 11 .
- the execution operations associated with an execution stage of an execution pipeline may further include execution, by an execution unit, of arithmetic instruction codes such as integer or floating point instruction codes.
- arithmetic instruction codes such as integer or floating point instruction codes.
- instruction code is intended to encompass any unit of work that can be understood and executed by an execution unit, such as a floating point execution unit, arithmetic logic unit, or load/store execution unit.
- An instruction code may be a micro-operation.
- the execution operations associated with the execution pipeline stage may also include execution, by an execution unit, of a memory instruction code such as a memory read or memory write instruction code.
- a multi-die processor 1702 such as described herein may be utilized on a processing system such as the processing system 1700 illustrated in FIG. 17 .
- System 1700 may be used, for example, to execute one or more instructions of an instruction set.
- a processing system includes any processing system that has a processor, such as, for example; a digital signal processor (DSP), a microcontroller, an application specific integrated circuit (ASIC), or a microprocessor, where such processor is a multi-die processor as described in the embodiments discussed above.
- DSP digital signal processor
- ASIC application specific integrated circuit
- the processing system 1700 is representative of processing systems based on the Itanium® and Itanium® 2 microprocessors as well as the Pentium®, Pentium® Pro, Pentium® II, Pentium® III, and Pentium® 4 microprocessors, all of which are available from Intel Corporation. Other systems (including personal computers (PCs) having other microprocessors, engineering workstations, personal digital assistants and other hand-held devices, set-top boxes and the like) may also be used. At least one embodiment of system 1700 may execute a version of the WindowsTM operating system available from Microsoft Corporation, although other operating systems and graphical user interfaces, for example, may also be used.
- WindowsTM operating system available from Microsoft Corporation, although other operating systems and graphical user interfaces, for example, may also be used.
- Processing system 1700 includes a memory system 1705 and a processor 1702 .
- Memory system 1705 may store instructions 1740 and data 1741 for controlling the operation of the processor 1702 .
- Memory system 1705 is intended as a generalized representation of memory and may include a variety of forms of memory, such as a hard drive, CD-ROM, random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), flash memory and related circuitry.
- Memory system 1705 may store instructions 1740 and/or data 1741 represented by data signals that may be executed by the processor 1702 .
- the processing system 1700 includes an interface portion 1704 .
- the interface portion 1704 may be coupled to only one or both of the dice 102 , 104 .
- the interface portion 1704 is to generate inter-component signals between the processor 1702 and another component of the system 1700 .
- the interface portion 1704 may generate inter-component signals between the processor 1702 and the memory system 1705 .
- the interface portion 1704 may generate signals between the processor 1702 and memory system 1705 in order to perform a memory transaction such as a data-retrieval read operation from memory or a data write to memory.
- the interface portion 1704 may also generate signals between the processor 1702 and another system component 1707 , such as an RF unit, keyboard, external memory device, monitor, mouse or the like.
Abstract
Disclosed are a multi-die processor apparatus and system. Processor logic to execute one or more instructions is allocated among two or more face-to-faces stacked dice. The processor includes a conductive interface between the stacked dice to facilitate die-to-die communication.
Description
- 1. Technical Field
- The present disclosure relates generally to information processing systems and, more specifically, to processors whose logic is partitioned among a plurality of stacked dice.
- 2. Background Art
- Electronic devices such as cellular telephones and notebook computers typically contain a number of integrated circuit (IC) packages mounted to a printed circuit board (PCB). IC packages typically include a single IC die on a substrate or leadframe. The die and substrate are encapsulated in a material such as plastic. The encapsulated packages are then mounted to another substrate such as a PCB. Various packaging approaches have been employed to improve performance for such electronic devices.
- Multichip modules (MCM) are IC packages that can contain two or more “bare” or unpackaged integrated circuit dice interconnected on a common substrate. The size of the electronic device that uses MCMs can be reduced because MCMs typically have a number of individual IC dice mounted within a single package in a laterally adjacent manner.
- System on a Chip (SoC) technology is the packaging of most or all of the necessary electronic circuits and parts for a “system” (such as a cell phone or digital camera) on a single IC die. For example, a system-on-a-chip for a sound-detecting device may include an audio receiver, an analog-to-digital converter, a microprocessor, memory, and input/output control logic on a single IC die.
- Another type of IC package configuration that attempts to decrease the footprint and volume of the IC package is known as a Stacked Chip Scale Package (Stacked-CSP). The Stacked-CSP is essentially a space-efficient MCM, where multiple die are stacked (in a face-to-back orientation) and integrated into a single package. Stacked-CSP packaging allows manufacturers of mobile phones and other portable devices to make their products smaller by vertically stacking heterogeneous dice, such as stacking flash and SRAM (Static Random Access Memory) dice, within a single package. By utilizing Stacked-CSP products that vertically mount two or more heterogeneous IC dice in a single package, wireless devices may be generated to have lower cost, weight and board space than devices made of traditional single-die packages.
- The present invention may be understood with reference to the following drawings in which like elements are indicated by like numbers. These drawings are not intended to be limiting but are instead provided to illustrate selected embodiments of an apparatus and system for a multiple-die processor for which logic of the processor is partitioned among the multiple die.
-
FIG. 1 is a block diagram of at least one embodiment of a multi-die processor exhibiting disclosed features. -
FIG. 2 is a block diagram illustrating further detail for at least one embodiment of a multi-die processor. - FIGS. 3 is a data flow diagram illustrating at least one embodiment of an illustrative instruction execution pipeline.
-
FIG. 4 is a data flow diagram illustrating a load access data path. -
FIG. 5 is a data flow diagram illustrating a schedule-execute data path. -
FIG. 6 is a block diagram illustrating conductive elements on the face side of multiple dice. -
FIG. 7 is a block diagram illustrating a sample scalar floor plan for logic of an illustrative scalar single-die processor. -
FIG. 8 is a block diagram illustrating at least one embodiment of partitioning and allocation of the logic illustrated inFIG. 7 between two face-to-face stacked dice. -
FIG. 9 is a plan view of a reduced-length load access data path for at least one embodiment of a multi-die processor. -
FIG. 10 is a plan view of a reduced-length schedule-execute data path for at least one embodiment of a multi-die processor. -
FIG. 11 is a block diagram illustrating a zigzag data path among blocks of processor logic partitioned between two face-to-face stacked dice according to at least one embodiment. -
FIG. 12 is a block diagram illustrating at least one embodiment of a split array partitioned between two face-to-face stacked dice. -
FIG. 13 is a block diagram of at least one embodiment of split execution logic partitioned between two face-to-face stacked dice. -
FIG. 14 is a block diagram of at least one embodiment of split scheduling logic partitioned between two face-to-face stacked dice. -
FIG. 15 is a block diagram of at least one embodiment of split floating point execution logic partitioned between two face-to-face stacked dice. -
FIG. 16 is a block diagram of at least one embodiment of split address generation logic partitioned between two face-to-face stacked dice. -
FIG. 17 is a block diagram illustrating at least one embodiment of a processing system that includes a multi-die processor. - Described herein are selected embodiments of a multi-die processor apparatus and system. In the following description, numerous specific details such as inter-component communication mechanisms, specified pipeline stages, overlap configurations for split logic, and the like have been set forth to provide a more thorough understanding of the present invention. It will be appreciated, however, by one skilled in the art that the invention may be practiced without such specific details. Additionally, some well-known structures, circuits, and the like have not been shown in detail to avoid unnecessarily obscuring the present invention.
- Disclosed herein is a packaging approach to stack multiple dice that, together, implement a processor device in a single package. For example, efficiencies in processor performance (as measured, for instance, by instructions per clock cycle) and heat and power management may be realized by splitting the logic of a processor core among two stacked dice that work together to cooperatively execute instructions.
-
FIG. 1 is a block diagram of amulti-die processor core 100 to execute one or more instructions. The processor core may be a general or special purpose microprocessor, a digital signal processor, a graphics processor, a communications processor, a network processor or any other processor that executes one or more instructions.FIG. 1 illustrates that theprocessor package 100 includes afirst die 102 and asecond die 104. The first and second dice operate cooperatively to execute instructions. That is, neither die 102, 104 is a fully functional processor; the functionality of a processor is “split” between the twodice - At least one embodiment of each of the first die 102 and second die 104 has a face side and a back side. By “face” it is intended to refer to the side of the die with an integrated circuit formed on it. This face side may be referred to as the side of the die having active silicon. The “back side” of a die is the side having inactive matter (such as silicon substrate) that may be coupled to another structure, such as a heat sink, C4 I/O bumps, a substrate, or the like.
- From
FIG. 1 it can be seen that at least one embodiment of thefirst die 102 andsecond die 104 are coupled together in a face-to-face orientation. Rather than epoxy or other adhesive-based coupling mechanism, the face sides of the first die 102 and second die 104 may be coupled together via a conductive interface. For at least one embodiment,conductive elements 106 on the face side of thefirst die 102 are coupled toconductive elements 108 on the face side of thesecond die 104. For example, correspondingconductive elements FIG. 2 ) - Brief reference to
FIG. 6 further illustrates the range of choices for placement ofconductive elements second die 604.FIG. 6 illustrates the face sides afirst die 602 and asecond die 604. It should be noted that theconductive elements respective dice FIG. 6 illustrates a plurality ofconductive elements die die conductive elements -
FIG. 6 illustrates that each die 602, 604 has a perimeter region 606 and a central region 608. Theconductive elements 106 for thefirst die 602 may be placed within the central region of the die 608 a and/or along the perimeter region of the die 606 a. Similarly, theconductive elements 108 for thesecond die 604 may also be placed anywhere along the face of thedie 604, including the central region and/or the perimeter region of thedie 604. As a result, there are no “keep out” areas on the face side of either die. Accordingly, the first andsecond dice -
FIG. 2 is a block diagram illustrating further detail for at least oneembodiment 200 of a multi-die processor made in accordance with disclosed techniques. The face side offirst die 102 includes anactive silicon layer 203 and one or more metal layers 214. The face side ofsecond die 104 also includes anactive silicon layer 205 and one or more metal layers 216. -
FIG. 2 further illustrates that the back side of each of thedice inactive bulk silicon first die 102 may also include one or more backside vias 233 to provide a communication pathway from theactive silicon 203 on the face of thefirst die 102 to metal input/output (I/O) solder bumps 212. Thevias 233 may be made of, for example, copper. - The I/
O 212 bumps provide a means for communicating with structures outside themulti-die processor 200, such as an interface portion of a processing system (see 1704,FIG. 17 ). The interface portion may be coupled to the I/O bumps 212, which may reside on both dice. Alternatively, the I/O bumps 212 may reside on only one of the two dice, such as on thefirst die 102 as illustrated inFIG. 2 . Accordingly, the interface portion may be coupled to only one of the multiple dice. The interface portion, in communication with the I/O bumps 212, may provide a mechanism for themulti-die processor 200 to communicate with other components of a processing system. For example, the interface portion and the I/O bumps 212 may facilitate communication between theprocessor 200 and a memory system (see 1750,FIG. 17 ) in order to process memory access requests from the processor. -
FIG. 2 illustrates that the back-side layer ofinactive silicon 230 for thesecond die 104 may be coupled to aheat sink 210. -
FIG. 3 is a block flow diagram illustrating asample execution pipeline 300 for a processor. Theillustrative execution pipeline 300 illustrated inFIG. 3 includes the following stages:instruction pointer generation 302, instruction fetch 304,instruction decode 306, registerrename 308,scheduling 312,execution 313, andinstruction retirement 314. Thepipeline 300 illustrated inFIG. 3 is illustrative only; the techniques described herein may be used on any processor. For an embodiment in which the processor utilizes anexecution pipeline 300, the stages of apipeline 300 may appear in different order than that depicted inFIG. 3 or may be split into many more, different or other stages. - The techniques disclosed herein may be utilized on a processor whose
pipeline 300 may include different or additional pipeline stages to those illustrated inFIG. 3 . For example, alternative embodiments of thepipeline 300 may include additional pipeline stages for rotation, expansion, exception detection, etc. In addition, a EPIC-type (Explicitly Parallel Instruction Computing) processor may include different pipeline stages, such as a word-line decode stage, architectural rename stage, etc., than appear in the pipeline for a processor that includes variable-length instructions in its instruction set. Also, some processors may not include adecode phase 306 to decode instructions into micro-operations. -
FIGS. 4 and 5 illustrate that, during execution of an instruction, various data paths through thepipeline 300 may be followed. Some of these paths may constitute performance-critical data loops. For example,FIG. 4 illustrates that, during execution of a load instruction, a loadaccess data path 420 may be followed from a load/store execution unit (not shown) during anexecution pipeline stage 313, to adata cache 365. Such loadaccess data path 420 is a performance-critical loop because of data dependencies inherent in software code—some later instructions cannot be executed until data from an earlier load instruction is retrieved from memory and is made available to the later instruction. - Similarly, other performance-critical loops may occur during a processor's execution of instructions. For example,
FIG. 5 illustrates that a schedule-executedata path 520 may also be a performance-critical loop. Again, such schedule-executedata path 520 may be a performance-critical loop due to data dependencies inherent in software code—some later instructions (consumers) cannot be executed until an earlier instruction (producer) has calculated a register value to be used by the consumer instruction. - A multi-die processor, such as, for example, the
embodiments FIGS. 1 and 2 , respectively, may be partitioned among multiple dice such that critical data paths are shortened. For example, theload access path 420 and the schedule-executepath 520 illustrated inFIGS. 4 and 5 , respectively, may be shortened by judicious partitioning and allocation of processor logic among a first and second face-to-face die. -
FIG. 7 illustrates, for example, a sample scalar processor layout wherein logic of the processor is allocated to a single die 700.FIG. 7 illustrates thedata paths FIGS. 4 and 5 , respectively. The loadaccess data path 420 begins at thegeneral register file 702, proceeds through a load/store execution unit in the execution unit block oflogic 704, and proceeds to thedata cache 706 to obtain the load data. The time delay to traverse theexecution unit logic 704 to get to thedata cache 706 is indicated byreference numeral 732. Once the load data is retrieved from thedata cache 706, the data is returned to the front of theexecution logic 704. The time delay to get the desired load data from thedata cache 706 to the front of theexecution logic 704 is indicated byreference numeral 730. - Regarding the schedule-execute
data path 520,FIG. 7 illustratessuch path 520 on the sample scalar processor logic 700.FIG. 7 illustrates that data forsuch path 520 starts atschedule logic 708, proceeds through thegeneral register file 702, the executelogic 704, thedata cache 706, and back to thescheduling logic 708. -
FIG. 8 illustrates at least one embodiment of processor logic, such as the sample processor logic illustrated inFIG. 7 , which has been allocated between twoseparate dice FIG. 8 , it should be understood thatFIG. 8 is presented only for the general purpose of illustrating that instruction-processing logic for a processor may be split among two or more dice. The particular allocation of processor logic between the first andsecond dice FIG. 8 should not be taken to be limiting. - At least one allocation scheme for splitting processor logic between two
dice FIG. 8 . As is explained immediately below, other advantages may also be realized from such an approach. -
FIG. 8 illustrates that the instruction-processing logic of the scalar processor die 700 illustrated inFIG. 7 may be partitioned such that theexecution logic 704 is partitioned to afirst die 802 and thedata cache 706 is partitioned to asecond die 804. Theexecution logic 704 may be placed on the face side of thefirst die 802 and thedata cache 706 may be placed on the face side of thesecond die 804. Thedata cache logic 706 and theexecution logic 704 may further be partitioned such that, when thefirst die 802 is stacked on top of thesecond die 804, thedata cache logic 706 overlaps theexecution logic 704. For instance, the logic may be partitioned such that, when the twodice front edge 830 of thedata cache 706 overlaps thefront edge 820 of theexecution logic 704. - Further partitioning of scalar processor logic is also illustrated in
FIG. 8 . For example, thescheduling logic 708 illustrated inFIG. 7 has been partitioned into twologic portions FIG. 8 illustrates that an L3 cache may be partitioned among the two dice, as may be a memory controller (“mem”), an L2 cache, retirement logic (“ret”), integer register file (“ireg”), floating point register file (“freg”), input/output logic (“I/O”), fetch logic (“fetch”), floating point execution logic (“fp”), and instruction queue logic (“iq”). - Turning to
FIG. 9 , thenew data path 420 a for theload access loop 420 illustrated inFIGS. 4 and 7 is shorter and more time-efficient under the multi-die approach illustrated inFIG. 8 . Specifically,FIG. 9 illustrates the loadaccess data path 420 for an embodiment wherein the logic of a processor has been split such that thedata cache 706 of asecond die 904 overlaps theexecution logic 704 on afirst die 902 such that thefront edge 830 of thedata cache 706 overlaps thefront edge 820 of theexecution logic 704. (Although not so illustrated inFIG. 9 , one of skill in the art will recognize that thegeneral register file 702 may also be split between the twodice FIG. 8 ). -
FIG. 9 illustrates that the new loadaccess data path 420 a is more efficient in that, after load data has been retrieved from thedata cache 706, the data need not be transmitted along the full length of theexecution logic 704 in order to arrive at thefront edge 820 of theexecution logic 704. Instead, the data may be transmitted directly to thefront edge 820 of theexecution logic 704 via the conductiveintra-die interface 275. In this manner, the data need not take time to traverse theexecution logic 704 and, thus, the time delay that to get the desired load data from thedata cache 706 to the front of the execution logic 704 (see 730,FIG. 7 ) is avoided or reduced. Similar time delay reduction may be realized for the portion of the data path indicated byreference numeral 732 inFIG. 7 . - The embodiment illustrated in
FIG. 9 may also provide beneficial heat-management effects, in addition to the timing efficiency discussed above. That is, for at least some processor embodiments, the execution units inexecution logic 704 represent relatively “hot” logic blocks. That is, they have significant internal transistor switching and therefore generate relatively more heat than other blocks. In contrast, for at least some embodiments adata cache 706 is a relatively “cool” logic block that does not generate as much heat as a “hot” block. Accordingly, the embodiment illustrated inFIG. 9 enjoys the additional benefit of stacking a “hot” block and a “cool” block such that they at least partially overlap. Such hot/cool stacking may minimize detrimental thermal effects that otherwise might ensue from stacking “hot” blocks in a multi-die processor such that they overlap each other. That is, the cool block may effectively be used as a heat sink to dissipate heat from the hot block. -
FIG. 10 illustrates that the partitioning of thescheduling logic 708 into two portions lo 708 a, 708 b that are each allocated todifferent dice data path 520 illustrated inFIGS. 5 and 7 .FIG. 10 illustrates that the new schedule-executedata path 520 a also enjoys the efficiency of die-to-die communication through theintra-die interface 275. Each of the intra-die communications involves only the time involved with communicating through the die-to-die interface 275 betweenlogic elements -
FIG. 11 illustrates an embodiment of amulti-die processor 1100 that has been partitioned such that logic portions for thestages FIG. 3 ) have been allocated between two die 1102, 1104. The logic for the pipeline stages have been allocated between the two die 1102, 1104 such that communication signals for the pipeline follows azigzag path 1106 between thefirst die 1102 and thesecond die 1104 via the die-to-die interface 275 to access the allocated portions of the pipeline state logic. This type of allocation may be advantageous in some embodiments because the die-to-die communications along the zigzag path are more efficient in that they may reduce path length and therefore improve performance. - Additionally,
FIG. 11 illustrates that at least oneembodiment 1100 of a multi-die processor may include I/O bumps 212 and aheat sink 210, as is discussed above. -
FIG. 12 illustrates at least one embodiment of a multi-die processor wherein anarray 1208 of a scalar floor plan 1200 is split and allocated between twodice array 1208 may be any array, including a general register file, a cache, a floating point register file, or a microcode ROM (read-only memory).FIG. 12 illustrates that the array may be split into afirst portion 1208 a and asecond portion 1208 b that may each be allocated to adifferent die logic portions FIG. 2 ) to facilitate communication between the overlappingportions - It should be noted, of course, that the two
portions portions -
FIG. 13 illustrates at least one other embodiment of a multi-die processor layout.FIG. 13 illustrates that a floor plan for ascalar processor 1300 includes aninteger execution unit 1306, a load/store execution unit 1308, and a floatingpoint execution unit 1310. The execution unit logic blocks 1306, 1308 and 1310 may be allocated between the face sides of two dice, 1302 and 1304.FIG. 13 illustrates that the execution unit logic blocks 1306, 1308, 1310 maybe allocated such that theinteger execution unit 1306 and load/store execution unit 1308 are allocated to afirst die 1302 and the floatingpoint execution unit 1310 may be allocated to asecond die 1304 such that they at least partially overlap each other when thefirst die 1302 and second die 1304 are stacked. This type of allocation may be advantageous in some embodiments at least because splitting theexecution units dice -
FIG. 14 illustrates at least one other embodiment of a multi-die processor layout.FIG. 14 illustrates that a floor plan for ascalar processor 1400 includesscheduling logic 1408. Thescheduling logic 1408 may be partitioned into two logic portions, 1408 a and 1408 b. Thescheduling logic 1408 may be partitioned, for example, into anarithmetic scheduling portion 1408 a and a memoryrequest scheduling portion 1408 b. The twologic portions first die 1402 and asecond die 1404, respectively, such that they at least partially overlap each other when thefirst die 1402 and second die 1404 are stacked. Such overlapping portions may operate to cooperatively accomplish scheduling operations for execution of an instruction. This type of allocation may be advantageous in some embodiments at least because signal bandwidth for thescheduling logic 1408 may be increased. Accordingly, it enhances performance by allowing more than one signal to be transmitted between thescheduling portions -
FIG. 15 illustrates at least one other embodiment of a multi-die processor layout.FIG. 15 illustrates that a floor plan for ascalar processor 1500 includes floatingpoint execution logic 1508. The floatingpoint execution logic 1508 may be partitioned into two logic portions, 1508 a and 1508 b. Thelogic 1508 may be partitioned, for example, into a non-SIMD (single-instruction-multiple-data) floating pointexecution logic potion 1508 a and a SIMD floating pointexecution logic portion 1508 b. The twologic portions first die 1502 and asecond die 1504, respectively, such that they at least partially overlap each other when thefirst die 1502 and second die 1504 are stacked. This type of allocation may be advantageous in some embodiments at least for those reasons, discussed above, related to increased transistor density and increased signal bandwidth. -
FIG. 16 illustrates at least one other embodiment of a multi-die processor layout.FIG. 16 illustrates that a floor plan for ascalar processor 1600 includesaddress generation 1608. Theaddress generation logic 1608 may be partitioned into two logic portions, 1608 a and 1608 b. The twologic portions first die 1602 and asecond die 1604, respectively, such that they at least partially overlap each other when thefirst die 1602 and second die 1604 are stacked. Again, this type of allocation may be advantageous in some embodiments at least for those reasons, discussed above, related to increased transistor density and increased signal bandwidth. - While
FIGS. 8-16 illustrate various embodiments of a multi-die processor wherein the instruction-processing logic of the processor is allocated among two die, such embodiments are provided for illustrative purposes and should not be taken to be limiting. For example, any one or more of the particular partitioning and/or allocation embodiments illustrated inFIGS. 8-16 may be practiced alone, or in conjunction with other such embodiments. That is, in a single multi-die processor the partitioning illustrated inFIG. 9 , for instance, may be practiced along with the partitioning illustrated inFIGS. 12, 13 , 14, 15 and/or 16. - Also for example, while the illustrated embodiments indicate a two-die processor, with each die having a logic portion of the processor disposed thereon, the logic of a processor may be partitioned among a plurality of dice. For example, face-to-face die may overlap such that a portion of a first top die and a portion of a second top die overlap a third bottom die. The partitioned logic on the multiple dice, whatever the number, cooperatively operates to execute one or more instructions.
- That is, as disclosed herein the logic portions allocated to respective multiple dice may be invoked to perform one or more execution operations associated with an instruction. The logic portions operate to cooperatively accomplish execution operations, such as those operations indicated for an execution pipeline (see, for example,
sample pipeline 300 illustrated inFIG. 3 ). These execution operations may include, but are not limited, address generation, instruction pointer generation, fetching instructions, decoding instructions into micro-operations, renaming registers to eliminate certain types of data dependencies, dispatch, scheduling, execution, and retirement. As such, the execution operations may include sub-instruction level tasks performed in response to an instruction. - The logic portions may be allocated among the multiple dice such that certain functions are split. That is, address generation unit logic may be split into a first portion and a second portion, with the first portion being allocated to a first die and a second portion being allocated to a second die. The first and second logic portions may at least partially overlap and may act together to cooperatively perform the operations of an address generation unit. Similarly, a scheduling unit may be split, as may an array such as a general register file, a cache, a floating point register file or a microcode memory array. A memory controller may also be split, as may a cache, a translation lookaside buffer, decode logic, rename logic, fetch logic, retirement logic, and floating point execution unit logic.
- As is indicated above, logic portions may also be allocated such that, rather splitting a block of logic, the intact logic blocks for successive pipeline stages are allocated among the multiple dice of the processor. Such allocation of the logic for pipeline stages may result in a
zigzag communication path 1106 through the die-to-die interface 275 as illustrated inFIG. 11 . - The execution operations associated with an execution stage of an execution pipeline may further include execution, by an execution unit, of arithmetic instruction codes such as integer or floating point instruction codes. As used herein, the term “instruction code” is intended to encompass any unit of work that can be understood and executed by an execution unit, such as a floating point execution unit, arithmetic logic unit, or load/store execution unit. An instruction code may be a micro-operation.
- The execution operations associated with the execution pipeline stage may also include execution, by an execution unit, of a memory instruction code such as a memory read or memory write instruction code.
- The foregoing discussion discloses selected embodiments of a multi-die processor. A
multi-die processor 1702 such as described herein may be utilized on a processing system such as theprocessing system 1700 illustrated inFIG. 17 .System 1700, may be used, for example, to execute one or more instructions of an instruction set. For purposes of this disclosure, a processing system includes any processing system that has a processor, such as, for example; a digital signal processor (DSP), a microcontroller, an application specific integrated circuit (ASIC), or a microprocessor, where such processor is a multi-die processor as described in the embodiments discussed above. Theprocessing system 1700 is representative of processing systems based on the Itanium® and Itanium® 2 microprocessors as well as the Pentium®, Pentium® Pro, Pentium® II, Pentium® III, and Pentium® 4 microprocessors, all of which are available from Intel Corporation. Other systems (including personal computers (PCs) having other microprocessors, engineering workstations, personal digital assistants and other hand-held devices, set-top boxes and the like) may also be used. At least one embodiment ofsystem 1700 may execute a version of the Windows™ operating system available from Microsoft Corporation, although other operating systems and graphical user interfaces, for example, may also be used. -
Processing system 1700 includes amemory system 1705 and aprocessor 1702.Memory system 1705 may storeinstructions 1740 anddata 1741 for controlling the operation of theprocessor 1702.Memory system 1705 is intended as a generalized representation of memory and may include a variety of forms of memory, such as a hard drive, CD-ROM, random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), flash memory and related circuitry.Memory system 1705 may storeinstructions 1740 and/ordata 1741 represented by data signals that may be executed by theprocessor 1702. - The
processing system 1700 includes aninterface portion 1704. Rather than the die-to-die interface 275 between thefirst die 102 and second die 104 of theprocessor 1702, theinterface portion 1704 may be coupled to only one or both of thedice interface portion 1704 is to generate inter-component signals between theprocessor 1702 and another component of thesystem 1700. For example, theinterface portion 1704 may generate inter-component signals between theprocessor 1702 and thememory system 1705. For instance, theinterface portion 1704 may generate signals between theprocessor 1702 andmemory system 1705 in order to perform a memory transaction such as a data-retrieval read operation from memory or a data write to memory. Theinterface portion 1704 may also generate signals between theprocessor 1702 and anothersystem component 1707, such as an RF unit, keyboard, external memory device, monitor, mouse or the like. - In the preceding description, various aspects of an apparatus and system for a multi-die processor are disclosed. For purposes of explanation, specific numbers, examples, systems and configurations were set forth in order to provide a more thorough understanding. However, it is apparent to one skilled in the art that the described apparatus and system may be practiced without the specific details. It will be obvious to those skilled in the art that changes and modifications can be made without departing from the present invention in its broader aspects. While particular embodiments of the present invention have been shown and described, the appended claims are to encompass within their scope all such changes and modifications that fall within the true scope of the present invention.
Claims (47)
1. An apparatus comprising:
a first die having a first face side and a first back side, the first die comprising a first logic portion;
a second die having a second face side and a second back side, the second die comprising a second logic portion;
said first and second dice being coupled together with their faces opposed to each other;
wherein said first logic portion and said second logic portion are to cooperatively execute an instruction.
2. The apparatus of claim 1 , wherein
said first and second dice are further coupled such that the first logic portion and said second logic portion at least partially overlap.
3. The apparatus of claim 1 , wherein:
to cooperatively execute an instruction is further to cooperatively accomplish sub-instruction level tasks in response to an instruction.
4. The apparatus of claim 1 , further comprising:
a conductive inter-die interface between the opposing faces of the first and second dice.
5. The apparatus of claim 4 , wherein:
the inter-die interface is disposed between a subset of the face side of the first die and a subset of the face side of the second die.
6. The apparatus of claim 5 , wherein:
the subset of the face side of the first die is a central region.
7. The apparatus of claim 5 , wherein:
the subset of the face side of the second die is a central region.
8. The apparatus of claim 5 , wherein:
the subset of the face side of the first die is a perimeter region.
9. The apparatus of claim 5 , wherein:
the subset of the face side of the second die is a perimeter region.
10. The apparatus of claim 4 , further comprising:
an interface portion, said interface portion being operatively coupled to at least one of said first logic portion and said second logic portion to generate inter-component signals between the processor and a component.
11. The apparatus of claim 10 , wherein: the interface portion is coupled to the first die.
12. The apparatus of claim 10 , wherein:
the component is a memory system.
13. The apparatus of claim 1 , wherein:
said first logic portion and said second logic portion collectively form address generation logic.
14. The apparatus of claim 1 , wherein:
said first logic portion and said second logic portion collectively form scheduling logic.
15. The apparatus of claim 14 , wherein:
said first logic portion comprises arithmetic scheduling logic and wherein said second logic portion comprises memory request scheduling logic.
16. The apparatus of claim 1 , wherein:
said first logic portion comprises a first portion of an array and wherein said second logic portion comprises a second portion of the array.
17. The apparatus of claim 16 , wherein:
wherein said array is a register file array.
18. The apparatus of claim 16 , wherein:
said array is a microcode memory array.
19. The apparatus of claim 1 , wherein
said first logic portion comprises a hot logic block and said second logic portion comprises a cool logic block.
20. The apparatus of claim 19 , wherein
said first logic portion at least partially overlaps said second logic portion.
21. The apparatus of claim 19 , wherein:
said first logic portion further comprises an execution unit and wherein said second logic portion further comprises a data cache.
22. The apparatus of claim 1 , wherein:
said first logic portion comprises a first execution unit and said second logic portion comprises a second execution unit.
23. The apparatus of claim 22 , wherein:
said first execution unit comprises an integer execution unit and said second execution unit comprises a floating point execution unit.
24. The apparatus of claim 22 , wherein:
said first execution unit comprises floating point execution unit and said second execution unit comprises a single-instruction-multiple-data (SIMD) execution unit
25. The apparatus of claim 1 , wherein:
the first logic portion is disposed on the face side of the first die.
26. The apparatus of claim 1 , wherein:
the second logic portion is disposed on the face side of the second die.
27. The apparatus of claim 1 , wherein:
said first logic portion comprises logic to execute a first pipeline stage to execute the instruction; and
said second logic portion comprises logic to execute a second pipeline stage to execute the instruction.
28. The apparatus of claim 27 , wherein:
logic blocks for additional pipeline stages are disposed on said first and second dice such that a signal path for the pipeline follows a zigzag path between the first and second dice.
29. A processor comprising:
a first partition on a first die;
a second partition on a second die; and
execution logic to invoke the first partition and the second partition to perform an execution operation associated with an instruction.
30. The processor of claim 29 , wherein:
said execution operation further comprises a scheduling operation.
31. The processor of claim 29 , wherein:
said execution operation further comprises an address generation operation.
32. The processor of claim 29 , wherein:
said execution logic, in response the instruction, is further to invoke a partition on the first die to invoke a second execution operation associated with the instruction and is to invoke a partition on the second die to perform a third execution operation associated with the instruction.
33. The processor of claim 32 , wherein:
said second execution operation further comprises an operation associated with an execute stage of an instruction pipeline.
34. The processor of claim 32 , wherein:
said second execution operation further comprises an instruction pointer generation operation.
35. The processor of claim 29 , wherein:
said execution operation further comprises an instruction fetching operation.
36. The processor of claim 29 , wherein:
said second execution operation further comprises a decoding operation.
37. The processor of claim 29 , wherein:
said second execution operation further comprises a renaming operation.
38. The processor of claim 29 , wherein:
said second execution operation further comprises a retirement operation.
39. An apparatus, comprising:
a first die comprising:
an execution unit; and
a first array fraction; and
a second die comprising:
a second array fraction coupled to said first register file fraction by die-to-die couplings to cooperatively operate as an array in conjunction with said first array fraction.
40. The apparatus of claim 39 , wherein:
said first die further comprises a first scheduling fraction; and
said second die further comprises a second scheduling fraction coupled to said first scheduling fraction by die-to-die couplings to cooperatively operate as a scheduling unit in conjunction with said first scheduling fraction.
41. The apparatus of claim 39 , wherein:
said first die further comprises a first address generation fraction; and
said second die further comprises a second address generation fraction;
wherein said first address generation fraction is coupled to said second address generation fraction to cooperatively operate as an address generation unit in conjunction with the second address generation fraction.
42. The apparatus of claim 39 , wherein:
the array is a register file.
43. The apparatus of claim 39 , wherein:
the array is a microcode memory array.
44. The apparatus of claim 39 , wherein a microprocessor comprises said first die comprising said first partition and said second die comprising said second partition as well as an interface disposed on said first die, and further wherein said apparatus is a system further comprising:
a memory coupled to the interface portion of the microprocessor, said memory to store an instruction which when executed by the microprocessor causes said microprocessor to invoke said first partition on said first die and said second partition on said second die
45. The system of claim 38 , further comprising:
an additional system component comprising an RF unit.
46. The apparatus of claim 1 , wherein
said first logic portion comprises a low power-density region and said second logic portion comprises a high power-density region.
47. The apparatus of claim 46 , wherein
said first logic portion at least partially overlaps said second logic portion.
Priority Applications (8)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US10/738,680 US20050127490A1 (en) | 2003-12-16 | 2003-12-16 | Multi-die processor |
PCT/US2004/039236 WO2005062190A1 (en) | 2003-12-16 | 2004-11-24 | Multi-die processor |
JP2006543849A JP2007514317A (en) | 2003-12-16 | 2004-11-24 | Equipment and processor |
TW093136574A TWI284410B (en) | 2003-12-16 | 2004-11-26 | Multi-die processor |
CN2004100955659A CN1630080B (en) | 2003-12-16 | 2004-11-29 | Multi-die processor |
US12/365,171 US8860199B2 (en) | 2003-12-16 | 2009-02-04 | Multi-die processor |
JP2010225469A JP2011029663A (en) | 2003-12-16 | 2010-10-05 | Apparatus and processor |
JP2013173544A JP5847773B2 (en) | 2003-12-16 | 2013-08-23 | Equipment and processor |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US10/738,680 US20050127490A1 (en) | 2003-12-16 | 2003-12-16 | Multi-die processor |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US12/365,171 Division US8860199B2 (en) | 2003-12-16 | 2009-02-04 | Multi-die processor |
Publications (1)
Publication Number | Publication Date |
---|---|
US20050127490A1 true US20050127490A1 (en) | 2005-06-16 |
Family
ID=34654251
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/738,680 Abandoned US20050127490A1 (en) | 2003-12-16 | 2003-12-16 | Multi-die processor |
US12/365,171 Expired - Fee Related US8860199B2 (en) | 2003-12-16 | 2009-02-04 | Multi-die processor |
Family Applications After (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US12/365,171 Expired - Fee Related US8860199B2 (en) | 2003-12-16 | 2009-02-04 | Multi-die processor |
Country Status (5)
Country | Link |
---|---|
US (2) | US20050127490A1 (en) |
JP (3) | JP2007514317A (en) |
CN (1) | CN1630080B (en) |
TW (1) | TWI284410B (en) |
WO (1) | WO2005062190A1 (en) |
Cited By (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070126105A1 (en) * | 2005-12-06 | 2007-06-07 | Elpida Memory Inc. | Stacked type semiconductor memory device and chip selection circuit |
US20070290365A1 (en) * | 2006-06-20 | 2007-12-20 | Infineon Technologies Ag | Electronic Device Including a Component Stack and Connecting Elements, and Connecting Elements, and Method for Producing the Electronic Device |
US20080122061A1 (en) * | 2006-11-29 | 2008-05-29 | Texas Instruments Incorporated | Semiconductor chip embedded in an insulator and having two-way heat extraction |
US20080150088A1 (en) * | 2006-12-20 | 2008-06-26 | Reed Paul A | Method for incorporating existing silicon die into 3d integrated stack |
US20090224388A1 (en) * | 2008-03-04 | 2009-09-10 | International Business Machines Corporation | Semiconductor chip stacking for redundancy and yield improvement |
US20100015732A1 (en) * | 2007-11-29 | 2010-01-21 | International Business Machines Corporation | Semiconductor chip repair by stacking of a base semiconductor chip and a repair semiconductor chip |
US7692946B2 (en) | 2007-06-29 | 2010-04-06 | Intel Corporation | Memory array on more than one die |
US20100140750A1 (en) * | 2008-12-10 | 2010-06-10 | Qualcomm Incorporated | Parallel Plane Memory and Processor Coupling in a 3-D Micro-Architectural System |
US8418115B1 (en) * | 2010-05-11 | 2013-04-09 | Xilinx, Inc. | Routability based placement for multi-die integrated circuits |
US9041220B2 (en) | 2013-02-13 | 2015-05-26 | Qualcomm Incorporated | Semiconductor device having stacked memory elements and method of stacking memory elements on a semiconductor device |
CN108027633A (en) * | 2015-09-21 | 2018-05-11 | 高通股份有限公司 | Using packaging part inner sensor circuit and the method that temperature is alleviated are provided for computing device |
WO2019079625A1 (en) * | 2017-10-20 | 2019-04-25 | Xcelsis Corporation | 3d compute circuit with high density z-axis interconnects |
US10580735B2 (en) | 2016-10-07 | 2020-03-03 | Xcelsis Corporation | Stacked IC structure with system level wiring on multiple sides of the IC die |
US10580757B2 (en) | 2016-10-07 | 2020-03-03 | Xcelsis Corporation | Face-to-face mounted IC dies with orthogonal top interconnect layers |
US10586786B2 (en) | 2016-10-07 | 2020-03-10 | Xcelsis Corporation | 3D chip sharing clock interconnect layer |
US10593667B2 (en) | 2016-10-07 | 2020-03-17 | Xcelsis Corporation | 3D chip with shielded clock lines |
US10600691B2 (en) | 2016-10-07 | 2020-03-24 | Xcelsis Corporation | 3D chip sharing power interconnect layer |
US10600735B2 (en) | 2016-10-07 | 2020-03-24 | Xcelsis Corporation | 3D chip sharing data bus |
US10600780B2 (en) | 2016-10-07 | 2020-03-24 | Xcelsis Corporation | 3D chip sharing data bus circuit |
US10607136B2 (en) | 2017-08-03 | 2020-03-31 | Xcelsis Corporation | Time borrowing between layers of a three dimensional chip stack |
US10672743B2 (en) | 2016-10-07 | 2020-06-02 | Xcelsis Corporation | 3D Compute circuit with high density z-axis interconnects |
US10672744B2 (en) | 2016-10-07 | 2020-06-02 | Xcelsis Corporation | 3D compute circuit with high density Z-axis interconnects |
US10672663B2 (en) | 2016-10-07 | 2020-06-02 | Xcelsis Corporation | 3D chip sharing power circuit |
US10672745B2 (en) | 2016-10-07 | 2020-06-02 | Xcelsis Corporation | 3D processor |
US10930593B2 (en) * | 2019-03-13 | 2021-02-23 | Samsung Electronics Co., Ltd. | Package on package and package connection system comprising the same |
US10937778B2 (en) | 2018-06-18 | 2021-03-02 | Commissariat à l'énergie atomique et aux énergies alternatives | Integrated circuit comprising macros and method of fabricating the same |
US11238206B1 (en) | 2021-03-26 | 2022-02-01 | Xilinx, Inc. | Partition wire assignment for routing multi-partition circuit designs |
US11289333B2 (en) | 2016-10-07 | 2022-03-29 | Xcelsis Corporation | Direct-bonded native interconnects and active base die |
US11599299B2 (en) | 2019-11-19 | 2023-03-07 | Invensas Llc | 3D memory circuit |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050127490A1 (en) | 2003-12-16 | 2005-06-16 | Black Bryan P. | Multi-die processor |
US7777330B2 (en) * | 2008-02-05 | 2010-08-17 | Freescale Semiconductor, Inc. | High bandwidth cache-to-processing unit communication in a multiple processor/cache system |
CN103681639B (en) * | 2012-09-25 | 2017-02-08 | 格科微电子(上海)有限公司 | A system-level packaging structure and a packaging method thereof |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5376825A (en) * | 1990-10-22 | 1994-12-27 | Seiko Epson Corporation | Integrated circuit package for flexible computer system alternative architectures |
US6093969A (en) * | 1999-05-15 | 2000-07-25 | Lin; Paul T. | Face-to-face (FTF) stacked assembly of substrate-on-bare-chip (SOBC) modules |
US6229158B1 (en) * | 1999-05-25 | 2001-05-08 | Intel Corporation | Stacked die integrated circuit device |
US6411561B2 (en) * | 2000-05-26 | 2002-06-25 | Hitachi, Ltd. | Semiconductor device including multi-chip |
US6605875B2 (en) * | 1999-12-30 | 2003-08-12 | Intel Corporation | Integrated circuit die having bond pads near adjacent sides to allow stacking of dice without regard to dice size |
US6844624B1 (en) * | 2003-06-26 | 2005-01-18 | Renesas Technology Corp. | Multichip module |
US6917219B2 (en) * | 2003-03-12 | 2005-07-12 | Xilinx, Inc. | Multi-chip programmable logic device having configurable logic circuitry and configuration data storage on different dice |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH04196159A (en) * | 1990-11-26 | 1992-07-15 | Kawasaki Steel Corp | Semiconductor device |
JPH05283607A (en) * | 1992-02-03 | 1993-10-29 | Hitachi Ltd | Semiconductor integrated circuit device and computer system utilizing same |
EP0721662A1 (en) | 1993-09-30 | 1996-07-17 | Kopin Corporation | Three-dimensional processor using transferred thin film circuits |
JPH1022449A (en) * | 1996-07-02 | 1998-01-23 | Matsushita Electric Ind Co Ltd | Semiconductor device |
US5909587A (en) * | 1997-10-24 | 1999-06-01 | Advanced Micro Devices, Inc. | Multi-chip superscalar microprocessor module |
JPH11168185A (en) * | 1997-12-03 | 1999-06-22 | Rohm Co Ltd | Laminated substrate body and semiconductor device |
JP2000223657A (en) * | 1999-02-03 | 2000-08-11 | Rohm Co Ltd | Semiconductor device and semiconductor chip used for the same |
JP2000227457A (en) * | 1999-02-05 | 2000-08-15 | Rohm Co Ltd | Semiconductor device |
US6559531B1 (en) * | 1999-10-14 | 2003-05-06 | Sun Microsystems, Inc. | Face to face chips |
US20050127490A1 (en) | 2003-12-16 | 2005-06-16 | Black Bryan P. | Multi-die processor |
-
2003
- 2003-12-16 US US10/738,680 patent/US20050127490A1/en not_active Abandoned
-
2004
- 2004-11-24 JP JP2006543849A patent/JP2007514317A/en active Pending
- 2004-11-24 WO PCT/US2004/039236 patent/WO2005062190A1/en active Application Filing
- 2004-11-26 TW TW093136574A patent/TWI284410B/en not_active IP Right Cessation
- 2004-11-29 CN CN2004100955659A patent/CN1630080B/en not_active Expired - Fee Related
-
2009
- 2009-02-04 US US12/365,171 patent/US8860199B2/en not_active Expired - Fee Related
-
2010
- 2010-10-05 JP JP2010225469A patent/JP2011029663A/en active Pending
-
2013
- 2013-08-23 JP JP2013173544A patent/JP5847773B2/en not_active Expired - Fee Related
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5376825A (en) * | 1990-10-22 | 1994-12-27 | Seiko Epson Corporation | Integrated circuit package for flexible computer system alternative architectures |
US6093969A (en) * | 1999-05-15 | 2000-07-25 | Lin; Paul T. | Face-to-face (FTF) stacked assembly of substrate-on-bare-chip (SOBC) modules |
US6229158B1 (en) * | 1999-05-25 | 2001-05-08 | Intel Corporation | Stacked die integrated circuit device |
US6605875B2 (en) * | 1999-12-30 | 2003-08-12 | Intel Corporation | Integrated circuit die having bond pads near adjacent sides to allow stacking of dice without regard to dice size |
US6411561B2 (en) * | 2000-05-26 | 2002-06-25 | Hitachi, Ltd. | Semiconductor device including multi-chip |
US6917219B2 (en) * | 2003-03-12 | 2005-07-12 | Xilinx, Inc. | Multi-chip programmable logic device having configurable logic circuitry and configuration data storage on different dice |
US6844624B1 (en) * | 2003-06-26 | 2005-01-18 | Renesas Technology Corp. | Multichip module |
Cited By (59)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8709871B2 (en) | 2005-12-06 | 2014-04-29 | Junji Yamada | Stacked type semiconductor memory device and chip selection circuit |
US8076764B2 (en) * | 2005-12-06 | 2011-12-13 | Elpida Memory Inc. | Stacked type semiconductor memory device and chip selection circuit |
US20070126105A1 (en) * | 2005-12-06 | 2007-06-07 | Elpida Memory Inc. | Stacked type semiconductor memory device and chip selection circuit |
US7800237B2 (en) * | 2006-06-20 | 2010-09-21 | Infineon Technologies Ag | Electronic device including a component stack and connecting elements, and connecting elements, and method for producing the electronic device |
US20070290365A1 (en) * | 2006-06-20 | 2007-12-20 | Infineon Technologies Ag | Electronic Device Including a Component Stack and Connecting Elements, and Connecting Elements, and Method for Producing the Electronic Device |
US20080122061A1 (en) * | 2006-11-29 | 2008-05-29 | Texas Instruments Incorporated | Semiconductor chip embedded in an insulator and having two-way heat extraction |
US20080150088A1 (en) * | 2006-12-20 | 2008-06-26 | Reed Paul A | Method for incorporating existing silicon die into 3d integrated stack |
US8110899B2 (en) * | 2006-12-20 | 2012-02-07 | Intel Corporation | Method for incorporating existing silicon die into 3D integrated stack |
US8059441B2 (en) | 2007-06-29 | 2011-11-15 | Intel Corporation | Memory array on more than one die |
US7692946B2 (en) | 2007-06-29 | 2010-04-06 | Intel Corporation | Memory array on more than one die |
US20100149849A1 (en) * | 2007-06-29 | 2010-06-17 | Mohammed Taufique | Memory array on more than one die |
TWI405212B (en) * | 2007-06-29 | 2013-08-11 | Intel Corp | Memory array on more than one die and method for accessing it |
US8796047B2 (en) | 2007-11-29 | 2014-08-05 | International Business Machines Corporation | Semiconductor chip repair by stacking of a base semiconductor chip and a repair semiconductor chip |
US20100015732A1 (en) * | 2007-11-29 | 2010-01-21 | International Business Machines Corporation | Semiconductor chip repair by stacking of a base semiconductor chip and a repair semiconductor chip |
US8679861B2 (en) | 2007-11-29 | 2014-03-25 | International Business Machines Corporation | Semiconductor chip repair by stacking of a base semiconductor chip and a repair semiconductor chip |
US8597960B2 (en) * | 2008-03-04 | 2013-12-03 | International Business Machines Corporation | Semiconductor chip stacking for redundancy and yield improvement |
US8686559B2 (en) | 2008-03-04 | 2014-04-01 | International Business Machines Corporation | Semiconductor chip stacking for redundancy and yield improvement |
US20090224388A1 (en) * | 2008-03-04 | 2009-09-10 | International Business Machines Corporation | Semiconductor chip stacking for redundancy and yield improvement |
US20100140750A1 (en) * | 2008-12-10 | 2010-06-10 | Qualcomm Incorporated | Parallel Plane Memory and Processor Coupling in a 3-D Micro-Architectural System |
CN102217066A (en) * | 2008-12-10 | 2011-10-12 | 高通股份有限公司 | Parallel plane memory and processor coupling in a 3-d micro-architectural system |
WO2010068785A1 (en) * | 2008-12-10 | 2010-06-17 | Qualcomm Incorporated | Parallel plane memory and processor coupling in a 3-d micro-architectural system |
US8418115B1 (en) * | 2010-05-11 | 2013-04-09 | Xilinx, Inc. | Routability based placement for multi-die integrated circuits |
US9041220B2 (en) | 2013-02-13 | 2015-05-26 | Qualcomm Incorporated | Semiconductor device having stacked memory elements and method of stacking memory elements on a semiconductor device |
US9153461B2 (en) | 2013-02-13 | 2015-10-06 | Qualcomm Incorporated | Semiconductor device having stacked memory elements and method of stacking memory elements on a semiconductor device |
CN108027633A (en) * | 2015-09-21 | 2018-05-11 | 高通股份有限公司 | Using packaging part inner sensor circuit and the method that temperature is alleviated are provided for computing device |
US10586786B2 (en) | 2016-10-07 | 2020-03-10 | Xcelsis Corporation | 3D chip sharing clock interconnect layer |
US11152336B2 (en) | 2016-10-07 | 2021-10-19 | Xcelsis Corporation | 3D processor having stacked integrated circuit die |
US10580757B2 (en) | 2016-10-07 | 2020-03-03 | Xcelsis Corporation | Face-to-face mounted IC dies with orthogonal top interconnect layers |
US11824042B2 (en) | 2016-10-07 | 2023-11-21 | Xcelsis Corporation | 3D chip sharing data bus |
US10593667B2 (en) | 2016-10-07 | 2020-03-17 | Xcelsis Corporation | 3D chip with shielded clock lines |
US10600691B2 (en) | 2016-10-07 | 2020-03-24 | Xcelsis Corporation | 3D chip sharing power interconnect layer |
US10600735B2 (en) | 2016-10-07 | 2020-03-24 | Xcelsis Corporation | 3D chip sharing data bus |
US10600780B2 (en) | 2016-10-07 | 2020-03-24 | Xcelsis Corporation | 3D chip sharing data bus circuit |
US11823906B2 (en) | 2016-10-07 | 2023-11-21 | Xcelsis Corporation | Direct-bonded native interconnects and active base die |
US10672743B2 (en) | 2016-10-07 | 2020-06-02 | Xcelsis Corporation | 3D Compute circuit with high density z-axis interconnects |
US10672744B2 (en) | 2016-10-07 | 2020-06-02 | Xcelsis Corporation | 3D compute circuit with high density Z-axis interconnects |
US10672663B2 (en) | 2016-10-07 | 2020-06-02 | Xcelsis Corporation | 3D chip sharing power circuit |
US10672745B2 (en) | 2016-10-07 | 2020-06-02 | Xcelsis Corporation | 3D processor |
US11557516B2 (en) | 2016-10-07 | 2023-01-17 | Adeia Semiconductor Inc. | 3D chip with shared clock distribution network |
US11289333B2 (en) | 2016-10-07 | 2022-03-29 | Xcelsis Corporation | Direct-bonded native interconnects and active base die |
US10886177B2 (en) | 2016-10-07 | 2021-01-05 | Xcelsis Corporation | 3D chip with shared clock distribution network |
US10892252B2 (en) | 2016-10-07 | 2021-01-12 | Xcelsis Corporation | Face-to-face mounted IC dies with orthogonal top interconnect layers |
US10580735B2 (en) | 2016-10-07 | 2020-03-03 | Xcelsis Corporation | Stacked IC structure with system level wiring on multiple sides of the IC die |
US11881454B2 (en) | 2016-10-07 | 2024-01-23 | Adeia Semiconductor Inc. | Stacked IC structure with orthogonal interconnect layers |
US10950547B2 (en) | 2016-10-07 | 2021-03-16 | Xcelsis Corporation | Stacked IC structure with system level wiring on multiple sides of the IC die |
US10978348B2 (en) | 2016-10-07 | 2021-04-13 | Xcelsis Corporation | 3D chip sharing power interconnect layer |
US10970627B2 (en) | 2017-08-03 | 2021-04-06 | Xcelsis Corporation | Time borrowing between layers of a three dimensional chip stack |
US11790219B2 (en) | 2017-08-03 | 2023-10-17 | Adeia Semiconductor Inc. | Three dimensional circuit implementing machine trained network |
US11176450B2 (en) | 2017-08-03 | 2021-11-16 | Xcelsis Corporation | Three dimensional circuit implementing machine trained network |
US10762420B2 (en) | 2017-08-03 | 2020-09-01 | Xcelsis Corporation | Self repairing neural network |
US10719762B2 (en) | 2017-08-03 | 2020-07-21 | Xcelsis Corporation | Three dimensional chip structure implementing machine trained network |
US10607136B2 (en) | 2017-08-03 | 2020-03-31 | Xcelsis Corporation | Time borrowing between layers of a three dimensional chip stack |
TWI745626B (en) * | 2017-10-20 | 2021-11-11 | 美商塞爾席斯公司 | 3d compute circuit with high density z-axis interconnects |
WO2019079625A1 (en) * | 2017-10-20 | 2019-04-25 | Xcelsis Corporation | 3d compute circuit with high density z-axis interconnects |
US10937778B2 (en) | 2018-06-18 | 2021-03-02 | Commissariat à l'énergie atomique et aux énergies alternatives | Integrated circuit comprising macros and method of fabricating the same |
US10930593B2 (en) * | 2019-03-13 | 2021-02-23 | Samsung Electronics Co., Ltd. | Package on package and package connection system comprising the same |
TWI799580B (en) * | 2019-03-13 | 2023-04-21 | 南韓商三星電子股份有限公司 | Package on package and package connection system comprising the same |
US11599299B2 (en) | 2019-11-19 | 2023-03-07 | Invensas Llc | 3D memory circuit |
US11238206B1 (en) | 2021-03-26 | 2022-02-01 | Xilinx, Inc. | Partition wire assignment for routing multi-partition circuit designs |
Also Published As
Publication number | Publication date |
---|---|
JP2007514317A (en) | 2007-05-31 |
JP2013258425A (en) | 2013-12-26 |
WO2005062190A1 (en) | 2005-07-07 |
TWI284410B (en) | 2007-07-21 |
JP2011029663A (en) | 2011-02-10 |
TW200532895A (en) | 2005-10-01 |
CN1630080A (en) | 2005-06-22 |
US20090138688A1 (en) | 2009-05-28 |
US8860199B2 (en) | 2014-10-14 |
CN1630080B (en) | 2010-05-05 |
JP5847773B2 (en) | 2016-01-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8860199B2 (en) | Multi-die processor | |
Ozaki et al. | Cool mega-arrays: Ultralow-power reconfigurable accelerator chips | |
Puttaswamy et al. | Thermal herding: Microarchitecture techniques for controlling hotspots in high-performance 3D-integrated processors | |
US8386690B2 (en) | On-chip networks for flexible three-dimensional chip integration | |
Gwennap | Digital 21264 sets new standard | |
Diefendorff | Power4 focuses on memory bandwidth | |
US20040019765A1 (en) | Pipelined reconfigurable dynamic instruction set processor | |
US7994822B2 (en) | Semiconductor device for synchronous communication between stacked LSI | |
US20100140750A1 (en) | Parallel Plane Memory and Processor Coupling in a 3-D Micro-Architectural System | |
US7680972B2 (en) | Micro interrupt handler | |
CN102193779A (en) | MPSoC (multi-processor system-on-chip)-oriented multithread scheduling method | |
JP5412662B2 (en) | Three-dimensional laminated structure computer system with low-capacity through electrode | |
Chacin et al. | COOL interconnect low power interconnection technology for scalable 3D LSI design | |
Swaminathan | The next frontier: Enabling Moore’s Law using heterogeneous integration | |
Matsumoto et al. | Cool System scalable 3-D stacked heterogeneous Multi-Core/Multi-Chip architecture for ultra low-power digital TV applications | |
Loh et al. | The next era for chiplet innovation | |
Rea et al. | PowerPC™ RAD750™-a microprocessor for now and the future | |
Neela et al. | Challenges in 3DIC implementation of a design using current CAD tools | |
CN111209246A (en) | Micro programmable on-chip computer based on multi-chip packaging technology | |
US11822475B2 (en) | Integrated circuit with 3D partitioning | |
KR20170067986A (en) | Processor, and computing method comprising the same and method for driving the processor | |
Linderman et al. | A dependable high performance wafer scale architecture for embedded signal processing | |
Buckhanan et al. | Bridging the HPC processor-memory gap with quilt packaging | |
Chen et al. | Design and chip implementation of a heterogeneous multi-core DSP | |
Swaminathan et al. | AMD Instinct TM MI250X Accelerator enabled by Elevated Fanout Bridge Advanced Packaging Architecture |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: INTEL CORPORATION, CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:BLACK, BRYAN P.;SAMRA, NICHOLAS G.;WEBB, M. CLAIR;REEL/FRAME:015380/0733;SIGNING DATES FROM 20040512 TO 20040526 |
|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION |