WO2010132995A1 - Configurable module and memory subsystem - Google Patents
Configurable module and memory subsystem Download PDFInfo
- Publication number
- WO2010132995A1 WO2010132995A1 PCT/CA2010/000769 CA2010000769W WO2010132995A1 WO 2010132995 A1 WO2010132995 A1 WO 2010132995A1 CA 2010000769 W CA2010000769 W CA 2010000769W WO 2010132995 A1 WO2010132995 A1 WO 2010132995A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- mcd
- pair
- memory module
- memory
- communication
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C5/00—Details of stores covered by group G11C11/00
- G11C5/02—Disposition of storage elements, e.g. in the form of a matrix array
- G11C5/025—Geometric lay-out considerations of storage- and peripheral-blocks in a semiconductor storage device
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C7/00—Arrangements for writing information into, or reading information out from, a digital store
- G11C7/10—Input/output [I/O] data interface arrangements, e.g. I/O data control circuits, I/O data buffers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F13/14—Handling requests for interconnection or transfer
- G06F13/16—Handling requests for interconnection or transfer for access to memory bus
- G06F13/1668—Details of memory controller
- G06F13/1684—Details of memory controller using multiple buses
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F13/38—Information transfer, e.g. on bus
- G06F13/40—Bus structure
- G06F13/4004—Coupling between buses
- G06F13/4022—Coupling between buses using switching circuits, e.g. switching matrix, connection or expansion network
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
- G06F13/38—Information transfer, e.g. on bus
- G06F13/42—Bus transfer protocol, e.g. handshake; Synchronisation
- G06F13/4204—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus
- G06F13/4234—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus being a memory bus
- G06F13/4243—Bus transfer protocol, e.g. handshake; Synchronisation on a parallel bus being a memory bus with synchronous protocol
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C16/00—Erasable programmable read-only memories
- G11C16/02—Erasable programmable read-only memories electrically programmable
- G11C16/04—Erasable programmable read-only memories electrically programmable using variable threshold transistors, e.g. FAMOS
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C5/00—Details of stores covered by group G11C11/00
- G11C5/02—Disposition of storage elements, e.g. in the form of a matrix array
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C5/00—Details of stores covered by group G11C11/00
- G11C5/06—Arrangements for interconnecting storage elements electrically, e.g. by wiring
Definitions
- the invention relates generally to memory subsystems. More specifically, the invention relates to a memory subsystem that allows a user to configure the number of memory channels, memory speed, and memory capacity.
- CPU central processing unit
- Host systems often use a hierarchy of memories with different speeds and capacities. Higher capacity memories typically have longer latency and therefore are slower than lower capacity memories.
- a CPU communicates directly with highspeed registers and a cache memory, transferring small amounts of data at a rate comparable to the CPU speed.
- the cache memory prefetches a larger block of data from a dynamic random access memory (DRAM), but at a rate slower than the CPU speed.
- DRAM dynamic random access memory
- the cache takes advantage of the sequential nature of many data requests by the CPU to minimize CPU stalls or situations where the CPU cannot complete a task for lack of necessary data.
- the DRAM also prefetches a larger block of data from a hard disk drive (HDD).
- HDD memories are significantly slower than DRAMs and other solid-state memories because of their mechanical nature.
- each level of memory hierarchy provides several orders of magnitude increase in memory capacity with a single order of magnitude increase in latency. The significant increase in latency between the DRAM and HDD memory hierarchy levels necessitates a memory solution with performance and capacity in between what is available from DRAM and HDD technologies so that data flow is balanced throughout the memory hierarchy.
- the invention features a memory module with a plurality of memory- containing devices (MCDs).
- MCDs memory- containing devices
- Each MCD has an input port, an output port, a memory, and a bridge in communication with the input port, the output port, and the memory.
- the bridge is adapted to receive a command and a data packet from the input port and adapted to execute in response to the command at least one of a transfer of a portion of the data packet from the input port to the memory, a transfer of a portion of the data packet from the input port to the output port, and a transfer of a portion of a memory packet from the memory to the output port.
- a circuit board has the MCDs mounted thereto and has a plurality of conducting paths for serial communication between at least two of the MCDs.
- the invention features a configurable memory subsystem with a memory module having a first MCD pair and a second MCD pair mounted thereto.
- Each of the MCD pairs has a first MCD and a second MCD.
- Each MCD has an input port, an output port, a memory, and a bridge in communication with the input port, the output port, and the memory.
- the bridge is adapted to receive a command and a data packet from the input port and adapted to execute in response to the command at least one of a transfer of a portion of the data packet from the input port to the memory, a transfer of a portion of the data packet from the input port to the output port, and a transfer of a portion of a memory packet from the memory to the output port.
- Each of the MCD pairs has a pair input in communication with the input port of the first MCD and in communication with a plurality of input conductors on an edge of the circuit board.
- the output port of the first MCD is in communication with the input port of the second MCD.
- the output port of the second MCD is in communication with a pair output and in communication with a plurality of output conductors on the edge of the circuit board.
- a loop-back device is in communication with the memory module. The loop-back device receives the command and the data packet from the pair output of the first MCD pair and transmits the command and the data packet to the pair input of the second MCD pair.
- the invention features a configurable memory subsystem with a forward memory module and a reversed memory module.
- Each of the forward and reversed memory modules has a circuit board divided into an upper half and a lower half.
- Each of the upper and lower halves includes a plurality of MCD pairs mounted thereto.
- Each of the MCD pairs has a pair input for receiving packets and a pair output for transmitting packets.
- the upper half has an outer MCD pair and an inner MCD pair.
- Each MCD pair is disposed adjacent to a same unpopulated MCD pair mounting location.
- the lower half has an outer MCD pair and an inner MCD pair.
- the outer MCD pair is disposed adjacent to the inner MCD pair.
- the inner MCD pair is disposed adjacent to an unpopulated MCD pair mounting location adjacent to the upper half.
- the reversed memory module is disposed adjacent to the forward memory module, wherein a linear order of MCD pairs on the reversed memory module is opposite to a linear order of the forward memory module.
- the at least one of the MCD pairs on the upper half of the forward memory module being in serial communication with at least one of the MCD pairs on the lower half of the reversed memory module.
- the at least one of the MCD pairs on the lower half of the forward memory module being in serial communication with the at least one of the MCD pairs on the upper half of the reversed memory module.
- FIG. 1 is a perspective view of an embodiment of a configurable memory subsystem according to the invention.
- FIG. 2 is a schematic view of an MCD.
- FIG. 3 is a cross-sectional view of FIG. 1 taken along A-A', further illustrating the direction of one of the communication channels.
- FIG. 4 is schematic view of the front-side and back-side of a memory module showing communication between four MCD pairs.
- FIG. 5A is a schematic view of the front-side and back-side of a memory module showing alternating placement of MCD pairs.
- FIG. 5B is a schematic view of the loop-back signals of one channel of the memory module in FIG. 5 A for an eight-bit bus.
- FIG. 6A is a schematic view of the front-side and back-side of a memory module showing alternating placement of MCD pairs, and alternating bus order of two of the MCD pairs.
- FIG. 6B is a schematic view of the loop-back signals of one channel of the memory module in FIG. 6A for an eight-bit bus.
- FIG. 7 is a schematic view of the front-side and back-side of a memory module showing alternating and staggered placement of MCD pairs.
- FIG. 8 is a schematic view of the front-side and back-side of a memory module showing alternating and rotated placement of MCD pairs.
- FIG. 9 is a schematic view of an embodiment of a loop-back device including four MCD pairs.
- FIG. 10 is a schematic view of the front-side and back-side of a reversible memory module showing MCD mounting locations.
- FIGs. 1 IA, 1 IB and 11C illustrate an embodiment of a two-channel memory subsystem with one of three loop-back devices.
- FIGs. 12A, 12B and 12C illustrate an embodiment of a four-channel interleaved memory subsystem with one of three loop-back devices.
- FIG. 13 illustrates an embodiment of a single-channel memory subsystem with one of three loop-back devices.
- FIG. 14 illustrates an embodiment of a reversible memory subsystem showing one module in a forward position and one module in the reversed position.
- Embodiments of configurable modules and memory subsystems described herein provide a user the flexibility to configure the number of memory channels, memory speed and memory capacity or any combination thereof.
- the user includes, but is not limited to, the product consumer or any person or entity involved with manufacturing the product.
- a personal computer includes a motherboard with four memory modules attached by electrical sockets. The user configures the system by removing memory modules and replacing them in other sockets or replacing them with different memory modules having different attributes. The user reverses the orientation of the memory module to alter the connections between the memory module and the motherboard. Alternatively, the user reprograms a controller that communicates with the memory modules.
- the controller provides communication between the host system and the memory modules, the latter configured in one or more daisy-chained rings.
- communication refers generally to electrical voltage conduction over a circuit board trace, although other communication techniques such as current conduction, fiber-optic conduction, or inductive coupling are contemplated in other embodiments.
- the controller Upon powering up the host system, the controller sends command packets to the interconnected memory modules to learn the characteristics of the memory modules. The controller then sends and receives commands and optionally data packets to the rings.
- the rings originate and terminate at the controller and the controller can configure the number of rings.
- the various configurable memory subsystems described herein are, for example, systems ranging from high-capacity storage class memories, high-speed personal computer memories or low-power memories for portable devices.
- the memory subsystems can replace HDDs with a solid-state drive (SSD) thereby reducing both power consumption and data access latency, and improving life-time and shock resistance.
- SSD solid-state drive
- Using a combination of various memory types, speed and capacities of the memory subsystems described herein enables new memory applications.
- FIG. 1 shows a perspective view of an embodiment of a memory subsystem 10 with four memory modules 12a, 12b, 12c and 12d (generally 12) and illustrates the direction of flow of command and data packets for two channels.
- the memory subsystem 10 can contain any number of memory modules 12.
- Each memory module 12 has a circuit board 14 with conductors 16a on a front-side 22 and conductors 16b (generally 16) on a back-side 24 for communication with another circuit board 14 such as a motherboard.
- the conductors 16 can communicate with another circuit board 14 through a socket connection, a soldered connection to the other board, or through a variety of means as are known by those in the art.
- the circuit board 14 has a plurality of MCDs 20a attached to a front side 22 and MCDs 20b attached to a back-side 24.
- MCDs 20a and MCDs 20b are either similarly constructed or have different pin order or different types, quantities and capacities of components included therein. Examples of MCD 20 include a multi- chip device and a monolithic device (having one component).
- FIG. 1 depicts a two- channel system.
- the first channel (CHO) receives a command and optionally a data packet on bus 42 from the controller 40.
- the packet on bus 42 flows through the first memory module 12a and exits as a packet on bus 44.
- the packet on bus 44 flows through memory modules 12b, 12c and 12d in succession, returns to 12d through a loop-back path 46, and flows through memory modules 12d, 12c and 12b in succession resulting in a packet on bus 48.
- the packet on bus 48 flows through memory module 12a, exits as a packet on bus 50 and returns to the controller 40 thereby completing a daisy-chained ring.
- the second channel (CHl) follows a similar path to the first channel CHO, beginning as a packet on bus 52, flowing through memory module 12a, exiting as a packet on bus 54 and flowing though memory modules 12b, 12c and 12d in succession.
- the channel CHl then returns to 12d through a loop-back path 56. Subsequently, the channel CHl proceeds through memory modules 12d, 12c and 12b in succession, entering memory module 12a as a packet on bus 58 and exiting as a packet on bus 60 and communicating with the controller 40.
- the controller 40 configures the width of the command and data packets.
- Various embodiments illustrated throughout the figures show an eight-bit wide bus. Although one bit, four bit and eight bit widths are typically used, any width is within the scope of the invention.
- a narrow bus width reduces the physical size of the system, however a wider bus width increases the system bandwidth or rate at which data flows through the system.
- the representative memory module 12 includes conductors 16 on the edge of the circuit board 14.
- the form factor and physical shape of the memory module 12 conforms to a DDR2 SDRAM 200-pin small outline dual in-line memory module (SO- DIMM) standard.
- the memory module 12 conforms to a 240-pin DDR2 SDRAM SO-DIMM form factor.
- FIG. 2 illustrates the general function of the MCD 20.
- a bridge 32 receives a command and optionally a data packet at the input port 26.
- the command packet instructs the bridge 32 to perform one or more operations that include writing a portion of the data packet into the memory 30, transferring a portion of the data packet to the output port 28 or transferring a portion of a memory packet from the memory 30 to the output port 28.
- the bridge 32 allows the command and data packets to propagate through all of the devices on the ring.
- Each ring corresponds to a channel, for example CHO and CHl as illustrated in FIG. 1.
- the bridge 32 permits data to be written to any MCD 20 or broadcast to all the MCDs. Similarly data can be written to one MCD 20 in the channel while simultaneously reading data from another MCD 20 further down-stream, which is at a location that is closer to where the channel returns to the controller 40.
- the controller 40 Upon power up, the controller 40 assigns a unique number to each MCD 20 on each channel. The controller 40 also interrogates each MCD 20 to determine the characteristics of the memory 30. For example, different MCDs 20 can have different amounts of memory capacity, a variety of memory speeds and memory types. In one embodiment each of the MCDs 20 has eight stacked NAND Flash NVMs with each memory supporting eight bit data transfers to and from the memory 30. Other embodiments use such non- volatile memory types as NOR Flash, EEPROM, ferroelectric RAM (FRAM), magneto-resistive RAM (MRAM), phase change RAM, MEMS RAM, ROMs and others known to those in the art. In another embodiment, one or more of the MCDs 20 include volatile memories such as SRAM, and DRAM.
- each MCD 20 includes a mixture of volatile and non- volatile memories, a mixture of memories of different capacity and a mixture of memories of different speeds.
- Memory speed typically refers to sustained data throughput determined by a clock rate, although memory speed can also refer to latency.
- the memory 30 is manufactured from a combination of CMOS, MOS and bipolar designs whether based on silicon, carbon, gallium arsenide (GaAs) and other semiconductor technologies, known to those in the art.
- Assembly of the MCD 20 is based on either a monolithic device or a multi-chip device.
- a multi-chip device can be based on technologies that employ stacked components, components placed on a common substrate ("side-by-side"), system in package (SiP), package in package (PiP) and package on package (PoP) technologies.
- a component can be a CMOS design in a semiconductor technology for example.
- the bridge 32 is a device as disclosed in commonly owned U.S. Patent Application Serial No. 12/401,963 filed on March 11, 2009 entitled "A COMPOSITE MEMORY HAVING A BRIDGING DEVICE FOR CONNECTING DISCRETE MEMORY DEVICES TO A SYSTEM", which is hereby incorporated by reference, wherein the bridge 32 uses a HYPERLINK® NAND (HLNANDTM) Flash interface.
- HPNANDTM HYPERLINK® NAND
- the HYPERLINK® topology includes an HL2 version that uses a source synchronous clock and an HLl version that uses a parallel clock.
- the HL2 version is preferable over the HLl version to optimize the benefits of reduce signal loading however both versions are envisioned within the scope of this invention.
- FIG. 3 shows a first MCD 20a on a front-side 22 of a circuit board 14 and a second MCD 20b on a back-side 24 of the circuit board 14.
- the MCDs 20 are in serial communication with each other and form an MCD pair.
- the output port 28a of MCD 20a is positioned proximate to the input port 26b of the second MCD 20b to reduce the communication path and hence the capacitive loading. Proximately located means that the lengths of all of the bit paths within a port are averaged together, rather than optimizing the length of one bit more than another bit.
- FIG. 3 further illustrates the direction of channel communication from FIG. 1.
- a packet on bus 42 from CHO communicates with the memory module 12 through the input conductors 16a.
- the input conductors 16a and the output conductors 16b make contact with a socket that is attached to a motherboard, wherein the socket has electrical connections with the controller 40.
- the socket attaches to a daughter board, which in turn attaches to the motherboard.
- a daughterboard arrangement gives the user further control over configuring how the memory modules 12 communicate with each other.
- the input conductors 16a and output conductors 16b attach to either a daughter board or a mother-board by a soldered connection.
- either or both of the input conductors 16a and the output conductors 16b are disposed on one or both sides of the circuit board 14.
- the input port 26a of the first MCD 20a receives the packet on bus 42, and the output port 28a of the first MCD 20a transmits at least the command portion of the packet on bus 42 through conductors in the circuit board 14 to the input port 26b of the second MCD 20b.
- the second MCD 20b subsequently transmits from the output port 28b and through output conductors 16b at least the command portion of the packet received at the input port 26b of MCD 20b.
- FIG. 4 shows the front-side 22 and the back-side 24 of the memory module 12.
- the channel directions conform to the directions shown in FIG. 1.
- the packet on bus 42 communicates with the input port 26a of the first MCD 20a mounted on the front-side 22 of module 12.
- MCD 20a transmits from the output port 28a to the input port 26b of MCD 20b on the back-side 24 of the module 12.
- MCD 20b on the back-side 24 of the module 12 transmits a packet on bus 44 to the next module.
- the input port 26a of MCD 20a on the front-side 22 of the module 12 receives a packet on bus 48.
- FIG. 5A is an alternative embodiment to that shown in FIG. 4 wherein the communication paths between the conductors 16, and the MCDs 20 shown in FIG.
- a packet on bus 48 communicates with the input port 26b of the MCD 20b on the back-side 24 of the memory module 62
- the output port 28b of the MCD 20b on the back-side 24 of the memory module 62 communicates with the input port 26a of the MCD 20a on the front-side 22 of the memory module 62
- the output port 28a of the MCD 20a on the front-side 22 of the memory module 62 transmits a packet on bus 50 to the controller 40.
- a similar path for CHl is shown in FIG. 5A.
- FIG. 5B illustrates the loop-back connection 46 of FIG. 1 further illustrated in subsequent figures. More particularly, FIG. 5B illustrates a bus order of the output port 28b of the MCD 20b, transmitting a packet on bus 44 for CHO, and mounted on the back-side 24 of the memory module 62 and the bus order of the input port 26b of the MCD 20b receiving a packet on bus 48 for CHO, and mounted on the back-side 24 of the memory module 62.
- the output port 28b transmitting a packet on bus 44 and the input port 26b receiving a packet on bus 48 are each configured to have eight bits shown as 70, 71, 72, 73, 74, 75, 76 and 77, where the bit 70 corresponds to the least significant bit (LSB) and the bit 77 corresponds to the most significant bit (MSB).
- the bus order shown in FIG. 5B requires signals to cross over each other, however it has the advantage that each bit 70-77 has a similar length and therefore a similar capacitive loading on the driver of the output port 28b.
- each bit of the bus 56 has a similar length and therefore a similar capacitive loading on the driver of the output port 28b.
- a similar loading on all of the bits is advantageous because the driver is designed to drive the bit with the highest capacitive loading, which requires a larger peak current consumption and driver area on MCD 20b.
- a memory device typically has limited drive capability when the process used to fabricate the memory is optimized for memory density rather than high transmission speed.
- FIG. 6A is an alternative embodiment to that shown in FIG. 5A wherein the bus order of the input port 26b receiving a packet on bus 48 is reversed from the bus order of the output port 28b transmitting a packet on bus 44, and the bus order of the input port 26b receiving a packet on bus 58 is reversed from the bus order of the output port 28b transmitting a packet on bus 54.
- the embodiment of FIG. 6A reduces routing complexity by eliminating cross-over of any of the bits 70-77; however one bit 70 has the shortest length and another bit 77 has the longest length. The increased bit length of the longest bit 77 illustrated in FIG. 6B (compared to FIG.
- FIG. 7 shows an embodiment wherein the MCDs 20a on the front-side 22 of the memory module 66 are not collinear to the other MCDs 20a, and the MCDs 20b on the back-side 24 of the memory module 66 are not collinear to the other MCDs 20b.
- the MCDs 20a and MCDs 20b are mounted to the memory module 66 in a staggered arrangement.
- This embodiment advantageously permits the mounting of an additional component (not shown) in proximity with the output port 28a of the MCD 20a or the output port 28b of the MCD 20b, for example a series resistor or a capacitor to filter a signal transient response.
- FIG. 8 shows an embodiment with the MCDs 20a and the MCDs 20b rotated on a memory module 68.
- This MCD rotation advantageously facilitates signal routing between the output port 28a of MCD 20a and the input port 26b of MCD 20b, and signal routing between the output port 28b of MCD 20b and the input port 26a of MCD 20a.
- the rotation provides the advantage of reducing the length of loop-back paths 46 and 56 shown in FIG. 1.
- FIG. 9 shows a modification of the memory module 64 of FIG. 6A that is equally applicable to any of the aforementioned embodiments.
- the illustrated memory module 69 combines the loop-back paths 46 and 56 shown in FIG. 1. Specifically, the output port 28b of the MCD 20b mounted on the back-side 24 communicates with the input port 26b of the MCD 20b mounted on the back-side 24 through path 80. In addition, the output port 28b of the MCD 20b mounted on the back-side 24 communicates with the input port 26b of the MCD 20b mounted on the back-side 24 through path 82.
- FIG. 10 shows a front-side 22 and a back-side 24 of a reversible memory module 90.
- Each locations 91, 92, 93, 94, 95 and 96 provide mounting sites for an MCD pair, with each MCD pair including an MCD 20a mounted on the front-side 22 of the memory module 90 and an MCD pair 20b mounted on the back-side of the memory module 90.
- the MCD pair mounting locations 91, 92 and 93 defines an upper half of the memory module 90
- the MCD pair mounting locations 94, 95 and 96 defines a lower half of the memory module 90.
- the upper half of the memory module 90 has an outer MCD pair mounted at MCD pair mounting location 91 and being adjacent to a same unpopulated MCD pair mounting location 92.
- the upper half of the memory module 90 also has an inner MCD pair mounted at MCD pair mounting location 93 and being adjacent to the same unpopulated MCD pair mounting location 92.
- the lower half of the memory module 90 has an outer MCD pair mounted at MCD pair mounting location 96, adjacent to an inner MCD pair mounted at MCD pair mounting location 95.
- the inner MCD pair is adjacent to an unpopulated MCD pair mounting location 94.
- the mounting locations of the MCD pairs on the memory module 90 facilitates memory system loop-back connections when the memory module 90 is placed adjacent to a reversed memory module 90b hereinafter described in further detail in FIG. 14.
- This embodiment is preferably made with a form factor and physical shape conforming to the 240-pin DDR2 SDRAM SO-DIMM standard, although other various form factors and shapes are contemplated.
- the 240-pin DDR2 SDRAM SO- DIMM standard is a cost effective industry standard format.
- Another embodiment alters the embodiment shown in FIG. 10 by replicating the group of MCD pair mounting locations 91-93 and MCD pairs mounted thereto on the upper half of memory module 90 by the same number of times that the group of MCD pair mounting locations 94-96 and MCD pairs mounted thereto on the lower half of the module 90 are replicated.
- FIG. 1 IA shows the two-channel configurable memory subsystem 10 configured as shown in FIG. 1.
- an MCD pair 100 in FIG. 1 IA represents the first MCD 20a in communication with the second MCD 20b as shown in FIG. 4.
- Pair input 102 in FIG. 1 IA represents the input port 26a of the first MCD 20a on the front-side 22 of memory module 12 shown in FIG. 4.
- Pair output 104 in FIG. 1 IA represents the output port 28b of the second MCD 20b on the back-side 24 of the memory module 12.
- the loop-back paths 46 and 56 shown in the configurable memory subsystem 10 can be conduction paths on the same circuit board that provides communication paths between the memory modules 12 and the host controller 40. Any one or more of the memory modules 12 in FIG. 1 IA can be modified or replaced with the aforementioned embodiments of memory modules.
- FIG. 1 IB shows a configurable memory subsystem 110 with the memory module 12d of FIG. 1 IA removed, resulting in an empty memory module location 118, and with the memory module 12c of FIG. 1 IA replaced by a termination module 112.
- the termination module 112 provides the loop-back function previously provided by the loop-back paths 46 and 56 that are disabled upon removal of the memory module 12d.
- the pair output 104 of MCD pair 100 on memory module 12b communicates with the pair input 102 of MCD pair 100 on memory module 12b through the loop-back path 1 14.
- the pair output 104 of another MCD pair 100 on memory module 12b communicates with the pair input 102 of another MCD pair 100 on memory module 12b through the loop-back path 116.
- FIG. 1 IB shows a configurable memory subsystem 110 with the memory module 12d of FIG. 1 IA removed, resulting in an empty memory module location 118, and with the memory module 12c of FIG. 1 IA replaced by a termination module 112.
- the termination module 112
- FIG. 11C shows another variation to the configurable memory subsystem of FIG. 1 IB with the termination module 112 replaced by the memory module 69 shown in FIG. 9 and related variants described herein. Replacing the memory module 12b in FIG. 11 C with the memory module 69 results in the same memory capacity as the memory subsystem 110 shown in FIG. HB.
- an MCD pair 100 can contain ROM to be used to "boot" a host system or load an operating system upon application of power.
- the memory subsystem 120 can be configured with shorter rings for each channel so that MCD pairs 100 with memories requiring slower clock rates can be used without increasing the ring latency.
- Ring latency for CHO is the elapsed time between the controller 40 transmitting a packet on bus 42 to the memory module 12a and the controller 40 receiving a packet on bus 50 from the memory module 12a.
- a memory subsystem using slow memories can be reconfigured for use with an upgraded host system that requires shorter ring latency.
- the shortest ring latency can be achieved by replacing the memory module 12a with the memory module 69.
- FIG. 12A shows a four-channel interleaved memory subsystem 140.
- the memory subsystem 140 has twice as many channels with each channel having half the memory capacity.
- a user can increase the number of memory channels to take advantage of a host upgrade from a single processor system to a multiprocessor system that executes several processes in parallel.
- Representative channel CHO receives a packet on bus 42 from the controller 40.
- An MCD pair 100 on the memory module 12a receives the packet on bus 42 at the pair input 102 and transmits a packet on bus 44 from the pair output 104 of the MCD pair 100 on the memory module 12a.
- An MCD pair 100 on the memory module 12c receives the packet on bus 44 at the pair input 102 and transmits a packet through a loopback path 46 to the input port 102 of another MCD pair 100 on the memory module 12c.
- the other MCD pair 100 on the memory module 12c transmits a packet on bus 48 to the input port 102 of another MCD pair 100 on the memory module 12a.
- the other MCD pair 100 transmits a packet on bus 50 from the output port 104 to the controller 40.
- Representative interleaved channel CH2 receives a packet on bus 142 from the controller 40.
- An MCD pair 100 on the memory module 12b receives the packet on bus 142 at pair input 102 and transmits a packet 144 from the pair output 104 of the MCD pair 100 on the memory module 12b.
- An MCD pair 100 on the memory module 12d receives the packet on bus 144 at the pair input 102 and transmits a packet through a loopback path 146 to the input port 102 of another MCD pair 100 on the memory module 12d.
- the other MCD pair 100 on the memory module 12d transmits a packet on bus 148 to the input port 102 of another MCD pair 100 on the memory module 12b.
- FIG. 12B shows a configurable memory subsystem 160 with the memory modules 12c and 12d of FIG. 12A each replaced by a termination module 112.
- the configurable memory module 160 shown in FIG. 12B reduces the memory capacity of the configurable memory module 140 shown in FIG. 12A by half and reduces the ring delay by half while retaining the same number of channels.
- FIG. 12C shows a configurable memory subsystem 180 with the memory modules ] 2c and 12d of FIG. 12A each replaced by a memory module 69.
- the configurable memory module 180 retains the same memory capacity, ring delay and number of channels of the configurable memory module 140 shown in FIG. 12A.
- a user can simply reconfigure the configurable memory subsystem 180 by replacing the memory modules 12a and 12b with the memory modules 69 to reduce the memory subsystem memory capacity by half and the ring delay by half without altering the number of channels.
- a user can only replace memory module 12a with a memory module 69 to change the characteristics of CHO and CHl without affecting the characteristics of CH2 and CH3.
- FIG. 13 shows a single-channel embodiment of a memory subsystem 190 that is a modification to the two-channel embodiment shown in FIG. 1 IA.
- the configurable memory subsystem 190 has twice the memory capacity, twice the ring delay and half the number of channels of the memory subsystem 10 shown in FIG. 1 IA.
- the previously described modifications to the memory module 12 are applicable to the configurable memory subsystem 190.
- FIG. 14 shows a reversible memory subsystem 200 using the reversible memory module 90 shown in FIG. 10.
- a full memory subsystem 200 can be implemented with only one type of memory module 90 without the need for a special termination module 112 or a memory module 69 as shown in FIGs. 12B and 12C, respectively.
- the memory module 90 is reversed as shown by memory module 90b to complete the loop-back of each channel.
- the channel CHO receives a packet on bus 42 from the controller 40 at the pair input 102 of the MCD pair 100 on the memory module 90.
- the output port 104 of the MCD pair 100 on the memory module 90 transmits a packet on bus 44 to the input port 102 of the MCD pair 100 on the memory module 90b.
- the MCD pair 100 on the memory module 90b transmits a packet from the output port 104 to the input port 102 of another MCD pair 100 on the memory module 90b.
- the output port 104 of the other MCD pair 100 on the memory module 90b transmits a memory packet on bus 48 to the input port 102 of another MCD pair 100 on the memory module 90.
- the output port 104 of the other MCD pair 100 on the memory module 90 transmits a packet on bus 50 to the controller 40.
- the second channel CHl follows a similar path to that described for CHO.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Mathematical Physics (AREA)
- Computer Hardware Design (AREA)
- Dram (AREA)
- Logic Circuits (AREA)
- Techniques For Improving Reliability Of Storages (AREA)
- Memory System (AREA)
Abstract
Description
Claims
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201080002916.0A CN102187396B (en) | 2009-05-22 | 2010-05-20 | Configurable module and memory subsystem |
CA2738350A CA2738350A1 (en) | 2009-05-22 | 2010-05-20 | Configurable module and memory subsystem |
DE112010002059T DE112010002059T5 (en) | 2009-05-22 | 2010-05-20 | Configurable module and memory subsystem |
EP10777275A EP2443629A4 (en) | 2009-05-22 | 2010-05-20 | Configurable module and memory subsystem |
JP2012511109A JP5681704B2 (en) | 2009-05-22 | 2010-05-20 | Configurable module and memory subsystem |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US18069309P | 2009-05-22 | 2009-05-22 | |
US61/180,693 | 2009-05-22 | ||
US12/770,376 | 2010-04-29 | ||
US12/770,376 US8503211B2 (en) | 2009-05-22 | 2010-04-29 | Configurable module and memory subsystem |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2010132995A1 true WO2010132995A1 (en) | 2010-11-25 |
Family
ID=43124444
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CA2010/000769 WO2010132995A1 (en) | 2009-05-22 | 2010-05-20 | Configurable module and memory subsystem |
Country Status (8)
Country | Link |
---|---|
US (2) | US8503211B2 (en) |
EP (1) | EP2443629A4 (en) |
JP (1) | JP5681704B2 (en) |
KR (1) | KR20120030328A (en) |
CN (1) | CN102187396B (en) |
CA (1) | CA2738350A1 (en) |
DE (1) | DE112010002059T5 (en) |
WO (1) | WO2010132995A1 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014509025A (en) * | 2011-03-11 | 2014-04-10 | マイクロン テクノロジー, インク. | System, apparatus, memory controller, and method for memory initialization |
Families Citing this family (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20100138428A (en) * | 2009-06-25 | 2010-12-31 | 삼성전자주식회사 | Memory module for preventing esd and system including the same |
US9204550B2 (en) | 2011-09-30 | 2015-12-01 | Smart Modular Technologies, Inc. | Extended capacity memory system with load relieved memory and method of manufacture thereof |
TWI581267B (en) * | 2011-11-02 | 2017-05-01 | 諾瓦晶片加拿大公司 | Flash memory module and memory subsystem |
US9128662B2 (en) * | 2011-12-23 | 2015-09-08 | Novachips Canada Inc. | Solid state drive memory system |
US8994425B2 (en) * | 2012-08-03 | 2015-03-31 | Altera Corporation | Techniques for aligning and reducing skew in serial data signals |
KR102104917B1 (en) * | 2013-02-04 | 2020-04-27 | 삼성전자주식회사 | Semiconductor package |
US9588937B2 (en) | 2013-02-28 | 2017-03-07 | International Business Machines Corporation | Array of processor core circuits with reversible tiers |
US9368489B1 (en) | 2013-02-28 | 2016-06-14 | International Business Machines Corporation | Interconnect circuits at three-dimensional (3-D) bonding interfaces of a processor array |
KR102365111B1 (en) | 2014-07-07 | 2022-02-18 | 삼성전자주식회사 | Memory module set, semiconductor memory device and semiconductor memory system including the same |
US10184220B1 (en) * | 2015-12-08 | 2019-01-22 | James R McGilvery | Boat fender holder |
GB2560904B (en) * | 2017-03-27 | 2020-06-10 | Ge Aviat Systems Ltd | Processor performance monitor |
US10762006B2 (en) * | 2017-03-31 | 2020-09-01 | Intel Corporation | Techniques to dynamically enable memory channels on a compute platform |
US11004475B2 (en) | 2018-03-28 | 2021-05-11 | Micron Technology, Inc. | Methods and apparatuses for aligning read data in a stacked semiconductor device |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6144576A (en) * | 1998-08-19 | 2000-11-07 | Intel Corporation | Method and apparatus for implementing a serial memory architecture |
EP1422717A1 (en) * | 2002-11-21 | 2004-05-26 | Infineon Technologies AG | Memory system and memory subsystem |
US20040260859A1 (en) | 2000-06-09 | 2004-12-23 | Samsung Electronics Co., Ltd. | Memory module with improved data bus performance |
US20050268061A1 (en) * | 2004-05-31 | 2005-12-01 | Vogt Pete D | Memory channel with frame misalignment |
US20060095592A1 (en) * | 2004-10-29 | 2006-05-04 | International Business Machines Corporation | Multi-channel memory architecture for daisy chained arrangements of nodes with bridging between memory channels |
US20070124532A1 (en) * | 2005-04-21 | 2007-05-31 | Bennett Jon C | Interconnection system |
US20080028160A1 (en) | 2006-07-26 | 2008-01-31 | Gerald Keith Bartley | Carrier having daisy chain of self timed memory chips |
WO2008067652A1 (en) * | 2006-12-06 | 2008-06-12 | Mosaid Technologies Incorporated | Address assignment and type recognition of serially interconnected memory devices of mixed type |
US20080201548A1 (en) | 2007-02-16 | 2008-08-21 | Mosaid Technologies Incorporated | System having one or more memory devices |
US20090119451A1 (en) * | 2007-11-07 | 2009-05-07 | Christopher Haywood | Redriven/Retimed Registered Dual Inline Memory Module |
Family Cites Families (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP3757973B2 (en) * | 1996-06-07 | 2006-03-22 | 株式会社日立製作所 | Signal transmission device |
JP3455040B2 (en) * | 1996-12-16 | 2003-10-06 | 株式会社日立製作所 | Source clock synchronous memory system and memory unit |
US6061263A (en) * | 1998-12-29 | 2000-05-09 | Intel Corporation | Small outline rambus in-line memory module |
JP3820843B2 (en) * | 1999-05-12 | 2006-09-13 | 株式会社日立製作所 | Directional coupled memory module |
US6625687B1 (en) * | 2000-09-18 | 2003-09-23 | Intel Corporation | Memory module employing a junction circuit for point-to-point connection isolation, voltage translation, data synchronization, and multiplexing/demultiplexing |
JP4173970B2 (en) * | 2002-03-19 | 2008-10-29 | 株式会社ルネサステクノロジ | Memory system and memory module |
JP4094370B2 (en) * | 2002-07-31 | 2008-06-04 | エルピーダメモリ株式会社 | Memory module and memory system |
US7102907B2 (en) * | 2002-09-09 | 2006-09-05 | Micron Technology, Inc. | Wavelength division multiplexed memory module, memory system and method |
US7308524B2 (en) * | 2003-01-13 | 2007-12-11 | Silicon Pipe, Inc | Memory chain |
US7165153B2 (en) * | 2003-06-04 | 2007-01-16 | Intel Corporation | Memory channel with unidirectional links |
DE10345549B3 (en) * | 2003-09-30 | 2005-04-28 | Infineon Technologies Ag | Integrated memory circuit e.g. dynamic random-access memory, has interface system with different interface circuits used for providing several alternative operating modes |
US7212423B2 (en) * | 2004-05-31 | 2007-05-01 | Intel Corporation | Memory agent core clock aligned to lane |
US7539800B2 (en) * | 2004-07-30 | 2009-05-26 | International Business Machines Corporation | System, method and storage medium for providing segment level sparing |
US7224595B2 (en) * | 2004-07-30 | 2007-05-29 | International Business Machines Corporation | 276-Pin buffered memory module with enhanced fault tolerance |
JP4274140B2 (en) * | 2005-03-24 | 2009-06-03 | 日本電気株式会社 | Memory system with hot swap function and replacement method of faulty memory module |
US7353316B2 (en) * | 2006-03-24 | 2008-04-01 | Micron Technology, Inc. | System and method for re-routing signals between memory system components |
US7480201B2 (en) * | 2006-07-26 | 2009-01-20 | International Business Machines Corporation | Daisy chainable memory chip |
US7904639B2 (en) * | 2006-08-22 | 2011-03-08 | Mosaid Technologies Incorporated | Modular command structure for memory and memory system |
DE102006051514B4 (en) * | 2006-10-31 | 2010-01-21 | Qimonda Ag | Memory module and method for operating a memory module |
US7957173B2 (en) * | 2008-10-14 | 2011-06-07 | Mosaid Technologies Incorporated | Composite memory having a bridging device for connecting discrete memory devices to a system |
US8134852B2 (en) * | 2008-10-14 | 2012-03-13 | Mosaid Technologies Incorporated | Bridge device architecture for connecting discrete memory devices to a system |
-
2010
- 2010-04-29 US US12/770,376 patent/US8503211B2/en not_active Expired - Fee Related
- 2010-05-20 CN CN201080002916.0A patent/CN102187396B/en not_active Expired - Fee Related
- 2010-05-20 WO PCT/CA2010/000769 patent/WO2010132995A1/en active Application Filing
- 2010-05-20 JP JP2012511109A patent/JP5681704B2/en not_active Expired - Fee Related
- 2010-05-20 CA CA2738350A patent/CA2738350A1/en not_active Abandoned
- 2010-05-20 DE DE112010002059T patent/DE112010002059T5/en not_active Withdrawn
- 2010-05-20 KR KR1020117009822A patent/KR20120030328A/en not_active Application Discontinuation
- 2010-05-20 EP EP10777275A patent/EP2443629A4/en not_active Withdrawn
-
2013
- 2013-08-02 US US13/957,713 patent/US8767430B2/en not_active Expired - Fee Related
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6144576A (en) * | 1998-08-19 | 2000-11-07 | Intel Corporation | Method and apparatus for implementing a serial memory architecture |
US20040260859A1 (en) | 2000-06-09 | 2004-12-23 | Samsung Electronics Co., Ltd. | Memory module with improved data bus performance |
EP1422717A1 (en) * | 2002-11-21 | 2004-05-26 | Infineon Technologies AG | Memory system and memory subsystem |
US20050268061A1 (en) * | 2004-05-31 | 2005-12-01 | Vogt Pete D | Memory channel with frame misalignment |
US20060095592A1 (en) * | 2004-10-29 | 2006-05-04 | International Business Machines Corporation | Multi-channel memory architecture for daisy chained arrangements of nodes with bridging between memory channels |
US20070124532A1 (en) * | 2005-04-21 | 2007-05-31 | Bennett Jon C | Interconnection system |
US20080028160A1 (en) | 2006-07-26 | 2008-01-31 | Gerald Keith Bartley | Carrier having daisy chain of self timed memory chips |
WO2008067652A1 (en) * | 2006-12-06 | 2008-06-12 | Mosaid Technologies Incorporated | Address assignment and type recognition of serially interconnected memory devices of mixed type |
US20080201548A1 (en) | 2007-02-16 | 2008-08-21 | Mosaid Technologies Incorporated | System having one or more memory devices |
US20090119451A1 (en) * | 2007-11-07 | 2009-05-07 | Christopher Haywood | Redriven/Retimed Registered Dual Inline Memory Module |
Non-Patent Citations (1)
Title |
---|
See also references of EP2443629A4 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014509025A (en) * | 2011-03-11 | 2014-04-10 | マイクロン テクノロジー, インク. | System, apparatus, memory controller, and method for memory initialization |
US8856482B2 (en) | 2011-03-11 | 2014-10-07 | Micron Technology, Inc. | Systems, devices, memory controllers, and methods for memory initialization |
US9251068B2 (en) | 2011-03-11 | 2016-02-02 | Micron Technology, Inc. | Systems, devices, memory controllers, and methods for memory initialization |
Also Published As
Publication number | Publication date |
---|---|
US8767430B2 (en) | 2014-07-01 |
KR20120030328A (en) | 2012-03-28 |
JP5681704B2 (en) | 2015-03-11 |
DE112010002059T5 (en) | 2012-09-13 |
CA2738350A1 (en) | 2010-11-25 |
EP2443629A4 (en) | 2013-04-03 |
US20100296256A1 (en) | 2010-11-25 |
JP2012527661A (en) | 2012-11-08 |
US20130322173A1 (en) | 2013-12-05 |
CN102187396B (en) | 2015-05-27 |
EP2443629A1 (en) | 2012-04-25 |
US8503211B2 (en) | 2013-08-06 |
CN102187396A (en) | 2011-09-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8767430B2 (en) | Configurable module and memory subsystem | |
US11200181B2 (en) | Asymmetric-channel memory system | |
US7640386B2 (en) | Systems and methods for providing memory modules with multiple hub devices | |
US20190266112A1 (en) | High capacity memory system with improved command-address and chip-select signaling mode | |
JP5116760B2 (en) | System and method for providing a remote prefetch buffer | |
US7952944B2 (en) | System for providing on-die termination of a control signal bus | |
US20180090187A1 (en) | Memory module and system supporting parallel and serial access modes | |
US7710144B2 (en) | Controlling for variable impedance and voltage in a memory system | |
US7895374B2 (en) | Dynamic segment sparing and repair in a memory system | |
US8089813B2 (en) | Controllable voltage reference driver for a memory system | |
US10380056B2 (en) | Memory with alternative command interfaces | |
US11853600B2 (en) | Memory systems, modules, and methods for improved capacity | |
WO2012064670A1 (en) | Area-efficient multi-modal signaling interface | |
US10831963B1 (en) | Apparatus and method of parallel architecture for NVDIMM |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 201080002916.0 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 10777275 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 421/MUMNP/2011 Country of ref document: IN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2012511109 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2738350 Country of ref document: CA |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2010777275 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 20117009822 Country of ref document: KR Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 112010002059 Country of ref document: DE Ref document number: 1120100020590 Country of ref document: DE |