CA2109642A1 - Generalized shared memory in a cluster architecture for a computing system - Google Patents

Generalized shared memory in a cluster architecture for a computing system

Info

Publication number
CA2109642A1
CA2109642A1 CA002109642A CA2109642A CA2109642A1 CA 2109642 A1 CA2109642 A1 CA 2109642A1 CA 002109642 A CA002109642 A CA 002109642A CA 2109642 A CA2109642 A CA 2109642A CA 2109642 A1 CA2109642 A1 CA 2109642A1
Authority
CA
Canada
Prior art keywords
node
shared
cache
inter
local
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
Application number
CA002109642A
Other languages
French (fr)
Inventor
John C. Hunter
John A. Wertz
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Bull HN Information Systems Inc
Original Assignee
Bull HN Information Systems Inc
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Bull HN Information Systems Inc filed Critical Bull HN Information Systems Inc
Publication of CA2109642A1 publication Critical patent/CA2109642A1/en
Abandoned legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F12/00Accessing, addressing or allocating within memory systems or architectures
    • G06F12/02Addressing or allocation; Relocation
    • G06F12/08Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
    • G06F12/0802Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
    • G06F12/0806Multiuser, multiprocessor or multiprocessing cache systems
    • G06F12/0813Multiuser, multiprocessor or multiprocessing cache systems with a network or matrix configuration

Abstract

ABSTRACT OF THE DISCLOSURE
Cluster architectures are improved by introducing a Generalized Shared Memory, which is maintained in a consistent state by a hardware-based coherency mechanism that operates on shared objects, wherever they happen to be located.
This increases both the performance and the versatility of the architectures by permitting the composition of private vs. shared memory to be of arbitrary size and dynamically variable on different computer nodes in the cluster.

Description

.

I:X1'1~ ; M~lL NO. I13425548585 GENERALIZED S}IARED MEMORY IN A CLUSTER
2 ARCllITECTURE FOR A COMl'UTER SYSTEM
4 Field of the Invelltion This invel)tioll relales lo tlle dala processill~ arts and, more palliclllally, to a 6 genelalized shared memol-y for use in a cluster arcllitecture~
8 Back~round ofthe Invelltioll g Generally, a cluster computer arcllitectllle incllldes at least one, and typically o a plurality of, central processing Ullit(S) (CPU) and local memory, 1/0, etc. resident at each of a plurality of nodes. In one well re~arded cluster arcllitecture, pllysical 12 memory address space is pelmanently divided into two regions. tl~e lower half is 13 local memory (most si6nificant bit = "O"), whicll is accessible only by the 14 processors in that node, while the upper half (most significant bit = "1") defines `
shaled memory, wllicll is physically centralized and accessible by all nodes. There 16 iS no haldware SUppOlt to maintaill cohelellcy amoll~ copies of sllared variables 17 contained in difrerent nodes. Collerency is thlls leît entilely to software.
18 The Bull HN Shaled Buffer Arclliteclllre (SBA) expands UpOIl this 1~ arcllitectllre by illhodllcing a hardwale mecllallism that maintaills collerellcy among copies of data that ori~inated in tlle physically centralized sllared memory but are 21 con~nined in tlle cache memories of differellt nodes in the clustel. This improves 22 Ihe perfolmance of Ille archilechlle by pelmitting tlle differellt nodes to perfolm 23 some operations on shal ed data in parallel and with shol1el- access times.
24 A variation of SBA, tlle Bull HN Dishibuted Shared Buffer Architecture (DSBA) uses distribll~ed sllared memoly inslead of n cenbalized shared memoly, a26 configuratioll whicll has a number of advanlages in terllls of perfolmance and 27 availability as well as compatibilily willl some existing computer products. The address division between private and shared memory remains the saune, but the 2 shared memory is replicated at each computer, thlls permitting simultauleous read 3 access by all nodes. A haldware collerency mecllanism ensul-es that the data in the 4 shared memories remains collerellt.
s These variations of tlle knowll arcllitectule all employ a rigid partition 6 between private and shaled memoly. This chalacteristic makes it difficult, if not 7 impossible, to configure different sizes of private and shared memoly to meet the 8 needs of the application mix, both witllill and amollg nodes in the cluster. Private g memoly can nevel be used for shaled applications, and if shal-ed memory is used for o private purposes, it is subject to urulecessaly coherency traffic and potential integrity problems. The present invention overcomes this resb iction.

13 Objects of the lnvention 14 lt iS therefore a broad object of tl-is invention to achieve an improved cluster architecture. .
It is a more specific object of this invelltioll to provide a cluster arcllitectllle 17 employing a generalized shared memory incolporating disconti~uous shaled objects 8 whicll can be of any size an(l be shaled by an all)i~l aly nlllnl)er of no(les in a cluster.
19 It is anotller specific ol)jecl of this invelltiolI lo provide sucll a clusler architecture within wllich shared memoly is available vely quickly to all elements in .
21 the CIllSter.

23 SummalY of the Inventioll 24 Briefly, these and other objects of the invention are achieved in a cluster architectllre by inhoducill~ fl Gcne~ulized-SI1aredMemo)y, wllich iS Illailllail~ed ill a 26 consistent state by a hardware-based cohelellcy mechalIism that operates on shared 27 objects, whelever they l-appen to be locale(l. Tllis incleases bolll the perfolmance and ~lle versatility of the arcl-itectules by pelmitting the composition Or privflte ~s.
2 shared memory to be of arbih-aly size and dyllamically variable on difrerent 3 computer nodes in tl~e cluslel. ~;
4 ` -s Description of lhe Drawing 6 The subject matter of the inven~ioll is palticularly pointe(J out and distinctly 7 claimed in the concluding poltion of the specification. The invelltion, however, 8 botll as to organizatioll and melllod of operalion, may best be undelstood by g reference to tlle following description taken in conjunction with the subjoined o claims and tbe accompanying dlawillg of wllicl FIG. I is a high level block diagram repl esenting a prior alt cluster -2 architecture; i~
13 FIG. 2 is a high level block diagram representing a cluster architecture14 accordin8 to the present invention;
s FIG. 3 is a more detailed block diagram palticularly showing an intermediate 6 block diagram of an Exten~al C:ollerency Unit component of the cluster architecture ~-according to the present invention; ~ -8 FIG. 4 is a plimaly cache stale bansi~ioll diagram of a specirlc computer ~-19 family chosen as an example;
FIG. 5 is an Extemal Cohelency Unil stale bransition dia~ram as it can be 21 incoll)olated into tlle salne specific complltel family; flnd 22 FIG. 6 illustrates all exempla~y exlen~al col~erency Ullit associative memory 23 enhy.

Description of the Preferred Embodiment~
First, consider the high level block diagram shown in FIG. I which illustrates 3 a cluster architecture which may be deemed state-of-the-art. In this exemplary ~-4 system, sixteen central processing units (CPUs) I are configured into four nodes of s four CPUs each. Each CPU has the property of coherence in its communications 6 with other system components and typically incorporates a primary cache 2, and . : .
7 each node includes a local shared memory 3 which communicates through the 8 primary caches with the CPUs I in the node via a node-local communications g system such as a node-local bus 4. In addition, each node-local bus 4 is coupled to o a cluster shared memory 5. Each of the local shared memories 3 and the cluster 1l shared memory 5 may be the same size, say 2 gigabytes. The memory addressing 12 structure is such that a memory address having a "0" most significant bit (MSB) 13 issued from a given CPU addresses information stored in the local shared memory 14 of the corresponding node, and a memory address having a "1" MSB addresses ls information stored in the cluster shared memory. Inter-node communication is 16 carried out through VO units 12 and inter-node bus 17, and, if there is sufficient distance between nodes, a communications path 20 may be employed, all as well 8 known in the art.
19 ~his architecture enjoys the advantage that information can be exchanged 20 throughout the cluster via the cluster shared memory 5 by suitable manipulation.
21 However, there are also cer~ain drawbacks to this architecture; for example:
22 A) the division between private and shared memory is permanently fixed 23 which can bç very inefficient for a mix of those applications which, on the ~4 one hand, may work best with private memory and those, on the other hand, 2s which may work best with shared memory;
26 B) there is a single point of failure at the cluster shared memory;

C) each node must have its own operating system; the memory management 2 of all the nodes must agree on individual accesses to the cluster shared 3 memory; and 4 D) the memory mana~ement of all the nodes must agree on each individual s access to the cluster shared memory.
6 As previously discussed, there are variations to the architecture shown in 7 FIG. 1 which achieve incremental improvement in system performance. The Shared 8 Buffer Architecture (SBA) variation incorporates a hardware mechanism that g maintains coherency among copies of data that originated in the centralized shared o memory but are contained in the cache memories of different nodes in the cluster -This increases the performance of the architecture by permitting the different nodes 12 to perform some operations on shared data in parallel and with shorter access times.
13 The Distributed Shared Buffer Architecture (DSBA) uses distributed shared 14 memory instead of a centralized shared memory, a configuration which has a ls number of advantages in terms of performance and availability as well as 16 compatibility with existing computer products. The address division between private and shared memory remains the same, but the shared memory is replicated at 8 each computer, thus permitting simultaneous read access by all nodes. A hardware 19 coherency mechanism ensures that the data in the shared memories remains coherent.
21 The subject invention, however, represents a fiundamental departure in 22 cluster architecture. In order to understand the subject invention, the concept of a 23 Shared O~jec~ (SO,) which is an important aspect of the invention must first be 24 appreciated. An SO can be of any size and be shared by an arbitrary number of 25 nodes in a c!uster. The collection of SOs constitutes a Generalized Shared Memo~
26 which is maintained in a consistent state by a hardware-based coherency mechanism 27 that operates selectively on SOs, wherever they happen to be located.

~' ~''' ''''`''''`.'-The SO is defined as an object that has the capability to be shared. It is known to all nodes of a cluster. A descriptor defines the extent of the SO and the 3 processes that are permitted to share access. An SO can be realized in a variety of 4 ways, depending on the target operating system: as a file (e.g~, Bull HN's GCOS~) 8 s operating system), or a stream (e.g., the UNIX(~) operating system). Tl-e SO is 6 separately instantiated in the virtual space of each node to ma~;e it eligible to be 7 shared by processes on different nodes. An SO can then be independently 8 physically instantiated (e.g., on demand) in the memories of different nodes of a g distributed shared memory cluster. There can be multiple physical instantiations of a shared object within a centralized shared cluster memory; for example, one which Il can support pages shared arnong UNIX processes. An SO has:
12 A) a unique name with arbitrary length representation;
13 B) a unique identifier (WID) with fixed length representation;
14 C) addressable intemal elements; and D) reference permissions.
16 The DSBA is an environrnent which offers the simplest way to understand 17 the real-time operation of SOs. Referring to FIG. 2, each node in the cluster (only 18 two nodes of two CPUs I each are shown for simplicity) contains an Extemal 19 Coherency Unit (ECU) 10 that: (a) snoops its own node-local bus 4 for cornrnands on shared cache-lines that are also present in other nodes, and (b) conditionally 21 forwards these cornrnands (using a unique identifier for the cache-line) to the other 22 nodes via, for example, a separate inter-node ECU bus 11. (Point-to-point 23 connections are also possible, usinL~ a directory in each ECU to keep track of which 24 nodes have copies.) All other ECUs: (a) snoop the inter-node ECU bus 11 for commands that affect cache-lines resident in their own local memories, (b) translate 26 these to their own physical tags and (c) inject the translated commands into their 27 own node-local buses. In this conceptual example, the ECUs 10 could use ordinary , . ., .,, . ~:

~ 21096~2 coherence units for implementation and cache-ta~ directories witll a unique 2 identifier and the coherence state (e.g., modified, exclusive, shared, invalid) for 3 each shared cache-line that is presently instantiated in that node.
In exemplary operation, if a CPU in one node in a cluster attempts to alter s the contents of a shared cache-line, that CPU must obtain exclusive ownership 6 before proceeding. The command is snooped by the local ECU 10, and if this node 7 does not have exclusive ownership, then the command is transmitted to the ECUs in 8 the other nodes. Each node with a valid unmodified copy invalidates all local 9 copies (e.g., in its primary caches 2 and/or local main memory 3). If a node o contains the cache-line in a modified state, it is siphoned to the requesting node and invalidated in the originating node. The cache-line is then owned exclusively and 2 can be updated. The collection of coherency operations guarantees that any process 3 executing on any node that references a shared object will receive the most recent 4 data.
With this architecture, V0 12 operates correctly to and from local physical 6 memory that has been assigned to contain all or a portion of an SO. V0 reads from 7 shared memory will automatically siphon the most recent data from the shared 8 memories of other nodes. When V0 writes into shared memory, exclusive 19 ownership at that node will be obtained automatically as data alTives, and after the 20 V0 is complete, when the data is referenced by a processor in another node, it will 21 be automatically siphoned to that memory. The benefits resulting from generalized 22 shared memory are as follows~
~ -23 More Efficient Use of Physical Memory Within and Among Nodes:
24 (1) the physical instantiations of plivate and shared memory regions -2s can be discontiguous and of any size as long as their sum does not 26 exceed physical memory size;

2109642 ~
., , (2) different nodes can have different private/shared compositions at 2 any given time; and ~ -3 (3) total memory can exceed the physical addressing capability of a 4 singole operating system/platform. ~ ~ ~
s Simpler Cluster Software: ~-6 (1) shared memory is managed independently by each node and does ;~
7 not require a global memory manager or cooperation among nodes to 8 manage shared regions.
g Improved Availability:
~o (I) no single point of failure, as with the centralized shared memory of SBA; and 12 (2) when a node malfunctions, the private memory regions can be 13 made accessible to another node for recovery by forcing private 14 memory to be sharable.
Is Less Coherency Traffic:
16 ` (1) inter-node coherency traffic is generated only when non-17 exclusively owned shared objects are written or when shared objects are referenced that are not present; and 19 (2) coherency traffic is excluded from nodes that do not contain copies of shared objects.
21 Finally, the performance of dislrib2~ted generalized shared memory can 22 exceed that of one centrali7ed because the mechanism proposed herein perrnits23 shared data to receive the speed benefits of caching while remaining a traditional 24 addressable memory. That is, it pennits separate nodes to simultaneously read their 25 own copies of the same shared variable as fast as each local bus can perform. The 26 centralized shared memory in Abrojo and SBA becomes a bottleneck if multiple `
2~ reads are issued at the same time.

.- ~:,''"'"''''..''' ~ .

There are two basic approaches which can be taken to detect the "shared"
2 capability; viz.: the "hardware scenario" and the "software scenario".
3 The Hardware Scenario: A Page Table Descriptor Word (PTW) carries a bit to 4 define a page as shared. The PTW is present in the CPU when a shared re~ion s is referenced and it places the "shared" bit on the node-local bus to signal tl-e ;
6 local ECU. This permits the ECU to operate as a cache regarding the physical 7 tags for shared pages with only the most often accessed tags actually retained in 8 the ECU cache; less often referenced tags are retrieved from memory by the g ECU.
o Advantages: Simple, fast, low cost, and software independent. -~ Disadvantage: Requires a change in CPU and bus hardware.
12 The Software Scenario: Every PTW (actually an ECU Entry) for currently 13 instantiated shared pages are containe~ in a buffer memory in the ECU which 14 then detects shared cache-lines by snooping (or by using a directory with direct connection of the ECUs). The memory manager sofnvare updates the PTW
6 buffer in the ECU each time a shared page is instantiated or eliminated.Advantages: No change required in central system hardware. No extra bit 8 required in PTW. Potentially implen.entable on existing harchvare.
19 Disadvantage: Requires a change in memoly manager software and increases hardware cost.
21 Attention is now directed to the more detailed exemplary block diagram of 22 FIG. 3. The ECU 10 contains an associative directory 13 whose function is to 23 continuously snoop the addresses on both the node-local bus 4 and the inter-node 24 ECU bus 11. Assume that this directory contains an ECU Entry for each shared 2s page that is currently physically instantiated in its local memory 3. Referring ~6 briefly to FIG. 6, the ECU Entry contains the local physical address of the pa~e, a ^~7 unique identifier and a set of coherency state bits (~ - 3 bits) for each cache-line in , :: -:

the page (e.~., there are 64 contiguous 64-byte cache-lines in a 4096 byte page).
2 The total size an ECU Entry should be between 24 and 32 bytes per shared page.
3 Referring again to FIG. 3, the ECU 10 ignores commands for non-shared4 cache-line requests which proceed directly to address the identified memory in the non-shared physical space 14 of the local memory 3. However, when the ECU
6 identifies a comrnand for a shared cache-line, it examines its coherence state ~ -7 (modified, exclwsive, shared, or invalid), which is stored in the ECU directory 16, to s determine if any inter-node action is required to maintain coherence. If inter-node g action is required, the physical address is trans1ated to the appropriate unique o identifier and transmitted, together with the appropriate comrnand over the ECU bus to other nodes in the cluster.
12 For example, if a CPU wishes to obtain an exclusive copy of a shared cache-13 line, it places a suitable command (e~g., RTW - "read with intent to write" - in the 14 exemplary system to be explained more fully below) on its node-local bus which will cause each ECU 10 to take the following set of possible actions:
6 (1) if the cache-line state is exclusive or modified, it will be retrieved directly from local memory 3 and sent to the requesting processor with no remote 8 action taken because no other copies exist;
19 ('~) if the state is invalid, an RTW command will be transmitted over the inter-node ECU bus 11 to other nodes. A remote node that contains the line 21 in exclusive or modified state will transmit (siphon) the cache line over the 22 inter-node ECU bus to the requester. If several nodes contain the line in the 23 shared state, all attempt to send it, but ECU bus conventional priority logic 24 will choose one and cancel the others. All remote copies will be set to the invalid state. `

-::` 2109642 (3) if the state is shared> the INV cornmand will be sent to other nodes, 2 which cause them to set their states to invalid (siphonin~ is not necessaJy 3 because a current copy already exists in the requesting node).
4 There are other inter-node actions that are required to maintain coherency s with different combinations of inputs from buses and cache-line states; these are 6 sun~narized below. This coherency procedure, which operates at the hardware 7 level, guarantees than any process executing in any node that references a shared 8 page will see the most recent data for the whole page.
g Paee In Scenario o Economy dictates that shared pages be physically instantiated only in nodes that actually reference them - for example, at the time of the occurrence of a ~2 reference that causes a page fault (i.e., invalid PTW). In conventional systems, the 13 memory manager instantiates the physical page and PTW and then causes an V0 to 14 bring in the content of the page; e.g., via conventional VO communications channel s 17. ln a distributed memory system, however, the page contents might already exist 6 in other nodes. Therefore, the memory manager must, in addition to normal page7 instantiation services, force a search for the page content by (1) storing an ECU
8 Page Entry in the ECU with all cache-lines marl;ed invalid, and (2) attempting19 another reference to the location that caused the page fault - this time with a valid 20 PTW. This will result in either (1) a siphon of the requested cache-line from21 another node into the page frame in memory or (2) determination that the page does 22 not exist in the shared memory of any node. In case (2), the memory management 23 software must determine which node has V0 access to the memory (e.g., disk) that 24 contains the page and ~en request (e.g., via special interrupt) that it perform V0 to 2s bring in the page from disk to its own memory. After the V0 is complete, the 26 original software can be restarted at the point where the reference occurred and the 27 ECUs will be able to complete the reference by siphoning.

Pa~e-Out Scenario 2 It is desirable to permit individual nodes to make their own pa~e replacement 3 decisions independently of other nodes. Thus, a shared page that has been 4 instantiated in several nodes should be able to be removed from local memory in those nodes where, for example, it is seldom referenced. A modified shared pa~e 6 that has not been referenced for some time can evicted by signaling (e.g., via special 7 interrupt) the home node to write it to disl; 18 (FIG. 3). Memo~y mana6ement 8 software in the home node must physically instantiate the page if it is not present 9 and then initiate V0 which will gather the most recent complete copy of the page o from other nodes as it references each cache line in the page. The coherency states of cache-lines in remote nodes may be left intact or alternatively may be set to the 12 invalid state in the remote nodes. The latter choice ma~es them more quickly 13 eligible for deletion in those nodes where they are seldom referenced. (The PTW
14 might indicate a modified page, but memory management softu~are can safely s discard it if the ECU can be queried to determine that every cache-line is invalid.) 16 Nodes that subsequently reference the page would receive data from the copy in the 7 home node by siphoning.
8 Attention is now directed to FlGs. 4 and 5 which are state diagrams for the 19 interfaces among a CPU 1, its plimary cache 2, the node-local bus 4, the ECU 10 and the inter-node ECU bus 11 to therefore define the ECU 10 as it may be 21 incorporated into a homogeneous GCOS 8 environment. In this environment, the 22 various terms have the following meanings~
23 Level I Bus is the coupling between a CPU I and its pri nary cache 2.
24 Level 2 Bus is the node-local bus 4.
Level 3 Bus is the inter-node ECU bus 11.
26 Data movement commands:
'7 Il - data received from Bus 1.

21096~2 01 - data output to Bus 1.
2 I2 - data received from Bus 2.
3 02 - data output to Bus 2.
4 I3 - data received from Bus 3.
03 - data output to Bus 3. - -6 Interface commands~
7 <RDl> - read into processor. ~ ~-8 <WRl> - write from processor.
9 cRARI> - read-alter-rewrite atomic operation. ~ ~ -o As to the Level I Bus (FIG. 5)~
<RD2> - request cache-line from Level 2 Bus.
2 <RTW2> - read cache-line with exclusivity. -13 <INV2> - order all Level I processes to invalidate cache-line.
14 <WR2> - write cache-line data on the Level 2 Bus.
s As to the Level 2 Bus (FIG. 6)~
6 <RD2> - request cache-line from Level 2 Bus.
17 ~RTW2> - read cache-line with exclusivity from other CPU.
lS <INV2> - invalidate cache-line in all caches on local Level 2 19 Bus.
<WR:2> - remove cache-line from primary cache or siphon.
21 <RD3>-requestcache-linefromLevel 3 Bus. ;
22 <RTW3> - read cache-line with exclusivity from other node.
23 <INV3> - order all ECUs to invalidate cache-line.
24 <WR3> - write cache-line to Level 3 Bus. -~
2s Those skilled in the art will understand that the invention is readily 26 applicable to hierarchial interconnections of any depth; i.e., there could be a Level 27 4, Level 5, etc. ;~

Thus, while the principles of the invention have now been made clear in an 2 illustrative embodiment, there will be immediately obvious to those slcilled in the art 3 many modifications of structure, arrangements, proportions, the elements, materials, 4 and components, used in the practice of the invention which are particularly adapted s for specific environments and operating requirements without departing from those 6 principles.

' ~- ' ,, .,~ :~ '`' -. `... ' 1 4 : ~ ;

Claims (9)

1. A computer cluster architecture comprising:
A) a plurality of nodes;
B) at least one central processing unit resident at each said node, each said central processing unit having the property of coherency;
C) a node-local communications means resident at each said node; each said node-local communications means coupling all said primary caches resident in the same node with said node-local communications means;
D) a local main memory resident at each said node, each said local main memory having physical space assignable as shared physical space and non-shared physical space;
E) an external coherency unit resident at each said node, each said external coherency unit being coupled to the node-local communications means and to the local main memory resident in the same node with said external coherency unit;
F) inter-node communication means coupling all said external coherency units;
G) each said external coherency unit comprising:
1) monitoring means adapted to monitori both said inter-node communication means and said node-local communications means resident in the same node with said external coherency unit; and
2) coherency means adapted to:
a) respond to said monitoring means sensing a cache-line request appearing on said node-local communications means in the same node with said external coherency unit and determining that such cache-line request is non-shared by directing said non-shared cache-line request to said non-shared physical space of said local main memory; and b) respond to said monitoring means sensing a cache-line request appearing on said node-local communications means in the same node with said external coherency unit and determining that such cache-line request is shared by examining its coherence state to further determine if inter-node action is required to service the request and, if such inter-node action is required, transmitting a unique identifier and a coherency command over said inter-node communication means to all other said external coherency units;
whereby, the most recent instantiation of the requested data available in the shared physical spaces among all said local memories in the cluster are provided to the one of said external coherency units requesting the same.

The computer cluster architecture of Claim 1 in which each said external coherency unit assigns a coherency state to each unit of information stored in the shared memory spaces of said cluster, said coherency states comprising:
A) exclusive indicating that the copy of the requested information present in the shared memory space of said local main memory resident in the same node as said external coherency unit is the only copy extant in the cluster;
B) modified indicating that the copy of the requested information present in thelocal main memory resident in the same node as said external coherency unit has been updated by a central processing unit in said same node;
C) invalid indicating that the copy of the requested information present in the local main memory resident in the same node as said external coherency unit either does not exist or is known to be out-of-date; and D) shared indicating that the copy of the requested information present in the local main memory resident in the same node as said external coherency unit is one of a plurality of current copies of the requested information in a plurality of nodes.
3. The computer cluster architecture of Claim 2 in which each said coherency means in each said external coherency unit responds to a local cache-line request associated with a block of shared information as follows:
A) if the requested cache-line state is exclusive or modified, the requested block is retrieved directly from the local main memory resident in the requesting node with no remote action taken because no current copies exist elsewhere in the cluster;
B) if the requested cache-line state is invalid, a read-with-intent-to-write command is transmitted over said communications means to all other said nodes, and a remote node that contains the requested cache-line information in the exclusive or modified state responds thereto by siphoning the requested cache-line information over said communications means to the requesting node; and C) if the requested cache-line state is shared, an invalidating command is sent to all other nodes such that the requested cache-line state is set to invalid insuch all other nodes.
4. The computer cluster architecture of Claim 1 in which said inter-node communications means comprises an inter-node external coherency unit bus.
5. The computer cluster architecture of Claim 2 in which said inter-node communications means comprises an inter-node external coherency unit bus.
6. The computer cluster architecture of Claim 3 in which said inter-node communications means comprises an inter-node external coherency unit bus.
7. The computer cluster architecture of Claim 1 in which said inter-node communications means comprises direct coupling means.
8. The computer cluster architecture of Claim 2 in which said inter-node communications means comprises direct coupling means.
9. The computer cluster architecture of Claim 3 in which said inter-node communications means comprises direct coupling means.
CA002109642A 1992-12-23 1993-11-22 Generalized shared memory in a cluster architecture for a computing system Abandoned CA2109642A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US07/993,884 US5394555A (en) 1992-12-23 1992-12-23 Multi-node cluster computer system incorporating an external coherency unit at each node to insure integrity of information stored in a shared, distributed memory
US07/993,884 1992-12-23

Publications (1)

Publication Number Publication Date
CA2109642A1 true CA2109642A1 (en) 1994-06-24

Family

ID=25540038

Family Applications (1)

Application Number Title Priority Date Filing Date
CA002109642A Abandoned CA2109642A1 (en) 1992-12-23 1993-11-22 Generalized shared memory in a cluster architecture for a computing system

Country Status (5)

Country Link
US (1) US5394555A (en)
EP (1) EP0603801A3 (en)
JP (1) JPH06243035A (en)
AU (1) AU675982B2 (en)
CA (1) CA2109642A1 (en)

Families Citing this family (220)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5535116A (en) * 1993-05-18 1996-07-09 Stanford University Flat cache-only multi-processor architectures
US5761739A (en) * 1993-06-08 1998-06-02 International Business Machines Corporation Methods and systems for creating a storage dump within a coupling facility of a multisystem enviroment
FR2707778B1 (en) * 1993-07-15 1995-08-18 Bull Sa Processor node.
FR2707776B1 (en) 1993-07-15 1995-08-18 Bull Sa Method for managing memories of a computer system, memory computer system and recording medium implementing the method.
US5604882A (en) * 1993-08-27 1997-02-18 International Business Machines Corporation System and method for empty notification from peer cache units to global storage control unit in a multiprocessor data processing system
US5636365A (en) * 1993-10-05 1997-06-03 Nec Corporation Hierarchical buffer memories for selectively controlling data coherence including coherence control request means
US5522057A (en) * 1993-10-25 1996-05-28 Intel Corporation Hybrid write back/write through cache having a streamlined four state cache coherency protocol for uniprocessor computer systems
JPH07129468A (en) * 1993-11-04 1995-05-19 Matsushita Electric Ind Co Ltd Main memory controller
US5577204A (en) * 1993-12-15 1996-11-19 Convex Computer Corporation Parallel processing computer system interconnections utilizing unidirectional communication links with separate request and response lines for direct communication or using a crossbar switching device
US5751995A (en) * 1994-01-04 1998-05-12 Intel Corporation Apparatus and method of maintaining processor ordering in a multiprocessor system which includes one or more processors that execute instructions speculatively
JP3410535B2 (en) * 1994-01-20 2003-05-26 株式会社日立製作所 Parallel computer
JP3604176B2 (en) * 1994-09-14 2004-12-22 株式会社東芝 Virtual space management method and distributed processing system
JP3132749B2 (en) * 1994-12-05 2001-02-05 インターナショナル・ビジネス・マシーンズ・コーポレ−ション Multiprocessor data processing system
US5893147A (en) * 1994-12-22 1999-04-06 Intel Corporation Method and apparatus for distinguishing system memory data from alternative memory data in a shared cache memory
DE69621263T2 (en) * 1995-01-23 2002-12-19 Compaq Computer Corp Distributed data cache for multiprocessor system with cache
US5893921A (en) * 1995-02-10 1999-04-13 International Business Machines Corporation Method for maintaining memory coherency in a computer system having a cache utilizing snoop address injection during a read transaction by a dual memory bus controller
US5692149A (en) * 1995-03-16 1997-11-25 Samsung Electronics Co., Ltd. Block replacement method in cache only memory architecture multiprocessor
AU5386796A (en) * 1995-04-11 1996-10-30 Kinetech, Inc. Identifying data in a data processing system
US5860108A (en) * 1995-04-26 1999-01-12 Nec Corporation Method and clustered multi-processor system for controlling a clock phase for clusters
JP2786124B2 (en) * 1995-04-26 1998-08-13 甲府日本電気株式会社 Shared memory multiprocessor system
US5634110A (en) * 1995-05-05 1997-05-27 Silicon Graphics, Inc. Cache coherency using flexible directory bit vectors
US5623632A (en) * 1995-05-17 1997-04-22 International Business Machines Corporation System and method for improving multilevel cache performance in a multiprocessing system
US6049889A (en) * 1995-06-07 2000-04-11 Digital Equipment Corporation High performance recoverable communication method and apparatus for write-only networks
DE19524023B4 (en) * 1995-06-30 2004-02-05 Fujitsu Siemens Computers Gmbh Multiprocessor system with a very large number of microprocessors
FR2737031B1 (en) * 1995-07-21 1997-08-14 Bull Sa METHOD FOR TRANSFERRING DATA IN A MULTINODAL COMPUTER SYSTEM
US5758119A (en) * 1995-08-23 1998-05-26 International Business Machines Corp. System and method for indicating that a processor has prefetched data into a primary cache and not into a secondary cache
US5740399A (en) * 1995-08-23 1998-04-14 International Business Machines Corporation Modified L1/L2 cache inclusion for aggressive prefetch
US5765208A (en) * 1995-09-29 1998-06-09 Motorola, Inc. Method of speculatively executing store instructions prior to performing snoop operations
US5590301A (en) * 1995-10-06 1996-12-31 Bull Hn Information Systems Inc. Address transformation in a cluster computer system
US6728258B1 (en) * 1995-11-15 2004-04-27 Hitachi, Ltd. Multi-processor system and its network
US6122637A (en) * 1995-11-30 2000-09-19 Yohe; Thomas Patrick Apparatus and method for increased data access in an object oriented caching system
US6012085A (en) * 1995-11-30 2000-01-04 Stampede Technolgies, Inc. Apparatus and method for increased data access in a network file object oriented caching system
US5682514A (en) * 1995-11-30 1997-10-28 Stampede Technologies, Inc. Apparatus and method for increased data access in a network file oriented caching system
US5835943A (en) * 1995-11-30 1998-11-10 Stampede Technologies, Inc. Apparatus and method for increased data access in a network file oriented caching system
US6223260B1 (en) * 1996-01-25 2001-04-24 Unisys Corporation Multi-bus data processing system in which all data words in high level cache memories have any one of four states and all data words in low level cache memories have any one of three states
US5813034A (en) * 1996-01-25 1998-09-22 Unisys Corporation Method and circuitry for modifying data words in a multi-level distributed data processing system
WO1997030399A1 (en) 1996-02-20 1997-08-21 Intergraph Corporation High-availability super server
US5838631A (en) 1996-04-19 1998-11-17 Integrated Device Technology, Inc. Fully synchronous pipelined ram
US5920891A (en) * 1996-05-20 1999-07-06 Advanced Micro Devices, Inc. Architecture and method for controlling a cache memory
US5940870A (en) * 1996-05-21 1999-08-17 Industrial Technology Research Institute Address translation for shared-memory multiprocessor clustering
US6141692A (en) * 1996-07-01 2000-10-31 Sun Microsystems, Inc. Directory-based, shared-memory, scaleable multiprocessor computer system having deadlock-free transaction flow sans flow control protocol
US5873117A (en) 1996-07-01 1999-02-16 Sun Microsystems, Inc. Method and apparatus for a directory-less memory access protocol in a distributed shared memory computer system
US5864671A (en) * 1996-07-01 1999-01-26 Sun Microsystems, Inc. Hybrid memory access protocol for servicing memory access request by ascertaining whether the memory block is currently cached in determining which protocols to be used
US5937431A (en) * 1996-07-12 1999-08-10 Samsung Electronics Co., Ltd. Multi- node, multi-level cache- only memory architecture with relaxed inclusion
US7058696B1 (en) 1996-11-22 2006-06-06 Mangosoft Corporation Internet-based shared file service with native PC client access and semantics
US5909540A (en) * 1996-11-22 1999-06-01 Mangosoft Corporation System and method for providing highly available data storage using globally addressable memory
US6148377A (en) * 1996-11-22 2000-11-14 Mangosoft Corporation Shared memory computer networks
US6026474A (en) * 1996-11-22 2000-02-15 Mangosoft Corporation Shared client-side web caching using globally addressable memory
US6647393B1 (en) 1996-11-22 2003-11-11 Mangosoft Corporation Dynamic directory service
US5987506A (en) * 1996-11-22 1999-11-16 Mangosoft Corporation Remote access and geographically distributed computers in a globally addressable storage environment
US20060129627A1 (en) * 1996-11-22 2006-06-15 Mangosoft Corp. Internet-based shared file service with native PC client access and semantics and distributed version control
US6021261A (en) * 1996-12-05 2000-02-01 International Business Machines Corporation Method and system for testing a multiprocessor data processing system utilizing a plurality of event tracers
US6078997A (en) * 1996-12-09 2000-06-20 Intel Corporation Directory-based coherency system for maintaining coherency in a dual-ported memory system
US5848434A (en) * 1996-12-09 1998-12-08 Intel Corporation Method and apparatus for caching state information within a directory-based coherency memory system
US6324623B1 (en) 1997-05-30 2001-11-27 Oracle Corporation Computing system for implementing a shared cache
US6574720B1 (en) 1997-05-30 2003-06-03 Oracle International Corporation System for maintaining a buffer pool
US5946711A (en) * 1997-05-30 1999-08-31 Oracle Corporation System for locking data in a shared cache
US6078994A (en) * 1997-05-30 2000-06-20 Oracle Corporation System for maintaining a shared cache in a multi-threaded computer environment
US6044438A (en) * 1997-07-10 2000-03-28 International Business Machiness Corporation Memory controller for controlling memory accesses across networks in distributed shared memory processing systems
US6092155A (en) 1997-07-10 2000-07-18 International Business Machines Corporation Cache coherent network adapter for scalable shared memory processing systems
JPH1185659A (en) * 1997-09-03 1999-03-30 Hitachi Ltd Disk controller and storage device using the same
DE69715203T2 (en) 1997-10-10 2003-07-31 Bull Sa A data processing system with cc-NUMA (cache coherent, non-uniform memory access) architecture and cache memory contained in local memory for remote access
US6148375A (en) * 1998-02-13 2000-11-14 International Business Machines Corporation Hierarchical bus simple COMA architecture for shared memory multiprocessors having a bus directly interconnecting caches between nodes
US6115320A (en) 1998-02-23 2000-09-05 Integrated Device Technology, Inc. Separate byte control on fully synchronous pipelined SRAM
GB9806184D0 (en) * 1998-03-23 1998-05-20 Sgs Thomson Microelectronics A cache coherency mechanism
US6128690A (en) * 1998-03-24 2000-10-03 Compaq Computer Corporation System for remote memory allocation in a computer having a verification table contains information identifying remote computers which are authorized to allocate memory in said computer
US6341339B1 (en) * 1998-03-26 2002-01-22 Compaq Computer Corporation Apparatus and method for maintaining data coherence within a cluster of symmetric multiprocessors
JP3866426B2 (en) * 1998-11-05 2007-01-10 日本電気株式会社 Memory fault processing method in cluster computer and cluster computer
US6931430B1 (en) * 1998-05-13 2005-08-16 Thomas W. Lynch Maintaining coherency in a symbiotic computing system and method of operation thereof
US6338117B1 (en) * 1998-08-28 2002-01-08 International Business Machines Corporation System and method for coordinated hierarchical caching and cache replacement
JP2000076205A (en) * 1998-08-28 2000-03-14 Hitachi Ltd Multiprocessor
US7013305B2 (en) 2001-10-01 2006-03-14 International Business Machines Corporation Managing the state of coupling facility structures, detecting by one or more systems coupled to the coupling facility, the suspended state of the duplexed command, detecting being independent of message exchange
US6163801A (en) * 1998-10-30 2000-12-19 Advanced Micro Devices, Inc. Dynamic communication between computer processes
US6314491B1 (en) 1999-03-01 2001-11-06 International Business Machines Corporation Peer-to-peer cache moves in a multiprocessor data processing system
FR2792745B1 (en) * 1999-04-26 2001-06-15 Bull Sa MODULAR INTERCONNECTION ARCHITECTURE FOR EXPANDABLE MULTIPROCESSOR MACHINE, IMPLEMENTING A MULTI-LEVEL VIRTUAL BUS HIERARCHY AND THE SAME BASE BRICK FOR ALL LEVELS
US6446086B1 (en) 1999-06-30 2002-09-03 Computer Sciences Corporation System and method for logging transaction records in a computer system
US6952741B1 (en) 1999-06-30 2005-10-04 Computer Sciences Corporation System and method for synchronizing copies of data in a computer system
US7069406B2 (en) * 1999-07-02 2006-06-27 Integrated Device Technology, Inc. Double data rate synchronous SRAM with 100% bus utilization
US7340426B1 (en) 1999-07-30 2008-03-04 Computer Sciences Corporation Event-triggered transaction processing for electronic data interchange
US6970844B1 (en) 1999-08-27 2005-11-29 Computer Sciences Corporation Flow designer for establishing and maintaining assignment and strategy process maps
US6961708B1 (en) 1999-08-27 2005-11-01 Computer Sciences Corporation External interface for requesting data from remote systems in a generic fashion
US6339812B1 (en) 1999-09-30 2002-01-15 Silicon Graphics, Inc. Method and apparatus for handling invalidation requests to processors not present in a computer system
US6381681B1 (en) * 1999-09-30 2002-04-30 Silicon Graphics, Inc. System and method for shared memory protection in a multiprocessor computer
US7693731B1 (en) 1999-09-30 2010-04-06 Computer Sciences Corporation Business process framework for reinsurance
US7359863B1 (en) 1999-09-30 2008-04-15 Computer Sciences Corporation Condition component framework for reinsurance
US6651157B1 (en) * 1999-10-15 2003-11-18 Silicon Graphics, Inc. Multi-processor system and method of accessing data therein
US7363264B1 (en) 1999-10-29 2008-04-22 Computer Sciences Corporation Processing business transactions using dynamic database packageset switching
US7526487B1 (en) 1999-10-29 2009-04-28 Computer Sciences Corporation Business transaction processing systems and methods
US7693844B1 (en) 1999-10-29 2010-04-06 Computer Sciences Corporation Configuring processing relationships among entities of an organization
US6925468B1 (en) 1999-10-29 2005-08-02 Computer Sciences Corporation Configuring systems for generating business transaction reports using processing relationships among entities of an organization
US7356541B1 (en) 1999-10-29 2008-04-08 Computer Sciences Corporation Processing business data using user-configured keys
US7353196B1 (en) 1999-10-29 2008-04-01 Computer Sciences Corporation Configuring dynamic database packageset switching for use in processing business transactions
US7571171B1 (en) 1999-10-29 2009-08-04 Computer Sciences Corporation Smart trigger for use in processing business transactions
US7546304B1 (en) 1999-10-29 2009-06-09 Computer Sciences Corporation Configuring keys for use in processing business data
JP2001167077A (en) 1999-12-09 2001-06-22 Nec Kofu Ltd Data access method for network system, network system and recording medium
US7050453B1 (en) * 2000-02-17 2006-05-23 Apple Computer, Inc. Method and apparatus for ensuring compatibility on a high performance serial bus
US7509420B2 (en) * 2000-02-18 2009-03-24 Emc Corporation System and method for intelligent, globally distributed network storage
US7430515B1 (en) 2000-06-23 2008-09-30 Computer Sciences Corporation System and method for externalization of formulas for assessing damages
US7418400B1 (en) 2000-06-23 2008-08-26 Computer Sciences Corporation Internet-enabled system and method for assessing damages
US7095426B1 (en) 2000-06-23 2006-08-22 Computer Sciences Corporation Graphical user interface with a hide/show feature for a reference system in an insurance claims processing system
US7430514B1 (en) 2000-06-23 2008-09-30 Computer Sciences Corporation System and method for processing insurance claims using a table of contents
US7343307B1 (en) 2000-06-23 2008-03-11 Computer Sciences Corporation Dynamic help method and system for an insurance claims processing system
US7571107B1 (en) 2000-06-23 2009-08-04 Computer Sciences Corporation System and method for externalization of rules for assessing damages
US7398219B1 (en) 2000-06-23 2008-07-08 Computer Sciences Corporation System and method for displaying messages using a messages table
US8635410B1 (en) 2000-07-20 2014-01-21 Silicon Graphics International, Corp. System and method for removing data from processor caches in a distributed multi-processor computer system
US6799248B2 (en) 2000-09-11 2004-09-28 Emc Corporation Cache management system for a network data node having a cache memory manager for selectively using different cache management methods
US8000985B2 (en) * 2000-10-02 2011-08-16 Computer Sciences Corporation Computerized method and system of displaying a roadway configuration relating to an accident
DE10049498A1 (en) * 2000-10-06 2002-04-11 Philips Corp Intellectual Pty Digital home network with distributed software system having virtual memory device for management of all storage devices within network
US7096269B2 (en) * 2001-03-30 2006-08-22 Hitachi, Ltd. Path selection methods for storage based remote copy
US7398302B2 (en) * 2001-03-30 2008-07-08 Hitachi, Ltd. Remote copy with path selection and prioritization
US7963899B2 (en) * 2001-07-13 2011-06-21 The Proctor & Gamble Company Continuous in-line pleating apparatus and process
US20030023803A1 (en) * 2001-07-24 2003-01-30 Zatorski Richard A. Bus bridge circuit including audio logic and an addressable register for storing an address bit used when the audio logic accesses digital data, and method for initializing a chip set including the bus bridge circuit
JP4043225B2 (en) * 2001-12-10 2008-02-06 株式会社ルネサステクノロジ Cache apparatus and method
WO2003056500A1 (en) * 2001-12-24 2003-07-10 Digimarc Id Systems, Llc Covert variable information on id documents and methods of making same
EP1467834A4 (en) 2001-12-24 2005-04-06 Digimarc Id Systems Llc Laser etched security features for identification documents and methods of making same
US7694887B2 (en) * 2001-12-24 2010-04-13 L-1 Secure Credentialing, Inc. Optically variable personalized indicia for identification documents
KR100441712B1 (en) * 2001-12-29 2004-07-27 엘지전자 주식회사 Extensible Multi-processing System and Method of Replicating Memory thereof
WO2003088144A2 (en) 2002-04-09 2003-10-23 Digimarc Id Systems, Llc Image processing techniques for printing identification cards and documents
US7824029B2 (en) * 2002-05-10 2010-11-02 L-1 Secure Credentialing, Inc. Identification card printer-assembler for over the counter card issuing
US7281055B2 (en) * 2002-05-28 2007-10-09 Newisys, Inc. Routing mechanisms in systems having multiple multi-processor clusters
US7103636B2 (en) * 2002-05-28 2006-09-05 Newisys, Inc. Methods and apparatus for speculative probing of a remote cluster
US7702528B2 (en) * 2002-09-09 2010-04-20 Computer Sciences Corporation Computerized method and system for determining breach of duty in premises liability for an accident
US20040054557A1 (en) * 2002-09-09 2004-03-18 Stefan Wahlbin Computerized method and system for estimating premises liability for an accident
US20040054556A1 (en) * 2002-09-09 2004-03-18 Stephan Wahlbin Computerized method and system for determining causation in premises liability for an accident
US20040054558A1 (en) * 2002-09-09 2004-03-18 Stefan Wahlbin Computerized method and system for determining claimant status in premises liability for an accident
US7672860B2 (en) * 2002-09-09 2010-03-02 Computer Sciences Corporation Computerized method and system for determining the contribution of defenses to premises liability for an accident
US20040085357A1 (en) * 2002-10-31 2004-05-06 Childress Allen B. Method of generating a graphical display of a business rule and associated business rule elements
US7689442B2 (en) * 2002-10-31 2010-03-30 Computer Science Corporation Method of generating a graphical display of a business rule with a translation
US20040088199A1 (en) * 2002-10-31 2004-05-06 Childress Allen B. Method of forming a business rule
US20040088195A1 (en) * 2002-10-31 2004-05-06 Childress Allen B. Method of modifying a business rule
US7451148B2 (en) * 2002-10-31 2008-11-11 Computer Sciences Corporation Method of modifying a business rule while tracking the modifications
US7676387B2 (en) * 2002-10-31 2010-03-09 Computer Sciences Corporation Graphical display of business rules
US7296121B2 (en) * 2002-11-04 2007-11-13 Newisys, Inc. Reducing probe traffic in multiprocessor systems
US7346744B1 (en) 2002-11-04 2008-03-18 Newisys, Inc. Methods and apparatus for maintaining remote cluster state information
US7003633B2 (en) * 2002-11-04 2006-02-21 Newisys, Inc. Methods and apparatus for managing probe requests
US8185602B2 (en) * 2002-11-05 2012-05-22 Newisys, Inc. Transaction processing using multiple protocol engines in systems having multiple multi-processor clusters
US6934814B2 (en) * 2002-11-05 2005-08-23 Newisys, Inc. Cache coherence directory eviction mechanisms in multiprocessor systems which maintain transaction ordering
US6925536B2 (en) * 2002-11-05 2005-08-02 Newisys, Inc. Cache coherence directory eviction mechanisms for unmodified copies of memory lines in multiprocessor systems
US6920532B2 (en) * 2002-11-05 2005-07-19 Newisys, Inc. Cache coherence directory eviction mechanisms for modified copies of memory lines in multiprocessor systems
US7577755B2 (en) * 2002-11-19 2009-08-18 Newisys, Inc. Methods and apparatus for distributing system management signals
AU2003298731A1 (en) 2002-11-26 2004-06-18 Digimarc Id Systems Systems and methods for managing and detecting fraud in image databases used with identification documents
US7660725B2 (en) * 2002-11-27 2010-02-09 Computer Sciences Corporation Computerized method and system for estimating an effect on liability based on the stopping distance of vehicles
US7818187B2 (en) * 2002-11-27 2010-10-19 Computer Sciences Corporation Computerized method and system for estimating liability
US20040102984A1 (en) * 2002-11-27 2004-05-27 Stefan Wahlbin Computerized method and system for estimating liability using recorded vehicle data
US20040103005A1 (en) * 2002-11-27 2004-05-27 Stefan Wahlbin Computerized method and system for estimating monetary damages due to injuries in an accident from liability estimated using a computer system
US7725334B2 (en) * 2002-11-27 2010-05-25 Computer Sciences Corporation Computerized method and system for estimating liability for an accident using dynamic generation of questions
US7792690B2 (en) * 2002-11-27 2010-09-07 Computer Sciences Corporation Computerized method and system for estimating an effect on liability of the speed of vehicles in an accident and time and distance traveled by the vehicles
US7809586B2 (en) * 2002-11-27 2010-10-05 Computer Sciences Corporation Computerized method and system for estimating an effect on liability using a comparison of the actual speed of a vehicle in an accident and time and distance traveled by the vehicles in a merging vehicle accident
US7702529B2 (en) * 2002-11-27 2010-04-20 Computer Sciences Corporation Computerized method and system for estimating an effect on liability using claim data accessed from claim reporting software
US7895063B2 (en) * 2002-11-27 2011-02-22 Computer Sciences Corporation Computerized method and system for creating pre-configured claim reports including liability in an accident estimated using a computer system
US7805321B2 (en) * 2002-11-27 2010-09-28 Computer Sciences Corporation Computerized method and system for estimating liability for an accident from an investigation of the accident
US8751753B1 (en) 2003-04-09 2014-06-10 Guillermo J. Rozas Coherence de-coupling buffer
GB2416416B (en) * 2003-04-11 2006-11-22 Sun Microsystems Inc Multi-node computer system implementing global access state dependent transactions
US20050005075A1 (en) * 2003-04-11 2005-01-06 Sun Microsystems, Inc. Multi-node computer system employing multiple memory response states
US20050044174A1 (en) * 2003-04-11 2005-02-24 Sun Microsystems, Inc. Multi-node computer system where active devices selectively initiate certain transactions using remote-type address packets
EP1614064B1 (en) 2003-04-16 2010-12-08 L-1 Secure Credentialing, Inc. Three dimensional data storage
US20040215494A1 (en) * 2003-04-24 2004-10-28 Wahlbin Stefan L. Method and system for determining monetary amounts in an insurance processing system
US7334089B2 (en) * 2003-05-20 2008-02-19 Newisys, Inc. Methods and apparatus for providing cache state information
US7386626B2 (en) * 2003-06-23 2008-06-10 Newisys, Inc. Bandwidth, framing and error detection in communications between multi-processor clusters of multi-cluster computer systems
US7337279B2 (en) * 2003-06-27 2008-02-26 Newisys, Inc. Methods and apparatus for sending targeted probes
US7395347B2 (en) * 2003-08-05 2008-07-01 Newisys, Inc, Communication between and within multi-processor clusters of multi-cluster computer systems
US7895064B2 (en) * 2003-09-02 2011-02-22 Computer Sciences Corporation Graphical input display in an insurance processing system
US20050060205A1 (en) * 2003-09-02 2005-03-17 Woods Randall K. Systems and methods for a graphical input display in an insurance processing system
US20050108063A1 (en) * 2003-11-05 2005-05-19 Madill Robert P.Jr. Systems and methods for assessing the potential for fraud in business transactions
US20050192850A1 (en) * 2004-03-01 2005-09-01 Lorenz Scott K. Systems and methods for using data structure language in web services
US8898246B2 (en) * 2004-07-29 2014-11-25 Hewlett-Packard Development Company, L.P. Communication among partitioned devices
US7046560B2 (en) * 2004-09-02 2006-05-16 Micron Technology, Inc. Reduction of fusible links and associated circuitry on memory dies
US20060059021A1 (en) * 2004-09-15 2006-03-16 Jim Yulman Independent adjuster advisor
US20070168639A1 (en) * 2006-01-17 2007-07-19 Mccalpin John D Data processing system and method for selecting a scope of broadcast of an operation by reference to a translation table
US20060200469A1 (en) * 2005-03-02 2006-09-07 Lakshminarayanan Chidambaran Global session identifiers in a multi-node system
US8392668B2 (en) * 2005-04-04 2013-03-05 Hewlett-Packard Development Company, L.P. Distributed-state-information-based distributed computing systems and methods and protocols for managing distributed state information
US7478201B2 (en) * 2005-05-24 2009-01-13 International Business Machines Corporation Data processing system, cache system and method for passively scrubbing a domain indication
US7395376B2 (en) * 2005-07-19 2008-07-01 International Business Machines Corporation Method, apparatus, and computer program product for a cache coherency protocol state that predicts locations of shared memory blocks
US7360032B2 (en) * 2005-07-19 2008-04-15 International Business Machines Corporation Method, apparatus, and computer program product for a cache coherency protocol state that predicts locations of modified memory blocks
US7398360B2 (en) * 2005-08-17 2008-07-08 Sun Microsystems, Inc. Multi-socket symmetric multiprocessing (SMP) system for chip multi-threaded (CMT) processors
US8185576B2 (en) 2006-03-14 2012-05-22 Altnet, Inc. Filter for a distributed network
US8856782B2 (en) 2007-03-01 2014-10-07 George Mason Research Foundation, Inc. On-demand disposable virtual work system
US7975109B2 (en) 2007-05-30 2011-07-05 Schooner Information Technology, Inc. System including a fine-grained memory and a less-fine-grained memory
US8010390B2 (en) * 2007-06-04 2011-08-30 Computer Sciences Corporation Claims processing of information requirements
US8010391B2 (en) 2007-06-29 2011-08-30 Computer Sciences Corporation Claims processing hierarchy for insured
US8000986B2 (en) 2007-06-04 2011-08-16 Computer Sciences Corporation Claims processing hierarchy for designee
US8010389B2 (en) * 2007-06-04 2011-08-30 Computer Sciences Corporation Multiple policy claims processing
CN100489814C (en) * 2007-08-01 2009-05-20 杭州华三通信技术有限公司 Shared buffer store system and implementing method
US7991630B2 (en) * 2008-01-18 2011-08-02 Computer Sciences Corporation Displaying likelihood values for use in settlement
US8229945B2 (en) 2008-03-20 2012-07-24 Schooner Information Technology, Inc. Scalable database management software on a cluster of nodes using a shared-distributed flash memory
US8732386B2 (en) * 2008-03-20 2014-05-20 Sandisk Enterprise IP LLC. Sharing data fabric for coherent-distributed caching of multi-node shared-distributed flash memory
US9098698B2 (en) 2008-09-12 2015-08-04 George Mason Research Foundation, Inc. Methods and apparatus for application isolation
US8161247B2 (en) * 2009-06-26 2012-04-17 Microsoft Corporation Wait loss synchronization
US8812796B2 (en) 2009-06-26 2014-08-19 Microsoft Corporation Private memory regions and coherence optimizations
US8370577B2 (en) 2009-06-26 2013-02-05 Microsoft Corporation Metaphysically addressed cache metadata
US8356166B2 (en) * 2009-06-26 2013-01-15 Microsoft Corporation Minimizing code duplication in an unbounded transactional memory system by using mode agnostic transactional read and write barriers
US8250331B2 (en) 2009-06-26 2012-08-21 Microsoft Corporation Operating system virtual memory management for hardware transactional memory
US8489864B2 (en) * 2009-06-26 2013-07-16 Microsoft Corporation Performing escape actions in transactions
US8229907B2 (en) * 2009-06-30 2012-07-24 Microsoft Corporation Hardware accelerated transactional memory system with open nested transactions
US8839422B2 (en) 2009-06-30 2014-09-16 George Mason Research Foundation, Inc. Virtual browsing environment
US8539465B2 (en) 2009-12-15 2013-09-17 Microsoft Corporation Accelerating unbounded memory transactions using nested cache resident transactions
US8533440B2 (en) * 2009-12-15 2013-09-10 Microsoft Corporation Accelerating parallel transactions using cache resident transactions
US9092253B2 (en) * 2009-12-15 2015-07-28 Microsoft Technology Licensing, Llc Instrumentation of hardware assisted transactional memory system
US8402218B2 (en) 2009-12-15 2013-03-19 Microsoft Corporation Efficient garbage collection and exception handling in a hardware accelerated transactional memory system
US8856593B2 (en) 2010-04-12 2014-10-07 Sandisk Enterprise Ip Llc Failure recovery using consensus replication in a distributed flash memory system
US9164554B2 (en) 2010-04-12 2015-10-20 Sandisk Enterprise Ip Llc Non-volatile solid-state storage system supporting high bandwidth and random access
US8700842B2 (en) 2010-04-12 2014-04-15 Sandisk Enterprise Ip Llc Minimizing write operations to a flash memory-based object store
US9047351B2 (en) 2010-04-12 2015-06-02 Sandisk Enterprise Ip Llc Cluster of processing nodes with distributed global flash memory using commodity server technology
US8868487B2 (en) 2010-04-12 2014-10-21 Sandisk Enterprise Ip Llc Event processing in a flash memory-based object store
US8402106B2 (en) * 2010-04-14 2013-03-19 Red Hat, Inc. Asynchronous future based API
US8954385B2 (en) 2010-06-28 2015-02-10 Sandisk Enterprise Ip Llc Efficient recovery of transactional data stores
EP2630574A1 (en) * 2010-10-22 2013-08-28 ST-Ericsson SA Improving storage lifetime using data swapping
US8463762B2 (en) 2010-12-17 2013-06-11 Microsoft Corporation Volumes and file system in cluster shared volumes
US8732346B2 (en) 2010-12-17 2014-05-20 Microsoft Corporation Coordination of direct I/O with a filter
US8694733B2 (en) 2011-01-03 2014-04-08 Sandisk Enterprise Ip Llc Slave consistency in a synchronous replication environment
US8874515B2 (en) 2011-04-11 2014-10-28 Sandisk Enterprise Ip Llc Low level object version tracking using non-volatile memory write generations
US9652469B2 (en) 2011-06-04 2017-05-16 Microsoft Technology Licensing, Llc Clustered file service
US8627431B2 (en) 2011-06-04 2014-01-07 Microsoft Corporation Distributed network name
WO2013082437A1 (en) 2011-12-02 2013-06-06 Invincia, Inc. Methods and apparatus for control and detection of malicious content using a sandbox environment
US9135064B2 (en) 2012-03-07 2015-09-15 Sandisk Enterprise Ip Llc Fine grained adaptive throttling of background processes
US8918587B2 (en) * 2012-06-13 2014-12-23 International Business Machines Corporation Multilevel cache hierarchy for finding a cache line on a remote node
KR101905055B1 (en) * 2012-10-22 2018-10-08 인텔 코포레이션 High performance interconnect physical layer
US10949200B2 (en) 2013-06-16 2021-03-16 President And Fellows Of Harvard College Methods and apparatus for executing data-dependent threads in parallel
US9454485B2 (en) 2013-08-01 2016-09-27 Lenovo Enterprise Solutions (Singapore) Pte. Ltd. Sharing local cache from a failover node
US20180095906A1 (en) * 2016-09-30 2018-04-05 Intel Corporation Hardware-based shared data coherency

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5297269A (en) * 1990-04-26 1994-03-22 Digital Equipment Company Cache coherency protocol for multi processor computer system

Also Published As

Publication number Publication date
EP0603801A3 (en) 1994-10-05
EP0603801A2 (en) 1994-06-29
US5394555A (en) 1995-02-28
AU675982B2 (en) 1997-02-27
AU5244393A (en) 1994-07-07
JPH06243035A (en) 1994-09-02

Similar Documents

Publication Publication Date Title
US5394555A (en) Multi-node cluster computer system incorporating an external coherency unit at each node to insure integrity of information stored in a shared, distributed memory
US10073778B1 (en) Caching in multicore and multiprocessor architectures
US6105113A (en) System and method for maintaining translation look-aside buffer (TLB) consistency
JP2662603B2 (en) Method and apparatus for filtering invalidation requests
US5897664A (en) Multiprocessor system having mapping table in each node to map global physical addresses to local physical addresses of page copies
JP3722415B2 (en) Scalable shared memory multiprocessor computer system with repetitive chip structure with efficient bus mechanism and coherence control
EP0817059B1 (en) Auxiliary translation lookaside buffer for assisting in accessing data in remote address spaces
US5303362A (en) Coupled memory multiprocessor computer system including cache coherency management protocols
US5897656A (en) System and method for maintaining memory coherency in a computer system having multiple system buses
US6049847A (en) System and method for maintaining memory coherency in a computer system having multiple system buses
US5809536A (en) Method for reducing the number of coherency cycles within a directory-based cache coherency memory system uitilizing a memory state cache
JPH10143482A (en) Multiprocessor system for executing efficient write operation
JPH09185549A (en) System/method for hybrid numa/coma cache storage for selection between cache storage modes
JPH09185550A (en) System and method for hybrid numa coma cache storage for selection between cache storage modes
US6560681B1 (en) Split sparse directory for a distributed shared memory multiprocessor system
WO1995025306A2 (en) Distributed shared-cache for multi-processors
US20020078304A1 (en) System and method for allocating a directory entry for use in multiprocessor-node data processing systems
US5802563A (en) Efficient storage of data in computer system with multiple cache levels
EP1224553B1 (en) Multi-processor system and method of accessing data therein
US6138216A (en) Microprocessor cache consistency
JPH07152647A (en) Shared memory multiprocessor
US20050257010A1 (en) Method and system for migrating memory segments
JPH11316750A (en) Maintenance method for cache coherency and page allocation releasing method
Tamir et al. Hierarchical coherency management for shared virtual memory multicomputers
JPH0822418A (en) Virtual address space managing device

Legal Events

Date Code Title Description
FZDE Dead