公开(公告)号：CA2285847A1

公开(公告)日：2000-05-02

申请号：CA2285847

申请日：1999-10-13

Applicant: IBM

Inventor： GLASCO DAVID BRIAN ,  BAUMGARTNER YOANNA ,  CARPENTER GARY DALE ,  ELMAN ANNA ,  FIELDS JAMES STEPHEN ,  DEAN MARK EDWARD

IPC: G06F12/08 ,  G06F15/167

Abstract: A non-uniform memory access (NUMA) computer system includes a plurality of processing nodes coupled to a node interconnect. The plurality of processing nodes include at least a remote processing node, which contains a processor having an associated cache hierarchy, and a home processing node. The home processing node includes a shared system memory containing a plurality of memory granules and a coherence directory that indicates possible coherence states of copies of memory granules among the plurality of memory granules that are stored within at least one processing node other than the home processing node. If the processor within the remote processing node has a reservation for a memory granule among the plurality of memory granules that is not resident within the associated cache hierarchy, the coherence directory associates the memory granule with a coherence state indicating that the reserved memory granule may possibly be held non-exclusively at the remote processing node. In this manner, the coherence mechanism can be utilized to manage processor reservations even in cases in which a reserving processor's cache hierarchy does not hold a copy of the reserved memory granule.

公开(公告)号：CA2280125A1

公开(公告)日：2000-03-29

申请号：CA2280125

申请日：1999-08-12

Applicant: IBM

Inventor： IACHETTA RICHARD NICHOLAS JR ,  CARPENTER GARY DALE ,  DEAN MARK EDWARD ,  GLASCO DAVID BRIAN

IPC: G06F15/163 ,  G06F12/08 ,  G06F15/173

Abstract: A non-uniform memory access (NUMA) data processing system includes a node interconnect to which at least a first processing node and a second processing node are coupled. The first and the second processing nodes each include a local interconnect, a processor coupled to the local interconnect, a system memory coupled to the local interconnect, and a node controller interposed between the local interconnect and the node interconnect. In order to reduce communication latency, the node controller of the first processing node speculatively transmits request transactions received from the local interconnect of the first processing node to the second processing node via the node interconnect. In one embodiment, the node controller of the first processing node subsequently transmits a status signal to the node controller of the second processing node in order to indicate how the request transaction should be processed at the second processing node.

公开(公告)号：HK23796A

公开(公告)日：1996-02-16

申请号：HK23796

申请日：1996-02-08

Applicant: IBM

Inventor： BEGUN RALPH MURRAY ,  BLAND PATRICK MAURICE ,  DEAN MARK EDWARD

IPC: G06F12/08

Abstract: A microcomputer system includes a microprocessor, a cache memroy, and a cache controller all coupled to a local bus. The local bus is coupled to a sytem bus, connecting the remaining system components, through latches. When writing data, the microcomputer can perform a posted write to a unit on the system bus by writing the data into the latches and then, on receipt of a ready signal from the cache controller, continuing its operations without waiting for the data to pas to its destination. A problem arises if the data is posted to a unit with a data width less than that of the microprocessor. In this case, the data should be sent in multiple cycles, but the read signal is generated before the data width of the unit is known and the microprocessor then continues its operations and can not, therefore, transmit the data correctly. To solve this problem, a logic unit is added to monitor the ready signal and the output of a decoder which detects non-cacheable addresses (which are of data width different from the microprocessor data width). If a non-cacheable address is detected, the read signal from the cache controller through the logic unit is withheld from the microprocessor, which now waits to continue processing beyond the write until a ready signal is received from the addressed unit on the system bus.

公开(公告)号：IN176088B

公开(公告)日：1996-01-20

申请号：IN443DE1989

申请日：1989-05-19

Applicant: IBM

Inventor： BLEND PATRICK MAURICE ,  DEAN MARK EDWARD ,  BEGUN RALPH MURRAY

IPC: G06F7/00

公开(公告)号：DE68922784T2

公开(公告)日：1995-11-30

申请号：DE68922784

申请日：1989-03-03

Applicant: IBM

Inventor： BEGUN RALPH MURRAY ,  BLAND PATRICK MAURICE ,  DEAN MARK EDWARD

IPC: G06F9/44 ,  G06F9/445 ,  G06F13/36 ,  G06F13/362

Abstract: A multi-bus microcomputer system includes a cache subsystem and an arbitration supervisor. A CPU is provided with a PREEMPT signal source which generates a preempt signal in CPU cycles extending beyond a specified duration. The preempt signal is effective at any device having access to the bus to initiate an orderly termination of the bus usage. When that device signals its termination of bus usage, the arbitration supervisor changes the state of an ARB/GRANT conductor, which had been in the grant phase, to the arbitration phase. During the arbitration phase each of the devices (other than the CPU) cooperates in an arbitration mechanism for bus usage during the next grant phase. On the other hand, the CPU, having asserted preempt, responds to a signal indicating initiation of the arbitration phase by immediately accessing the system bus.

公开(公告)号：DE68924368D1

公开(公告)日：1995-11-02

申请号：DE68924368

申请日：1989-05-25

Applicant: IBM

Inventor： BEGUN RALPH MURRAY ,  BLAND PATRICK MAURICE ,  DEAN MARK EDWARD

IPC: G06F12/08

Abstract: In a dual bus microcomputer system using a cache memory and a cache controller, the timing requirements placed on non-cache memory components by the cache controller are more stringent than the timing requirements placed on the non-cache memory components by the microprocessor. A logic circuit operates on the cache write enable (CWE) signals, and delays those signals in the event of a cache read miss. Delaying the CWE signals relaxes the timing requirements placed on non-cache memory components and at the same time does not impact wait state parameters for read miss operations.

公开(公告)号：ES2075045T3

公开(公告)日：1995-10-01

申请号：ES89302136

申请日：1989-03-03

Applicant: IBM

Inventor： BLAND PATRICK MAURICE ,  DEAN MARK EDWARD

IPC: G11C11/401 ,  G06F13/36 ,  G06F13/362 ,  G11C7/10 ,  G11C7/00

Abstract: A computer system includes a page memory in which a row address accompanied by a row address strobe (RAS) is followed by a column address accompanied by a column address strobe (CAS) to read data from a memory location during a memory cycle. When, in a following memory cycle, a further location from the same page is to be accessed, the row address and the RAS remain constant and a new column address is used with the CAS being precharged by switching it to its OFF state and then returning it to its ON state. This is normally done at the start of the following memory cycle. In the present system, the data is read and latched shortly after arrival of the column address and CAS in the first of the memory cycles so that the CAS recharge can take place at the end of the first memory cycle and before the start of the following memory cycle.

公开(公告)号：ES2075044T3

公开(公告)日：1995-10-01

申请号：ES89302135

申请日：1989-03-03

Applicant: IBM

Inventor： BEGUN RALPH MURRAY ,  BLAND PATRICK MAURICE ,  DEAN MARK EDWARD

IPC: G06F12/08

Abstract: A microcomputer system includes a microprocessor, a cache memroy, and a cache controller all coupled to a local bus. The local bus is coupled to a sytem bus, connecting the remaining system components, through latches. When writing data, the microcomputer can perform a posted write to a unit on the system bus by writing the data into the latches and then, on receipt of a ready signal from the cache controller, continuing its operations without waiting for the data to pas to its destination. A problem arises if the data is posted to a unit with a data width less than that of the microprocessor. In this case, the data should be sent in multiple cycles, but the read signal is generated before the data width of the unit is known and the microprocessor then continues its operations and can not, therefore, transmit the data correctly. To solve this problem, a logic unit is added to monitor the ready signal and the output of a decoder which detects non-cacheable addresses (which are of data width different from the microprocessor data width). If a non-cacheable address is detected, the read signal from the cache controller through the logic unit is withheld from the microprocessor, which now waits to continue processing beyond the write until a ready signal is received from the addressed unit on the system bus.

公开(公告)号：FI95175B

公开(公告)日：1995-09-15

申请号：FI891787

申请日：1989-04-14

Applicant: IBM

Inventor： BEGUN RALPH MURRAY ,  BLAND PATRICK MAURICE ,  DEAN MARK EDWARD

IPC: G06F9/38 ,  G06F12/08

Abstract: In a microcomputer system comprising a microprocessor and a cache subsystem and operable in a pipelined mode, there is potential incompatibility between pipelined operations and dynamic bus sizing as the cache subsystem operates with a fixed size data width and dynamic bus sizing allows the system to operate with devices of differing data width. This incompatibility is accommodated by the present system by defining certain addresses as cacheable addresses and other addresses as non-cacheable addresses and ensuring that addresses of devices of data width different from the cache data width are non-cacheable. An address decoder provides a control signal indicating whether or not a generated address is within the cacheable range. This control signal controls a next address signal applied to the microprocessor, which signal allows the processor to proceed to a following cycle prior to the end of the current cycle. Whenever a non-cacheable address is detected, the next address signal is suppressed.

公开(公告)号：AT125058T

公开(公告)日：1995-07-15

申请号：AT89302136

申请日：1989-03-03

Applicant: IBM

Inventor： BLAND PATRICK MAURICE ,  DEAN MARK EDWARD

IPC: G11C11/401 ,  G06F13/36 ,  G06F13/362 ,  G11C7/10 ,  G11C7/00

Abstract: A computer system includes a page memory in which a row address accompanied by a row address strobe (RAS) is followed by a column address accompanied by a column address strobe (CAS) to read data from a memory location during a memory cycle. When, in a following memory cycle, a further location from the same page is to be accessed, the row address and the RAS remain constant and a new column address is used with the CAS being precharged by switching it to its OFF state and then returning it to its ON state. This is normally done at the start of the following memory cycle. In the present system, the data is read and latched shortly after arrival of the column address and CAS in the first of the memory cycles so that the CAS recharge can take place at the end of the first memory cycle and before the start of the following memory cycle.