Abstract: One embodiment provides a method of managing power in a computer system. A device of the computer system is operated at a selected power-state. The power consumption of the computer system is monitored. If the power consumption of the computer system is approaching or has exceeded a power cap selected for the computer system, then a request to reduce the power-state for the device is generated in response. The operating system is used to service the request to reduce the power-state according to the priority of the request. The reduced power state is forced out-of-band following the request to reduce the power-state if the request is not immediately serviceable by the operating system. Different approaches can be taken to force the reduced power state, using, for example, system management mode or a platform environment control interface.

Abstract: A method, computer program product and system for controlling the maximum turbo mode of a processor in a turbo boost state. The method comprises limiting a maximum turbo mode available to the processor by over-reporting the amount of current drawn by the processor to the current monitoring feedback line to the processor, wherein the processor uses the over-reported current to maintain operation of the processor within performance specifications of the processor. An automatic calibration routine may be used to determine nominal amounts of current over-reporting that may be used to prevent the processor performance from exceeding the maximum turbo mode. In one embodiment, a digital potentiometer is included in the voltage regulator circuit to over-report the current as instructed.

Abstract: A method identifies a plurality of PCI devices in a computer system by an associated PCI device handle, wherein each of the PCI devices is also associated with a default EFI device driver. The method further identifies a target PCI device to be disabled from within the plurality of PCI devices, provides a dummy driver that enables fewer functions for the target PCI device than would the default EFI device driver, and binds the dummy driver to the target PCI device instead of binding the default EFI device driver associated with the target PCI device. The dummy driver may be used to effectively disable the target PCI device so that the POST does not hang up or completes faster without loading the default EFI device driver.

Abstract: Methods, computers, and products for managing power consumption of a computer, the computer including a computer processor and managing power consumption of a computer includes: dynamically during operation of the computer, setting, by an in-band power manager in dependence upon performance metrics of the computer processor, a current performance state (‘p-state’) of the computer processor; and providing, by the in-band power manager to an out-of-band power manager, the current p-state of the computer processor.

Abstract: Methods, computers, and products for managing power consumption of a computer, the computer including a computer processor and managing power consumption of a computer includes: dynamically during operation of the computer, setting, by an in-band power manager in dependence upon performance metrics of the computer processor, a current performance state (‘p-state’) of the computer processor; and providing, by the in-band power manager to an out-of-band power manager, the current p-state of the computer processor.

Abstract: A method for self-contained testing within a DRAM comprises the DRAM receiving an instruction from an external processor to test a memory core on the DRAM, and the DRAM self-testing the memory core with one or more BIST pattern stored in a multipurpose register on the DRAM. Optionally, the step of self-testing may include writing the BIST pattern into all locations of the memory core, reading each location of the memory core, and comparing the content read from each location of the memory core with the BIST pattern, wherein a negative comparison indicates a failure has occurred. In a further option, the method may further comprise, after testing the DRAM, initializing the DRAM with an INIT pattern stored in the multipurpose register on the DRAM.

Abstract: Remotely administering a server, the server including non-volatile memory upon which is disposed one or more digital images representing the server, the server also including one or more components each of which includes non-volatile memory in which is disposed one or more digital images representing the component, where the server is connected for data communications to a management module, and remotely administering the server includes: retrieving, by the management module from the server, the digital images representing the server and the digital images representing the installed components; generating, by the management module with the digital images representing the server and the digital images representing the installed components, a graphical representation of the server with the installed components; and presenting, by the management module to a user through a GUI, the graphical representation of the server with the installed components.

Abstract: Memory is logically partitioned into two regions. A first region may be a similar size relative to the second region or the first region may be a small subset of the memory. The first region of memory is initialized and an operating system utilizes the first region. A system handler simulates a physical hot add of the second region. After the simulated physical hot add, the operating system may utilize the second region as if it were newly added physical memory and/or may utilize both the first region and second region.

Abstract: Remotely administering a server, the server including non-volatile memory upon which is disposed one or more digital images representing the server, the server also including one or more components each of which includes non-volatile memory in which is disposed one or more digital images representing the component, where the server is connected for data communications to a management module, and remotely administering the server includes: retrieving, by the management module from the server, the digital images representing the server and the digital images representing the installed components; generating, by the management module with the digital images representing the server and the digital images representing the installed components, a graphical representation of the server with the installed components; and presenting, by the management module to a user through a GUI, the graphical representation of the server with the installed components.

Abstract: A computer system that initializes a fraction of the computer system's memory for execution of video during booting of the computer system is provided. The computer system can include a first portion of BIOS code on a ROM device, wherein the first portion includes instructions for initializing the fraction. The computer system further can include a second portion of BIOS code that copies itself to the fraction upon completion of initialization of the fraction, wherein the second portion executes on the fraction and wherein the second portion initializes system memory and initializes a video buffer. The computer system further can include a copy of the second portion located on the ROM device, wherein the copy of the second portion executes until video buffer initialization is completed but before all of the system memory is initialized. Further, the video buffer displays video before all of the computer system's memory is initialized.

Abstract: Updating a BIOS boot block security module in a plurality of compute nodes of a multinode computer including, for each compute node in the multinode computer, upon powering on the compute node: determining whether a new image of security module is available; if a new image of the security module is available, commanding the reset control module of the compute node to block an attempt by the security module to propagate a reset signal on the scalability bus to other compute nodes in the multinode computer; updating to the new image of the security module; upon completion of the update, resetting the compute node including attempting to propagate a reset signal on the scalability bus to other compute nodes in the multinode computer; and blocking the attempt to propagate the reset signal on the scalability bus.

Abstract: Methods and products for managing power consumption of a computer and computers for which power consumption is managed. The computer includes the computer including a computer processor and embodiments of the present invention include providing, by an in-band power manger to an out-of-band power manager, a proposed performance state (‘p-state’) for the computer processor; determining, by the out-of-band power manager, in dependence upon a power setpoint and currently-measured operating metrics of the computer processor, whether to approve the proposed p-state; and if the out-of-band power manager approves the proposed p-state, setting operating parameters of the computer processor according to the approved p-state.

Abstract: A method, computer program product and system for controlling the maximum turbo mode of a processor in a turbo boost state. The method comprises limiting a maximum turbo mode available to the processor by over-reporting the amount of current drawn by the processor to the current monitoring feedback line to the processor, wherein the processor uses the over-reported current to maintain operation of the processor within performance specifications of the processor. An automatic calibration routine may be used to determine nominal amounts of current over-reporting that may be used to prevent the processor performance from exceeding the maximum turbo mode. In one embodiment, a digital potentiometer is included in the voltage regulator circuit to over-report the current as instructed.

Abstract: Memory is logically partitioned into two regions. A first region may be a similar size relative to the second region or the first region may be a small subset of the memory. The first region of memory is initialized and an operating system utilizes the first region. A system handler simulates a physical hot add of the second region. After the simulated physical hot add, the operating system may utilize the second region as if it were newly added physical memory and/or may utilize both the first region and second region.

Abstract: Detection of a reset failure in a multinode data processing system is provided by a diagnostic circuit in each of a plurality of the server nodes of the system. Each diagnostic circuit is coupled to a code fetch chain of its corresponding node. At reset, prior to a node processor retrieving startup code from the code fetch chain, the diagnostic circuit provides diagnostic signals to the code fetch chain. A problem in the code fetch chain is detected from a response to the diagnostic signals. When a problem is detected, a node failure status for the problem node may be signaled to the other nodes. The multinode system may be configured in response to signaled node failure status, such as by dropping failed nodes and replacing a failed primary node with a secondary node if necessary.

Abstract: This invention generally provides a method for speeding up system boot time, by initializing a subset of memory during the system firmware test/initialization, and allowing the system to boot an operating system with this subset of installed memory. While the system is completing the operating system boot with the subset of installed memory, a remainder of the installed system memory is being initialized/tested. When the initialization the remainder of system memory is completed (and after the OS has booted), the SMI handler is invoked. The SMI handler then simulates a physical memory “Hot Add” event, and reports the event to the OS. This allows much of the memory initialization/test activity to occur in parallel with the firmware initialization/test and operating system startup processes.

Abstract: A computing device includes a processor, firmware, a hardware component, and a debugging module. The firmware stores error decoding logic particular to the computing device. The hardware component detects an error in the computing device, and responsively issues an interrupt and halts the processor such that the processor cannot execute any more computer-readable code. The debugging module loads the logic from the firmware at reset and executes the logic responsive to the interrupt. The debugging module does not use the processor to execute the logic, the firmware is not part of the debugging module, and the debugging module is not part of the hardware component. The firmware may also store a mapping between registers of the hardware component and field-replaceable hardware units of the computing device, which the debugging module loads at reset and uses when executing the error decoding logic to determine which unit has caused the error.

Abstract: A method for self-contained testing within a DRAM comprises the DRAM receiving an instruction from an external processor to test a memory core on the DRAM, and the DRAM self-testing the memory core with one or more BIST pattern stored in a multipurpose register on the DRAM. Optionally, the step of self-testing may include writing the BIST pattern into all locations of the memory core, reading each location of the memory core, and comparing the content read from each location of the memory core with the BIST pattern, wherein a negative comparison indicates a failure has occurred. In a further option, the method may further comprise, after testing the DRAM, initializing the DRAM with an INIT pattern stored in the multipurpose register on the DRAM.

Abstract: One embodiment provides a method of managing power in a computer system. A device of the computer system is operated at a selected power-state. The power consumption of the computer system is monitored. If the power consumption of the computer system is approaching or has exceeded a power cap selected for the computer system, then a request to reduce the power-state for the device is generated in response. The operating system is used to service the request to reduce the power-state according to the priority of the request. The reduced power state is forced out-of-band following the request to reduce the power-state if the request is not immediately serviceable by the operating system. Different approaches can be taken to force the reduced power state, using, for example, system management mode or a platform environment control interface.

Abstract: Detection of a reset failure in a multinode data processing system is provided by a diagnostic circuit in each of a plurality of the server nodes of the system. Each diagnostic circuit is coupled to a code fetch chain of its corresponding node. At reset, prior to a node processor retrieving startup code from the code fetch chain, the diagnostic circuit provides diagnostic signals to the code fetch chain. A problem in the code fetch chain is detected from a response to the diagnostic signals. When a problem is detected, a node failure status for the problem node may be signaled to the other nodes. The multinode system may be configured in response to signaled node failure status, such as by dropping failed nodes and replacing a failed primary node with a secondary node if necessary.