摘要:
Each component binary in a heterogeneous program is translated from a platform-specific instruction set into a set of intermediate representation (IR) instructions that are platform-neutral. The IR instructions are grouped into IR code blocks, the IR code blocks into IR procedures, and the IR procedures into IR components to create an intermediate representation hierarchy for the program. An application program interface is provided that permits user access to the IR hierarchy for instrumentation, optimization, navigation, and manipulation of the IR hierarchy. The transformed IR hierarchy is then translated into platform-specific instructions and output as a modified binary. The user can designate a different platform for the output translation of a code block than the platform for which the code block was originally written. Prologue and epilog code is added to contiguous blocks that are translated into different architectures. The modified binary can be iterated through the translation and transformation process to produce multiple versions of the heterogeneous program.
摘要:
Each component binary in a heterogeneous program is translated from a platform-specific instruction set into a set of intermediate representation (IR) instructions that are platform-neutral. The IR instructions are grouped into IR code blocks, the IR code blocks into IR procedures, and the IR procedures into IR components to create an intermediate representation hierarchy for the program. An application program interface is provided that permits user access to the IR hierarchy for instrumentation, optimization, navigation, and manipulation of the IR hierarchy. The transformed IR hierarchy is then translated into platform-specific instructions and output as a modified binary. The user can designate a different platform for the output translation of a code block than the platform for which the code block was originally written. Prologue and epilog code is added to contiguous blocks that are translated into different architectures. The modified binary can be iterated through the translation and transformation process to produce multiple versions of the heterogeneous program.
摘要:
Pre-defined performance tools enable instrumentation and optimization of a heterogeneous program by modifying a platform-neutral intermediate representation (IR) of the program. The intermediate representation is hierarchy of base elements that correspond to instructions, code blocks, procedures and components within the program. Each base element exports an application program interface that provides for navigation, querying and modification of the corresponding element in the hierarchy. An application program interface for the entire intermediate representation includes the application program interfaces for the individual elements along with additional functionality. A user can request instrumentation and optimization of any level of the intermediate representation hierarchy and the performance tools perform the required modifications to the intermediate representation. When a modification requires inserting instructions, the performance tools add prologue and epilog code to connect the inserted instructions with the original intermediate representation instructions. When the intermediate representation instructions are translated into platform-specific instructions and executed, the modifications made to the intermediate representation provide the instrumentation and/or optimization of the heterogeneous program.
摘要:
A shared library optimization process relocates a code block in a platform-neutral intermediate representation (IR) of a library into the IR of another library. The shared library optimization process preserves the original control flow that passed through the relocated code block using forwarding pointers and address substitution. The shared library optimization process can also relocate code blocks from multiple shared libraries into an existing shared library, or from one shared library into an empty shared library. The shared libraries to be optimized are specified by a user, as is the type of optimization to be performed.
摘要:
An output translator provides for cross module representations of components within a heterogeneous program by translating modifying a platform-neutral intermediate representation (IR) of the program into platform-specific instructions for different architectures. The intermediate representation is hierarchy of base elements that correspond to instructions, code blocks, procedures and components within the program. Blocks of instructions that were originally written for one architecture can be translated from the intermediate representation into platform-specific instructions for a different architecture. The output translator provides any necessary code to interface contiguous code blocks that are emitted in different instruction sets.
摘要:
A system and method for automatically updating software components on a running computer system without requiring any interruption of service. A software module is hotpatched by loading a patch into memory and modifying an instruction in the original module to jump to the patch. A coldpatching technique places a coldpatch version of the module on disk for subsequent loading by processes, after hotpatching occurred. The coldpatch has the entry points to its functions at the same relative locations within the module as the hotpatch, which facilitates subsequent hotpatching. A hotpatch and coldpatch are automatically generated by deriving differences between changed and original binary files, and establishing the point to insert the jump. Validation is performed to ensure that the hotpatch is applied to the correct version, and that the coldpatch is replacing the correct version. Version management is also provided to control the number of patches via support rules.
摘要:
Rapid determination of placement for code and data for optimal access of the code and data during execution of software applications. An application that is to be efficiently arranged in a software carrier medium is instrumented such that the status of whether a particular unit of code or data has been accessed during a time interval is recorded in a time-ordered bit sequence for each such unit. A sort routine is used to sort the order of time-ordered bit sequences within an array of such sequences. The sort routine invokes a comparison function that operates upon the temporal information stored in the two time-ordered bit sequences that are being compared for the routine. The order of the code and data in the software application is reordered in accordance with the results of the sort routine.