Analysis system, analysis method and program

Patent metadata

Field	Value
Publication number	US-9606890-B2
Application number	US-201514855508-A
Country	US
Kind code	B2
Filing date	Sep 16, 2015
Priority date	Sep 26, 2014
Publication date	Mar 28, 2017
Grant date	Mar 28, 2017

Abstract

Analysis system, analysis method and program. The system includes: trace means for acquiring a command issued by software executed in an information processing system and a physical address of a memory used by the command as trace data, and recording the trace data to storage means; event detecting means for detecting an event caused to occur by the software and acquiring event information; conversion means for converting the event information to a memory access pattern configured with a plurality of commands for accessing the memory and a plurality of physical addresses; and memory accessing means for accessing the memory using the converted memory access pattern, causing the trace means to acquire trace data and record the trace data to the storage means.

First claim

I claim:

1. A method for analyzing the state of use of a memory by software executed in an information processing system, the method comprising the steps of:
acquiring, by trace means, a command issued by the software and a physical address of the memory used by the command as trace data, and recording the trace data to storage means;
detecting an event caused to occur by the software and acquiring event information;
converting the acquired event information to a memory access pattern configured with a plurality of commands for accessing the memory and a plurality of physical addresses; and
accessing the memory using the converted memory access pattern, causing the trace means to acquire trace data and record the trace data to the storage means.

2. The method according to claim 1 , further comprising the steps of:
acquiring a plurality of pieces of trace data recorded in the storage means;
detecting trace data of the event from among the acquired plurality of pieces of trace data; and
extracting the event information from the detected trace data.

3. The method according to claim 2 , wherein
the event is a page fault event; and
the method further comprises a step of converting the physical addresses included in at least one piece of trace data to virtual addresses to be used by the software to access the memory, using the event information extracted at the extraction step.

4. The method according to claim 3 , further comprising a step of associating the physical addresses with the virtual addresses obtained by converting the physical addresses to generate a conversion table.

5. The method according to claim 1 , wherein
the memory is managed being divided in a plurality of banks and is accessed with the use of bank addresses for specifying the banks, row addresses for specifying a plurality of rows, and column addresses for specifying a plurality of columns; and
at the conversion step, the event information is converted to the memory access pattern by expressing the event information with the bank address, the row address and a plurality of the column addresses, and combining the bank address, the row address and each of the plurality of column addresses to generate a plurality of addresses.

6. The method according to claim 5 , wherein
the event information includes process identification information for identifying, as a process, software that performs processing using the memory, and page frame information to be used by the software to access the memory; and
at the conversion step, the process identification information is expressed with the bank address and the row address, and the page frame information is expressed with the plurality of the column addresses.

7. The method according to claim 1 , wherein the information processing system is configured with a plurality of nodes, each node being provided with the trace means; and at the recording step, a memory accessed by the software having caused the event to occur is accessed, using the converted memory access pattern, the trace means of a node having the memory is caused to acquire trace data, and the trace data is caused to be recorded to the storage means.

Assignees

• IBM

Inventors

• Munetoh Seiji

Classifications

• G06F11/3037

  where the computing system component is a memory, e.g. virtual memory, cache (accessing, addressing or allocating within memory systems or architectures G06F12/00; checking stores for correct operation G11C29/00) · CPC title

• G06F11/3476

  Data logging (G06F11/14, G06F11/2205 take precedence) · CPC title

• G06F3/0653

  Monitoring storage devices or systems · CPC title

• G06F11/3495

  for systems · CPC title

• G06F12/145

  the protection being virtual, e.g. for virtual blocks or segments before a translation mechanism · CPC title