WO2006070451A1 - メモリ診断方法 - Google Patents
メモリ診断方法 Download PDFInfo
- Publication number
- WO2006070451A1 WO2006070451A1 PCT/JP2004/019600 JP2004019600W WO2006070451A1 WO 2006070451 A1 WO2006070451 A1 WO 2006070451A1 JP 2004019600 W JP2004019600 W JP 2004019600W WO 2006070451 A1 WO2006070451 A1 WO 2006070451A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- memory
- data
- storage memory
- cyclic redundancy
- read
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0751—Error or fault detection not based on redundancy
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0706—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation the processing taking place on a specific hardware platform or in a specific software environment
- G06F11/0727—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation the processing taking place on a specific hardware platform or in a specific software environment in a storage system, e.g. in a DASD or network based storage system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0706—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation the processing taking place on a specific hardware platform or in a specific software environment
- G06F11/073—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation the processing taking place on a specific hardware platform or in a specific software environment in a memory management context, e.g. virtual memory or cache management
Definitions
- the present invention relates to a storage device having a memory diagnosis function and a memory diagnosis method in the storage device.
- SDRAM provided in the hard disk device for temporarily storing data exchanged between the host computer and the disk.
- This SDRAM which is interposed between the disk and the host computer, is also diagnosed as to whether data can be written, held, and read normally when the hard disk drive is started.
- FIG. 1 is a diagram showing a configuration of a hard disk device, a host computer, and an intervening SDRAM.
- the host computer 6 writes to and reads from the storage medium 7.
- SDRAM2 exists in the meantime, and temporarily stores write data and read data.
- the storage medium 7 is a magnetic disk and includes a magnetic head assembly necessary for access.
- the first storage memory 4 is a mechanism (FIFO) for complementing the difference in writing and reading speed between the host computer 6 and the SDRAM 2.
- the second storage memory 5 is a mechanism (FIFO) for complementing the difference in writing and reading speed between the storage medium 7 and the SDRAM 2.
- the memory manager 1 is hardware that manages writing to and reading from the SDRAM2.
- the microprocessor 3 manages the entire hard disk device, and starts and stops the spindle motor.
- FIG. 2 is an operation flow from the start of the hard disk device to the ready state. is there.
- the microprocessor 3 starts operating after the power is turned on.
- step S2 the first storage memory 4, the second storage memory 5, SDRAM 2 and the like are initialized.
- step S3 an initial diagnosis of the SDRAM 2 is performed.
- step S4 initialization of the microprocessor 3 and servo controller registers is performed.
- step S5 the spindle motor is activated and the disk reaches steady rotation. After these steps, the hard disk drive is ready for use.
- the microprocessor 3 generates 2-byte diagnostic data, and sends the diagnostic data to the memory manager 1 that manages the writing and reading of the SDRAM2.
- the memory manager 1 receives the diagnostic data from the microprocessor 3 and writes the diagnostic data on the address of the SDRAM 2 designated by the microprocessor 3.
- the memory manager 1 reads the address force data that was just written.
- the read data is sent from the memory manager 1 to the microprocessor 3, and the write data and read data are compared in the microprocessor 3.
- Patent Document 1 Japanese Patent Laid-Open No. 7-271679
- the memory diagnosis is performed on the basis of the firmware deployed in the microprocessor, and the diagnostic data is written and read in units of 2 bytes, so the processing speed is slow.
- the reliability is inferior to the case of diagnosing the entire area.
- the object of the present invention is to be performed by a microprocessor based on conventional firmware! /, By performing a batch write and read of large-scale data using hardware. It is possible to increase the speed of the memory diagnosis and perform it in parallel with other processing by opening the microprocessor. In addition, the entire memory area is diagnosed accordingly, and the reliability of memory diagnosis is improved. Means for solving the problem
- a storage medium capable of writing and reading data from a host computer, and temporary storage of write data and read data to the storage medium
- An SDRAM that stores data
- a memory manager that manages writing and reading to and from the SDRAM
- a first storage memory that complements a difference in processing speed between writing and reading between the host computer and the SDRAM
- the memory manager includes a comparison unit, The diagnostic data held in one storage memory is written to the SDRAM, and the written data is written to the SDRAM.
- the comparison unit compares the diagnostic data in the first storage memory and the read data in the second storage memory, the comparison result is If they do not match, an abnormality is notified.
- a storage medium capable of writing and reading data from a host computer, and an SDRAM for temporarily storing write data and read data on the storage medium
- a memory manager that manages writing and reading to and from the SDRAM, a first storage memory that complements a difference in processing speed between writing and reading between the host computer and the SDRAM, and the storage medium
- a method for diagnosing SDRAM in a storage device comprising a second storage memory that compensates for a difference in processing speed between writing and reading between the memory and the SDRAM, the diagnostic data in the first storage memory Is written to the SDRAM, the write data is read out from the SDRAM, stored in the second storage memory, and the diagnostic data in the first storage memory is stored by the comparison unit in the memory manager. And the read data in the second storage memory are compared, and if the comparison result does not match, an abnormality is notified.
- a storage medium capable of writing and reading data from a host computer, and temporary writing and reading of data to and from the storage medium.
- SDRAM that performs general storage
- a memory manager that manages writing and reading of the data to and from the SDRAM
- a first storage that complements the processing speed difference between the host computer and the SDRAM
- the memory manager is configured to perform cyclic redundancy.
- a cyclic redundancy code calculation unit that calculates a code and a comparison unit that compares the cyclic redundancy code, and is preset in the first storage memory
- Write data generated based on fixed diagnostic data is written to the SDRAM, the write data is read from the SDRAM, and the cyclic redundancy code calculation unit writes the write data to the SDRAM
- a cyclic redundancy code based on the write data is calculated, and the calculated cyclic redundancy code is written to the SDRAM together with the write data.
- the comparison unit compares the cyclic redundancy code based on the write data with the cyclic redundancy code based on the read data, and compares If the results do not match, a notice is given of the abnormality.
- the memory manager uses a seed to which a predetermined number is added or subtracted each time the write data is written or the read data is read.
- the cyclic redundancy code calculation unit adds the calculated seed to the write data or the read data.
- different cyclic redundancy codes are obtained from the same write data or read data.
- the write data is composed of a plurality of the diagnostic data.
- a storage medium capable of writing and reading data from a host computer, and temporary writing and reading of data to and from the storage medium.
- SDRAM that performs general storage, a memory manager that manages writing and reading of the data to and from the SDRAM, and a first that complements the processing speed difference between writing and reading between the host computer and the SDRAM
- a method for diagnosing SDRAM in a storage device comprising: a storage memory; and a second storage memory that compensates for a difference in processing speed between writing and reading between the previous storage medium and the SDRAM. The write data generated based on predetermined diagnostic data set in the first storage memory is written to the SDRAM.
- the cyclic redundancy code calculation unit in the memory manager calculates a cyclic redundancy code based on the write data, and writes the calculated cyclic redundancy code together with the write data to the SDRAM.
- the cyclic redundancy code written by the memory manager is read from the SDRAM, the write data is also read by the SDRAM manager, and read by the cyclic redundancy code calculation unit in the memory manager.
- a cyclic redundancy code based on the read data is calculated, and a comparison unit in the memory manager compares the cyclic redundancy code based on the write data with the cyclic redundancy code based on the read data. If there is a discrepancy, an abnormality is notified.
- the storage memory is
- the memory manager uses a seed to which a predetermined number is added or subtracted each time the write data is written or the read data is read.
- the cyclic redundancy code calculation unit adds the calculated seed to the write data or the read data.
- different cyclic redundancy codes are obtained from the same write data or read data.
- the initial write data is composed of a plurality of the diagnostic data.
- the memory diagnosis function of the present invention reduces the diagnosis time by writing, reading, and comparing the diagnosis data stored in the storage memory with respect to the SDRAM by using the node software. Enable diagnosis. In addition, since memory diagnosis can be performed without using a microprocessor, other operations can be performed in parallel. Brief Description of Drawings
- FIG. 1 is a diagram showing a configuration of a hard disk device, a host computer, and an intervening SDRAM.
- FIG. 3 is a configuration diagram of a hard disk device having a memory diagnosis function in the first embodiment of the present invention.
- FIG. 4 A diagram of an arbitration mechanism that processes requests to the memory manager.
- FIG. 5 is an operation flow of memory diagnosis in the first embodiment of the present invention.
- FIG. 6 is a configuration diagram of a hard disk device having a memory diagnosis function in a second embodiment of the present invention.
- FIG. 7 is an operation flow of memory diagnosis in the second embodiment of the present invention.
- FIG. 8 is a diagram showing data written to SDRAM in the second embodiment of the present invention.
- FIG. 9 is an explanatory diagram of cyclic redundancy codes.
- FIG. 10 is a diagram showing an operation flow of the node disk device having the memory diagnosis function of the present invention and an operation flow of the node disk device having the memory diagnosis function of the present invention. Explanation of symbols [0023] 1 Memory Manager
- FIG. 3 is a configuration diagram of a hard disk device having a memory diagnosis function according to the first embodiment of the present invention.
- a memory manager 1 that manages writing and reading to and from the SDRAM 2 is connected to a host computer 6, a storage medium 7, and a microphone processor 3 that manages the entire hard disk device.
- a first storage memory 4 exists between the memory manager 1 and the host computer 6.
- a second storage memory 5 exists between the memory manager 1 and the storage medium 7.
- the inside of the dotted line in FIG. 3 is the memory diagnostic device.
- the first storage memory 4 and the second storage memory 5 are each constituted by a FIFO.
- the hard disk device includes SDRAM2 for temporarily storing write and read data, a FIFO for compensating for the difference in data transfer speed between the storage medium 7 and SDRAM2, and a host computer 6
- a FIFO is provided to compensate for the difference in data transfer speed between the SDRAM and SDRAM2.
- FIG. 4 is a diagram of an arbitration mechanism that processes requests to the memory manager 1.
- the memory manager 1 processes various requests such as writing to and reading from the SDRAM 2 using an arbitration mechanism.
- memory management As a request for the first storage memory 4, the data power of the first storage memory 4 also generates a cyclic redundancy code (CRC) and writes to the SDRAM 2 or a request R1 for writing or reading the data.
- the arbitration mechanism in FIG. 4 processes the generated request in a clockwise direction. For example, if data in the first storage memory 4 is to be written to the SDRAM 2, a request R3 is generated, and the request R4, the request R5, and the request R1 are checked after the processing of the request R3. If no requests have occurred, it is looping idle.
- the first storage memory 4 stores diagnostic data. This data may be generated by the microprocessor 3 or transferred from the host computer 6. The diagnostic data is written into the SDRAM 2 via the memory manager 1 and immediately read out to the memory manager 1. The read data is stored in the second storage memory 5. Then, the SDRAM 2 is diagnosed by comparing the data in the first storage memory 4 and the data in the second storage memory 5 by the comparison unit in the memory manager 1.
- FIG. 5 is an operation flow of memory diagnosis in the first embodiment of the present invention.
- the start position and end position of the memory diagnosis are determined by the microprocessor 3 operating based on the firmware (processing step Pl-1).
- the microprocessor 3 activates the initial diagnosis function of the memory manager 1 (processing step Pl-2).
- Processing step P1-1 and processing step P1-2 are implemented by firmware, and thereafter are implemented by the memory manager 1 hardware.
- the memory manager 1 writes the diagnostic data in the first storage memory 4 to the SDRAM 2 (processing step Pl-3). At this time, the request R3 is generated and processed in the arbitration mechanism of FIG. Next, immediately read the data on the written address, Store in the storage memory 5 (processing step Pl-4). At this time, the request R4 is generated and processed in the arbitration mechanism of FIG.
- the comparison unit in the memory manager compares the data in the first storage memory 4 with the data in the second storage memory 5 (processing step P1-5). Immediately, it is determined that there is an abnormality, and the diagnosis of SDRAM2 ends (processing step Pl-6). If they match, it is confirmed whether or not the pointer position is the end position (processing step Pl-7).
- processing step Pl-8 If it is not the end position, the SDRAM2 pointer is incremented (processing step Pl-8), and the process returns to writing diagnostic data to the SDRAM2 (processing step P1-3).
- the loop from processing step P1-3 to processing step P1-8 is repeated until the position of the pointer reaches the end position.
- the memory diagnostic function ends normally (processing step P1-9).
- the SDRAM2 has an area of 8 megabytes, and the first storage memory 4 and the second storage memory 5 have a size of 128 bytes, respectively, to diagnose the entire area of the SDRAM2,
- the loop of writing, reading and comparison (loop from processing step P1-3 to processing step P1-8) is 65536 times.
- the write, read, and compare loops of every 2 bytes are a significant improvement because the entire loop of SDRAM2 requires 4 mega loops.
- the overhead time for writing, reading, and comparison can be omitted, resulting in a significant time reduction.
- the memory diagnosis that has been performed by the microprocessor 3 based on the conventional firmware is performed, and large-scale data is collectively written and read by the hardware. Therefore, it is possible to increase the speed of the memory diagnosis and to perform it in parallel with other processes by opening the microprocessor 3. As a result, the entire area of the SDRAM 2 can be diagnosed and the reliability can be improved.
- the hard disk device is used as an example of the storage device.
- the present invention is also applicable to devices and products having a memory.
- FIG. 6 is a configuration diagram of a hard disk device having a memory diagnosis function according to the second embodiment of the present invention.
- the memory manager 1 that manages writing and reading to the SDRAM 2 is a manager that manages the host computer 6, the storage medium 7, and the entire hard disk device.
- a first storage memory 4 exists between the memory manager 1 and the host computer 6, and a second storage memory 5 exists between the memory manager 1 and the storage medium 7.
- the inside of the dotted line in FIG. 6 is the hard disk device.
- the first storage memory 4 and the second storage memory 5 are each constituted by a FIFO.
- the hard disk device includes SDRAM2 for temporarily storing write and read data, a FIFO for compensating for the difference in data transfer speed between the storage medium 7 and SDRAM2, and a host computer 6
- a FIFO is provided to compensate for the difference in data transfer speed between the SDRAM and SDRAM2.
- the first storage memory 4 stores diagnostic data. This data may be generated by the microphone processor 3 or may be transferred from the host computer 6. This diagnostic data is repeatedly written to the SDRAM 2 via the memory manager 1 until the write end position is reached. At this time, a cyclic redundancy code is calculated from the seed and 512-byte write data (the calculation method will be described later) and written to a part of SDRAM2.
- the seed is 4-byte data that is initialized to 0 when the initial diagnosis is started and incremented every time 512 bytes of diagnostic data is written from the start position. Also, one sector on SDRAM2 is composed of 512 bytes. If the first storage memory 4 and the second storage memory 5 are 128 bytes each, then writing to and reading from SDRAM2 Four data transfers will be performed.
- the memory manager 1 reads one sector of data and the corresponding cyclic redundancy code from the write start position, and calculates the read data redundancy of the one sector again.
- the comparison unit in the memory manager 1 compares the read cyclic redundancy code with the calculated cyclic redundancy code to diagnose the SDRAM 2.
- FIG. 7 is an operation flow of memory diagnosis in the second embodiment of the present invention.
- the start position and end position of the memory diagnosis are determined by the microprocessor 3 operated by the firmware (processing step P2-l).
- the microprocessor is memory money Start the initial diagnosis function of Jar 1 (Process P2-2).
- Processing step P2-1 and processing step P2-2 are implemented by firmware, and thereafter are implemented by the memory manager 1 hardware.
- the memory manager 1 sequentially writes the diagnostic data in the first storage memory 4 to the SDRAM 2 while incrementing the pointer (processing step P2-5) (processing step P2-3).
- a request R3 for writing from the first storage memory 4 to the SDRAM 2 is generated in the arbitration mechanism of FIG.
- the cyclic redundancy code calculation unit in the memory manager 1 calculates the cyclic redundancy code from the write data for one sector and the seed corresponding thereto, and sequentially writes it in the start position of the SDRAM 2.
- the arbitration mechanism makes four requests R3, and finally requests R1.
- FIG. 8 is a diagram showing data written to the SDRAM 2 in the second embodiment of the present invention.
- the 128-byte diagnostic data stored in the first storage memory 4 is written sequentially from the memory diagnosis start position (data write start position in Fig. 8) determined in process step P2-1.
- the seed is incremented each time one sector is written, and a cyclic redundancy code is calculated in combination with the data for one sector.
- the corresponding cyclic redundancy code is sequentially written from the position that is the data write end position.
- the diagnostic data finally reaches the data write end position, and the corresponding cyclic redundancy code reaches the CRC write end position.
- the initial diagnosis function of the memory manager 1 reads the written data and the cyclic redundancy code while incrementing the pointer (processing step P2-10) (processing step P2-7).
- the arbitration mechanism generates R4 requests for reading data from the SDRAM 2 to the second storage memory 5, and finally generates a request R2 for reading the cyclic redundancy code.
- this cyclic redundancy code is sent to the comparator in the memory manager 11.
- the cyclic redundancy code calculation unit in the memory manager 1 calculates the cyclic redundancy code from the seed corresponding to the read data (processing step P2-8), and sends the cyclic redundancy code to the comparison unit.
- the two cyclic redundancy codes sent to the comparison unit are compared (processing step P2-ll), and if the two cyclic redundancy codes match, it is confirmed whether the end position is reached (processing step P2-ll). 13) . If the two cyclic redundancy codes do not match, the initial diagnosis function of the memory manager 1 detects an abnormality and ends (processing step P2-12).
- FIG. 9 is an explanatory diagram of cyclic redundancy codes.
- the 128-byte diagnostic data does not change, and the same data ( ⁇ ⁇ ⁇ in Fig. 9) is repeatedly written from the first storage memory 4 to SDRAM2.
- the 4 byte seed is incremented once.
- Memory Manager 1 does not need to reserve a 516-byte area for division. Divides 132-byte data, which is a concatenation of 4-byte seed and 128-byte diagnostic data, by a constant.
- the comparison unit in the memory manager 1 compares the 4-byte cyclic redundancy code against the 512-byte area diagnosis, so it is efficient! Compared to comparing 521 bytes of data directly, when comparing 4 bytes of cyclic redundancy code, the amount of comparison is 1Z128. In addition, it is not necessary to hold 512 bytes of data read from SDRAM2.
- the arbitration mechanism starts simultaneously with the start of the initial diagnostic function. When no request is generated, the arbitration mechanism loops in the idle state as shown in Fig. 4. Therefore, a write request from the first storage memory 4 to SDRAM2 R3 is generated. After writing 128 bytes of diagnostic data from the first storage memory 4, the requests R3, R1, R2, R4, and R5 have not occurred, so the request R3 that has occurred is processed again. Request R3 is generated four times until the processing of 512 bytes for one sector is completed.
- a request R1 for generating and writing a cyclic redundancy code is generated, and a cyclic redundancy code for the written 512-byte data is generated and written. Thereafter, the processing of 4 requests R3 and 1 request R1 is repeated 100 times in total to write 100 sectors of diagnostic data and the corresponding cyclic redundancy code.
- the memory diagnosis is performed by hardware, and a relatively large size of data is written and read at a time. A cocoon can be realized. This makes it possible to diagnose the entire area of the SDRAM 2 and improve the reliability.
- the memory diagnosis that has been performed by the microprocessor based on the conventional firmware is performed by the node software, thereby making it possible to open the microprocessor and perform other processing.
- FIG. 10 is a diagram showing an operation flow of the node disk device having the memory diagnostic function of the present invention and an initial operation sequence of the conventional hard disk device not having the memory diagnostic function of the present invention.
- the thick arrow shown at the bottom is the initial operation sequence of the node disk device having the memory diagnostic function of the present invention, and the thin arrow at the upper side has the memory diagnostic function of the present invention.
- the operation flow of the device is shown. [0051] There is no difference between step SI in which the firmware is expanded in the microprocessor 3 and step S2 in which the first storage memory 4, the second storage memory 5 and the SDRAM 2 are initialized. In step S3 where the initial diagnosis of the SDRAM 2 is performed, the initial diagnosis by the hard disk device having no memory diagnosis function of the present invention is completed earlier.
- step S4 the firmware deployed in the microprocessor 3 and the servo controller are initialized without waiting for the completion of step S3. Can move to S4.
- step S5 the spindle motor, there is no difference between the two. As a result, the time can be reduced by the time difference T1.
- the second embodiment of the present invention has been described using an example applied to SDRAM of a hard disk device, but can also be applied to a general memory.
- a hard disk device has been described as an example of a storage device.
- the present invention temporarily stores data between the host system and a host system.
- the present invention can be applied to a general device or product having a memory functioning as a memory, or a device or system using a replaceable storage medium such as a magnetic tape device, an optical disk device, or a magneto-optical disk device.
- the power described in a host computer such as a personal computer as a host system, and in the case of a product incorporating a storage device such as a hard disk video recorder, is a host processor (such as CPU), and controls data transfer to the storage device. What you do falls under the higher system.
- a host processor such as CPU
- diagnostic data stored in the storage memory is written into, read out from, and compared to the SDRAM by using nodeware.
- a storage device having a memory diagnosis function that shortens the diagnosis time and enables diagnosis of the entire area of the SDRAM is provided. Further, the memory device having the memory diagnosis function can perform the memory diagnosis without occupying the microprocessor, so that other operations can be performed in parallel.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Techniques For Improving Reliability Of Storages (AREA)
- For Increasing The Reliability Of Semiconductor Memories (AREA)
Abstract
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2006550519A JP4435180B2 (ja) | 2004-12-28 | 2004-12-28 | メモリ診断方法 |
PCT/JP2004/019600 WO2006070451A1 (ja) | 2004-12-28 | 2004-12-28 | メモリ診断方法 |
US11/820,618 US20070266277A1 (en) | 2004-12-28 | 2007-06-20 | Memory diagnostic method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2004/019600 WO2006070451A1 (ja) | 2004-12-28 | 2004-12-28 | メモリ診断方法 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US11/820,618 Continuation US20070266277A1 (en) | 2004-12-28 | 2007-06-20 | Memory diagnostic method |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2006070451A1 true WO2006070451A1 (ja) | 2006-07-06 |
Family
ID=36614578
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2004/019600 WO2006070451A1 (ja) | 2004-12-28 | 2004-12-28 | メモリ診断方法 |
Country Status (3)
Country | Link |
---|---|
US (1) | US20070266277A1 (ja) |
JP (1) | JP4435180B2 (ja) |
WO (1) | WO2006070451A1 (ja) |
Families Citing this family (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7996731B2 (en) | 2005-11-02 | 2011-08-09 | Advanced Micro Devices, Inc. | Error detection in high-speed asymmetric interfaces |
JP5406689B2 (ja) * | 2009-12-10 | 2014-02-05 | 富士通テン株式会社 | 制御装置、及び、制御方法 |
JP6717059B2 (ja) * | 2016-06-06 | 2020-07-01 | オムロン株式会社 | 制御システム |
US11551028B2 (en) | 2017-04-04 | 2023-01-10 | Hailo Technologies Ltd. | Structured weight based sparsity in an artificial neural network |
US11238334B2 (en) | 2017-04-04 | 2022-02-01 | Hailo Technologies Ltd. | System and method of input alignment for efficient vector operations in an artificial neural network |
US10387298B2 (en) | 2017-04-04 | 2019-08-20 | Hailo Technologies Ltd | Artificial neural network incorporating emphasis and focus techniques |
US11544545B2 (en) | 2017-04-04 | 2023-01-03 | Hailo Technologies Ltd. | Structured activation based sparsity in an artificial neural network |
US11615297B2 (en) | 2017-04-04 | 2023-03-28 | Hailo Technologies Ltd. | Structured weight based sparsity in an artificial neural network compiler |
US11874900B2 (en) | 2020-09-29 | 2024-01-16 | Hailo Technologies Ltd. | Cluster interlayer safety mechanism in an artificial neural network processor |
US11263077B1 (en) | 2020-09-29 | 2022-03-01 | Hailo Technologies Ltd. | Neural network intermediate results safety mechanism in an artificial neural network processor |
US11811421B2 (en) | 2020-09-29 | 2023-11-07 | Hailo Technologies Ltd. | Weights safety mechanism in an artificial neural network processor |
US11221929B1 (en) * | 2020-09-29 | 2022-01-11 | Hailo Technologies Ltd. | Data stream fault detection mechanism in an artificial neural network processor |
US11237894B1 (en) | 2020-09-29 | 2022-02-01 | Hailo Technologies Ltd. | Layer control unit instruction addressing safety mechanism in an artificial neural network processor |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05224833A (ja) * | 1992-02-10 | 1993-09-03 | Fujitsu Ltd | 外部記憶装置のデータ保証方法 |
JPH08115268A (ja) * | 1994-10-13 | 1996-05-07 | Toshiba Corp | メモリ回路装置 |
JPH11306798A (ja) * | 1998-04-22 | 1999-11-05 | Oki Electric Ind Co Ltd | メモリ装置のテスト容易化回路 |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5077656A (en) * | 1986-03-20 | 1991-12-31 | Channelnet Corporation | CPU channel to control unit extender |
US7269841B1 (en) * | 1992-12-09 | 2007-09-11 | Sedna Patent Services, Llc | Digital cable headend for cable television delivery system |
US5953418A (en) * | 1995-06-14 | 1999-09-14 | David Hall | Providing selective data broadcast receiver addressability |
US5802311A (en) * | 1995-06-15 | 1998-09-01 | David Hall | Using adaptive redundant packet retrieval to improve reliability in a unidirectional data broadcasting system |
US5673279A (en) * | 1995-11-06 | 1997-09-30 | Sun Microsystems, Inc. | Verification of network transporter in networking environments |
JPH1050056A (ja) * | 1996-07-31 | 1998-02-20 | Sharp Corp | 半導体記憶装置 |
US5829047A (en) * | 1996-08-29 | 1998-10-27 | Lucent Technologies Inc. | Backup memory for reliable operation |
KR100474985B1 (ko) * | 1997-06-23 | 2005-07-01 | 삼성전자주식회사 | 메모리로직복합반도체장치 |
US6085285A (en) * | 1997-11-13 | 2000-07-04 | International Business Machines Corporation | Intermixing different devices along a single data communication link by placing a strobe signal in a parity bit slot |
US6467060B1 (en) * | 1998-06-26 | 2002-10-15 | Seagate Technology Llc | Mass storage error correction and detection system, method and article of manufacture |
-
2004
- 2004-12-28 JP JP2006550519A patent/JP4435180B2/ja not_active Expired - Fee Related
- 2004-12-28 WO PCT/JP2004/019600 patent/WO2006070451A1/ja not_active Application Discontinuation
-
2007
- 2007-06-20 US US11/820,618 patent/US20070266277A1/en not_active Abandoned
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH05224833A (ja) * | 1992-02-10 | 1993-09-03 | Fujitsu Ltd | 外部記憶装置のデータ保証方法 |
JPH08115268A (ja) * | 1994-10-13 | 1996-05-07 | Toshiba Corp | メモリ回路装置 |
JPH11306798A (ja) * | 1998-04-22 | 1999-11-05 | Oki Electric Ind Co Ltd | メモリ装置のテスト容易化回路 |
Also Published As
Publication number | Publication date |
---|---|
JP4435180B2 (ja) | 2010-03-17 |
JPWO2006070451A1 (ja) | 2008-06-12 |
US20070266277A1 (en) | 2007-11-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20070266277A1 (en) | Memory diagnostic method | |
US5392302A (en) | Address error detection technique for increasing the reliability of a storage subsystem | |
US7206219B2 (en) | Slave and master of serial peripheral interface, system thereof, and method thereof | |
US9195538B1 (en) | System and method for in-line error correction for storage systems | |
US7865784B1 (en) | Write validation | |
JPH08203294A (ja) | メモリ | |
US5630054A (en) | Method and apparatus for storing and retrieving error check information | |
US7809899B2 (en) | System for integrity protection for standard 2n-bit multiple sized memory devices | |
JPS5848288A (ja) | デ−タ記憶レジスタのアドレス方法及び装置 | |
JP4130811B2 (ja) | 試験装置及び試験方法 | |
JPH10269150A (ja) | メモリ初期化方法 | |
JPH1078853A (ja) | 記憶装置 | |
JP2000065904A (ja) | 半導体試験装置 | |
EP0365114A2 (en) | Interface arrangement for interfacing a data storage device with a data handling system | |
JPH11102328A (ja) | メモリ診断システム | |
CN118155698A (zh) | 数据测试方法、电子设备及存储介质 | |
JP2868003B1 (ja) | 磁気ディスク装置 | |
JP3022688B2 (ja) | 補助記憶装置 | |
JP3675748B2 (ja) | データ記憶制御装置 | |
JP2001144628A (ja) | データ転送装置 | |
JPH0896310A (ja) | 磁気テープ用エラー訂正符号の記録方法および装置 | |
CN114756485A (zh) | 一种基于fpga平台的ddr读写仲裁及测试方法 | |
JPH1153267A (ja) | メモリデータ誤り自動訂正回路 | |
JPH08123628A (ja) | ディスクアレイ装置 | |
KR950015087A (ko) | 메모리 진단장치 및 방법 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2006550519 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 11820618 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWP | Wipo information: published in national office |
Ref document number: 11820618 Country of ref document: US |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 04807955 Country of ref document: EP Kind code of ref document: A1 |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: 4807955 Country of ref document: EP |