Method and means for defective bit processing in a fault tolerant bubble storage system
A method and means for modifying up to k bits in each byte of (n+k) bits to be accessed with respect to counterpart minor loops of a sequentially accessed byte-organized major/minor loop memory array. The combination of bad byte location pointers and a byte mask is used to delete bits on bad bytes extracted from the memory or to insert zeroes into defective loop locations when bad bytes are written into the memory. Alternatively, the ith bit in each byte of (n+1) bits to be accessed may be modified through the utilization of the value of the ith bit written into the byte position of n+1st bit, while the ith bit of a defective byte position is zeroed. As a bad byte is extracted, the value of the n+1st bit is substituted for the ith bit. A dense list, rather than a sparse bad byte pointer list, is used.
Latest IBM Patents:
- GUARD RAILING OF A LARGE LANGUAGE MODEL USING PRODUCTION DATA AUGMENTATION
- Augmented reality projection of predicted vehicle movements
- Selecting an executable operation from a lookup table identified by a conditional instruction and indexed based on combined results of concurrently-computed relational operators
- Completion of data gaps in a dataset
- Self-supervision in table question answering
Type: Grant
Filed: Jul 7, 1980
Date of Patent: Mar 3, 1981
Assignee: International Business Machines Corporation (Armonk, NY)
Inventors: Abraham M. Gindi (San Jose, CA), Magdi R. Orfali (San Jose, CA), Arvind M. Patel (San Jose, CA)
Application Number: 6/166,718
International Classification: G06F 1108; G06F 1100; G11C 1908;
