Exception preserving parallel data processing of string and unstructured text

A computer program product, the computer program product comprising a computer readable storage medium having program instructions for performing parallel data wide accesses on an unstructured text embodied therewith, the program instructions executable by a processor coupled to a memory to cause the processor to perform:

• creating a plurality of separate string index pointers, each separate string index pointer pointing at a beginning of a plurality of string segments of the unstructured text;
• loading into a vector register of the memory using a no-fault vector load instruction, the plurality of separate string segments of the unstructured text based on the plurality of separate string index pointers;
• automatically testing access permissions of a first byte of each separate string segment of the plurality of separate string segments;
• determining whether one of the plurality of separate string segments includes an end indication;
• updating a corresponding pointer of the plurality of separate string index pointers to point to a next string segment of the plurality of separate string segments when one of the plurality of separate string segments is determined to not include the end indication;
• validating a remaining portion of the unstructured text by accessing and loading a last character position identified by the end indication into the vector register when the one of the plurality of separate string segments is determined to include the end indication, the last character position comprising a null terminating character;
• appending a load instruction to the last character position verifying that the one of the plurality of separate string segments was contained in accessible storage;
• loading into the vector register of the memory the next string segment of the unstructured text based on the corresponding pointer, the vector register holding or storing at least ‘m’ characters from the unstructured text, wherein if less than ‘m’ characters are available, the vector register is filled with a default value, wherein when the default value is encountered, processing continues without providing any data to any region in memory that a program does not have permission to access;
• and
• automatically testing access permissions of a first byte of the next string segment, where
• if the first byte is not in a read-permitted location in the memory, a permission violation exception is taken and processing is terminated, and
• if the first byte is in the read-permitted location in the memory, contents of the vector register containing the next string segment are scanned for a string termination character and a determination is made as to whether the next string segment includes the end indication.

A system, comprising a processor and a memory storing program instructions for performing parallel data wide accesses on an unstructured text thereon, the program instructions executable by a processor to cause the system to perform:

• creating a plurality of separate string index pointers, each separate string index pointer pointing at a beginning of a plurality of string segments of the unstructured text;
• loading into a vector register of the memory using a no-fault vector load instruction, the plurality of separate string segments of the unstructured text based on the plurality of separate string index pointers;
• automatically testing access permissions of a first byte of each separate string segment of the plurality of separate string segments;
• determining whether one of the plurality of separate string segments includes an end indication;
• updating a corresponding pointer of the plurality of separate string index pointers to point to a next string segment of the plurality of separate string segments when one of the plurality of separate string segments is determined to not include the end indication;
• validating a remaining portion of the unstructured text by accessing and loading a last character position identified by the end indication into the vector register when the one of the plurality of separate string segments is determined to include the end indication, the last character position comprising a null terminating character;
• appending a load instruction to the last character position verifying that the one of the plurality of separate string segments was contained in accessible storage;
• loading into the vector register of the memory the next string segment of the unstructured text based on the corresponding pointer, the vector register holding or storing at least ‘m’ characters from the unstructured text, wherein if less than ‘m’ characters are available, the vector register is filled with a default value, wherein when the default value is encountered, processing continues without providing any data to any region in memory that a program does not have permission to access;
• and
• automatically testing access permissions of a first byte of the next string segment, where
• if the first byte is not in a read-permitted location in the memory, a permission violation exception is taken and processing is terminated, and
• if the first byte is in the read-permitted location in the memory, contents of the vector register containing the next string segment are scanned for a string termination character and a determination is made as to whether the next string segment includes the end indication.