Hardware accelerator based method and device for string searching
View Patent ↗A method for searching within a data block for a data chunk having a predefined value, the method includes: fetching, by a processor, a data block search instruction; fetching, a data unit that includes multiple data chunks; wherein at least one data chunk within the data unit belongs to the data block; deciding whether to use a mask for data chunk level masking; searching, by a hardware accelerator, for a valid data chunk within the fetched data unit that has the predefined value; wherein the searching comprising applying a mask; wherein a valid data chunk in an non-masked data chunk that belongs to the data block; and determining whether to update the value of the mask and whether to fetch a new data unit that belongs to the data block.
1. A method for searching within a data block for a data chunk having a predefined value, the method comprises:
fetching, by a processor, a data block search instruction;
fetching a data unit that comprises multiple data chunks, wherein at least one data chunk within the data unit belongs to the data block;
deciding whether to use a mask for data chunk level masking or for bit level masking;
searching, by a hardware accelerator, for a valid data chunk within the fetched data unit that has the predefined value, wherein the searching comprising applying a mask; wherein a valid data chunk is a non-masked data chunk that belongs to the data block; and
determining whether to update the value of the mask and whether to fetch a new data unit that belongs to the data block.
2. The method according to claim 1 wherein the searching comprises duplicating a mask portion, the mask portion having a size determined based on a size of the data chunk, such as to provide a mask, if the mask is utilized for bit level masking.
3. The method according to claim 2 wherein the fetching, by a processor, comprises fetching an instruction that comprises a data unit size value.
4. The method according to claim 2 wherein the fetching, by a processor, comprises fetching an instruction that comprises a data unit size location field.
5. The method according to claim 2 further comprising updating the mask if the mask is utilized for data chunk level masking.
6. The method according to claim 5 wherein a size of the mask is determined based on a size of the data chunk.
7. The method according to claim 5 wherein the updating comprises rotating the mask by a data chunk size.
8. The method according to claim 2 wherein the searching comprises performing in parallel multiple match operations between multiple data chunks that belong to the data unit and multiple reference data chunks.
9. The method according to claim 1 wherein the searching comprises performing in parallel multiple match operations between multiple data chunks that belong to the data unit and multiple reference data chunks.
10. The method according to claim 1 wherein the searching comprises duplicating a mask portion, the mask portion having a size determined based on a size of the data chunk, such as to provide a mask, if the mask is utilized for bit level masking.
11. The method according to claim 1 wherein the fetching, by a processor, comprises fetching an instruction that comprises a data unit size value.
12. The method according to claim 1 wherein the fetching, by a processor, comprises fetching an instruction that comprises a data unit size location field.
13. A device for locating a predefined value data chunk within a data block, the device comprises:
a memory unit adapted to store data units and a processor adapted to fetch an instruction and selectively generate control signals in response to the fetched instruction; and
a hardware accelerator, coupled to the memory unit, wherein the hardware accelerator is adapted to:
(i) receive control signals from the processor,
(ii) fetch a data unit that comprises multiple data chunks, wherein at least one data chunk within the data unit belongs to the data block,
(iii) decide whether to use a mask for data chunk level masking or for bit level masking;
(iv) search for a valid data chunk within the fetched data unit that has the predefined value, wherein the searching comprising applying a mask, wherein a valid data chunk is a non-masked data chunk that belongs to the data block; and
(v) determine whether to update the value of the mask and whether to fetch a new data unit that belongs to the data block.
14. The device according to claim 13 wherein the hardware accelerator is adapted to duplicate a mask portion having a size that is responsive to a size of the data chunk, such as to provide a mask, if the mask is utilized for bit level masking.
15. The device according to claim 14 wherein the processor is adapted to fetch an instruction that comprises a data unit size value.
16. The device according to claim 14 wherein the processor is adapted to fetch an instruction that comprises a data unit size location field.
17. The device according to claim 14 wherein the hardware accelerator is adapted to update the mask if the mask is utilized for data chunk level masking.
18. The device according to claim 14 wherein the hardware accelerator is adapted to update the mask in response to a size of the data chunk.
19. The device according to claim 14 wherein the hardware accelerator is adapted to rotate the mask by a data chunk size.
20. The device according to claim 14 wherein the hardware accelerator is adapted to perform in parallel multiple match operations between multiple data chunks that belong to the data unit and multiple reference data chunks.