A data compressor with a hash computing hardware configured to evaluate the hash value for the current hash key extracted from a source data string, obtain a hash line corresponding to the hash value from a hash table, and perform hash key comparison to find at least one matching hash key. The hash line includes a prefix address column that stores a prefix address. Each entry of the hash line is provided to store a hash key and an offset. The hash computing hardware evaluates an address of the at least one matching hash key by combining the prefix address and an offset of the at least one matching hash key, and the offset of the at least one matching hash key is obtained from an entry storing the at least one matching hash key.

A data compressor with a hash computing hardware configured to evaluate the hash value for the current hash key extracted from a source data string, obtain a hash line corresponding to the hash value from a hash table, and perform hash key comparison to find at least one matching hash key. The hash line includes a prefix address column that stores a prefix address. Each entry of the hash line is provided to store a hash key and an offset. The hash computing hardware evaluates an address of the at least one matching hash key by combining the prefix address and an offset of the at least one matching hash key, and the offset of the at least one matching hash key is obtained from an entry storing the at least one matching hash key.

A method of reducing the storage requirements of blockchain metadata via dictionary-style compression includes receiving a request to add a transaction block to a blockchain. The method further includes determining an identifier (ID) of a dictionary block most recently stored on the blockchain. The method further includes compressing, by a processing device, one or more transactions of the transaction block based on the dictionary block to generate a compressed transaction block. The method further includes adding the ID of the dictionary block to the compressed transaction block. The method further includes providing the compressed transaction block, including the ID of the dictionary block, for storage on the blockchain.

According to one embodiment, a dividing circuit divides a first bit string into second bit strings and outputs the divided second bit strings. The dividing circuit includes first, second, and third blocks. The first block executes first operation for each bit of a third bit string in the first bit string. The first operation is to calculate a head bit of a succeeding symbol when one bit is assumed to be a head of one symbol. The second block executes second operation for each bit of the third bit string for a set number of times. The second operation is to overwrite boundary information associated with one bit with boundary information associated with a bit indicated by the boundary information. The third block divides the third bit string immediately before a second bit indicated by boundary information associated with a first bit of the third bit string.

A system for encoding and decoding data-tokens. In some examples, the system may be configured to encode and decode integers. In other cases, the system may be configured to encode and decode symbols or bytes of data.

A system for encoding and decoding data-tokens. In some examples, the system may be configured to encode and decode integers. In other cases, the system may be configured to encode and decode symbols or bytes of data.

A database-management system evaluates a query that retrieves and transforms encoded symbols stored in a database. If the stored symbols assume a relatively small set of distinct values, the system initially performs the transformation on every value in the set. During execution of subsequent queries, rather than performing the transformation upon every stored symbol fetched from the database, the system merely returns the previously derived encoded transformation results that correspond to the decoded value of each fetched symbol. If the symbols stored in the database span a relatively large set of distinct values, the system does not initially perform the transformation upon every value in the set. Instead, the first time the system fetches a symbol that has a particular value, it saves that symbol's encoded transformation result and reuses that result the next time it fetches an encoded symbol with the same value.

Technologies for dividing work across one or more accelerator devices include a compute device. The compute device is to determine a configuration of each of multiple accelerator devices of the compute device, receive a job to be accelerated from a requester device remote from the compute device, and divide the job into multiple tasks for a parallelization of the multiple tasks among the one or more accelerator devices, as a function of a job analysis of the job and the configuration of each accelerator device. The compute engine is further to schedule the tasks to the one or more accelerator devices based on the job analysis and execute the tasks on the one or more accelerator devices for the parallelization of the multiple tasks to obtain an output of the job.

Technologies for dividing work across one or more accelerator devices include a compute device. The compute device is to determine a configuration of each of multiple accelerator devices of the compute device, receive a job to be accelerated from a requester device remote from the compute device, and divide the job into multiple tasks for a parallelization of the multiple tasks among the one or more accelerator devices, as a function of a job analysis of the job and the configuration of each accelerator device. The compute engine is further to schedule the tasks to the one or more accelerator devices based on the job analysis and execute the tasks on the one or more accelerator devices for the parallelization of the multiple tasks to obtain an output of the job.

Described herein are systems and methods for compressing or otherwise reducing the memory requirements for storing and computing the model parameters in recurrent neural language models. Embodiments include space compression methodologies that share the structured parameters at the input embedding layer, the output embedding layers, or both of a recurrent neural language model to significantly reduce the size of model parameters, but still compactly represent the original input and output embedding layers. Embodiments of the methodology are easy to implement and tune. Experiments on several data sets show that embodiments achieved similar perplexity and BLEU score results while only using a fraction of the parameters.