Abstract: Technologies for performing speculative decompression include a managed node to decode a variable size code at a present position in compressed data with a deterministic decoder and concurrently perform speculative decodes over a range of subsequent positions in the compressed data, determine the position of the next code, determine whether the position of the next code is within the range, and output, in response to a determination that the position of the next code is within the range, a symbol associated with the deterministically decoded code and another symbol associated with a speculatively decoded code at the position of the next code.

Abstract: Techniques and apparatus for parallel decompression are described. In one embodiment, for example, an apparatus to provide parallel decompression may include at least one memory and logic for a decompression component, at least a portion of the logic comprised in hardware coupled to the at least one memory, the logic to determine decompression information of a compressed data unit, annotate the compressed data unit with at least a portion of the decompression information to generate an annotated data unit, parallel-decode the annotated data unit to generate a plurality of intermediate elements, and decode and merge the plurality of intermediate elements to generate a decompressed data unit. Other embodiments are described and claimed.

Abstract: Processor instructions for determining two minimum and two maximum values and associated apparatus and methods. The instructions include various 2MIN instructions for determining the two smallest values among three or four input values and 2MAX instructions for determining the two largest values among three or four input values. The 2MIN instructions employ two operands, with the first operand in some of the variations storing concatenated min1 and min2 values in a first register and a scr2 comparison value or two src2 concatenated src2 values in a second register. Comparators are used to implement hardware logic for determining whether the scr2 value(s) is/are less than each of min1 and min2. Based on the hardware logic, the smallest two values among min1, min2, and src2 (or both src2 values) are stored as concatenated values in the first register. The 2MAX instructions are implemented in a similar manner, except the comparisons are whether the scr2 value(s) is/are greater than each of max1 and max2 values.

Abstract: In an embodiment, a processor comprises a plurality of processing cores and a compression accelerator to compress an input stream comprising a first data block and a second data block. The compression accelerator comprises a first compression engine to compress the first data block; and a second compression engine to update state data for the second compression engine using a sub-portion of the first data block; and after an update of the state data for the second compression engine using the sub-portion of the first data block, compress a second data block using the updated state data for the second compression engine. Other embodiments are described and claimed.

Abstract: A method of an aspect includes receiving an instruction. The instruction indicates a first source of a first packed data including state data elements ai, bi, ei, and fi for a current round (i) of a secure hash algorithm 2 (SHA2) hash algorithm. The instruction indicates a second source of a second packed data. The first packed data has a width in bits that is less than a combined width in bits of eight state data elements ai, bi, ci, di, ei, fi, gi, hi of the SHA2 hash algorithm. The method also includes storing a result in a destination indicated by the instruction in response to the instruction. The result includes updated state data elements ai+, bi+, ei+, and fi+ that have been updated from the corresponding state data elements ai, bi, ei, and fi by at least one round of the SHA2 hash algorithm.

Abstract: A processor includes a memory hierarchy, buffer, and a decompressor. The decompressor includes circuitry to read elements to be decompressed according to a compression scheme, parse the elements to identify literals and matches, and, with the literals and matches, generate an intermediate token stream formatted for software-based copying of the literals and matches to produce decompressed data. The intermediate token stream is to include a format for multiple tokens that are to be written in parallel with each other, and another format for tokens that include a data dependency upon themselves.

Abstract: A processor includes a plurality of registers, an instruction decoder to receive an instruction to process a KECCAK state cube of data representing a KECCAK state of a KECCAK hash algorithm, to partition the KECCAK state cube into a plurality of subcubes, and to store the subcubes in the plurality of registers, respectively, and an execution unit coupled to the instruction decoder to perform the KECCAK hash algorithm on the plurality of subcubes respectively stored in the plurality of registers in a vector manner.

Abstract: In one embodiment, an apparatus comprises a first compression engine to receive a first compressed data block from a second compression engine that is to generate the first compressed data block by compressing a first plurality of repeated instances of data that each have a length greater than or equal to a first length. The first compression engine is further to compress a second plurality of repeated instances of data of the first compressed data block that each have a length greater than or equal to a second length, the second length being shorter than the first length, wherein each compressed repeated instance of the first and second pluralities of repeated instances comprises a location and length of a data instance that is repeated. The apparatus further comprises a memory buffer to store the compressed first and second plurality of repeated instances of data.

Abstract: A method in one aspect may include receiving a multiply instruction. The multiply instruction may indicate a first source operand and a second source operand. A product of the first and second source operands may be stored in one or more destination operands indicated by the multiply instruction. Execution of the multiply instruction may complete without writing a carry flag. Other methods are also disclosed, as are apparatus, systems, and instructions on machine-readable medium.

Abstract: A compute device to generate deterministic compressed streams receives a current string to be matched to one or more prior instances of the current string, the current string being located within an input buffer and the one or more prior instances located within a history buffer. The compute device identifies a limited subset of index memory designated for storing pointers to the prior instances, identifying a reserved slop region in the index memory, and compares the current string to a prior instance, locating the at least one prior instance using at least one pointer to the at least one prior instance. The at least one pointer is stored within the limited subset of the index memory, and the compute device also prohibits use of any pointers stored in the reserved slop region of the index memory. Other embodiments are described and claimed.

Abstract: Technologies for flexibly compressing data include a computing device having an accelerator complex that is to receive a compression job request and schedule the compression job request for one or more hardware compression resources of the accelerator complex. The accelerator complex is further to perform the compression job request with the one or more hardware compression resources in response to scheduling the compression job request and to communicate uncompressed data and compressed data with an I/O subsystem of the computing device in response to performing the compression job request. Other embodiments are described and claimed.

Abstract: Technologies for high-ratio compression with heterogeneous history buffers include a computing device having an accelerator complex with a large history buffer and a small history buffer. The large history buffer has a larger size than the small history buffer. For example, the small history buffer may be 32 kilobytes and the large history buffer may be 64 kilobytes, 1 megabyte, or larger. The large history buffer is coupled to a large-buffer compare core that searches for matches in the large history buffer, finds a best match, and forwards the best match to a small-buffer compare core. The small-buffer compare core searches the small history buffer for matches, receives the match forwarded from the large-buffer compare core, and determines a best match from the matches in the small history buffer and the forwarded match. Other embodiments are described and claimed.

Abstract: Technologies for database acceleration include a computing device having a database accelerator. The database accelerator performs a decompress operation on one or more compressed elements of a compressed database to generate one or more decompressed elements. After decompression of the compressed elements, the database accelerator prepares the one or more decompressed elements to generate one or more prepared elements to be processed by an accelerated filter. The database accelerator then performs the accelerated filter on the one or more prepared elements to generate one or more output elements. Other embodiments are described and claimed.

Abstract: A method in one aspect may include receiving a multiply instruction. The multiply instruction may indicate a first source operand and a second source operand. A product of the first and second source operands may be stored in one or more destination operands indicated by the multiply instruction. Execution of the multiply instruction may complete without writing a carry flag. Other methods are also disclosed, as are apparatus, systems, and instructions on machine-readable medium.

Abstract: Technologies for heuristic Huffman code generation include a computing device that generates a weighted list of symbols for a data block. The computing device determines a threshold weight and identifies one or more lightweight symbols in the list that have a weight less than or equal to the threshold weight. The threshold weight may be the average weight of all symbols with non-zero weight in the list. The computing device generates a balanced sub-tree of nodes for the lightweight symbols, with each lightweight symbol associated with a leaf node. The computing device adds the remaining symbols and the root of the balanced sub-tree to a heap and generates a Huffman code tree by processing the heap. The threshold weight may be adjusted to tune performance and compression ratio. Other embodiments are described and claimed.

Abstract: A method is described. The method includes iteratively performing for each position in a result matrix stored in a third register, multiplying a value at a matrix position stored in a first register with a value at a matrix position stored in a second register to obtain a first multiplicative value, where the positions in the first register and the second register are determined by the position in the result matrix and performing an exclusive or (XOR) operation with the first multiplicative value and a value stored at a result matrix position stored in the third register to obtain a result value.

Abstract: Technologies for performing low-latency decompression include a managed node to parse, in response to a determination that a read tree descriptor does not match a cached tree descriptor, the read tree descriptor to construct one or more tables indicative of codes in compressed data. Each code corresponds to a different symbol. The managed node is further to decompress the compressed data with the one or more tables and store the one or more tables in association with the read tree descriptor in a cache memory for subsequent use.

Abstract: A method of one aspect may include receiving a rotate instruction. The rotate instruction may indicate a source operand and a rotate amount. A result may be stored in a destination operand indicated by the rotate instruction. The result may have the source operand rotated by the rotate amount. Execution of the rotate instruction may complete without reading a carry flag.

Abstract: A method of one aspect may include receiving a rotate instruction. The rotate instruction may indicate a source operand and a rotate amount. A result may be stored in a destination operand indicated by the rotate instruction. The result may have the source operand rotated by the rotate amount. Execution of the rotate instruction may complete without reading a carry flag.

Abstract: Instruction sets for variable length integer (varint) coding and associated methods and apparatus. The instructions sets include instructions for encoding and decoding varints, and may be included as a part of an instruction set architecture (ISA) for processors architectures such as x86 and Arm-based architectures, as well as other ISAs. In one aspect, the instructions include, a varint size encode instruction to encode a size of a varint, a varint encode instruction to encode a varint, a varint size decode instruction to decode a size of an encoded varint, and a varint decode instruction to decode an encoded varint. Varint encode size and encode instructions may be combined in a single instructions. Similarly, varint decode size and decode instructions may be combined in a single instruction. In one aspect, the instructions use a variable-length quantity (VLQ) encoding scheme under which varints are encoded into one or more VLQ octets.