Documentation
¶
Overview ¶
Package decode provides an x86_64 implementation of two Stream VByte decoding algorithms, a normal decoding approach and one that incorporates differential coding.
Index ¶
- func Get4uint32DeltaScalar(in []byte, out []uint32, ctrl uint8, prev uint32)
- func Get4uint32Scalar(in []byte, out []uint32, ctrl uint8)
- func Get8uint32(in []byte, out []uint32, ctrl uint16)
- func Get8uint32Delta(in []byte, out []uint32, ctrl uint16, prev uint32)
- func Get8uint32DeltaFast(in []byte, out []uint32, ctrl uint16, prev uint32)
- func Get8uint32DeltaFastAsm(in []byte, out []uint32, ctrl uint16, prev uint32, shuffle *[256][16]uint8, ...)
- func Get8uint32DeltaScalar(in []byte, out []uint32, ctrl uint16, prev uint32)
- func Get8uint32Fast(in []byte, out []uint32, ctrl uint16)
- func Get8uint32FastAsm(in []byte, out []uint32, ctrl uint16, shuffle *[256][16]uint8, ...)
- func Get8uint32Scalar(in []byte, out []uint32, ctrl uint16)
- func GetMode() shared.PerformanceMode
- func GetUint32DeltaScalar(in []byte, out []uint32, ctrl uint8, count int, prev uint32) int
- func GetUint32Scalar(in []byte, out []uint32, ctrl uint8, count int) int
- type Get8DeltaImpl
- type Get8Impl
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func Get4uint32DeltaScalar ¶ added in v0.3.0
Get4uint32DeltaScalar will decode 4 uint32 values from in into out and reconstruct the original values via differential coding. Prev provides a way for you to indicate the base value for this batch of 4. For example, when decoding the second batch of 4 integers out of, e.g. 8, you would provide a prev value of the last value in the first batch of 4 you decoded. This is done to ensure that the integers are correctly resolved to the correct diff. An example below.
Input: [ 10, 10, 10, 10 ] [ 10, 10, 10, 10 ] Output: [ 10, 20, 30, 40 ] [ 50, 60, 70, 80 ] Prev: 40
func Get4uint32Scalar ¶
Get4uint32Scalar will decode 4 uint32 values from in into out using the Stream VByte format. Returns the number of bytes read from the input buffer.
Note: It is your responsibility to ensure that the incoming slices have the appropriate sizes and data otherwise this func will panic.
func Get8uint32 ¶
Get8uint32 is a general func you can use to decode 8 uint32's at a time. It will use the fastest implementation available determined during package initialization. If your CPU supports special hardware instructions then it will use an accelerated version of Stream VByte. Otherwise, the scalar implementation will be used as the fallback.
func Get8uint32Delta ¶ added in v0.3.0
Get8uint32Delta is a general func you can use to decode 8 differentially coded uint32's at a time. It will use the fastest implementation available determined during package initialization. If your CPU supports special hardware instructions then it will use an accelerated version of Stream VByte. Otherwise, the scalar implementation will be used as the fallback.
func Get8uint32DeltaFast ¶ added in v0.3.0
Get8uint32DeltaFast binds to get8uint32DeltaFast which is implemented in assembly.
func Get8uint32DeltaFastAsm ¶ added in v0.3.0
func Get8uint32DeltaFastAsm( in []byte, out []uint32, ctrl uint16, prev uint32, shuffle *[256][16]uint8, lenTable *[256]uint8, )
Get8uint32DeltaFastAsm works similarly to get8uint32Fast with the exception that prior to writing the uncompressed integers out to the output slice, the original values are reconstructed from the diffs. The basic reconstruction algorithm is as follows:
Input: [A B C D] Input Shifted: [- A B C] Add above two: [A AB BC CD] Add Prev: [PA PAB PBC PCD] Input Shifted: [- - A AB] Add Shifted: [PA PAB PABC PABCD]
func Get8uint32DeltaScalar ¶ added in v0.3.0
Get8uint32DeltaScalar will decode 8 uint32 values from in into out and reconstruct the original values via differential coding. Prev provides a way for you to indicate the base value for this batch of 8. For example, when decoding the second batch of 8 integers out of, e.g. 16, you would provide a prev value of the last value in the first batch of 8 you decoded. This is done to ensure that the integers are correctly resolved to the correct diff. An example below.
Input: [ 10, 10, 10, 10, 10, 10, 10, 10 ] [ 10, 10, 10, 10, 10, 10, 10, 10 ] Output: [ 10, 20, 30, 40, 50, 60, 70, 80 ] [ 90, 100, 110, 120, 130, 140, 150, 160 ] Prev: 80
func Get8uint32Fast ¶ added in v0.3.0
Get8uint32Fast binds to get8uint32Fast which is implemented in assembly.
func Get8uint32FastAsm ¶ added in v0.3.0
func Get8uint32FastAsm( in []byte, out []uint32, ctrl uint16, shuffle *[256][16]uint8, lenTable *[256]uint8, )
Get8uint32FastAsm uses the provided 16-bit control to load the appropriate decoding shuffle masks and performs a shuffle operation on the provided input bytes. This in effect decompresses the input byte stream to uint32s. The result is written to the provided output slice.
func Get8uint32Scalar ¶
Get8uint32Scalar will decode 8 uint32 values from in into out using the Stream VByte format. Returns the number of bytes read from the input buffer.
Note: It is your responsibility to ensure that the incoming slices have the appropriate sizes and data otherwise this func will panic.
func GetMode ¶
func GetMode() shared.PerformanceMode
GetMode performs a check to see if the current ISA supports the below decoding funcs.
func GetUint32DeltaScalar ¶ added in v0.4.0
GetUint32DeltaScalar decodes up to 4 integers from in into out using the Stream VByte format. It will reconstruct the original non differentially encoded values.
Note: It is your responsibility to ensure that the incoming slices have the appropriate sizes and data otherwise this func will panic.
func GetUint32Scalar ¶ added in v0.3.0
GetUint32Scalar decodes up to 4 integers from in into out using the Stream VByte format.
Note: It is your responsibility to ensure that the incoming slices have the appropriate sizes and data otherwise this func will panic.
Source Files
Directories