AI Engine Intrinsics User Guide  (AIE) r2p23
 All Data Structures Namespaces Functions Variables Typedefs Groups Pages

Vector update intrinsic functions allow substitution of the lanes within a vector value. More...

Overview

Vector update intrinsic functions allow substitution of the lanes within a vector value.

Below the buffer sizes are as follows: V - 128 bit W - 256 bit X - 512 bit Y - 1024 bit

For more information see Integer Vector Types.

Note
All intrinsics require a compile time constant for the idx parameter.

load_hi and load_lo intrinsic functions

Updates the a 48/80 bit accumulator from a 64/128 bit integer vector. Uses the saturation mode set by the user. load_lo updates the lower lanes of the accumulator and load_hi the higher ones.

upd_hi and upd_lo intrinsic functions

Update the top half or bottom half of the lanes within a data type.

upd_v({W,X,Y} buf,int idx,V val) 128-bit intrinsic functions

upd_v(buf,0...7,val) update the successive 128-bit lanes into a 256/512/1024 bit vector.

In the following example, a large 32-way complex vector is updated 4 elements at-a-time using a 128-bit update. The notation "XX++" in comments shows the starting element index of the lane being updated. While the notation "XX.." shows that the lane update was initiated in a prior cycle. This example also shows that the updates can be pipelined.

const v4cint16 * input = d_in;
v32cint16 sbuff = undef_v32cint16();
...
sbuff = upd_v(sbuff,0, *input++); // 00++|____|____|____ ____|____|____|____
sbuff = upd_v(sbuff,1, *input++); // 00..|04++|____|____ ____|____|____|____

upd_w({X,Y} buf,int idx,W val) 256-bit intrinsic functions

upd_w(buf,0...3,val) update the successive 256-bit lanes into a 512/1024 bit vector.

The following example shows the update of a large 32-way complex vector 8 elements at-a-time using a 256-bit update. These updates are also pipelined.

const v8cint16 * input = d_in;
v32cint16 sbuff = undef_v32cint16();
...
sbuff = upd_w(sbuff,0, *input_++); // 00++|04++|____|____ ____|____|____|____
sbuff = upd_w(sbuff,1, *input_++); // 00..|04..|08++|12++ ____|____|____|____

upd_x(Y buf,int idx,X val) 512-bit intrinsic functions

upd_x(buf,0...1,val) update the successive 512-bit lanes into a 1024 bit vector.

The following example shows the update of a large 32-way complex vector 16 elements at-a-time using a 512-bit update. These updates are also pipelined.

const v16cint16 * input = d_in;
v32cint16 sbuff = undef_v32cint16();
...
sbuff = upd_x(sbuff,0, *input_++); // 00++|04++|08++|12++ ____|____|____|____
sbuff = upd_x(sbuff,1, *input_++); // 00..|04..|08..|12.. 16++|20++|24++|28++

Accumulator load operations

v8acc48 load_lo (v8acc48 acc, v4int64 vec)
 Load the 64 bit from the vector saturated in a 48 bit accumulator. This is done for the 4 lower lanes.
 
v8acc48 load_hi (v8acc48 acc, v4int64 vec)
 Load the 64 bit from the vector saturated in a 48 bit accumulator. This is done for the 4 higher lanes.
 
v4cacc48 load_lo (v4cacc48 acc, v2cint64 vec)
 Load the 64 bit from the vector saturated in a 48 bit accumulator. This is done for the 2 lower lanes.
 
v4cacc48 load_hi (v4cacc48 acc, v2cint64 vec)
 Load the 64 bit from the vector saturated in a 48 bit accumulator. This is done for the 2 higher lanes.
 
v4acc80 load_lo (v4acc80 acc, v2int128 vec)
 Load the 128 bit from the vector saturated in a 80 bit accumulator. This is done for the 2 lower lanes.
 
v4acc80 load_hi (v4acc80 acc, v2int128 vec)
 Load the 128 bit from the vector saturated in a 80 bit accumulator. This is done for the 2 higher lanes.
 
v2cacc80 load_lo (v2cacc80 acc, v1cint128 vec)
 Load the 128 bit from the vector saturated in a 80 bit accumulator. This is done for the 1 lower lanes.
 
v2cacc80 load_hi (v2cacc80 acc, v1cint128 vec)
 Load the 128 bit from the vector saturated in a 80 bit accumulator. This is done for the 1 higher lanes.
 

Vector Unpacking Operations

v16int16 unpack (v16int8 vec)
 Load and unpack vectors These intrinsics can only be used coupled with a memory access operation. Ex: v16int16 vec = unpack(<em>src);//where src is v16int8
 
v16int16 unpack (v16uint8 vec)
 Load and unpack vectors These intrinsics can only be used coupled with a memory access operation. Ex: v16int16 vec = unpack(<em>src);//where src is v16int8
 
v32int16 unpack (v32int8 vec)
 Load and unpack vectors These intrinsics can only be used coupled with a memory access operation. Ex: v16int16 vec = unpack(<em>src);//where src is v16int8
 
v32int16 unpack (v32uint8 vec)
 Load and unpack vectors These intrinsics can only be used coupled with a memory access operation. Ex: v16int16 vec = unpack(<em>src);//where src is v16int8
 

128-bit lane update into 256-bit vector

Update the successive 128-bit lanes in a 256-bit vector. idx parameter must be a compile time constant.

promotion v32int8 upd_v (v32int8, int, v16int8)
 Update a 256-bit vector 16 elements at a time using a 128-bit update.
 
promotion v32uint8 upd_v (v32uint8, int, v16uint8)
 Update a 256-bit vector 16 elements at a time using a 128-bit update.
 
promotion v16int16 upd_v (v16int16, int, v8int16)
 Update a 256-bit vector 8 elements at a time using a 128-bit update.
 
promotion v8cint16 upd_v (v8cint16, int, v4cint16)
 Update a 256-bit vector 4 elements at a time using a 128-bit update.
 
promotion v8int32 upd_v (v8int32, int, v4int32)
 Update a 256-bit vector 4 elements at a time using a 128-bit update.
 
promotion v4cint32 upd_v (v4cint32, int, v2cint32)
 Update a 256-bit vector 2 elements at a time using a 128-bit update.
 
promotion v8float upd_v (v8float, int, v4float)
 Update a 256-bit vector 4 elements at a time using a 128-bit update.
 
promotion v4cfloat upd_v (v4cfloat, int, v2cfloat)
 Update a 256-bit vector 2 elements at a time using a 128-bit update.
 

128-bit lane update into 512-bit vector

Update the successive 128-bit lanes in a 512-bit vector. idx parameter must be a compile time constant.

promotion v64int8 upd_v (v64int8, int, v16int8)
 Update a 512-bit vector 16 elements at a time using a 128-bit update.
 
promotion v64uint8 upd_v (v64uint8, int, v16uint8)
 Update a 512-bit vector 16 elements at a time using a 128-bit update.
 
promotion v32int16 upd_v (v32int16, int, v8int16)
 Update a 512-bit vector 8 elements at a time using a 128-bit update.
 
promotion v16cint16 upd_v (v16cint16, int, v4cint16)
 Update a 512-bit vector 4 elements at a time using a 128-bit update.
 
promotion v16int32 upd_v (v16int32, int, v4int32)
 Update a 512-bit vector 4 elements at a time using a 128-bit update.
 
promotion v8cint32 upd_v (v8cint32, int, v2cint32)
 Update a 512-bit vector 2 elements at a time using a 128-bit update.
 
promotion v16float upd_v (v16float, int, v4float)
 Update a 512-bit vector 4 elements at a time using a 128-bit update.
 
promotion v8cfloat upd_v (v8cfloat, int, v2cfloat)
 Update a 512-bit vector 2 elements at a time using a 128-bit update.
 

128-bit lane update into 1024-bit vector

Set 128-bit lanes in a 1024-bit vector.

promotion v128int8 upd_v (v128int8, int, v16int8)
 Update a 1024-bit vector 16 elements at a time using a 128-bit update.
 
promotion v128uint8 upd_v (v128uint8, int, v16uint8)
 Update a 1024-bit vector 16 elements at a time using a 128-bit update.
 
promotion v64int16 upd_v (v64int16, int, v8int16)
 Update a 1024-bit vector 8 elements at a time using a 128-bit update.
 
promotion v32cint16 upd_v (v32cint16, int, v4cint16)
 Update a 1024-bit vector 4 elements at a time using a 128-bit update.
 
promotion v32int32 upd_v (v32int32, int, v4int32)
 Update a 1024-bit vector 4 elements at a time using a 128-bit update.
 
promotion v16cint32 upd_v (v16cint32, int, v2cint32)
 Update a 1024-bit vector 2 elements at a time using a 128-bit update.
 
promotion v32float upd_v (v32float, int, v4float)
 Update a 1024-bit vector 4 elements at a time using a 128-bit update.
 
promotion v16cfloat upd_v (v16cfloat, int, v2cfloat)
 Update a 1024-bit vector 2 elements at a time using a 128-bit update.
 
promotion v128int8 yset_v (int, v16int8)
 Creates a new 1024-bit vector with 16 elements already set using a 128-bit set.
 
promotion v128uint8 yset_v (int, v16uint8)
 Creates a new 1024-bit vector with 16 elements already set using a 128-bit set.
 
promotion v64int16 yset_v (int, v8int16)
 Creates a new 1024-bit vector with 8 elements already set using a 128-bit set.
 
promotion v32cint16 yset_v (int, v4cint16)
 Creates a new 1024-bit vector with 4 elements already set using a 128-bit set.
 
promotion v32int32 yset_v (int, v4int32)
 Creates a new 1024-bit vector with 4 elements already set using a 128-bit set.
 
promotion v16cint32 yset_v (int, v2cint32)
 Creates a new 1024-bit vector with 2 elements already set using a 128-bit set.
 
promotion v32float yset_v (int, v4float)
 Creates a new 1024-bit vector with 4 elements already set using a 128-bit set.
 
promotion v16cfloat yset_v (int, v2cfloat)
 Creates a new 1024-bit vector with 2 elements already set using a 128-bit set.
 

256-bit lane update into 512-bit vector

Update 256-bit lanes in a 512-bit vector. idx parameter must be a compile time constant.

promotion v64int8 upd_w (v64int8, int, v32int8)
 Update a 512-bit vector 32 elements at a time using a 256-bit update.
 
promotion v64uint8 upd_w (v64uint8, int, v32uint8)
 Update a 512-bit vector 32 elements at a time using a 256-bit update.
 
promotion v32int16 upd_w (v32int16, int, v16int16)
 Update a 512-bit vector 16 elements at a time using a 256-bit update.
 
promotion v16cint16 upd_w (v16cint16, int, v8cint16)
 Update a 512-bit vector 8 elements at a time using a 256-bit update.
 
promotion v16int32 upd_w (v16int32, int, v8int32)
 Update a 512-bit vector 8 elements at a time using a 256-bit update.
 
promotion v8cint32 upd_w (v8cint32, int, v4cint32)
 Update a 512-bit vector 4 elements at a time using a 256-bit update.
 
promotion v16float upd_w (v16float, int, v8float)
 Update a 512-bit vector 8 elements at a time using a 256-bit update.
 
promotion v8cfloat upd_w (v8cfloat, int, v4cfloat)
 Update a 512-bit vector 4 elements at a time using a 256-bit update.
 

256-bit lane update into 1024-bit vector

Update 256-bit lanes in a 1024-bit vector.

promotion v128int8 upd_w (v128int8, int, v32int8)
 Update a 1024-bit vector 32 elements at a time using a 256-bit update.
 
promotion v128uint8 upd_w (v128uint8, int, v32uint8)
 Update a 1024-bit vector 32 elements at a time using a 256-bit update.
 
promotion v64int16 upd_w (v64int16, int, v16int16)
 Update a 1024-bit vector 16 elements at a time using a 256-bit update.
 
promotion v32cint16 upd_w (v32cint16, int, v8cint16)
 Update a 1024-bit vector 8 elements at a time using a 256-bit update.
 
promotion v32int32 upd_w (v32int32, int, v8int32)
 Update a 1024-bit vector 8 elements at a time using a 256-bit update.
 
promotion v16cint32 upd_w (v16cint32, int, v4cint32)
 Update a 1024-bit vector 4 elements at a time using a 256-bit update.
 
promotion v32float upd_w (v32float, int, v8float)
 Update a 1024-bit vector 8 elements at a time using a 256-bit update.
 
promotion v16cfloat upd_w (v16cfloat, int, v4cfloat)
 Update a 1024-bit vector 4 elements at a time using a 256-bit update.
 

512-bit lane update into 1024-bit vector

Update 512-bit lanes in a 1024-bit vector. idx parameter must be a compile time constant.

promotion v128int8 upd_x (v128int8, int, v64int8)
 Update a 1024-bit vector 64 elements at a time using a 512-bit update.
 
promotion v128uint8 upd_x (v128uint8, int, v64uint8)
 Update a 1024-bit vector 64 elements at a time using a 512-bit update.
 
promotion v64int16 upd_x (v64int16, int, v32int16)
 Update a 1024-bit vector 32 elements at a time using a 512-bit update.
 
promotion v32cint16 upd_x (v32cint16, int, v16cint16)
 Update a 1024-bit vector 16 elements at a time using a 512-bit update.
 
promotion v32int32 upd_x (v32int32, int, v16int32)
 Update a 1024-bit vector 16 elements at a time using a 512-bit update.
 
promotion v16cint32 upd_x (v16cint32, int, v8cint32)
 Update a 1024-bit vector 8 elements at a time using a 512-bit update.
 
promotion v32float upd_x (v32float, int, v16float)
 Update a 1024-bit vector 16 elements at a time using a 512-bit update.
 
promotion v16cfloat upd_x (v16cfloat, int, v8cfloat)
 Update a 1024-bit vector 8 elements at a time using a 512-bit update.
 

320/384-bit lane update into 640/768-bit accumulator

v16acc48 upd_lo (v16acc48, v8acc48)
 
v16acc48 upd_hi (v16acc48, v8acc48)
 
v8cacc48 upd_lo (v8cacc48, v4cacc48)
 
v8cacc48 upd_hi (v8cacc48, v4cacc48)
 
v8acc80 upd_lo (v8acc80, v4acc80)
 
v8acc80 upd_hi (v8acc80, v4acc80)
 
v4cacc80 upd_lo (v4cacc80, v2cacc80)
 
v4cacc80 upd_hi (v4cacc80, v2cacc80)
 

Float accumulator updates

v8accfloat upd_lo (v8accfloat, v4accfloat)
 
v16accfloat upd_lo (v16accfloat, v8accfloat)
 
v32accfloat upd_lo (v32accfloat, v16accfloat)
 
v8accfloat upd_hi (v8accfloat, v4accfloat)
 
v16accfloat upd_hi (v16accfloat, v8accfloat)
 
v32accfloat upd_hi (v32accfloat, v16accfloat)
 
v4caccfloat upd_lo (v4caccfloat, v2caccfloat)
 
v8caccfloat upd_lo (v8caccfloat, v4caccfloat)
 
v16caccfloat upd_lo (v4caccfloat, v8caccfloat)
 
v4caccfloat upd_hi (v4caccfloat, v2caccfloat)
 
v8caccfloat upd_hi (v8caccfloat, v4caccfloat)
 
v16caccfloat upd_hi (v4caccfloat, v8caccfloat)
 

Updates for configuration registers

mac_idx upd0 (mac_idx, unsigned int)
 
mac_idx upd1 (mac_idx, unsigned int)
 
pmx_idx upd0 (pmx_idx, unsigned int)
 
pmx_idx upd1 (pmx_idx, unsigned int)
 
pmx_idx upd2 (pmx_idx, unsigned int)
 

Function Documentation

v8acc48 load_hi ( v8acc48  acc,
v4int64  vec 
)

Load the 64 bit from the vector saturated in a 48 bit accumulator. This is done for the 4 higher lanes.

Parameters
accTarget accumulator
vecInput vector
Returns
The Updated accumulator
v4cacc48 load_hi ( v4cacc48  acc,
v2cint64  vec 
)

Load the 64 bit from the vector saturated in a 48 bit accumulator. This is done for the 2 higher lanes.

Parameters
accTarget accumulator
vecInput vector
Returns
The Updated accumulator
v4acc80 load_hi ( v4acc80  acc,
v2int128  vec 
)

Load the 128 bit from the vector saturated in a 80 bit accumulator. This is done for the 2 higher lanes.

Parameters
accTarget accumulator
vecInput vector
Returns
The Updated accumulator
v2cacc80 load_hi ( v2cacc80  acc,
v1cint128  vec 
)

Load the 128 bit from the vector saturated in a 80 bit accumulator. This is done for the 1 higher lanes.

Parameters
accTarget accumulator
vecInput vector
Returns
The Updated accumulator
v8acc48 load_lo ( v8acc48  acc,
v4int64  vec 
)

Load the 64 bit from the vector saturated in a 48 bit accumulator. This is done for the 4 lower lanes.

Parameters
accTarget accumulator
vecInput vector
Returns
The Updated accumulator
v4cacc48 load_lo ( v4cacc48  acc,
v2cint64  vec 
)

Load the 64 bit from the vector saturated in a 48 bit accumulator. This is done for the 2 lower lanes.

Parameters
accTarget accumulator
vecInput vector
Returns
The Updated accumulator
v4acc80 load_lo ( v4acc80  acc,
v2int128  vec 
)

Load the 128 bit from the vector saturated in a 80 bit accumulator. This is done for the 2 lower lanes.

Parameters
accTarget accumulator
vecInput vector
Returns
The Updated accumulator
v2cacc80 load_lo ( v2cacc80  acc,
v1cint128  vec 
)

Load the 128 bit from the vector saturated in a 80 bit accumulator. This is done for the 1 lower lanes.

Parameters
accTarget accumulator
vecInput vector
Returns
The Updated accumulator
v16int16 unpack ( v16int8  vec)

Load and unpack vectors These intrinsics can only be used coupled with a memory access operation. Ex: v16int16 vec = unpack(<em>src);//where src is v16int8

v16int16 unpack ( v16uint8  vec)

Load and unpack vectors These intrinsics can only be used coupled with a memory access operation. Ex: v16int16 vec = unpack(<em>src);//where src is v16int8

v32int16 unpack ( v32int8  vec)

Load and unpack vectors These intrinsics can only be used coupled with a memory access operation. Ex: v16int16 vec = unpack(<em>src);//where src is v16int8

v32int16 unpack ( v32uint8  vec)

Load and unpack vectors These intrinsics can only be used coupled with a memory access operation. Ex: v16int16 vec = unpack(<em>src);//where src is v16int8

mac_idx upd0 ( mac_idx  ,
unsigned  int 
)
pmx_idx upd0 ( pmx_idx  ,
unsigned  int 
)
mac_idx upd1 ( mac_idx  ,
unsigned  int 
)
pmx_idx upd1 ( pmx_idx  ,
unsigned  int 
)
pmx_idx upd2 ( pmx_idx  ,
unsigned  int 
)
v16acc48 upd_hi ( v16acc48  ,
v8acc48   
)
v8cacc48 upd_hi ( v8cacc48  ,
v4cacc48   
)
v8acc80 upd_hi ( v8acc80  ,
v4acc80   
)
v4cacc80 upd_hi ( v4cacc80  ,
v2cacc80   
)
v8accfloat upd_hi ( v8accfloat  ,
v4accfloat   
)
v16accfloat upd_hi ( v16accfloat  ,
v8accfloat   
)
v32accfloat upd_hi ( v32accfloat  ,
v16accfloat   
)
v4caccfloat upd_hi ( v4caccfloat  ,
v2caccfloat   
)
v8caccfloat upd_hi ( v8caccfloat  ,
v4caccfloat   
)
v16caccfloat upd_hi ( v4caccfloat  ,
v8caccfloat   
)
v16acc48 upd_lo ( v16acc48  ,
v8acc48   
)
v8cacc48 upd_lo ( v8cacc48  ,
v4cacc48   
)
v8acc80 upd_lo ( v8acc80  ,
v4acc80   
)
v4cacc80 upd_lo ( v4cacc80  ,
v2cacc80   
)
v8accfloat upd_lo ( v8accfloat  ,
v4accfloat   
)
v16accfloat upd_lo ( v16accfloat  ,
v8accfloat   
)
v32accfloat upd_lo ( v32accfloat  ,
v16accfloat   
)
v4caccfloat upd_lo ( v4caccfloat  ,
v2caccfloat   
)
v8caccfloat upd_lo ( v8caccfloat  ,
v4caccfloat   
)
v16caccfloat upd_lo ( v4caccfloat  ,
v8caccfloat   
)
promotion v32int8 upd_v ( v32int8  ,
int  ,
v16int8   
)

Update a 256-bit vector 16 elements at a time using a 128-bit update.

v32int8 sbuff = undef_v32int8();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 16 elements in buf will be updated by val.
valInput vector used to update buf
promotion v32uint8 upd_v ( v32uint8  ,
int  ,
v16uint8   
)

Update a 256-bit vector 16 elements at a time using a 128-bit update.

v32uint8 sbuff = undef_v32uint8();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 16 elements in buf will be updated by val.
valInput vector used to update buf
promotion v16int16 upd_v ( v16int16  ,
int  ,
v8int16   
)

Update a 256-bit vector 8 elements at a time using a 128-bit update.

v16int16 sbuff = undef_v16int16();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 8 elements in buf will be updated by val.
valInput vector used to update buf
promotion v8cint16 upd_v ( v8cint16  ,
int  ,
v4cint16   
)

Update a 256-bit vector 4 elements at a time using a 128-bit update.

v8cint16 sbuff = undef_v8cint16();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 4 elements in buf will be updated by val.
valInput vector used to update buf
promotion v8int32 upd_v ( v8int32  ,
int  ,
v4int32   
)

Update a 256-bit vector 4 elements at a time using a 128-bit update.

v8int32 sbuff = undef_v8int32();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 4 elements in buf will be updated by val.
valInput vector used to update buf
promotion v4cint32 upd_v ( v4cint32  ,
int  ,
v2cint32   
)

Update a 256-bit vector 2 elements at a time using a 128-bit update.

v4cint32 sbuff = undef_v4cint32();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 2 elements in buf will be updated by val.
valInput vector used to update buf
promotion v8float upd_v ( v8float  ,
int  ,
v4float   
)

Update a 256-bit vector 4 elements at a time using a 128-bit update.

v8float sbuff = undef_v8float();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 4 elements in buf will be updated by val.
valInput vector used to update buf
promotion v4cfloat upd_v ( v4cfloat  ,
int  ,
v2cfloat   
)

Update a 256-bit vector 2 elements at a time using a 128-bit update.

v4cfloat sbuff = undef_v4cfloat();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 2 elements in buf will be updated by val.
valInput vector used to update buf
promotion v64int8 upd_v ( v64int8  ,
int  ,
v16int8   
)

Update a 512-bit vector 16 elements at a time using a 128-bit update.

v64int8 sbuff = undef_v64int8();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 16 elements in buf will be updated by val.
valInput vector used to update buf
promotion v64uint8 upd_v ( v64uint8  ,
int  ,
v16uint8   
)

Update a 512-bit vector 16 elements at a time using a 128-bit update.

v64uint8 sbuff = undef_v64uint8();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 16 elements in buf will be updated by val.
valInput vector used to update buf
promotion v32int16 upd_v ( v32int16  ,
int  ,
v8int16   
)

Update a 512-bit vector 8 elements at a time using a 128-bit update.

v32int16 sbuff = undef_v32int16();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 8 elements in buf will be updated by val.
valInput vector used to update buf
promotion v16cint16 upd_v ( v16cint16  ,
int  ,
v4cint16   
)

Update a 512-bit vector 4 elements at a time using a 128-bit update.

v16cint16 sbuff = undef_v16cint16();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 4 elements in buf will be updated by val.
valInput vector used to update buf
promotion v16int32 upd_v ( v16int32  ,
int  ,
v4int32   
)

Update a 512-bit vector 4 elements at a time using a 128-bit update.

v16int32 sbuff = undef_v16int32();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 4 elements in buf will be updated by val.
valInput vector used to update buf
promotion v8cint32 upd_v ( v8cint32  ,
int  ,
v2cint32   
)

Update a 512-bit vector 2 elements at a time using a 128-bit update.

v8cint32 sbuff = undef_v8cint32();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 2 elements in buf will be updated by val.
valInput vector used to update buf
promotion v16float upd_v ( v16float  ,
int  ,
v4float   
)

Update a 512-bit vector 4 elements at a time using a 128-bit update.

v16float sbuff = undef_v16float();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 4 elements in buf will be updated by val.
valInput vector used to update buf
promotion v8cfloat upd_v ( v8cfloat  ,
int  ,
v2cfloat   
)

Update a 512-bit vector 2 elements at a time using a 128-bit update.

v8cfloat sbuff = undef_v8cfloat();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 2 elements in buf will be updated by val.
valInput vector used to update buf
promotion v128int8 upd_v ( v128int8  ,
int  ,
v16int8   
)

Update a 1024-bit vector 16 elements at a time using a 128-bit update.

v128int8 sbuff = undef_v128int8();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 7 . Used to choose which set of 16 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v128uint8 upd_v ( v128uint8  ,
int  ,
v16uint8   
)

Update a 1024-bit vector 16 elements at a time using a 128-bit update.

v128uint8 sbuff = undef_v128uint8();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 7 . Used to choose which set of 16 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v64int16 upd_v ( v64int16  ,
int  ,
v8int16   
)

Update a 1024-bit vector 8 elements at a time using a 128-bit update.

v64int16 sbuff = undef_v64int16();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 7 . Used to choose which set of 8 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v32cint16 upd_v ( v32cint16  ,
int  ,
v4cint16   
)

Update a 1024-bit vector 4 elements at a time using a 128-bit update.

v32cint16 sbuff = undef_v32cint16();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 7 . Used to choose which set of 4 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v32int32 upd_v ( v32int32  ,
int  ,
v4int32   
)

Update a 1024-bit vector 4 elements at a time using a 128-bit update.

v32int32 sbuff = undef_v32int32();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 7 . Used to choose which set of 4 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v16cint32 upd_v ( v16cint32  ,
int  ,
v2cint32   
)

Update a 1024-bit vector 2 elements at a time using a 128-bit update.

v16cint32 sbuff = undef_v16cint32();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 7 . Used to choose which set of 2 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v32float upd_v ( v32float  ,
int  ,
v4float   
)

Update a 1024-bit vector 4 elements at a time using a 128-bit update.

v32float sbuff = undef_v32float();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 7 . Used to choose which set of 4 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v16cfloat upd_v ( v16cfloat  ,
int  ,
v2cfloat   
)

Update a 1024-bit vector 2 elements at a time using a 128-bit update.

v16cfloat sbuff = undef_v16cfloat();
sbuff = upd_v (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 7 . Used to choose which set of 2 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v64int8 upd_w ( v64int8  ,
int  ,
v32int8   
)

Update a 512-bit vector 32 elements at a time using a 256-bit update.

v64int8 sbuff = undef_v64int8();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 32 elements in buf will be updated by val.
valInput vector used to update buf
promotion v64uint8 upd_w ( v64uint8  ,
int  ,
v32uint8   
)

Update a 512-bit vector 32 elements at a time using a 256-bit update.

v64uint8 sbuff = undef_v64uint8();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 32 elements in buf will be updated by val.
valInput vector used to update buf
promotion v32int16 upd_w ( v32int16  ,
int  ,
v16int16   
)

Update a 512-bit vector 16 elements at a time using a 256-bit update.

v32int16 sbuff = undef_v32int16();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 16 elements in buf will be updated by val.
valInput vector used to update buf
promotion v16cint16 upd_w ( v16cint16  ,
int  ,
v8cint16   
)

Update a 512-bit vector 8 elements at a time using a 256-bit update.

v16cint16 sbuff = undef_v16cint16();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 8 elements in buf will be updated by val.
valInput vector used to update buf
promotion v16int32 upd_w ( v16int32  ,
int  ,
v8int32   
)

Update a 512-bit vector 8 elements at a time using a 256-bit update.

v16int32 sbuff = undef_v16int32();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 8 elements in buf will be updated by val.
valInput vector used to update buf
promotion v8cint32 upd_w ( v8cint32  ,
int  ,
v4cint32   
)

Update a 512-bit vector 4 elements at a time using a 256-bit update.

v8cint32 sbuff = undef_v8cint32();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 4 elements in buf will be updated by val.
valInput vector used to update buf
promotion v16float upd_w ( v16float  ,
int  ,
v8float   
)

Update a 512-bit vector 8 elements at a time using a 256-bit update.

v16float sbuff = undef_v16float();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 8 elements in buf will be updated by val.
valInput vector used to update buf
promotion v8cfloat upd_w ( v8cfloat  ,
int  ,
v4cfloat   
)

Update a 512-bit vector 4 elements at a time using a 256-bit update.

v8cfloat sbuff = undef_v8cfloat();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 4 elements in buf will be updated by val.
valInput vector used to update buf
promotion v128int8 upd_w ( v128int8  ,
int  ,
v32int8   
)

Update a 1024-bit vector 32 elements at a time using a 256-bit update.

v128int8 sbuff = undef_v128int8();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 32 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v128uint8 upd_w ( v128uint8  ,
int  ,
v32uint8   
)

Update a 1024-bit vector 32 elements at a time using a 256-bit update.

v128uint8 sbuff = undef_v128uint8();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 32 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v64int16 upd_w ( v64int16  ,
int  ,
v16int16   
)

Update a 1024-bit vector 16 elements at a time using a 256-bit update.

v64int16 sbuff = undef_v64int16();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 16 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v32cint16 upd_w ( v32cint16  ,
int  ,
v8cint16   
)

Update a 1024-bit vector 8 elements at a time using a 256-bit update.

v32cint16 sbuff = undef_v32cint16();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 8 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v32int32 upd_w ( v32int32  ,
int  ,
v8int32   
)

Update a 1024-bit vector 8 elements at a time using a 256-bit update.

v32int32 sbuff = undef_v32int32();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 8 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v16cint32 upd_w ( v16cint32  ,
int  ,
v4cint32   
)

Update a 1024-bit vector 4 elements at a time using a 256-bit update.

v16cint32 sbuff = undef_v16cint32();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 4 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v32float upd_w ( v32float  ,
int  ,
v8float   
)

Update a 1024-bit vector 8 elements at a time using a 256-bit update.

v32float sbuff = undef_v32float();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 8 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v16cfloat upd_w ( v16cfloat  ,
int  ,
v4cfloat   
)

Update a 1024-bit vector 4 elements at a time using a 256-bit update.

v16cfloat sbuff = undef_v16cfloat();
sbuff = upd_w (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 3 . Used to choose which set of 4 elements in buf will be updated by val. idx must be a compile time constant
valInput vector used to update buf
promotion v128int8 upd_x ( v128int8  ,
int  ,
v64int8   
)

Update a 1024-bit vector 64 elements at a time using a 512-bit update.

v128int8 sbuff = undef_v128int8();
sbuff = upd_x (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 64 elements in buf will be updated by val.
valInput vector used to update buf
promotion v128uint8 upd_x ( v128uint8  ,
int  ,
v64uint8   
)

Update a 1024-bit vector 64 elements at a time using a 512-bit update.

v128uint8 sbuff = undef_v128uint8();
sbuff = upd_x (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 64 elements in buf will be updated by val.
valInput vector used to update buf
promotion v64int16 upd_x ( v64int16  ,
int  ,
v32int16   
)

Update a 1024-bit vector 32 elements at a time using a 512-bit update.

v64int16 sbuff = undef_v64int16();
sbuff = upd_x (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 32 elements in buf will be updated by val.
valInput vector used to update buf
promotion v32cint16 upd_x ( v32cint16  ,
int  ,
v16cint16   
)

Update a 1024-bit vector 16 elements at a time using a 512-bit update.

v32cint16 sbuff = undef_v32cint16();
sbuff = upd_x (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 16 elements in buf will be updated by val.
valInput vector used to update buf
promotion v32int32 upd_x ( v32int32  ,
int  ,
v16int32   
)

Update a 1024-bit vector 16 elements at a time using a 512-bit update.

v32int32 sbuff = undef_v32int32();
sbuff = upd_x (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 16 elements in buf will be updated by val.
valInput vector used to update buf
promotion v16cint32 upd_x ( v16cint32  ,
int  ,
v8cint32   
)

Update a 1024-bit vector 8 elements at a time using a 512-bit update.

v16cint32 sbuff = undef_v16cint32();
sbuff = upd_x (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 8 elements in buf will be updated by val.
valInput vector used to update buf
promotion v32float upd_x ( v32float  ,
int  ,
v16float   
)

Update a 1024-bit vector 16 elements at a time using a 512-bit update.

v32float sbuff = undef_v32float();
sbuff = upd_x (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 16 elements in buf will be updated by val.
valInput vector used to update buf
promotion v16cfloat upd_x ( v16cfloat  ,
int  ,
v8cfloat   
)

Update a 1024-bit vector 8 elements at a time using a 512-bit update.

v16cfloat sbuff = undef_v16cfloat();
sbuff = upd_x (sbuff, idx, val);
Parameters
bufVector which will be updated
idxNumber in the range 0 to 1 . Used to choose which set of 8 elements in buf will be updated by val.
valInput vector used to update buf
promotion v128int8 yset_v ( int  ,
v16int8   
)

Creates a new 1024-bit vector with 16 elements already set using a 128-bit set.

sbuff = vset(idx, val);
Parameters
idxNumber in the range 0 to 7 . Used to choose which set of 16 elements in buf will be set to val. idx must be a compile time constant
valInput vector
promotion v128uint8 yset_v ( int  ,
v16uint8   
)

Creates a new 1024-bit vector with 16 elements already set using a 128-bit set.

sbuff = vset(idx, val);
Parameters
idxNumber in the range 0 to 7 . Used to choose which set of 16 elements in buf will be set to val. idx must be a compile time constant
valInput vector
promotion v64int16 yset_v ( int  ,
v8int16   
)

Creates a new 1024-bit vector with 8 elements already set using a 128-bit set.

sbuff = vset(idx, val);
Parameters
idxNumber in the range 0 to 7 . Used to choose which set of 8 elements in buf will be set to val. idx must be a compile time constant
valInput vector
promotion v32cint16 yset_v ( int  ,
v4cint16   
)

Creates a new 1024-bit vector with 4 elements already set using a 128-bit set.

sbuff = vset(idx, val);
Parameters
idxNumber in the range 0 to 7 . Used to choose which set of 4 elements in buf will be set to val. idx must be a compile time constant
valInput vector
promotion v32int32 yset_v ( int  ,
v4int32   
)

Creates a new 1024-bit vector with 4 elements already set using a 128-bit set.

sbuff = vset(idx, val);
Parameters
idxNumber in the range 0 to 7 . Used to choose which set of 4 elements in buf will be set to val. idx must be a compile time constant
valInput vector
promotion v16cint32 yset_v ( int  ,
v2cint32   
)

Creates a new 1024-bit vector with 2 elements already set using a 128-bit set.

sbuff = vset(idx, val);
Parameters
idxNumber in the range 0 to 7 . Used to choose which set of 2 elements in buf will be set to val. idx must be a compile time constant
valInput vector
promotion v32float yset_v ( int  ,
v4float   
)

Creates a new 1024-bit vector with 4 elements already set using a 128-bit set.

sbuff = vset(idx, val);
Parameters
idxNumber in the range 0 to 7 . Used to choose which set of 4 elements in buf will be set to val. idx must be a compile time constant
valInput vector
promotion v16cfloat yset_v ( int  ,
v2cfloat   
)

Creates a new 1024-bit vector with 2 elements already set using a 128-bit set.

sbuff = vset(idx, val);
Parameters
idxNumber in the range 0 to 7 . Used to choose which set of 2 elements in buf will be set to val. idx must be a compile time constant
valInput vector