Difference between revisions of "Instruction Set/merge"
From Mill Computing Wiki
| Line 4: | Line 4: | ||
</div> | </div> | ||
| − | + | Bitwise merge. | |
| + | |||
| + | It works like a multiplex: the bits in the mask operand decide whether the corresponding bit of the result is taken from op0 or op1.<br /> | ||
| + | Or explained differently: it does what [[Instruction_Set/pick|pick]] does on the operand level on the bit level. | ||
| + | |||
---- | ---- | ||
<code style="font-size:130%"><b style="color:#050">merge</b>(<span style="color:#009">[[Domains#u|u]]</span> <span title="belt operand from ops window">mask</span>, <span style="color:#009">[[Domains#u|u]]</span> <span title="belt operand from ops window">bits0</span>, <span style="color:#009">[[Domains#u|u]]</span> <span title="belt operand from ops | <code style="font-size:130%"><b style="color:#050">merge</b>(<span style="color:#009">[[Domains#u|u]]</span> <span title="belt operand from ops window">mask</span>, <span style="color:#009">[[Domains#u|u]]</span> <span title="belt operand from ops window">bits0</span>, <span style="color:#009">[[Domains#u|u]]</span> <span title="belt operand from ops | ||
Revision as of 18:50, 20 December 2014
Bitwise merge.
It works like a multiplex: the bits in the mask operand decide whether the corresponding bit of the result is taken from op0 or op1.
Or explained differently: it does what pick does on the operand level on the bit level.
merge(u mask, u bits0, u bits1) → op r0
operands: like Identity [xx:x]
encoding:
merge(op op0)
,
exuArgs(op arg0, op arg1)
| Core | In Slots | Latencies |
|---|---|---|
| Tin | E0 | 1 |
| Copper | E0 E1 | 1 |
| Silver | E0 E1 E2 E3 | 1 |
| Gold | E0 E1 E2 E3 E4 E5 E6 E7 | 1 |
| Decimal8 | E0 E1 E2 E3 | 1 |
| Decimal16 | E0 E1 E2 E3 | 1 |
Instruction Set, alphabetical, Instruction Set by Category, Instruction Set, sortable, filterable
