Difference between revisions of "Instruction Set/widensf"
From Mill Computing Wiki
(Created page with "{{DISPLAYTITLE:widensf}} <div style="font-size:80%;line-height:90%;margin-bottom:2em">realizing exu stream Decode|exu bloc...") | |||
| (4 intermediate revisions by the same user not shown) | |||
| Line 1: | Line 1: | ||
{{DISPLAYTITLE:widensf}} | {{DISPLAYTITLE:widensf}} | ||
| − | <div style="font-size:80%;line-height:90%;margin-bottom:2em">[[Speculation| | + | <div style="font-size:80%;line-height:90%;margin-bottom:2em">[[Speculation|speculable]] [[Encoding|exu stream]] [[Decode|exu block]] [[Phasing|compute phase]] operation [[Domains|in the signed fixed point value domain]] <br /> |
'''native on:''' [[Cores|all]]<br /> | '''native on:''' [[Cores|all]]<br /> | ||
</div> | </div> | ||
| − | widen to | + | Double the scalar width of a signed fixed point value. |
| + | |||
| + | The reason a special widen for fixed point is needed is because the fixed point is defined relative to the most significant side. Widening increases precision, but doesn't extend the value range. | ||
| + | |||
| + | The natively available byte widths on all [[Cores]] are 1, 2, 4, 8, and on the high end also 16. | ||
| + | |||
---- | ---- | ||
| − | <code style="font-size:130%"><b style="color:#050">widensf</b>(<span style="color:#009">[[Domains#sf|sf]]</span> <span title="belt operand from ops window"> | + | <code style="font-size:130%"><b style="color:#050">widensf</b>(<span style="color:#009">[[Domains#sf|sf]]</span> <span title="belt operand from ops window">op0</span>, <i><span style="color:#009">[[Immediates#width|width]]</span> <span title="data width and vector length (exu)">width0</span></i>) → [[Domains#sf|sf]] r<sub>0</sub></code> |
<div style="font-size:80%;line-height:90%;margin-bottom:2em">'''operands:''' [[Operands#likeWiden|like Widen xx:2x]] | <div style="font-size:80%;line-height:90%;margin-bottom:2em">'''operands:''' [[Operands#likeWiden|like Widen xx:2x]] | ||
</div> | </div> | ||
| Line 14: | Line 19: | ||
! [[Cores|Core]] || [[Slot|In Slots]]|| [[Latency|Latencies]] | ! [[Cores|Core]] || [[Slot|In Slots]]|| [[Latency|Latencies]] | ||
|- | |- | ||
| − | | [[Cores/Tin/Encoding# | + | | [[Cores/Tin/Encoding#widensf|Tin]] || E0 || |
|- | |- | ||
| − | | [[Cores/Copper/Encoding# | + | | [[Cores/Copper/Encoding#widensf|Copper]] || E0 || |
|- | |- | ||
| − | | [[Cores/Silver/Encoding# | + | | [[Cores/Silver/Encoding#widensf|Silver]] || E0 E1 E2 E3 || |
|- | |- | ||
| − | | [[Cores/Gold/Encoding# | + | | [[Cores/Gold/Encoding#widensf|Gold]] || E0 || |
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
|} | |} | ||
| + | |||
| + | |||
| + | [[Instruction_Set|Instruction Set, alphabetical]], [[Instruction Set by Category]], [http://millcomputing.com/instructions.html?collapse=7#ops Instruction Set, sortable, filterable] | ||
Latest revision as of 14:10, 23 February 2021
speculable exu stream exu block compute phase operation in the signed fixed point value domain
native on: all
Double the scalar width of a signed fixed point value.
The reason a special widen for fixed point is needed is because the fixed point is defined relative to the most significant side. Widening increases precision, but doesn't extend the value range.
The natively available byte widths on all Cores are 1, 2, 4, 8, and on the high end also 16.
widensf(sf op0, width width0) → sf r0
operands: like Widen xx:2x
| Core | In Slots | Latencies |
|---|---|---|
| Tin | E0 | |
| Copper | E0 | |
| Silver | E0 E1 E2 E3 | |
| Gold | E0 |
Instruction Set, alphabetical, Instruction Set by Category, Instruction Set, sortable, filterable
