Skip to content
This repository was archived by the owner on Dec 22, 2021. It is now read-only.

Commit c46c5cf

Browse files
authored
Small cleanup of simd exec/instructions (#401)
- `extend` can have types as params, we don't need to take the bitwidth - bunch of places using `i_1^N` or `i_2^N` can be changed to `i^\ast`
1 parent 890c043 commit c46c5cf

File tree

1 file changed

+14
-14
lines changed

1 file changed

+14
-14
lines changed

document/core/exec/instructions.rst

Lines changed: 14 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -396,7 +396,7 @@ SIMD instructions are defined in terms of generic numeric operators applied lane
396396

397397
5. Let :math:`t_2` be the type :math:`\unpacked(t_1\K{x}N)`.
398398

399-
6. Let :math:`c_2` be the result of computing :math:`\extend^{sx^?}_{|t_1|,|t_2|}(i^\ast[x])`.
399+
6. Let :math:`c_2` be the result of computing :math:`\extend^{sx^?}_{t_1,t_2}(i^\ast[x])`.
400400

401401
7. Push the value :math:`t_2.\CONST~c_2` to the stack.
402402

@@ -408,7 +408,7 @@ SIMD instructions are defined in terms of generic numeric operators applied lane
408408
\\ \qquad
409409
\begin{array}[t]{@{}r@{~}l@{}}
410410
(\iff & t_2 = \unpacked(t_1\K{x}N) \\
411-
\wedge & c_2 = \extend^{sx^?}_{|t_1|,|t_2|}(\lanes_{t_1\K{x}N}(c_1)[x])
411+
\wedge & c_2 = \extend^{sx^?}_{t_1,t_2}(\lanes_{t_1\K{x}N}(c_1)[x])
412412
\end{array}
413413
\end{array}
414414
@@ -430,9 +430,9 @@ SIMD instructions are defined in terms of generic numeric operators applied lane
430430

431431
6. Pop the value :math:`\V128.\VCONST~c_2` from the stack.
432432

433-
7. Let :math:`i_2^\ast` be the sequence :math:`\lanes_{\shape}(c_2)`.
433+
7. Let :math:`i^\ast` be the sequence :math:`\lanes_{\shape}(c_2)`.
434434

435-
8. Let :math:`c` be the result of computing :math:`\lanes^{-1}_{\shape}(i_2^\ast \with [x] = c_1)`
435+
8. Let :math:`c` be the result of computing :math:`\lanes^{-1}_{\shape}(i^\ast \with [x] = c_1)`
436436

437437
9. Push :math:`\V128.\VCONST~c` on the stack.
438438

@@ -443,8 +443,8 @@ SIMD instructions are defined in terms of generic numeric operators applied lane
443443
\end{array}
444444
\\ \qquad
445445
\begin{array}[t]{@{}r@{~}l@{}}
446-
(\iff & i_2^\ast = \lanes_{\shape}(c_2)) \\
447-
\wedge & c = \lanes^{-1}_{\shape}(i_2^\ast \with [x] = c_1)
446+
(\iff & i^\ast = \lanes_{\shape}(c_2)) \\
447+
\wedge & c = \lanes^{-1}_{\shape}(i^\ast \with [x] = c_1)
448448
\end{array}
449449
\end{array}
450450
@@ -680,9 +680,9 @@ SIMD instructions are defined in terms of generic numeric operators applied lane
680680

681681
2. Pop the value :math:`\V128.\VCONST~c_1` from the stack.
682682

683-
3. Let :math:`i_1^N` be the sequence :math:`\lanes_{t_1\K{x}M}(c_1)[0 \slice N]`.
683+
3. Let :math:`i^\ast` be the sequence :math:`\lanes_{t_1\K{x}M}(c_1)[0 \slice N]`.
684684

685-
4. Let :math:`c` be the result of computing :math:`\lanes^{-1}_{t_2\K{x}N}((\extend^{\sx}_{t_1,t_2}(i_1))^N)`
685+
4. Let :math:`c` be the result of computing :math:`\lanes^{-1}_{t_2\K{x}N}(\extend^{\sx}_{t_1,t_2}(i^\ast))`
686686

687687
5. Push the value :math:`\V128.\VCONST~c` onto the stack.
688688

@@ -693,8 +693,8 @@ SIMD instructions are defined in terms of generic numeric operators applied lane
693693
\end{array}
694694
\\ \qquad
695695
\begin{array}[t]{@{}r@{~}l@{}}
696-
(\iff & i_1^N = \lanes_{t_1\K{x}M}(c_1)[0 \slice N] \\
697-
\wedge & c = \lanes^{-1}_{t_2\K{x}N}((\extend^{\sx}_{M,N}(i_1))^N)
696+
(\iff & i^\ast = \lanes_{t_1\K{x}M}(c_1)[0 \slice N] \\
697+
\wedge & c = \lanes^{-1}_{t_2\K{x}N}(\extend^{\sx}_{M,N}(i^\ast))
698698
\end{array}
699699
\end{array}
700700
@@ -706,9 +706,9 @@ SIMD instructions are defined in terms of generic numeric operators applied lane
706706

707707
2. Pop the value :math:`\V128.\VCONST~c_1` from the stack.
708708

709-
3. Let :math:`i_1^N` be the sequence :math:`\lanes_{t_1\K{x}M}(c_1)[N \slice N]`.
709+
3. Let :math:`i^\ast` be the sequence :math:`\lanes_{t_1\K{x}M}(c_1)[N \slice N]`.
710710

711-
4. Let :math:`c` be the result of computing :math:`\lanes^{-1}_{t_2\K{x}N}((\extend^{\sx}_{t_1,t_2}(i_1))^N)`
711+
4. Let :math:`c` be the result of computing :math:`\lanes^{-1}_{t_2\K{x}N}(\extend^{\sx}_{t_1,t_2}(i^\ast))`
712712

713713
5. Push the value :math:`\V128.\VCONST~c` onto the stack.
714714

@@ -719,8 +719,8 @@ SIMD instructions are defined in terms of generic numeric operators applied lane
719719
\end{array}
720720
\\ \qquad
721721
\begin{array}[t]{@{}r@{~}l@{}}
722-
(\iff & i_1^N = \lanes_{t_1\K{x}M}(c_1)[N \slice N] \\
723-
\wedge & c = \lanes^{-1}_{t_2\K{x}N}((\extend^{\sx}_{M,N}(i_1))^N)
722+
(\iff & i^\ast = \lanes_{t_1\K{x}M}(c_1)[N \slice N] \\
723+
\wedge & c = \lanes^{-1}_{t_2\K{x}N}(\extend^{\sx}_{M,N}(i^\ast))
724724
\end{array}
725725
\end{array}
726726

0 commit comments

Comments
 (0)