chore: update cycle counts in docs

docs/src/user_docs/assembly/u32_operations.md

@@ -103,11 +103,11 @@ If the error code is omitted, the default value of $0$ is assumed.
 | u32shl <br> - *(18 cycles)* <br> u32shl.*b* <br> - *(3 cycles)*   | [b, a, ...]    | [c, ...]      | $c \leftarrow (a \cdot 2^b) \mod 2^{32}$ <br> Undefined if $a \ge 2^{32}$ or $b > 31$                                          |
 | u32shr <br> - *(18 cycles)* <br> u32shr.*b* <br> - *(3 cycles)*   | [b, a, ...]    | [c, ...]      | $c \leftarrow \lfloor a/2^b \rfloor$ <br> Undefined if $a \ge 2^{32}$ or $b > 31$                                              |
 | u32rotl <br> - *(18 cycles)* <br> u32rotl.*b* <br> - *(3 cycles)* | [b, a, ...]    | [c, ...]      | Computes $c$ by rotating a 32-bit representation of $a$ to the left by $b$ bits. <br> Undefined if $a \ge 2^{32}$ or $b > 31$  |
-| u32rotr <br> - *(22 cycles)* <br> u32rotr.*b* <br> - *(3 cycles)* | [b, a, ...]    | [c, ...]      | Computes $c$ by rotating a 32-bit representation of $a$ to the right by $b$ bits. <br> Undefined if $a \ge 2^{32}$ or $b > 31$ |
+| u32rotr <br> - *(23 cycles)* <br> u32rotr.*b* <br> - *(3 cycles)* | [b, a, ...]    | [c, ...]      | Computes $c$ by rotating a 32-bit representation of $a$ to the right by $b$ bits. <br> Undefined if $a \ge 2^{32}$ or $b > 31$ |
 | u32popcnt <br> - *(33 cycles)*                                              | [a, ...]       | [b, ...]      | Computes $b$ by counting the number of set bits in $a$ (hamming weight of $a$). <br> Undefined if $a \ge 2^{32}$               |
-| u32clz <br> - *(37 cycles)*                                                     | [a, ...]    | [b, ...]      | Computes $b$ as a number of leading zeros of $a$. <br> Undefined if $a \ge 2^{32}$               |
+| u32clz <br> - *(42 cycles)*                                                     | [a, ...]    | [b, ...]      | Computes $b$ as a number of leading zeros of $a$. <br> Undefined if $a \ge 2^{32}$               |
 | u32ctz <br> - *(34 cycles)*                                                     | [a, ...]    | [b, ...]      | Computes $b$ as a number of trailing zeros of $a$. <br> Undefined if $a \ge 2^{32}$               |
-| u32clo <br> - *(36 cycles)*                                                     | [a, ...]    | [b, ...]      | Computes $b$ as a number of leading ones of $a$. <br> Undefined if $a \ge 2^{32}$               |
+| u32clo <br> - *(41 cycles)*                                                     | [a, ...]    | [b, ...]      | Computes $b$ as a number of leading ones of $a$. <br> Undefined if $a \ge 2^{32}$               |
 | u32cto <br> - *(33 cycles)*                                                     | [a, ...]    | [b, ...]      | Computes $b$ as a number of trailing ones of $a$. <br> Undefined if $a \ge 2^{32}$               |
 
 

stdlib/asm/math/u64.masm

@@ -568,7 +568,7 @@ end
 #! error.
 #! Stack transition looks as follows:
 #! [b, a_hi, a_lo, ...] -> [c_hi, c_lo, ...], where c = a << b mod 2^64.
-#! This takes 40 cycles.
+#! This takes 44 cycles.
 export.rotr
     push.31
     dup.1
@@ -609,7 +609,7 @@ end
 #! The input value is assumed to be represented using 32 bit limbs, but this is not checked.
 #! Stack transition looks as follows:
 #! [n_hi, n_lo, ...] -> [clz, ...], where clz is a number of leading zeros of value n.
-#! This takes 43 cycles.
+#! This takes 48 cycles.
 export.clz
     dup.0
     eq.0
@@ -650,7 +650,7 @@ end
 #! The input value is assumed to be represented using 32 bit limbs, but this is not checked.
 #! Stack transition looks as follows:
 #! [n_hi, n_lo, ...] -> [clo, ...], where clo is a number of leading ones of value n.
-#! This takes 42 cycles.
+#! This takes 47 cycles.
 export.clo
     dup.0
     eq.4294967295