fmt of non-decimal radix untangled #143730

pascaldekloe · 2025-07-10T12:51:01Z

Have the implementation match its decimal counterpart.

Digit table instead of conversion functions
Correct buffer size per radix
Elimination of dead code for negative
No trait abstraction for integers

Original Performance

    fmt::write_10ints_bin                                                393.03ns/iter      +/- 1.41
    fmt::write_10ints_hex                                                316.84ns/iter      +/- 1.49
    fmt::write_10ints_oct                                                327.16ns/iter      +/- 0.46

Patched Performance

    fmt::write_10ints_bin                                                392.31ns/iter      +/- 3.05
    fmt::write_10ints_hex                                                302.41ns/iter      +/- 5.48
    fmt::write_10ints_oct                                                322.01ns/iter      +/- 3.82

r? tgross35

pascaldekloe · 2025-07-10T12:58:31Z

https://rust.godbolt.org/z/K1hWxa7dP

bors · 2025-07-17T08:43:24Z

☔ The latest upstream changes (presumably #144044) made this pull request unmergeable. Please resolve the merge conflicts.

pascaldekloe · 2025-07-18T13:37:22Z

Can you either have a look or reassign @tgross35? Got a followup pending on this one...

tgross35 · 2025-07-18T19:02:16Z

I'll reassign for now, but if nobody beats me to it I'll take a look on Monday

r? libs

(leaving myself assigned so it stays on my list)

tgross35 · 2025-07-18T21:14:44Z

By the way; if you are interested, we could use some int (and float) formatting and parsing benchmarks at https://github.com/rust-lang/rustc-perf/blob/2120e3b7b8996e96858b88edefea371679a3d415/collector/runtime-benchmarks/fmt/src/main.rs (I just learned that is possible), which would mean they get run as part of our pretty extensive perf infra rather than you needing to post the results of local runs.

pascaldekloe · 2025-07-18T21:56:23Z

Interesting indeed @tgross35. The specialized benchmarks we have at the moment (such as write_u8_min and write_u64_max) won't scale. Think about how many benchmarks we need for fmt::LowerExp with edge cases on alignment and precision. I am curious to hear what you think of the _10ints_ concept in this patch. The idea is to have a balanced mix of common cases.

…t output

* correct buffer size * no trait abstraction * similar to decimal

tgross35

Few notes, haven't looked at patches 3 & 4 yet

tgross35 · 2025-07-23T08:59:01Z

library/core/src/fmt/num.rs

-    fn to_u64(&self) -> u64;
-    fn to_u128(&self) -> u128;
-}
+// Formatting of integers with a non-decimal radix.


Nit, might as well make it a doc comment

Suggested change

// Formatting of integers with a non-decimal radix.

/// Formatting of integers with a non-decimal radix.

tgross35 · 2025-07-23T09:08:51Z

library/core/src/fmt/num.rs

+    (fmt::$Trait:ident for $T:ident -> $Radix:ident) => {
        #[stable(feature = "rust1", since = "1.0.0")]
        impl fmt::$Trait for $T {
            fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-                $Radix.fmt_int(*self as $U, f)
+                $Radix.fmt_int(*self, f)
            }
        }
    };


Maybe int_base could do both signed and unsigned rather than having the impls in integer?

(fmt::$Trait:ident for ($Int:ident, $Uint:ident) -> $Radix:ident) => { #[stable(feature = "rust1", since = "1.0.0")] impl fmt::$Trait for $Uint { fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { $Radix.fmt_int(*self as $U, f) $Radix.fmt_int(*self, f) } } #[stable(feature = "rust1", since = "1.0.0")] impl fmt::$Trait for $Int { fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { fmt::$Trait::fmt(&self.cast_unsigned(), f) } } };

tgross35 · 2025-07-23T09:12:24Z

library/core/src/fmt/num.rs

+        #[stable(feature = "rust1", since = "1.0.0")]
+        impl fmt::Binary for $Int {
+            fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+                fmt::Binary::fmt(&(*self as $Uint), f)


Mentioned above but we now have .cast_unsigned() which is slightly cleaner than as $Uint

tgross35 · 2025-07-23T09:24:24Z

library/coretests/benches/fmt.rs

+#[bench]
+fn write_12ints_bin(bh: &mut Bencher) {
+    bh.iter(|| {
+        black_box(format!("{:b}", black_box(0_u8)));
+        black_box(format!("{:b}", black_box(100_i8)));
+        black_box(format!("{:b}", black_box(-100_i8)));
+
+        black_box(format!("{:b}", black_box(0_u32)));
+        black_box(format!("{:b}", black_box(1000_i32)));
+        black_box(format!("{:b}", black_box(-1000_i32)));
+
+        black_box(format!("{:b}", black_box(0_u64)));
+        black_box(format!("{:b}", black_box(10000_i64)));
+        black_box(format!("{:b}", black_box(-10000_i64)));
+
+        black_box(format!("{:b}", black_box(0_u128)));
+        black_box(format!("{:b}", black_box(100000_i128)));
+        black_box(format!("{:b}", black_box(-100000_i128)));
+    });
+}


I think it would be good to break up benchmarks by integer size because that can be pretty platform-dependent, e.g. u64 might be about the same as u32 on 64-bit platforms, but significantly slower on 32-bit.

Testing a few assorted values seems good though. To get an even balance it may also be good to check:

1 << n for n 0..U::BITS, both signed and unsigned

10.pow(n) for n = 0 up to the max that fits

I::MIN, I::MAX, U::MAX

These can be collected outside of bb.iter so the cost of computing values isn't included in the benchmarks

This is preexisting but it would also be better to create a String::with_capacity(enough_for_u128), within bb.iter just .clear() and write! into it. Keeps time to alloc/free from skewing benchmark results.

rustbot assigned tgross35 Jul 10, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels Jul 10, 2025

pascaldekloe force-pushed the fmt-radix-trim branch from 2281143 to d593c6f Compare July 10, 2025 16:05

This comment has been minimized.

Sign in to view

pascaldekloe force-pushed the fmt-radix-trim branch from d593c6f to f0f1c53 Compare July 10, 2025 16:31

This comment has been minimized.

Sign in to view

pascaldekloe force-pushed the fmt-radix-trim branch from f0f1c53 to 76b8c3a Compare July 10, 2025 17:18

pascaldekloe force-pushed the fmt-radix-trim branch from 76b8c3a to 110d35e Compare July 17, 2025 14:05

rustbot assigned ibraheemdev and unassigned tgross35 Jul 18, 2025

tgross35 self-assigned this Jul 18, 2025

pascaldekloe added 4 commits July 22, 2025 16:41

fmt of non-decimals is always non-negative due to the two's-complemen…

7fa846e

…t output

fmt benchmarks for binary, octal and hex

8fdc969

fmt with table lookup for binary, octal and hex

16fcd95

* correct buffer size * no trait abstraction * similar to decimal

fmt::DisplayInt abstraction obsolete with better macro

2c6210d

pascaldekloe force-pushed the fmt-radix-trim branch from 110d35e to 2c6210d Compare July 22, 2025 14:51

tgross35 reviewed Jul 23, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fmt of non-decimal radix untangled #143730

fmt of non-decimal radix untangled #143730

pascaldekloe commented Jul 10, 2025

Uh oh!

pascaldekloe commented Jul 10, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

bors commented Jul 17, 2025

Uh oh!

pascaldekloe commented Jul 18, 2025

Uh oh!

tgross35 commented Jul 18, 2025 •

edited

Loading

Uh oh!

tgross35 commented Jul 18, 2025

Uh oh!

pascaldekloe commented Jul 18, 2025

Uh oh!

tgross35 left a comment

Uh oh!

tgross35 Jul 23, 2025

Uh oh!

tgross35 Jul 23, 2025

Uh oh!

tgross35 Jul 23, 2025

Uh oh!

tgross35 Jul 23, 2025

Uh oh!

tgross35 Jul 23, 2025

Uh oh!

Uh oh!

	// Formatting of integers with a non-decimal radix.
	/// Formatting of integers with a non-decimal radix.

fmt of non-decimal radix untangled #143730

Are you sure you want to change the base?

fmt of non-decimal radix untangled #143730

Conversation

pascaldekloe commented Jul 10, 2025

Original Performance

Patched Performance

Uh oh!

pascaldekloe commented Jul 10, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

bors commented Jul 17, 2025

Uh oh!

pascaldekloe commented Jul 18, 2025

Uh oh!

tgross35 commented Jul 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tgross35 commented Jul 18, 2025

Uh oh!

pascaldekloe commented Jul 18, 2025

Uh oh!

tgross35 left a comment

Choose a reason for hiding this comment

Uh oh!

tgross35 Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

tgross35 Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

tgross35 Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

tgross35 Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

tgross35 Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

tgross35 commented Jul 18, 2025 •

edited

Loading