library: optimize truncate method

[xtqqczze]

Closes #76089

Slightly better code when the elements have drop glue, see #74172 (comment).

Slightly worse code for truncate(0), see #78884 (comment). However, this is mitigated by the new clippy::manual_clear lint which is warn-by-default.

[rustbot]

r? @Mark-Simulacrum

rustbot has assigned @Mark-Simulacrum.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: libs
• libs expanded to 7 candidates
• Random selection from Mark-Simulacrum, joboet

[ZuseZ4]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 0b2c3b8 (0b2c3b8072bcdf1845525eef84dd5e719d76118f, parent: 6368fd52cb9f230dfb156097625993e7a8891800)

[rust-timer]

Finished benchmarking commit (0b2c3b8): comparison URL.

Overall result: no relevant changes - no action needed

Benchmarking means the PR may be perf-sensitive. Consider adding rollup=never if this change is not fit for rolling up.

@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

This perf run didn't have relevant results for this metric.

Max RSS (memory usage)

This perf run didn't have relevant results for this metric.

Cycles

Results (primary 2.3%, secondary 2.4%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	2.3%	[2.3%, 2.3%]	1
Regressions ❌ (secondary)	2.4%	[2.4%, 2.4%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.3%	[2.3%, 2.3%]	1

Binary size

Results (primary -0.0%, secondary -0.0%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.0%	[-0.0%, -0.0%]	4
Improvements ✅ (secondary)	-0.0%	[-0.1%, -0.0%]	14
All ❌✅ (primary)	-0.0%	[-0.0%, -0.0%]	4

Bootstrap: 511.654s -> 511.381s (-0.05%)
Artifact size: 400.70 MiB -> 400.67 MiB (-0.01%)

[mati865]

The same, slightly improved assembly is generated if you remove or equal from the original condition. I.e. turnif new_len <= self.len() { into if new_len < self.len() { which is equiva to your change.

[xtqqczze]

The same, slightly improved assembly is generated if you remove or equal from the original condition. I.e. turnif new_len <= self.len() { into if new_len < self.len() { which is equiva to your change.

You’re correct that changing <= to < in the original condition yields the same improvement. I opted for the early-return style instead, for consistency and to better align with the intent expressed in the comments.

[Mark-Simulacrum]

Can you post the diff you're expecting to see for this? Looking at the playground link from the description (https://play.rust-lang.org/?version=nightly&mode=release&edition=2024&gist=2e44cfb473fe7ed00c00c3993baf9768) this appers to be the diff.

But this doesn't seem like it's really changing the performance? The core loop is entirely unchanged; the function header loses an extra branch, but that's about it... I doubt that matters that much. I don't think we should bother with tweaks like this personally without benchmarks to substantiate them (and even then it seems unlikely that we'd see an effect outside a microbenchmark).

--- /tmp/old  
+++ /tmp/new
@@ -1,4 +1,4 @@
-playground::foo:     # @playground::foo
+playground::mytr:    # @playground::mytr
 # %bb.0:
        pushq   %r15
        pushq   %r14
@@ -7,25 +7,23 @@
        pushq   %rax
        movq    16(%rdi), %r14
        cmpq    %r14, %rsi
-       ja      .LBB0_4
+       jae     .LBB1_3
 # %bb.1:
        movq    %rsi, %rbx
        movq    8(%rdi), %r15
        movq    %rsi, 16(%rdi)
-       je      .LBB0_4
-# %bb.2:
        movq    __rustc::__rust_dealloc@GOTPCREL(%rip), %r12

-.LBB0_3:                                # =>This Inner Loop Header: Depth=1
+.LBB1_2:                                # =>This Inner Loop Header: Depth=1
        movq    (%r15,%rbx,8), %rdi
        movl    $4, %esi
        movl    $4, %edx
        callq   *%r12
        incq    %rbx
        cmpq    %rbx, %r14
-       jne     .LBB0_3
+       jne     .LBB1_2

-.LBB0_4:
+.LBB1_3:
        addq    $8, %rsp
        popq    %rbx
        popq    %r12