Use transmute_unchecked in NonZero::new.

[reitermarkus]

Tracking issue: #120257

See #120521 (comment).

[rustbot]

r? @cuviper

rustbot has assigned @cuviper.
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

[Noratrieb]

cool!
@bors r+ rollup

[bors]

📌 Commit d70d320 has been approved by Nilstrieb

It is now in the queue for this repository.

[scottmcm]

Thanks! This should save a bunch of MIR.

Per https://rust.godbolt.org/z/Ecjsvd9MM, this makes it just

fn demo_new(_1: u32) -> f32 {
    debug n => _1;
    let mut _0: f32;

    bb0: {
        _0 = _1 as f32 (Transmute);
        return;
    }
}

instead of all this with the pointers

fn demo_old(_1: u32) -> f32 {
    debug n => _1;
    let mut _0: f32;
    let mut _2: *const f32;
    let mut _3: *const u32;
    scope 1 (inlined std::ptr::const_ptr::<impl *const u32>::cast::<f32>) {
        debug self => _3;
    }
    scope 2 (inlined std::ptr::read::<f32>) {
        debug src => _2;
        let mut _4: *const f32;
        scope 3 {
            scope 4 (inlined std::ptr::read::runtime::<f32>) {
                debug src => _4;
                scope 5 (inlined intrinsics::is_aligned_and_not_null::<f32>) {
                    debug ptr => _4;
                    scope 6 (inlined std::ptr::const_ptr::<impl *const f32>::is_null) {
                        debug self => _4;
                        let mut _5: *const u8;
                        scope 7 {
                            scope 8 (inlined std::ptr::const_ptr::<impl *const T>::is_null::runtime_impl) {
                                debug ptr => _5;
                                scope 9 (inlined std::ptr::const_ptr::<impl *const u8>::addr) {
                                    debug self => _5;
                                    scope 10 {
                                        scope 11 (inlined std::ptr::const_ptr::<impl *const u8>::cast::<()>) {
                                            debug self => _5;
                                        }
                                    }
                                }
                            }
                        }
                    }
                    scope 12 (inlined std::ptr::const_ptr::<impl *const f32>::is_aligned) {
                        debug self => _4;
                        scope 13 (inlined align_of::<f32>) {
                        }
                    }
                }
            }
        }
    }

    bb0: {
        StorageLive(_2);
        StorageLive(_3);
        _3 = &raw const _1;
        _2 = _3 as *const f32 (PtrToPtr);
        StorageDead(_3);
        StorageLive(_4);
        StorageLive(_5);
        _0 = (*_2);
        StorageDead(_5);
        StorageDead(_4);
        StorageDead(_2);
        return;
    }
}

[scottmcm]

Since the rollup showed perf changes, I'm curious

@rust-timer build 1476217

[rust-timer]

Finished benchmarking commit (1476217): comparison URL.

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	1.3%	[1.3%, 1.3%]	1
Improvements ✅ (primary)	-0.4%	[-0.8%, -0.2%]	6
Improvements ✅ (secondary)	-2.2%	[-2.2%, -2.2%]	1
All ❌✅ (primary)	-0.4%	[-0.8%, -0.2%]	6

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	7.4%	[4.3%, 10.5%]	2
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-3.2%	[-3.2%, -3.2%]	1
Improvements ✅ (secondary)	-5.0%	[-5.0%, -5.0%]	1
All ❌✅ (primary)	3.9%	[-3.2%, 10.5%]	3

Cycles

This benchmark run did not return any relevant results for this metric.

Binary size

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	0.1%	[0.0%, 0.4%]	8
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.1%	[-0.4%, -0.0%]	30
Improvements ✅ (secondary)	-1.2%	[-1.3%, -1.2%]	3
All ❌✅ (primary)	-0.0%	[-0.4%, 0.4%]	38

Bootstrap: 663.818s -> 665.912s (0.32%)
Artifact size: 308.01 MiB -> 308.06 MiB (0.01%)

[reitermarkus]

Seems like the -0.4% here pretty much undid the 0.5% regression, if I'm reading that correctly.