(Note: I also tested instead of chaining on single elements, chaining the slice half and half, and qualitatively it has the same result)

0-benchmark-result

Improvements with impl TrustedLen on Chain, and using TrustedLen in Vec::extend

name                            extend-before-1.log ns/iter  extend-after-1.log ns/iter    diff ns/iter   diff %
bench_chain_chain_collect       62,249                       32,370                             -29,879  -48.00%
bench_chain_collect             26,828                       25,386                              -1,442   -5.37%
bench_nest_chain_chain_collect  44,562                       41,087                              -3,475   -7.80%

/Additional/ improvements using `.fold()` too in Vec::extend

name                            extend-after-1.log ns/iter  extend-after-fold-2.log ns/iter    diff ns/iter   diff %
bench_chain_chain_collect       32,370                      2,762                                   -29,608  -91.47%
bench_chain_collect             25,386                      2,760                                   -22,626  -89.13%
bench_nest_chain_chain_collect  41,087                      2,744                                   -38,343  -93.32%

bench-chain.rs

#[bench]
fn bench_chain_collect(b: &mut test::Bencher) {
    let data = black_box([0; LEN]);
    b.iter(|| {
        data.iter().cloned().chain([1].iter().cloned()).collect::<Vec<_>>()
    });
}

#[bench]
fn bench_chain_chain_collect(b: &mut test::Bencher) {
    let data = black_box([0; LEN]);
    b.iter(|| {
        data.iter().cloned().chain([1].iter().cloned()).chain([2].iter().cloned()).collect::<Vec<_>>()
    });
}

#[bench]
fn bench_nest_chain_chain_collect(b: &mut test::Bencher) {
    let data = black_box([0; LEN]);
    b.iter(|| {
        data.iter().cloned().chain([1].iter().chain([2].iter()).cloned()).collect::<Vec<_>>()
    });
}