Testing Beamium (a Prometheus to Warp10 metrics forwarder) at commit 819ef8c:
cargo +nightly-2019-12-18 build
534,97s user 30,78s system 604% cpu 1:33,65 total
cargo +nightly-2019-12-17 build
517,03s user 26,19s system 577% cpu 1:34,11 total
cargo +nightly-2019-12-18 build --release
1319,95s user 29,98s system 515% cpu 4:22,00 total
cargo +nightly-2019-12-17 build --release
1356,64s user 28,18s system 496% cpu 4:39,15 total
cargo +nightly-2019-12-18 check
341,53s user 26,41s system 689% cpu 53,353 total
cargo +nightly-2019-12-17 check
338,01s user 22,66s system 654% cpu 55,137 total
On Rust stable, just in case (1.40.0):
cargo build
526,18s user 27,23s system 619% cpu 1:29,33 total
cargo build --release
1322,03s user 27,14s system 520% cpu 4:19,12 total
cargo check
354,53s user 23,55s system 630% cpu 1:00,01 total
Building substrate 5dece712a on AMD EPYC Rome (64 cores/128 threads), 256 GB machine.
cargo clean && time cargo +nightly-2019-12-17 build
Finished dev [unoptimized + debuginfo] target(s) in 6m 15s
real 6m17.934s
user 135m2.361s
sys 17m1.909s
cargo clean && time cargo +nightly-2019-12-18 build
Finished dev [unoptimized + debuginfo] target(s) in 6m 29s
real 6m32.575s
user 228m14.238s
sys 32m0.462s
cargo clean && time cargo +nightly-2019-12-17 build --release
Finished release [optimized] target(s) in 8m 38s
real 8m41.842s
user 411m59.930s
sys 23m30.786s
cargo clean && time cargo +nightly-2019-12-18 build --release
Finished release [optimized] target(s) in 8m 45s
real 8m48.119s
user 537m58.740s
sys 45m5.127s
cargo clean && time cargo +nightly-2019-12-17 check
Finished dev [unoptimized + debuginfo] target(s) in 4m 46s
real 4m49.572s
user 154m7.090s
sys 20m57.464s
cargo clean && time cargo +nightly-2019-12-18 check
Finished dev [unoptimized + debuginfo] target(s) in 4m 26s
real 4m29.848s
user 210m40.669s
sys 39m16.688s
Running cargo +nightly-2019-12-18 check with RUSTFLAGS="-Zthreads=64" I am getting:
Finished dev [unoptimized + debuginfo] target(s) in 20m 54s
real 20m57.729s
user 746m42.182s
sys 1532m35.727s
The second time I ran it I got
thread '<unnamed>' panicked at 'called `Result::unwrap()` on an `Err` value: "SendError(..)"', src/libcore/result.rs:1188:5
stack backtrace:
0: backtrace::backtrace::libunwind::trace
at /cargo/registry/src/github.com-1ecc6299db9ec823/backtrace-0.3.40/src/backtrace/libunwind.rs:88
1: backtrace::backtrace::trace_unsynchronized
at /cargo/registry/src/github.com-1ecc6299db9ec823/backtrace-0.3.40/src/backtrace/mod.rs:66
2: std::sys_common::backtrace::_print_fmt
at src/libstd/sys_common/backtrace.rs:84
3: <std::sys_common::backtrace::_print::DisplayBacktrace as core::fmt::Display>::fmt
at src/libstd/sys_common/backtrace.rs:61
4: core::fmt::write
at src/libcore/fmt/mod.rs:1025
5: std::io::Write::write_fmt
at src/libstd/io/mod.rs:1426
6: std::sys_common::backtrace::_print
at src/libstd/sys_common/backtrace.rs:65
7: std::sys_common::backtrace::print
at src/libstd/sys_common/backtrace.rs:50
8: std::panicking::default_hook::{{closure}}
at src/libstd/panicking.rs:193
9: std::panicking::default_hook
at src/libstd/panicking.rs:210
10: rustc_driver::report_ice
11: std::panicking::rust_panic_with_hook
at src/libstd/panicking.rs:475
12: rust_begin_unwind
at src/libstd/panicking.rs:375
13: core::panicking::panic_fmt
at src/libcore/panicking.rs:84
14: core::result::unwrap_failed
at src/libcore/result.rs:1188
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
error: internal compiler error: unexpected panic
note: the compiler unexpectedly panicked. this is a bug.
note: we would appreciate a bug report: https://github.com/rust-lang/rust/blob/master/CONTRIBUTING.md#bug-reports
note: rustc 1.41.0-nightly (3ed3b8bb7 2019-12-17) running on x86_64-unknown-linux-gnu
note: compiler flags: -Z threads=64 -C debuginfo=2 --crate-type lib
note: some of the compiler flags provided by cargo are hidden
query stack during panic:
end of query stack
Testing sv-parser at 6e887e2.
The testing environment is Intel Xeon Gold 6134 CPU @ 3.20GHz (16cores/32threads) / 1.5TB DRAM.
About 15-20% improved.
cargo clean && time cargo +nightly-2019-12-17 build
real 2m35.962s
user 5m6.225s
sys 0m21.336s
cargo clean && time cargo +nightly-2019-12-18 build
real 2m10.055s
user 5m40.738s
sys 0m26.435s
cargo clean && time cargo +nightly-2019-12-17 build --release
real 2m49.236s
user 16m11.879s
sys 1m29.441s
cargo clean && time cargo +nightly-2019-12-18 build --release
real 2m24.627s
user 17m15.008s
sys 1m26.060s
cargo clean && time cargo +nightly-2019-12-17 check
real 2m5.321s
user 2m55.150s
sys 0m13.223s
cargo clean && time cargo +nightly-2019-12-18 check
real 1m41.439s
user 3m26.564s
sys 0m15.532s
On my 2-cpu 8-thread server (and old xeon E5), all of my projects were faster with the 2019-12-18 nightly (across all projects, best was 1.51 times faster for check --release, worst as 1.01 times faster for another check --release). In no cases was it slower.
Adding another set of data points, I just grabbed a few repos people were using in this thread. Seeing awesome improvements, nice job to the team working on this.
cargo +nightly-2019-12-17 build --release 6592.05s user 372.31s system 2733% cpu 4:14.79 total
cargo +nightly-2019-12-18 build --release 7017.31s user 436.78s system 2791% cpu 4:27.08 total