Mike Gerwitz
0a8fb71c1b
This was an oversight. The difference is significant. I had my suspicions
about this when I noticed the huge difference in time between writing to
/dev/null vs. an actual file during profiling.
On one of our systems, here's the number of syscalls _before_ this change:
$ strace -c target/release/tameld --emit xmle -o foo foo.xmlo
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
85.05 4.966192 16 318473 write
7.23 0.421977 13 32298 lstat
6.53 0.381424 15 25113 read
0.75 0.043691 13 3350 readlink
0.25 0.014713 61 241 close
0.12 0.007167 30 241 openat
0.05 0.003175 151 21 munmap
0.01 0.000488 14 35 brk
0.01 0.000292 9 33 mmap
0.00 0.000266 38 7 mremap
0.00 0.000004 1 3 sigaltstack
0.00 0.000000 0 6 fstat
0.00 0.000000 0 1 poll
0.00 0.000000 0 11 mprotect
0.00 0.000000 0 7 rt_sigaction
0.00 0.000000 0 1 rt_sigprocmask
0.00 0.000000 0 6 6 access
0.00 0.000000 0 1 execve
0.00 0.000000 0 1 arch_prctl
0.00 0.000000 0 1 sched_getaffinity
0.00 0.000000 0 1 set_tid_address
0.00 0.000000 0 1 set_robust_list
0.00 0.000000 0 2 prlimit64
------ ----------- ----------- --------- --------- ----------------
100.00 5.839389 379854 6 total
And _after_:
$ strace -c target/release/tameld --emit xmle -o foo foo.xmlo
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
45.21 0.435010 13 32298 lstat
40.09 0.385752 15 25113 read
6.14 0.059113 21 2809 write
4.75 0.045687 14 3350 readlink
2.51 0.024115 100 241 close
0.84 0.008045 33 241 openat
0.26 0.002468 118 21 munmap
0.06 0.000580 17 35 brk
0.06 0.000566 17 33 mmap
0.03 0.000279 40 7 mremap
0.02 0.000181 16 11 mprotect
0.01 0.000087 15 6 6 access
0.01 0.000082 12 7 rt_sigaction
0.01 0.000075 13 6 fstat
0.00 0.000027 9 3 sigaltstack
0.00 0.000024 12 2 prlimit64
0.00 0.000018 18 1 execve
0.00 0.000016 16 1 poll
0.00 0.000013 13 1 sched_getaffinity
0.00 0.000012 12 1 rt_sigprocmask
0.00 0.000012 12 1 arch_prctl
0.00 0.000012 12 1 set_robust_list
0.00 0.000011 11 1 set_tid_address
------ ----------- ----------- --------- --------- ----------------
100.00 0.962185 64190 6 total
What a difference!
There's still a lot of other red flags in there; those can be addressed
separately.
This was originally written as I was learning Rust, and I suspect that I
didn't realize that File wasn't buffered at the time.
For the above link: times go from 1.23s pre-change to 0.85s after:
0.77user 0.44system 0:01.23elapsed 99%CPU (0avgtext+0avgdata 48520maxresident)k
0inputs+43952outputs (0major+12825minor)pagefaults 0swaps
0.69user 0.15system 0:00.85elapsed 98%CPU (0avgtext+0avgdata 48396maxresident)k
0inputs+43952outputs (0major+12823minor)pagefaults 0swaps
|
||
|---|---|---|
| .. | ||
| benches | ||
| build-aux | ||
| src | ||
| tests | ||
| .gitignore | ||
| Cargo.lock | ||
| Cargo.toml | ||
| Makefile.am | ||
| README.md | ||
| autogen.sh | ||
| bootstrap | ||
| configure.ac | ||
| rustfmt.toml | ||
README.md
TAME in Rust (TAMER)
TAME was written to help tame the complexity of developing comparative insurance rating systems. This project aims to tame the complexity and performance issues of TAME itself. TAMER is therefore more tame than TAME.
TAME was originally written in XSLT. For more information about the
project, see the parent README.md.
Building
To bootstrap from the source repository, run ./bootstrap.
To configure the build for your system, run ./configure. To build, run
make. To run tests, run make check.
You may also invoke cargo directly, which make will do for you using
options provided to configure.
Note that the default development build results in terrible runtime performance! See [#Build Flags][] below for instructions on how to generate a release binary.
Build Flags
The environment variable CARGO_BUILD_FLAGS can be used to provide
additional arguments to cargo build when invoked via make. This can be
provided optionally during configure and can be overridden when invoking
make. For example:
# release build
$ ./configure && make CARGO_BUILD_FLAGS=--release
$ ./configure CARGO_BUILD_FLAGS=--release && make
# dev build
$ ./configure && make
$ ./configure CARGO_BUILD_FLAGS=--release && make CARGO_BUILD_FLAGS=
Hacking
This section contains advice for those developing TAMER.
Running Tests
Developers should be using test-driven development (TDD). make check will
run all necessary tests.
Code Format
Rust provides rustfmt that can automatically format code for you. This
project mandates its use and therefore eliminates personal preference in
code style (for better or worse).
Formatting checks are run during make check and, on failure, will output
the diff that would be applied if you ran make fmt (or make fix); this
will run cargo fmt for you (and will use the binaries configured via
configure).
Since developers should be doing test-driven development (TDD) and therefore
should be running make check frequently, the hope is that frequent
feedback on formatting issues will allow developers to quickly adjust their
habits to avoid triggering formatting errors at all.
If you want to automatically fix formatting errors and then run tests:
$ make fmt check
Benchmarking
Benchmarks serve two purposes: external integration tests (which are subject
to module visibility constraints) and actual benchmarking. To run
benchmarks, invoke make bench.
Note that link-time optimizations (LTO) are performed on the binary for benchmarking so that its performance reflects release builds that will be used in production.
The configure script will automatically detect whether the test feature
is unstable (as it was as of the time of writing) and, if so, will
automatically fall back to invoking nightly (by running cargo +nightly bench).
If you do not have nightly, run you install it via rustup install nightly.