Fix benchmarks

This commit is contained in:
Brian Anderson 2012-12-13 18:23:56 -08:00
parent 786c143a70
commit 9677fff0f9
2 changed files with 1 additions and 208 deletions

View File

@ -378,7 +378,7 @@ fn validate(edges: ~[(node_id, node_id)],
log(info, ~"Verifying tree and graph edges...");
let status = do par::alli(tree) |u, v| {
let status = do par::alli(tree) |u, v, copy edges| {
let u = u as node_id;
if *v == -1i64 || u == root {
true

View File

@ -1,207 +0,0 @@
// Copyright 2012 The Rust Project Developers. See the COPYRIGHT
// file at the top-level directory of this distribution and at
// http://rust-lang.org/COPYRIGHT.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
// xfail-pretty (extra blank line is inserted in vec::mapi call)
// multi tasking k-nucleotide
extern mod std;
use std::map;
use std::map::HashMap;
use std::sort;
use io::ReaderUtil;
use cmp::Ord;
// given a map, print a sorted version of it
fn sort_and_fmt(mm: HashMap<~[u8], uint>, total: uint) -> ~str {
fn pct(xx: uint, yy: uint) -> float {
return (xx as float) * 100f / (yy as float);
}
pure fn le_by_val<TT: Copy Ord, UU: Copy Ord>(kv0: &(TT,UU),
kv1: &(TT,UU)) -> bool {
let (_, v0) = *kv0;
let (_, v1) = *kv1;
return v0 >= v1;
}
pure fn le_by_key<TT: Copy Ord, UU: Copy Ord>(kv0: &(TT,UU),
kv1: &(TT,UU)) -> bool {
let (k0, _) = *kv0;
let (k1, _) = *kv1;
return k0 <= k1;
}
// sort by key, then by value
fn sortKV<TT: Copy Ord, UU: Copy Ord>(orig: ~[(TT,UU)]) -> ~[(TT,UU)] {
return sort::merge_sort(sort::merge_sort(orig, le_by_key), le_by_val);
}
let mut pairs = ~[];
// map -> [(k,%)]
mm.each(fn&(key: ~[u8], val: uint) -> bool {
pairs.push((key, pct(val, total)));
return true;
});
let pairs_sorted = sortKV(pairs);
let mut buffer = ~"";
for pairs_sorted.each |kv| {
let (k,v) = *kv;
unsafe {
buffer += (fmt!("%s %0.3f\n", str::to_upper(str::raw::from_bytes(k)), v));
}
}
return buffer;
}
// given a map, search for the frequency of a pattern
fn find(mm: HashMap<~[u8], uint>, key: ~str) -> uint {
match mm.find(str::to_bytes(str::to_lower(key))) {
option::None => { return 0u; }
option::Some(num) => { return num; }
}
}
// given a map, increment the counter for a key
fn update_freq(mm: HashMap<~[u8], uint>, key: &[u8]) {
let key = vec::slice(key, 0, key.len());
mm.update(key, 1, |v,v1| { v+v1 });
}
// given a ~[u8], for each window call a function
// i.e., for "hello" and windows of size four,
// run it("hell") and it("ello"), then return "llo"
fn windows_with_carry(bb: &[u8], nn: uint,
it: fn(window: &[u8])) -> ~[u8] {
let mut ii = 0u;
let len = vec::len(bb);
while ii < len - (nn - 1u) {
it(vec::view(bb, ii, ii+nn));
ii += 1u;
}
return vec::slice(bb, len - (nn - 1u), len);
}
fn make_sequence_processor(sz: uint, from_parent: comm::Port<~[u8]>,
to_parent: comm::Chan<~str>) {
let freqs: HashMap<~[u8], uint> = map::HashMap();
let mut carry: ~[u8] = ~[];
let mut total: uint = 0u;
let mut line: ~[u8];
loop {
line = comm::recv(from_parent);
if line == ~[] { break; }
carry = windows_with_carry(carry + line, sz, |window| {
update_freq(freqs, window);
total += 1u;
});
}
let buffer = match sz {
1u => { sort_and_fmt(freqs, total) }
2u => { sort_and_fmt(freqs, total) }
3u => { fmt!("%u\t%s", find(freqs, ~"GGT"), ~"GGT") }
4u => { fmt!("%u\t%s", find(freqs, ~"GGTA"), ~"GGTA") }
6u => { fmt!("%u\t%s", find(freqs, ~"GGTATT"), ~"GGTATT") }
12u => { fmt!("%u\t%s", find(freqs, ~"GGTATTTTAATT"), ~"GGTATTTTAATT") }
18u => { fmt!("%u\t%s", find(freqs, ~"GGTATTTTAATTTATAGT"), ~"GGTATTTTAATTTATAGT") }
_ => { ~"" }
};
//comm::send(to_parent, fmt!("yay{%u}", sz));
comm::send(to_parent, buffer);
}
// given a FASTA file on stdin, process sequence THREE
fn main() {
let args = os::args();
let rdr = if os::getenv(~"RUST_BENCH").is_some() {
// FIXME: Using this compile-time env variable is a crummy way to
// get to this massive data set, but include_bin! chokes on it (#2598)
let path = Path(env!("CFG_SRC_DIR"))
.push_rel(&Path("src/test/bench/shootout-k-nucleotide.data"));
result::get(&io::file_reader(&path))
} else {
io::stdin()
};
// initialize each sequence sorter
let sizes = ~[1u,2u,3u,4u,6u,12u,18u];
let from_child = vec::map (sizes, |_sz| comm::Port() );
let to_parent = vec::mapi(sizes, |ii, _sz| comm::Chan(&from_child[ii]) );
let to_child = vec::mapi(sizes, |ii, sz| {
let ii = ii;
let sz = *sz;
do task::spawn_listener |from_parent| {
make_sequence_processor(sz, from_parent, to_parent[ii]);
}
});
// latch stores true after we've started
// reading the sequence of interest
let mut proc_mode = false;
while !rdr.eof() {
let line: ~str = rdr.read_line();
if str::len(line) == 0u { loop; }
match (line[0], proc_mode) {
// start processing if this is the one
('>' as u8, false) => {
match str::find_str_from(line, ~"THREE", 1u) {
option::Some(_) => proc_mode = true,
option::None => ()
}
}
// break our processing
('>' as u8, true) => { break; }
// process the sequence for k-mers
(_, true) => {
let line_bytes = str::to_bytes(line);
for sizes.eachi |ii, _sz| {
let mut lb = line_bytes;
comm::send(to_child[ii], lb);
}
}
// whatever
_ => { }
}
}
// finish...
for sizes.eachi |ii, _sz| {
comm::send(to_child[ii], ~[]);
}
// now fetch and print result messages
for sizes.eachi |ii, _sz| {
io::println(comm::recv(from_child[ii]));
}
}