Added benchmarks for memory and time requirements relative to buildin…

…g corpus
LucaCappelletti94 · Apr 14, 2024 · 8afa3f3 · 8afa3f3
1 parent b6c01fe
commit 8afa3f3
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 16 deletions.
diff --git a/benchmarks/README.md b/benchmarks/README.md
@@ -249,7 +249,7 @@ In the new edition we also provide a parallel version, which has the same memory
 | OLD     | 5     | 20,336    | 9,583,720,360 |
 | NEW     | 6     | 3,893,922 | 615,458,920   |
 | NEWPAR  | 6     | 163,489   | 615,458,920   |
-| OLD     | 6     | 22,206    | 10,211,711,214|
+| OLD     | 6     | 22,206    | 10,211,711,214|ments relative to building corpus)
 
 ## Benchmarks 5 April 2024, 08:00 PM
 The sixth benchmark was run on a 6-core machine with 32 GBs of RAM. We loaded the entirety of the taxons dataset into memory.

diff --git a/benchmarks/src/main.rs b/benchmarks/src/main.rs
@@ -10,7 +10,6 @@ use core::fmt::Debug;
 use mem_dbg::*;
 use ngrammatic::prelude::*;
 use rayon::prelude::*;
-use sux::dict::rear_coded_list::{RearCodedList, RearCodedListBuilder};
 
 /// Returns an iterator over the taxons in the corpus.
 fn iter_taxons() -> impl Iterator<Item = String> {
@@ -129,6 +128,10 @@ fn load_corpus_old(arity: usize) -> ngrammatic_old::Corpus {
     corpus
 }
 
+/// We allow dead code here because the version of the
+/// webgraph crate that is necessary for this benchmark
+/// is currently in nightly.
+#[allow(dead_code)]
 fn load_corpus_webgraph<NG>()
 where
     NG: Ngram<G = ASCIIChar> + Debug,
@@ -156,6 +159,10 @@ where
     );
 }
 
+/// We allow dead code here because the version of the
+/// webgraph crate that is necessary for this benchmark
+/// is currently in nightly.
+#[allow(dead_code)]
 fn load_corpus_rcl_webgraph<NG>()
 where
     NG: Ngram<G = ASCIIChar> + Debug,
@@ -199,18 +206,12 @@ where
 
 fn main() {
     env_logger::builder().try_init().unwrap();
-    // experiment::<UniGram<ASCIIChar>>();
-    // experiment::<BiGram<ASCIIChar>>();
-    // experiment::<TriGram<ASCIIChar>>();
-    // experiment::<TetraGram<ASCIIChar>>();
-    // experiment::<PentaGram<ASCIIChar>>();
-    // experiment::<HexaGram<ASCIIChar>>();
-    // experiment::<HeptaGram<ASCIIChar>>();
-    // experiment::<OctaGram<ASCIIChar>>();
-    use ngrammatic::prelude::*;
-    let mut animals: Vec<String> = iter_taxons().collect();
-
-    let corpus: Corpus<Vec<String>, TriGram<char>, Lowercase> = Corpus::par_from(animals);
-
-    corpus.mem_dbg(DbgFlags::default() | DbgFlags::CAPACITY | DbgFlags::HUMANIZE).unwrap();
+    experiment::<UniGram<ASCIIChar>>();
+    experiment::<BiGram<ASCIIChar>>();
+    experiment::<TriGram<ASCIIChar>>();
+    experiment::<TetraGram<ASCIIChar>>();
+    experiment::<PentaGram<ASCIIChar>>();
+    experiment::<HexaGram<ASCIIChar>>();
+    experiment::<HeptaGram<ASCIIChar>>();
+    experiment::<OctaGram<ASCIIChar>>();
 }