minor fix in num_bits and bump pthash

jermp · Jun 21, 2024 · 2d1c27b · 2d1c27b
1 parent 11dc8f2
commit 2d1c27b
Show file tree

Hide file tree

Showing 5 changed files with 12 additions and 9 deletions.
diff --git a/benchmarks/README.md b/benchmarks/README.md
@@ -30,7 +30,7 @@ The query times are relative to the following configuration:
 
 ### Space in bits/kmer
 
-| Dictionary |elegans ||| cod   ||| kesterl ||| human |||
+| Dictionary |elegans ||| cod   ||| kestrel ||| human |||
 |:------------------|:------:|:------:|:------:|:------:|:------:|:------:|:------:|:------:|:------:|:-----:|:-------:|:-----:|
 |                   | k=31 | k=47 | k=63 | k=31 | k=47 | k=63 | k=31 | k=47 | k=63 | k=31 | k=47 | k=63 |
 | SSHash, **regular**   | 5.86 | 4.29 | 3.51 | 7.84 | 5.17 | 4.26 | 7.53 | 4.67 | 3.76 | 8.70 | 5.65 | 4.64 |

diff --git a/external/pthash b/external/pthash
diff --git a/include/info.cpp b/include/info.cpp
@@ -24,8 +24,9 @@ uint64_t skew_index::print_info() const {
 }
 
 void dictionary::print_space_breakdown() const {
-    std::cout << "total index size: " << essentials::convert((num_bits() + 7) / 8, essentials::MB)
-              << " [MB]" << '\n';
+    const uint64_t num_bytes = (num_bits() + 7) / 8;
+    std::cout << "total index size: " << num_bytes << " [B] -- "
+              << essentials::convert(num_bytes, essentials::MB) << " [MB]" << '\n';
     std::cout << "SPACE BREAKDOWN:\n";
     std::cout << "  minimizers: " << static_cast<double>(m_minimizers.num_bits()) / size()
               << " [bits/kmer]\n";

diff --git a/include/skew_index.hpp b/include/skew_index.hpp
@@ -33,7 +33,9 @@ struct skew_index {
 
     uint64_t num_bits() const {
         uint64_t n =
-            (sizeof(min_log2) + sizeof(max_log2) + sizeof(log2_max_num_super_kmers_in_bucket)) * 8;
+            (sizeof(min_log2) + sizeof(max_log2) + sizeof(log2_max_num_super_kmers_in_bucket) +
+             2 * sizeof(size_t) /* for std::vector::size */) *
+            8;
         for (uint64_t partition_id = 0; partition_id != mphfs.size(); ++partition_id) {
             auto const& mphf = mphfs[partition_id];
             auto const& P = positions[partition_id];

diff --git a/src/common.hpp b/src/common.hpp
@@ -12,11 +12,11 @@ void random_kmer(char* kmer, uint64_t k) {
 }
 
 void load_dictionary(dictionary& dict, std::string const& index_filename, bool verbose) {
-    uint64_t num_bytes_read = essentials::load(dict, index_filename.c_str());
+    const uint64_t num_bytes_read = essentials::load(dict, index_filename.c_str());
     if (verbose) {
-        std::cout << "index size: " << essentials::convert(num_bytes_read, essentials::MB)
-                  << " [MB] (" << (num_bytes_read * 8.0) / dict.size() << " [bits/kmer])"
-                  << std::endl;
+        std::cout << "total index size: " << num_bytes_read << " [B] -- "
+                  << essentials::convert(num_bytes_read, essentials::MB) << " [MB] ("
+                  << (num_bytes_read * 8.0) / dict.size() << " [bits/kmer])" << std::endl;
         dict.print_info();
     }
 }
+4 −0		README.md
+12 −4		include/builders/internal_memory_builder_partitioned_phf.hpp
+17 −27		include/builders/internal_memory_builder_single_phf.hpp
+22 −1		include/builders/util.hpp
+3 −3		include/encoders/util.hpp
+19 −1		include/partitioned_phf.hpp
+12 −1		include/single_phf.hpp
+6 −2		include/utils/bucketers.hpp
+4 −3		src/example.cpp