Changes

Jump to: navigation, search

GPUSquad

9,518 bytes added, 16:26, 22 February 2018
m
Idea 2 - LZW Compression
PROFILING WITH THE ORIGINAL TEXT:
[[File:Flat profile:
 
Each sample counts as 0.01 seconds.
% cumulative self self total
time seconds seconds calls ns/call ns/call name
46.15 0.24 0.24 compress(std::string, int, std::string)
25.00 0.37 0.13 7954538 16.34 16.34 show_usage()
21.15 0.48 0.11 2091647 52.59 52.59 convert_int_to_bin(int)
7.69 0.52 0.04 2091647 19.12 35.47 std::__detail::_Map_base<std::string, std::pair<std::string const, int>, std::_Select1st<std::pair<std::string const, int> >, true, std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true> >::operator[](std::string const&)
0.00 0.52 0.00 3841 0.00 0.00 std::__detail::_Hashtable_iterator<std::pair<std::string const, int>, false, false> std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true>::_M_insert_bucket<std::pair<std::string, unsigned int> >(std::pair<std::string, unsigned int>&&, unsigned int, unsigned int)
0.00 0.52 0.00 256 0.00 0.00 std::__detail::_Hashtable_iterator<std::pair<std::string const, int>, false, false> std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true>::_M_insert_bucket<std::pair<std::string, int> >(std::pair<std::string, int>&&, unsigned int, unsigned int)
0.00 0.52 0.00 256 0.00 16.34 std::__detail::_Map_base<std::string, std::pair<std::string const, int>, std::_Select1st<std::pair<std::string const, int> >, true, std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true> >::operator[](std::string&&)
0.00 0.52 0.00 1 0.00 0.00 _GLOBAL__sub_I__Z18convert_int_to_bini]]
 
 
PROFILING WITH TWICE THE TEXT:
 
Flat profile:
 
Each sample counts as 0.01 seconds.
% cumulative self self total
time seconds seconds calls ns/call ns/call name
38.71 0.24 0.24 compress(std::string, int, std::string)
25.81 0.40 0.16 2091647 76.49 76.49 convert_int_to_bin(int)
22.58 0.54 0.14 7954538 17.60 17.60 show_usage()
9.68 0.60 0.06 2091647 28.69 46.29 std::__detail::_Map_base<std::string, std::pair<std::string const, int>, std::_Select1st<std::pair<std::string const, int> >, true, std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true> >::operator[](std::string const&)
3.23 0.62 0.02 convert_char_to_string(char const*, int)
0.00 0.62 0.00 3841 0.00 0.00 std::__detail::_Hashtable_iterator<std::pair<std::string const, int>, false, false> std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true>::_M_insert_bucket<std::pair<std::string, unsigned int> >(std::pair<std::string, unsigned int>&&, unsigned int, unsigned int)
0.00 0.62 0.00 256 0.00 0.00 std::__detail::_Hashtable_iterator<std::pair<std::string const, int>, false, false> std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true>::_M_insert_bucket<std::pair<std::string, int> >(std::pair<std::string, int>&&, unsigned int, unsigned int)
0.00 0.62 0.00 256 0.00 17.60 std::__detail::_Map_base<std::string, std::pair<std::string const, int>, std::_Select1st<std::pair<std::string const, int> >, true, std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true> >::operator[](std::string&&)
0.00 0.62 0.00 1 0.00 0.00 _GLOBAL__sub_I__Z18convert_int_to_bini
 
 
PROFILING WITH 4X THE TEXT:
 
Flat profile:
 
Each sample counts as 0.01 seconds.
% cumulative self self total
time seconds seconds calls us/call us/call name
40.48 1.02 1.02 compress(std::string, int, std::string)
30.16 1.78 0.76 31802927 0.02 0.02 show_usage()
22.62 2.35 0.57 8363660 0.07 0.07 convert_int_to_bin(int)
5.56 2.49 0.14 8363660 0.02 0.04 std::__detail::_Map_base<std::string, std::pair<std::string const, int>, std::_Select1st<std::pair<std::string const, int> >, true, std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true> >::operator[](std::string const&)
0.79 2.51 0.02 convert_char_to_string(char const*, int)
0.40 2.52 0.01 256 39.06 39.09 std::__detail::_Map_base<std::string, std::pair<std::string const, int>, std::_Select1st<std::pair<std::string const, int> >, true, std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true> >::operator[](std::string&&)
0.00 2.52 0.00 3841 0.00 0.00 std::__detail::_Hashtable_iterator<std::pair<std::string const, int>, false, false> std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true>::_M_insert_bucket<std::pair<std::string, unsigned int> >(std::pair<std::string, unsigned int>&&, unsigned int, unsigned int)
0.00 2.52 0.00 256 0.00 0.00 std::__detail::_Hashtable_iterator<std::pair<std::string const, int>, false, false> std::_Hashtable<std::string, std::pair<std::string const, int>, std::allocator<std::pair<std::string const, int> >, std::_Select1st<std::pair<std::string const, int> >, std::equal_to<std::string>, std::hash<std::string>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, false, false, true>::_M_insert_bucket<std::pair<std::string, int> >(std::pair<std::string, int>&&, unsigned int, unsigned int)
0.00 2.52 0.00 1 0.00 0.00 _GLOBAL__sub_I__Z18convert_int_to_bini
 
 
A useful hotspot for parallelization is not immediately obvious through profiling, since the main compress() function contains the bulk of the program logic, and its show_usage(, and convert_int_to_bin are simple functions, that are called frequently in one for loop.
 
What really affects the runtime of the program based on data is the extent to which long matching strings can be tokenized. When this happens, larger, and larger chunks of text can be processed and compressed in an iteration. This is reflected by the fact that pasting the same blocks of text over again to increase data size does not proportionally increase run time because the same tokens work for subsequent pasted blocks.
41
edits

Navigation menu