"A prebuild HTML of the documentation can be found at ../Tsunami-Simulation/docs_sphinx/prebuild."
[x] Can't find it in the submission; it is in the repo though.
"There are changes needed to be done to the F-Wave solver, [...]"
[x] Do you mean "no changes"?
"Otherwise, three loaded values would be lost if the arrays are aligned correctly. If the cache line is smaller than 128 bit"
[x] It should be bytes not bits for the cache line size, no?
I guess you'd need alignment of the base address and a zero padding in the fast dimension if the cells.
We have a dedicated optimization phase where this could be studied.
[x] Did you observer speedups with your approach?
#ifdef TEST
std::ifstream l_file( "resources/config.test.json" );
#endif // TEST
#ifndef TEST
[x] Its better to prepend you preprocessor defines with a unique identifier so that you don't run into issues where some external
include (or similar) defines TEST, e.g., something like PP_TSUNAMI_LAB_TEST.
// create a thread that runs the stations write function
std::thread writeStationsThread( writeStations, &l_stations, l_waveProp );
[ ... ]
[x] I guess we have a misunderstanding here.
The output frequency was supposed to be in relation to the simulation time and not wall clock time.
But asynchronous I/O is nice-to-have after all ;-)
(tsunami_lab) tsunami_19@LeChuck:~/04_two_dimensional/gerlach_hofer/Tsunami-Simulation/build$ cmake --config Debug --target sanitize ..
CMake Warning:
Ignoring extra path from command line:
"/home/tsunami_19/04_two_dimensional/gerlach_hofer/Tsunami-Simulation/build/Debug"
CMake Warning:
Ignoring extra path from command line:
"sanitize"
CMake Error: Unknown argument --target
CMake Error: Run 'cmake --help' for all supported options.
[x] Am I doing something wrong here?
(tsunami_lab) tsunami_19@LeChuck:~/04_two_dimensional/gerlach_hofer/Tsunami-Simulation/build$ ./sanitize_test
=================================================================
==287180==ERROR: AddressSanitizer: stack-use-after-return on address 0x7f0641abde9c at pc 0x556d052c7990 bp 0x7ffc9d799b20 sp 0x7ffc9d799b10
READ of size 4 at 0x7f0641abde9c thread T0
#0 0x556d052c798f in tsunami_lab::io::Stations::write(float const*) /home/tsunami_19/04_two_dimensional/gerlach_hofer/Tsunami-Simulation/src/io/Stations.cpp:92
#1 0x556d058107a1 in C_A_T_C_H_T_E_S_T_4 /home/tsunami_19/04_two_dimensional/gerlach_hofer/Tsunami-Simulation/src/io/Stations.test.cpp:51
#2 0x556d053f88c2 in Catch::TestInvokerAsFunction::invoke() const /home/tsunami_19/04_two_dimensional/gerlach_hofer/Tsunami-Simulation/submodules/Catch2/single_include/catch2/catch.hpp:14330
[x] The sanitizers are unhappy with the unit tests.
REQUIERED INPUT:
N_CELLS_X is the number of cells in x-direction.
[x] Typo.
/**
* Gets the combined height of bathymetry and water height
*
* @return combined height
*/
t_real const* getTotalHeight()
{
for( t_idx i = 1; i < m_yCells + 1; i++ )
{
[x] There's no need to have the implementation in the header.
"A prebuild HTML of the documentation can be found at ../Tsunami-Simulation/docs_sphinx/prebuild."
"There are changes needed to be done to the F-Wave solver, [...]"
"Otherwise, three loaded values would be lost if the arrays are aligned correctly. If the cache line is smaller than 128 bit"
[x] It should be bytes not bits for the cache line size, no? I guess you'd need alignment of the base address and a zero padding in the fast dimension if the cells. We have a dedicated optimization phase where this could be studied.
[x] Did you observer speedups with your approach?