Added splines for camera movement, fixed threading conflict with writ…

…e_frame(), more accurate runtime estimation, enabled FP16S by default

DOCUMENTATION.md

@@ -313,6 +313,39 @@
   }
   ```
 - To find suitable camera placement, run the simulation at low resolution in [`INTERACTIVE_GRAPHICS`](src/defines.hpp) mode, rotate/move the camera to the desired position, click the <kbd>Mouse</kbd> to disable mouse rotation, and press <kbd>G</kbd> to print the current camera settings as a copy-paste command in the console. <kbd>Alt</kbd>+<kbd>Tab</kbd> to the console and copy the camera placement command by selecting it with the mouse and right-clicking, then paste it into the [`main_setup()`](src/setup.cpp) function.
+- To fly the camera along a smooth path through a list of provided keyframe camera placements, use `catmull_rom` splines:
+  ```c
+  while(lbm.get_t()<=lbm_T) { // main simulation loop
+  	if(lbm.graphics.next_frame(lbm_T, 30.0f)) {
+  		const float t = (float)lbm.get_t()/(float)lbm_T;
+  		vector<float3> camera_positions = {
+  			float3(-0.282220f*(float)Nx,  0.529221f*(float)Ny,  0.304399f*(float)Nz),
+  			float3( 0.806921f*(float)Nx,  0.239912f*(float)Ny,  0.436880f*(float)Nz),
+  			float3( 1.129724f*(float)Nx, -0.130721f*(float)Ny,  0.352759f*(float)Nz),
+  			float3( 0.595601f*(float)Nx, -0.504690f*(float)Ny,  0.203096f*(float)Nz),
+  			float3(-0.056776f*(float)Nx, -0.591919f*(float)Ny, -0.416467f*(float)Nz)
+  		};
+  		vector<float> camera_rx = {
+  			 116.0f,
+  			  25.4f,
+  			 -10.6f,
+  			 -45.6f,
+  			 -94.6f
+  		};
+  		vector<float> camera_ry = {
+  			  26.0f,
+  			  33.3f,
+  			  20.3f,
+  			  25.3f,
+  			 -16.7f
+  		};
+  		const float camera_fov = 90.0f;
+  		lbm.graphics.set_camera_free(catmull_rom(camera_positions, t), catmull_rom(camera_rx, t), catmull_rom(camera_ry, t), camera_fov);
+  		lbm.graphics.write_frame(get_exe_path()+"export/");
+  	}
+  	lbm.run(1u, lbm_T);
+  }
+  ```
 - The visualization mode(s) can be specified as `lbm.graphics.visualization_modes` with the [`VIS_...`](src/defines.hpp) macros. You can also set the `lbm.graphics.slice_mode` (`0`=no slice, `1`=x, `2`=y, `3`=z, `4`=xz, `5`=xyz, `6`=yz, `7`=xy) and reposition the slices with `lbm.graphics.slice_x`/`lbm.graphics.slice_y`/`lbm.graphics.slice_z`.
 - Exported frames will automatically be assigned the current simulation time step in their name, in the format `bin/export/image-123456789.png`.
 - To convert the rendered `.png` images to video, use [FFmpeg](https://ffmpeg.org/):

README.md

@@ -182,6 +182,17 @@ The fastest and most memory efficient lattice Boltzmann CFD software, running on
   - fixed minor graphical artifacts in `raytrace_phi()`
   - fixed minor graphical artifacts in `ray_grid_traverse_sum()`
   - fixed wrong printed time step count on raindrop sample setup
+- [v2.19](https://github.com/ProjectPhysX/FluidX3D/releases/tag/v2.19) (07.09.2024) [changes](https://github.com/ProjectPhysX/FluidX3D/compare/v2.18...v2.19) (camera splines)
+  - the camera can now fly along a smooth path through a list of provided keyframe camera placements, [using Catmull-Rom splines](https://github.com/ProjectPhysX/FluidX3D/blob/master/DOCUMENTATION.md#video-rendering)
+  - more accurate remaining runtime estimation that includes time spent on rendering
+  - enabled FP16S memory compression by default
+  - printed camera placement using key <kbd>G</kbd> is now formatted for easier copy/paste
+  - added benchmark chart in Readme using mermaid gantt chart
+  - placed memory allocation info during simulation startup at better location
+  - fixed threading conflict between `INTERACTIVE_GRAPHICS` and `lbm.graphics.write_frame();`
+  - fixed maximum buffer allocation size limit for AMD GPUs and in Intel CPU Runtime for OpenCL
+  - fixed wrong `Re<Re_max` info printout for 2D simulations
+  - minor fix in `bandwidth_bytes_per_cell_device()`
 
 </details>
 

src/defines.hpp

@@ -10,8 +10,8 @@
 #define SRT // choose single-relaxation-time LBM collision operator; (default)
 //#define TRT // choose two-relaxation-time LBM collision operator
 
-//#define FP16S // compress LBM DDFs to range-shifted IEEE-754 FP16; number conversion is done in hardware; all arithmetic is still done in FP32
-//#define FP16C // compress LBM DDFs to more accurate custom FP16C format; number conversion is emulated in software; all arithmetic is still done in FP32
+#define FP16S // optional for 2x speedup and 2x VRAM footprint reduction: compress LBM DDFs to range-shifted IEEE-754 FP16; number conversion is done in hardware; all arithmetic is still done in FP32
+//#define FP16C // optional for 2x speedup and 2x VRAM footprint reduction: compress LBM DDFs to more accurate custom FP16C format; number conversion is emulated in software; all arithmetic is still done in FP32
 
 #define BENCHMARK // disable all extensions and setups and run benchmark setup instead
 

src/graphics.cpp

@@ -544,9 +544,11 @@ INT WINAPI WinMain(_In_ HINSTANCE hInstance, _In_opt_ HINSTANCE, _In_ PSTR, _In_
 			DispatchMessage(&msg);
 		}
 		// main loop ################################################################
+		camera.rendring_frame.lock(); // block rendering for other threads until finished
 		camera.update_state(fmax(1.0/(double)camera.fps_limit, frametime));
 		main_graphics();
 		update_frame(frametime);
+		camera.rendring_frame.unlock();
 		frametime = clock.stop();
 		sleep(1.0/(double)camera.fps_limit-frametime);
 		clock.start();
@@ -723,9 +725,11 @@ int main(int argc, char* argv[]) {
 	double frametime = 1.0;
 	while(running) {
 		// main loop ################################################################
+		camera.rendring_frame.lock(); // block rendering for other threads until finished
 		camera.update_state(fmax(1.0/(double)camera.fps_limit, frametime));
 		main_graphics();
 		update_frame(frametime);
+		camera.rendring_frame.unlock();
 		frametime = clock.stop();
 		sleep(1.0/(double)camera.fps_limit-frametime);
 		clock.start();
@@ -780,9 +784,11 @@ int main(int argc, char* argv[]) {
 	get_console_font_size(fontwidth, fontheight);
 	while(running) {
 		// main loop ################################################################
+		camera.rendring_frame.lock(); // block rendering for other threads until finished
 		camera.update_state(fmax(1.0/(double)camera.fps_limit, frametime));
 		main_graphics();
 		update_frame(frametime);
+		camera.rendring_frame.unlock();
 		frametime = clock.stop();
 		sleep(1.0/(double)camera.fps_limit-frametime);
 		clock.start();

src/graphics.hpp

@@ -8,6 +8,7 @@
 #include "defines.hpp"
 #include "utilities.hpp"
 #include <atomic>
+#include <mutex>
 
 extern vector<string> main_arguments; // console arguments
 extern std::atomic_bool running;
@@ -34,8 +35,11 @@ class Camera {
 	bool vr=false, tv=false; // virtual reality mode (enables stereoscopic rendering), VR TV mode
 	float eye_distance = 8.0f; // distance between cameras in VR mode
 	bool autorotation = false; // autorotation
-	bool key_update = true; // a key variable has been updated
 	bool lockmouse = false; // mouse movement won't change camera view when this is true
+	std::atomic_bool key_update = true; // a key variable has been updated
+	std::atomic_bool allow_rendering = false; // allows interactive redering if true
+	std::atomic_bool allow_labeling = true; // allows drawing label if true
+	std::mutex rendring_frame; // a frame for interactive graphics is currently rendered
 
 private:
 	float log_zoom=4.0f*log(zoom), target_log_zoom=log_zoom;

src/info.cpp

@@ -3,28 +3,11 @@
 
 Info info;
 
-void Info::initialize(LBM* lbm) {
-	this->lbm = lbm;
-#if defined(SRT)
-	collision = "SRT";
-#elif defined(TRT)
-	collision = "TRT";
-#endif // TRT
-#if defined(FP16S)
-	collision += " (FP32/FP16S)";
-#elif defined(FP16C)
-	collision += " (FP32/FP16C)";
-#else // FP32
-	collision += " (FP32/FP32)";
-#endif // FP32
-	cpu_mem_required = (uint)(lbm->get_N()*(ulong)bytes_per_cell_host()/1048576ull); // reset to get valid values for consecutive simulations
-	gpu_mem_required = lbm->lbm_domain[0]->get_device().info.memory_used;
-}
 void Info::append(const ulong steps, const ulong total_steps, const ulong t) {
 	if(total_steps==max_ulong) { // total_steps is not provided/used
 		this->steps = steps; // has to be executed before info.print_initialize()
 		this->steps_last = t; // reset last step count if multiple run() commands are executed consecutively
-		this->runtime_lbm_last = runtime_lbm; // reset last runtime if multiple run() commands are executed consecutively
+		this->runtime_total_last = this->runtime_total; // reset last runtime if multiple run() commands are executed consecutively
 		this->runtime_total = clock.stop();
 	} else { // total_steps has been specified
 		this->steps = total_steps; // has to be executed before info.print_initialize()
@@ -37,7 +20,8 @@ void Info::update(const double dt) {
 	this->runtime_total = clock.stop();
 }
 double Info::time() const { // returns either elapsed time or remaining time
-	return steps==max_ulong ? runtime_lbm : ((double)steps/(double)(lbm->get_t()-steps_last)-1.0)*(runtime_lbm-runtime_lbm_last); // time estimation on average so far
+	if(lbm==nullptr) return 0.0;
+	return steps==max_ulong ? runtime_total : ((double)steps/(double)(lbm->get_t()-steps_last)-1.0)*(runtime_total-runtime_total_last); // time estimation on average so far
 	//return steps==max_ulong ? runtime_lbm : ((double)steps-(double)(lbm->get_t()-steps_last))*runtime_lbm_timestep_smooth; // instantaneous time estimation
 }
 void Info::print_logo() const {
@@ -58,11 +42,27 @@ void Info::print_logo() const {
 	print("|                                  ");                 print("\\  \\ /  /", c);                print("                                  |\n");
 	print("|                                   ");                 print("\\  '  /", c);                 print("                                   |\n");
 	print("|                                    ");                 print("\\   /", c);                 print("                                    |\n");
-	print("|                                     ");                 print("\\ /", c);                 print("               FluidX3D Version 2.18 |\n");
+	print("|                                     ");                 print("\\ /", c);                 print("               FluidX3D Version 2.19 |\n");
 	print("|                                      ");                 print( "'", c);                 print("     Copyright (c) Dr. Moritz Lehmann |\n");
 	print("|-----------------------------------------------------------------------------|\n");
 }
-void Info::print_initialize() {
+void Info::print_initialize(LBM* lbm) {
+	info.allow_printing.lock(); // disable print_update() until print_initialize() has finished
+	this->lbm = lbm;
+#if defined(SRT)
+	collision = "SRT";
+#elif defined(TRT)
+	collision = "TRT";
+#endif // TRT
+#if defined(FP16S)
+	collision += " (FP32/FP16S)";
+#elif defined(FP16C)
+	collision += " (FP32/FP16C)";
+#else // FP32
+	collision += " (FP32/FP32)";
+#endif // FP32
+	cpu_mem_required = (uint)(lbm->get_N()*(ulong)bytes_per_cell_host()/1048576ull); // reset to get valid values for consecutive simulations
+	gpu_mem_required = lbm->lbm_domain[0]->get_device().info.memory_used;
 	const float Re = lbm->get_Re_max();
 	println("|-----------------.-----------------------------------------------------------|");
 	println("| Grid Resolution | "+alignr(57u, to_string(lbm->get_Nx())+" x "+to_string(lbm->get_Ny())+" x "+to_string(lbm->get_Nz())+" = "+to_string(lbm->get_N()))+" |");
@@ -91,10 +91,12 @@ void Info::print_initialize() {
 	println("'-----------------'-----------------------------------------------------------'");
 #endif // INTERACTIVE_GRAPHICS_ASCII
 	clock.start();
-	allow_rendering = true;
+	info.allow_printing.unlock();
 }
 void Info::print_update() const {
-	if(allow_rendering) reprint(
+	if(lbm==nullptr) return;
+	info.allow_printing.lock();
+	reprint(
 		"|"+alignr(8, to_uint((double)lbm->get_N()*1E-6/runtime_lbm_timestep_smooth))+" |"+ // MLUPs
 		alignr(7, to_uint((double)lbm->get_N()*(double)bandwidth_bytes_per_cell_device()*1E-9/runtime_lbm_timestep_smooth))+" GB/s |"+ // memory bandwidth
 		alignr(10, to_uint(1.0/runtime_lbm_timestep_smooth))+" | "+ // steps/s
@@ -103,16 +105,17 @@ void Info::print_update() const {
 	);
 #ifdef GRAPHICS
 	if(key_G) { // print camera settings
-		const string camera_position = "float3("+to_string(camera.pos.x/(float)lbm->get_Nx(), 6u)+"f*(float)Nx, "+to_string(camera.pos.y/(float)lbm->get_Ny(), 6u)+"f*(float)Ny, "+to_string(camera.pos.z/(float)lbm->get_Nz(), 6u)+"f*(float)Nz)";
-		const string camera_rx_ry_fov = to_string(degrees(camera.rx)-90.0, 1u)+"f, "+to_string(180.0-degrees(camera.ry), 1u)+"f, "+to_string(camera.fov, 1u)+"f";
-		const string camera_zoom = to_string(camera.zoom*(float)fmax(fmax(lbm->get_Nx(), lbm->get_Ny()), lbm->get_Nz())/(float)min(camera.width, camera.height), 6u)+"f";
-		if(camera.free) print_info("lbm.graphics.set_camera_free("+camera_position+", "+camera_rx_ry_fov+");");
-		else print_info("lbm.graphics.set_camera_centered("+camera_rx_ry_fov+", "+camera_zoom+");");
+		const string camera_position = "float3("+alignr(9u, to_string(camera.pos.x/(float)lbm->get_Nx(), 6u))+"f*(float)Nx, "+alignr(9u, to_string(camera.pos.y/(float)lbm->get_Ny(), 6u))+"f*(float)Ny, "+alignr(9u, to_string(camera.pos.z/(float)lbm->get_Nz(), 6u))+"f*(float)Nz)";
+		const string camera_rx_ry_fov = alignr(6u, to_string(degrees(camera.rx)-90.0, 1u))+"f, "+alignr(5u, to_string(180.0-degrees(camera.ry), 1u))+"f, "+alignr(5u, to_string(camera.fov, 1u))+"f";
+		const string camera_zoom = alignr(8u, to_string(camera.zoom*(float)fmax(fmax(lbm->get_Nx(), lbm->get_Ny()), lbm->get_Nz())/(float)min(camera.width, camera.height), 6u))+"f";
+		if(camera.free) println("\rlbm.graphics.set_camera_free("+camera_position+", "+camera_rx_ry_fov+");");
+		else println("\rlbm.graphics.set_camera_centered("+camera_rx_ry_fov+", "+camera_zoom+");          ");
 		key_G = false;
 	}
 #endif // GRAPHICS
+	info.allow_printing.unlock();
 }
 void Info::print_finalize() {
-	allow_rendering = false;
+	lbm = nullptr;
 	println("\n|---------'-------------'-----------'-------------------'---------------------|");
 }

src/info.hpp

@@ -1,24 +1,23 @@
 #pragma once
 
 #include "utilities.hpp"
+#include <mutex>
 
 class LBM;
 struct Info { // contains redundant information for console printing
 	LBM* lbm = nullptr;
-	bool allow_rendering = false; // allows interactive redering if true
-	bool allow_labeling = true; // allows drawing label if true
-	double runtime_lbm=0.0, runtime_total=0.0f; // lbm (compute) and total (compute + rendering + data evaluation) runtime
-	double runtime_lbm_timestep_last=1.0, runtime_lbm_timestep_smooth=1.0, runtime_lbm_last=0.0; // for printing simulation info
+	double runtime_lbm=0.0, runtime_total=0.0f, runtime_total_last=0.0; // lbm (compute) and total (compute + rendering + data evaluation) runtime
+	double runtime_lbm_timestep_last=1.0, runtime_lbm_timestep_smooth=1.0; // for printing simulation info
 	Clock clock; // for measuring total runtime
-	ulong steps=max_ulong, steps_last=0ull; // runtime_lbm_last and steps_last are there if multiple run() commands are executed consecutively
+	ulong steps=max_ulong, steps_last=0ull; // runtime_total_last and steps_last are there if multiple run() commands are executed consecutively
 	uint cpu_mem_required=0u, gpu_mem_required=0u; // all in MB
 	string collision = "";
-	void initialize(LBM* lbm);
+	std::mutex allow_printing; // to prevent threading conflicts when continuously printing updates to console
 	void append(const ulong steps, const ulong total_steps, const ulong t);
 	void update(const double dt);
 	double time() const; // returns either elapsed time or remaining time
 	void print_logo() const;
-	void print_initialize(); // enables interactive rendering
+	void print_initialize(LBM* lbm); // enables interactive rendering
 	void print_update() const;
 	void print_finalize(); // disables interactive rendering
 };