performdocking.cpp 45 KB
Newer Older
lvs's avatar
lvs committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
//// ------------------------
//// Correct time measurement
//// Moved from main.cpp to performdocking.cpp
//// to skip measuring build time
//#include <sys/time.h>
//// ------------------------
#include "performdocking.h"

//// --------------------------------
//// Altera OpenCL Helper Variables
//// --------------------------------
#include <assert.h>
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
#include <cstring>
17
18
#include "CL/opencl.h"

lvs's avatar
lvs committed
19
20
21
22

//// --------------------------------
//// Device memory buffers
//// --------------------------------
Leonardo Solis's avatar
Leonardo Solis committed
23

lvs's avatar
lvs committed
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68

/*								                  // Nr elements	// Nr bytes
cl_mem mem_atom_charges_const;		// float [MAX_NUM_OF_ATOMS];			// 90	 = 90	//360
cl_mem mem_atom_types_const;		// char  [MAX_NUM_OF_ATOMS];			// 90	 = 90	//360
cl_mem mem_intraE_contributors_const;	// char  [3*MAX_INTRAE_CONTRIBUTORS];		// 3*8128=28384 //24384	
cl_mem mem_VWpars_AC_const;		// float [MAX_NUM_OF_ATYPES*MAX_NUM_OF_ATYPES];	// 14*14 = 196  //784
cl_mem mem_VWpars_BD_const;		// float [MAX_NUM_OF_ATYPES*MAX_NUM_OF_ATYPES]; // 14*14 = 196	//784
cl_mem mem_dspars_S_const;		// float [MAX_NUM_OF_ATYPES];			// 14    = 14   //56
cl_mem mem_dspars_V_const;		// float [MAX_NUM_OF_ATYPES];			// 14    = 14   //56
cl_mem mem_rotlist_const;		// int   [MAX_NUM_OF_ROTATIONS];		// 4096  = 4096 //16384
cl_mem mem_ref_coords_x_const;		// float [MAX_NUM_OF_ATOMS];			// 90    = 90   //360
cl_mem mem_ref_coords_y_const;		// float [MAX_NUM_OF_ATOMS];			// 90    = 90   //360
cl_mem mem_ref_coords_z_const;		// float [MAX_NUM_OF_ATOMS];			// 90	 = 90   //360
cl_mem mem_rotbonds_moving_vectors_const;// float [3*MAX_NUM_OF_ROTBONDS];		// 3*32  = 96   //384
cl_mem mem_rotbonds_unit_vectors_const;	// float [3*MAX_NUM_OF_ROTBONDS];		// 3*32  = 96   //384
cl_mem mem_ref_orientation_quats_const;	// float [4*MAX_NUM_OF_RUNS];			// 4*100 = 400  //1600
*/

//// --------------------------------
//// Docking
//// --------------------------------
int docking_with_gpu(const Gridinfo*   mygrid,
	             /*const*/ float*  cpu_floatgrids,
                           Dockpars*   mypars,
		     const Liganddata* myligand_init,
		     const int*        argc,
		           char**      argv,
		           clock_t     clock_start_program)
/* The function performs the docking algorithm and generates the corresponding result files.
parameter mygrid:
		describes the grid
		filled with get_gridinfo()
parameter cpu_floatgrids:
		points to the memory region containing the grids
		filled with get_gridvalues_f()
parameter mypars:
		describes the docking parameters
		filled with get_commandpars()
parameter myligand_init:
		describes the ligands
		filled with get_liganddata()
parameters argc and argv:
		are the corresponding command line arguments parameter clock_start_program:
		contains the state of the clock tick counter at the beginning of the program
filled with clock() */
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
{
	// The get_xil_devices will return vector of Xilinx Devices 
	std::vector<cl::Device> devices = xcl::get_xil_devices();
	cl::Device device = devices[0];

	//Creating Context and Command Queue for selected Device 
	cl::Context context(device);
    	std::string device_name = device.getInfo<CL_DEVICE_NAME>(); 
	std::cout << "Found Device=" << device_name.c_str() << std::endl;

	#ifdef ENABLE_KRNL_GA
	cl::CommandQueue command_queue_ga			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_CONFORM
	cl::CommandQueue command_queue_conform			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_INTERE
	cl::CommandQueue command_queue_intere			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_INTRAE
	cl::CommandQueue command_queue_intrae			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_BT_USHORT_FLOAT
	cl::CommandQueue command_queue_prng_bt_ushort_float	(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_GG_UCHAR
	cl::CommandQueue command_queue_prng_gg_uchar		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_GG_FLOAT
	cl::CommandQueue command_queue_prng_gg_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS123_USHORT
	cl::CommandQueue command_queue_prng_ls123_ushort	(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS_FLOAT
	cl::CommandQueue command_queue_prng_ls_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS2_FLOAT
	cl::CommandQueue command_queue_prng_ls2_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS3_FLOAT
	cl::CommandQueue command_queue_prng_ls3_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS4_FLOAT
	cl::CommandQueue command_queue_prng_ls4_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS5_FLOAT
	cl::CommandQueue command_queue_prng_ls5_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS6_FLOAT
	cl::CommandQueue command_queue_prng_ls6_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS7_FLOAT
	cl::CommandQueue command_queue_prng_ls7_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS8_FLOAT
	cl::CommandQueue command_queue_prng_ls8_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS9_FLOAT
	cl::CommandQueue command_queue_prng_ls9_float		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_LS
	cl::CommandQueue command_queue_ls			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_LS2
	cl::CommandQueue command_queue_ls2			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_LS3
	cl::CommandQueue command_queue_ls3			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_LS4
	cl::CommandQueue command_queue_ls4			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_LS5
	cl::CommandQueue command_queue_ls5			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_LS6
	cl::CommandQueue command_queue_ls6			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif	
	#ifdef ENABLE_KRNL_LS7
	cl::CommandQueue command_queue_ls7			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_LS8
	cl::CommandQueue command_queue_ls8			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_LS9
	cl::CommandQueue command_queue_ls9			(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif
	#ifdef ENABLE_KRNL_IGL_ARBITER
	cl::CommandQueue command_queue_igl_arbiter		(context, device, CL_QUEUE_PROFILING_ENABLE);
	#endif

    	// import_binary() command will find the OpenCL binary file created using the 
    	// xocc compiler load into OpenCL Binary and return as Binaries
    	// OpenCL and it can contain many functions which can be executed on the
   	// device.
    	std::string binaryFile = xcl::find_binary_file(device_name,"Krnl_GA");
    	cl::Program::Binaries bins = xcl::import_binary_file(binaryFile);
    	devices.resize(1);
    	cl::Program program(context, devices, bins);

    	// This call will extract a kernel out of the program we loaded in the
    	// previous line. A kernel is an OpenCL function that is executed on the
    	// FPGA. This function is defined in the device/Krnl_GA.cl file.
	#ifdef ENABLE_KRNL_GA
	cl::Kernel kernel_ga			(program, "Krnl_GA");
	#endif
	#ifdef ENABLE_KRNL_CONFORM
	cl::Kernel kernel_conform		(program, "Krnl_Conform");
	#endif
	#ifdef ENABLE_KRNL_INTERE
	cl::Kernel kernel_intere		(program, "Krnl_InterE");
	#endif
	#ifdef ENABLE_KRNL_INTRAE
	cl::Kernel kernel_intrae		(program, "Krnl_IntraE");
	#endif
	#ifdef ENABLE_KRNL_PRNG_BT_USHORT_FLOAT
	cl::Kernel kernel_prng_bt_ushort_float	(program, "Krnl_Prng_BT_ushort_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_GG_UCHAR
	cl::Kernel kernel_prng_gg_uchar		(program, "Krnl_Prng_GG_uchar");
	#endif
	#ifdef ENABLE_KRNL_PRNG_GG_FLOAT
	cl::Kernel kernel_prng_gg_float		(program, "Krnl_Prng_GG_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS123_USHORT
	cl::Kernel kernel_prng_ls123_ushort	(program, "Krnl_Prng_LS123_ushort");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS_FLOAT
	cl::Kernel kernel_prng_ls_float		(program, "Krnl_Prng_LS_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS2_FLOAT
	cl::Kernel kernel_prng_ls2_float	(program, "Krnl_Prng_LS2_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS3_FLOAT
	cl::Kernel kernel_prng_ls3_float	(program, "Krnl_Prng_LS3_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS4_FLOAT
	cl::Kernel kernel_prng_ls4_float	(program, "Krnl_Prng_LS4_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS5_FLOAT
	cl::Kernel kernel_prng_ls5_float	(program, "Krnl_Prng_LS5_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS6_FLOAT
	cl::Kernel kernel_prng_ls6_float	(program, "Krnl_Prng_LS6_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS7_FLOAT
	cl::Kernel kernel_prng_ls7_float	(program, "Krnl_Prng_LS7_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS8_FLOAT
	cl::Kernel kernel_prng_ls8_float	(program, "Krnl_Prng_LS8_float");
	#endif
	#ifdef ENABLE_KRNL_PRNG_LS9_FLOAT
	cl::Kernel kernel_prng_ls9_float	(program, "Krnl_Prng_LS9_float");
	#endif
	#ifdef ENABLE_KRNL_LS
	cl::Kernel kernel_ls			(program, "Krnl_LS");
	#endif
	#ifdef ENABLE_KRNL_LS2
	cl::Kernel kernel_ls2			(program, "Krnl_LS2");
	#endif
	#ifdef ENABLE_KRNL_LS3
	cl::Kernel kernel_ls3			(program, "Krnl_LS3");
	#endif
	#ifdef ENABLE_KRNL_LS4
	cl::Kernel kernel_ls4			(program, "Krnl_LS4");
	#endif
	#ifdef ENABLE_KRNL_LS5
	cl::Kernel kernel_ls5			(program, "Krnl_LS5");
	#endif
	#ifdef ENABLE_KRNL_LS6
	cl::Kernel kernel_ls6			(program, "Krnl_LS6");
	#endif
	#ifdef ENABLE_KRNL_LS7
	cl::Kernel kernel_ls7			(program, "Krnl_LS7");
	#endif
	#ifdef ENABLE_KRNL_LS8
	cl::Kernel kernel_ls8			(program, "Krnl_LS8");
	#endif
	#ifdef ENABLE_KRNL_LS9
	cl::Kernel kernel_ls9			(program, "Krnl_LS9");
	#endif
	#ifdef ENABLE_KRNL_IGL_ARBITER
	cl::Kernel kernel_igl_arbiter		(program, "Krnl_IGL_Arbiter");
	#endif
lvs's avatar
lvs committed
254

255
256
257
	clock_t clock_start_docking;
	clock_t	clock_stop_docking;
	clock_t clock_stop_program_before_clustering;
lvs's avatar
lvs committed
258

259
260
261
	Liganddata myligand_reference;

 	//allocating GPU memory for floatgrids,
262
263
	size_t size_floatgrids_nbytes = sizeof(float) * (mygrid->num_of_atypes+2) *
					(mygrid->size_xyz[0]) * (mygrid->size_xyz[1]) * (mygrid->size_xyz[2]);
lvs's avatar
lvs committed
264

265
266
	size_t size_populations_nbytes = mypars->num_of_runs * mypars->pop_size * ACTUAL_GENOTYPE_LENGTH * sizeof(float);
	size_t size_populations_nelems = mypars->num_of_runs * mypars->pop_size * ACTUAL_GENOTYPE_LENGTH;
lvs's avatar
lvs committed
267

268
269
	size_t size_energies_nbytes = mypars->num_of_runs * mypars->pop_size * sizeof(float);
	size_t size_energies_nelems = mypars->num_of_runs * mypars->pop_size;
lvs's avatar
lvs committed
270

271
272
273
274
275
	//allocating and initializing CPU memory for initial population
	vector<float,aligned_allocator<float>> cpu_init_populations (size_populations_nelems, 0.0f);

	//allocating CPU memory for final population
	vector<float,aligned_allocator<float>> cpu_final_populations (size_populations_nelems);
lvs's avatar
lvs committed
276

277
278
279
280
281
282
283
284
	//allocating CPU memory for results
	vector<float,aligned_allocator<float>> cpu_energies (size_energies_nelems);

	//allocating CPU memory for resulting ligands
	vector<Ligandresult,aligned_allocator<Ligandresult>> cpu_result_ligands (mypars->num_of_runs);

	//allocating memory in CPU for reference orientation angles
	vector<float,aligned_allocator<float>> cpu_ref_ori_angles (mypars->num_of_runs*3);
lvs's avatar
lvs committed
285
286
287
288

	//generating initial populations and random orientation angles of reference ligand
	//(ligand will be moved to origo and scaled as well)
	myligand_reference = *myligand_init;
289
	gen_initpop_and_reflig(mypars, cpu_init_populations.data(), cpu_ref_ori_angles.data(), &myligand_reference, mygrid);
290
291
292
293
294
295
296
297
298
299
300

	//allocating memory in CPU for pseudorandom number generator seeds
	const unsigned int num_of_prng_blocks = 25;
	size_t size_prng_seeds_nelems = num_of_prng_blocks * mypars->num_of_runs;
	vector<unsigned int,aligned_allocator<unsigned int>> cpu_prng_seeds (size_prng_seeds_nelems);
	
	//initializing seed generator
	genseed(time(NULL));	

	//generating seeds (for each thread during GA)
	for (unsigned int i=0; i<size_prng_seeds_nelems; i++) {
lvs's avatar
lvs committed
301
302
		cpu_prng_seeds[i] = genseed(0u);
	}
303
304
305
306

	size_t size_evals_of_runs_nbytes = mypars->num_of_runs*sizeof(int);
	size_t size_evals_of_runs_nelems = mypars->num_of_runs;

lvs's avatar
lvs committed
307
	// allocating memory in CPU for evaluation counters
308
309
310
311
	vector<int,aligned_allocator<int>> cpu_evals_of_runs (size_evals_of_runs_nelems, 0);
	
	// allocating memory in CPU for generation counters
	vector<int,aligned_allocator<int>> cpu_gens_of_runs (size_evals_of_runs_nelems, 0);
lvs's avatar
lvs committed
312

313
314
	//preparing the constant data fields for the FPGA
	// -----------------------------------------------------------------------------------------------------
lvs's avatar
lvs committed
315
316
317
318
	// The original function does CUDA calls initializing const Kernel data.
	// We create a struct to hold those constants
	// and return them <here> (<here> = where prepare_const_fields_for_gpu() is called),
	// so we can send them to Kernels from <here>, instead of from calcenergy.cpp as originally.
319
320
321
	// -----------------------------------------------------------------------------------------------------
	kernelconstant_static  KerConstStatic;
	if (prepare_conststatic_fields_for_gpu(&myligand_reference, mypars, cpu_ref_ori_angles.data(), &KerConstStatic) == 1)
lvs's avatar
lvs committed
322
323
		return 1;

324
325
326
327
328
	//preparing parameter struct
	Dockparameters dockpars;
	dockpars.num_of_atoms  			= ((unsigned char)  myligand_reference.num_of_atoms);
	dockpars.num_of_atypes 			= ((unsigned char)  myligand_reference.num_of_atypes);
	dockpars.num_of_intraE_contributors 	= ((unsigned int)   myligand_reference.num_of_intraE_contributors);
329
330
331
332
333
334
335
336
337
	dockpars.gridsize_x    			= ((unsigned char)  mygrid->size_xyz[0]);
	dockpars.gridsize_y    			= ((unsigned char)  mygrid->size_xyz[1]);
	dockpars.gridsize_z    			= ((unsigned char)  mygrid->size_xyz[2]);
	dockpars.g1	       			= dockpars.gridsize_x ;
	dockpars.g2	       			= dockpars.gridsize_x * dockpars.gridsize_y;
	dockpars.g3	       			= dockpars.gridsize_x * dockpars.gridsize_y * dockpars.gridsize_z;
	dockpars.grid_spacing  			= ((float) mygrid->spacing);
	dockpars.rotbondlist_length 		= ((unsigned int) NUM_OF_THREADS_PER_BLOCK*(myligand_reference.num_of_rotcyc));
	dockpars.coeff_elec    			= ((float) mypars->coeffs.scaled_AD4_coeff_elec);
338
	dockpars.coeff_desolv  			= ((float) mypars->coeffs.AD4_coeff_desolv);
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
	dockpars.num_of_energy_evals 		= (unsigned int) mypars->num_of_energy_evals;
	dockpars.num_of_generations  		= (unsigned int) mypars->num_of_generations;
	dockpars.pop_size      			= (unsigned int) mypars->pop_size;
	dockpars.num_of_genes  			= (unsigned int)(myligand_reference.num_of_rotbonds + 6);
	dockpars.tournament_rate 		= (mypars->tournament_rate)/100;
	dockpars.crossover_rate  		= (mypars->crossover_rate)/100;
	dockpars.mutation_rate   		= (mypars->mutation_rate)/100;
	dockpars.abs_max_dang    		= mypars->abs_max_dang;
	dockpars.abs_max_dmov    		= mypars->abs_max_dmov;
	dockpars.lsearch_rate    		= mypars->lsearch_rate;
	dockpars.num_of_lsentities 		= (unsigned int) (mypars->lsearch_rate/100.0*mypars->pop_size + 0.5);
	dockpars.rho_lower_bound   		= mypars->rho_lower_bound;
	dockpars.base_dmov_mul_sqrt3 		= mypars->base_dmov_mul_sqrt3;
	dockpars.base_dang_mul_sqrt3 		= mypars->base_dang_mul_sqrt3;
	dockpars.cons_limit        		= (unsigned int) mypars->cons_limit;
	dockpars.max_num_of_iters  		= (unsigned int) mypars->max_num_of_iters;
	dockpars.qasp 				= mypars->qasp;
	dockpars.smooth 			= mypars->smooth;
lvs's avatar
lvs committed
357
358
359
360
361
362
363
364
365
366

	// these variables hold multiplications between kernel-constants
	// better calculate them here and then pass them to Krnl_GA
	const float two_absmaxdmov = 2.0 * dockpars.abs_max_dmov;
	const float two_absmaxdang = 2.0 * dockpars.abs_max_dang;

	// these variables hold multiplications between kernel-constants
	// better calculate them here and then pass them to Krnl_InterE and Krnl_InterE2
	const unsigned int mul_tmp2 = dockpars.num_of_atypes * dockpars.g3;
	const unsigned int mul_tmp3 = (dockpars.num_of_atypes + 1) * dockpars.g3;
367

368
	// -----------------------------------------------------------------------------------------------------
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
	#if 0
	// Hardware specific configuration (valid if 4 DDR banks are available, e.g.: AWS)
	// Specifiying precisely which DDR memory bank is 
	// being pointed to by a kernel globgal memory pointer.

	cl_mem_ext_ptr_t d_bank0_ext; // Ideally: DDR bank0 <-> Krnl_GA
	cl_mem_ext_ptr_t d_bank1_ext; // Ideally: DDR bank1 <-> Krnl_Conform
	cl_mem_ext_ptr_t d_bank2_ext; // Ideally: DDR bank2 <-> Krnl_InterE
	cl_mem_ext_ptr_t d_bank3_ext; // Ideally: DDR bank3 <-> Krnl_IntraE

	d_bank0_ext.flags = XCL_MEM_DDR_BANK0; d_bank0_ext.obj = NULL; d_bank0_ext.param = 0;
	d_bank1_ext.flags = XCL_MEM_DDR_BANK1; d_bank1_ext.obj = NULL; d_bank1_ext.param = 0;
	d_bank2_ext.flags = XCL_MEM_DDR_BANK2; d_bank2_ext.obj = NULL; d_bank2_ext.param = 0;
	d_bank3_ext.flags = XCL_MEM_DDR_BANK3; d_bank3_ext.obj = NULL; d_bank3_ext.param = 0;
	#endif
	// -----------------------------------------------------------------------------------------------------
385

386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
        // These commands will allocate memory on the FPGA. The cl::Buffer objects can
        // be used to reference the memory locations on the device. The cl::Buffer
	// object cannot be referenced directly and must be passed to other OpenCL
	// functions.

	// Krnl_GA buffers
	cl::Buffer mem_dockpars_conformations_current	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_WRITE, 
							size_populations_nbytes,	cpu_init_populations.data());
	cl::Buffer mem_dockpars_energies_current     	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_WRITE, 
							size_energies_nbytes,     	cpu_energies.data());
	cl::Buffer mem_evals_performed			(context, CL_MEM_USE_HOST_PTR | CL_MEM_WRITE_ONLY, 
							size_evals_of_runs_nbytes, 	cpu_evals_of_runs.data());
	cl::Buffer mem_gens_performed			(context, CL_MEM_USE_HOST_PTR | CL_MEM_WRITE_ONLY, 
							size_evals_of_runs_nbytes,	cpu_gens_of_runs.data());
	// Krnl_Conform buffers
	cl::Buffer mem_KerConstStatic_rotlist_const	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ROTATIONS*sizeof(int),	&KerConstStatic.rotlist_const[0]);
	cl::Buffer mem_KerConstStatic_ref_coords_const	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ATOMS*sizeof(cl_float3),	&KerConstStatic.ref_coords_const[0]);	
	cl::Buffer mem_KerConstStatic_rotbonds_moving_vectors_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,  
							MAX_NUM_OF_ROTBONDS*sizeof(cl_float3),	&KerConstStatic.rotbonds_moving_vectors_const[0]);
	cl::Buffer mem_KerConstStatic_rotbonds_unit_vectors_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ROTBONDS*sizeof(cl_float3),	&KerConstStatic.rotbonds_unit_vectors_const[0]);
	cl::Buffer mem_KerConstStatic_ref_orientation_quats_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_RUNS*sizeof(cl_float4),   	&KerConstStatic.ref_orientation_quats_const[0]);
	
	// Krnl_InterE buffers
	cl::Buffer mem_dockpars_fgrids			(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							size_floatgrids_nbytes,		cpu_floatgrids);	
	cl::Buffer mem_KerConstStatic_InterE_atom_charges_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ATOMS*sizeof(float),	&KerConstStatic.atom_charges_const[0]);
	cl::Buffer mem_KerConstStatic_InterE_atom_types_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ATOMS*sizeof(char),	&KerConstStatic.atom_types_const[0]);
	
	// Krnl_IntraE buffers
	cl::Buffer mem_KerConstStatic_IntraE_atom_charges_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ATOMS*sizeof(float),	&KerConstStatic.atom_charges_const[0]);
	cl::Buffer mem_KerConstStatic_IntraE_atom_types_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ATOMS*sizeof(char), &KerConstStatic.atom_types_const[0]);
	cl::Buffer mem_KerConstStatic_intraE_contributors_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_INTRAE_CONTRIBUTORS*sizeof(cl_char3), &KerConstStatic.intraE_contributors_const[0]);
	cl::Buffer mem_KerConstStatic_reqm_const	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							ATYPE_NUM*sizeof(float),	&KerConstStatic.reqm_const);	
	cl::Buffer mem_KerConstStatic_reqm_hbond_const	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							ATYPE_NUM*sizeof(float),	&KerConstStatic.reqm_hbond_const);
	cl::Buffer mem_KerConstStatic_atom1_types_reqm_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							ATYPE_NUM*sizeof(unsigned int),	&KerConstStatic.atom1_types_reqm_const);
	cl::Buffer mem_KerConstStatic_atom2_types_reqm_const
							(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							ATYPE_NUM*sizeof(unsigned int),	&KerConstStatic.atom2_types_reqm_const);
	cl::Buffer mem_KerConstStatic_VWpars_AC_const	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ATYPES*MAX_NUM_OF_ATYPES*sizeof(float),	&KerConstStatic.VWpars_AC_const[0]);
	cl::Buffer mem_KerConstStatic_VWpars_BD_const	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ATYPES*MAX_NUM_OF_ATYPES*sizeof(float),	&KerConstStatic.VWpars_BD_const[0]);
	cl::Buffer mem_KerConstStatic_dspars_S_const	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ATYPES*sizeof(float),			&KerConstStatic.dspars_S_const[0]);
	cl::Buffer mem_KerConstStatic_dspars_V_const	(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY,
							MAX_NUM_OF_ATYPES*sizeof(float),			&KerConstStatic.dspars_V_const[0]);
	#if !defined(SW_EMU)
	//allocating CPU memory for dummy data (one integer)
	vector<int,aligned_allocator<int>> cpu_dummy (1);	
456

457
458
459
460
	// IMPORTANT: enable this dummy global argument only for "hw" build.
	// https://forums.xilinx.com/t5/SDAccel/ERROR-KernelCheck-83-114-in-sdx-2017-4/td-p/818135
	cl::Buffer mem_dummy				(context, CL_MEM_USE_HOST_PTR | CL_MEM_READ_ONLY, sizeof(int), cpu_dummy.data();
	#endif
461

462
	// -----------------------------------------------------------------------------------------------------
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
	//Separate Read/write Buffer vector is needed to migrate data between host/device
	std::vector<cl::Memory> inBufVec, outBufVec;

	// Krnl_GA
	inBufVec.push_back(mem_dockpars_conformations_current); 	// RD & WR
    	inBufVec.push_back(mem_dockpars_energies_current);		// RD & WR
	// Krnl_Conform
	inBufVec.push_back(mem_KerConstStatic_rotlist_const);
	inBufVec.push_back(mem_KerConstStatic_ref_coords_const);
	inBufVec.push_back(mem_KerConstStatic_rotbonds_moving_vectors_const);
	inBufVec.push_back(mem_KerConstStatic_rotbonds_unit_vectors_const);
	inBufVec.push_back(mem_KerConstStatic_ref_orientation_quats_const);
	// Krnl_InterE
	inBufVec.push_back(mem_dockpars_fgrids);
	inBufVec.push_back(mem_KerConstStatic_InterE_atom_charges_const);
	inBufVec.push_back(mem_KerConstStatic_InterE_atom_types_const);
	// Krnl_IntraE
	inBufVec.push_back(mem_KerConstStatic_IntraE_atom_charges_const);
	inBufVec.push_back(mem_KerConstStatic_IntraE_atom_types_const);
	inBufVec.push_back(mem_KerConstStatic_intraE_contributors_const);
	inBufVec.push_back(mem_KerConstStatic_reqm_const);
	inBufVec.push_back(mem_KerConstStatic_reqm_hbond_const);
	inBufVec.push_back(mem_KerConstStatic_atom1_types_reqm_const);
	inBufVec.push_back(mem_KerConstStatic_atom2_types_reqm_const);
	inBufVec.push_back(mem_KerConstStatic_VWpars_AC_const);
	inBufVec.push_back(mem_KerConstStatic_VWpars_BD_const);
	inBufVec.push_back(mem_KerConstStatic_dspars_S_const);
	inBufVec.push_back(mem_KerConstStatic_dspars_V_const);

	// Krnl_GA
        outBufVec.push_back(mem_dockpars_conformations_current);	// RD & WR
	outBufVec.push_back(mem_dockpars_energies_current);		// RD & WR
	outBufVec.push_back(mem_evals_performed);
	outBufVec.push_back(mem_gens_performed);
Leonardo Solis's avatar
Leonardo Solis committed
497

498
499
500
501
502
503
504
	// -----------------------------------------------------------------------------------------------------
    	// These commands will load CPU-sources vectors from the host
   	// application and into cl::Buffer objects. 
	// The data will be be transferred from system memory 
	// over PCIe to the FPGA on-board DDR memory.
	command_queue_ga.enqueueMigrateMemObjects(inBufVec,0/* 0 means from host*/);
	// -----------------------------------------------------------------------------------------------------
lvs's avatar
lvs committed
505
506

	clock_start_docking = clock();
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525

	int narg;
	#ifdef ENABLE_KRNL_GA
	narg = 0;
	kernel_ga.setArg(narg++, mem_dockpars_conformations_current);
	kernel_ga.setArg(narg++, mem_dockpars_energies_current);
	kernel_ga.setArg(narg++, mem_evals_performed);
	kernel_ga.setArg(narg++, mem_gens_performed);
	kernel_ga.setArg(narg++, dockpars.pop_size);
	kernel_ga.setArg(narg++, dockpars.num_of_energy_evals);
	kernel_ga.setArg(narg++, dockpars.num_of_generations);
	kernel_ga.setArg(narg++, dockpars.tournament_rate);
	kernel_ga.setArg(narg++, dockpars.mutation_rate);
	kernel_ga.setArg(narg++, dockpars.abs_max_dmov);
	kernel_ga.setArg(narg++, dockpars.abs_max_dang);
	kernel_ga.setArg(narg++, two_absmaxdmov);
	kernel_ga.setArg(narg++, two_absmaxdang);
	kernel_ga.setArg(narg++, dockpars.crossover_rate);
	kernel_ga.setArg(narg++, dockpars.num_of_lsentities);
526
527
	kernel_ga.setArg(narg++, dockpars.num_of_genes);
	// Other kernel args are configured at every docking run
lvs's avatar
lvs committed
528
529
	#endif

530
531
532
533
534
535
536
537
538
539
	#ifdef ENABLE_KRNL_CONFORM
	narg = 0;
	kernel_conform.setArg(narg++, mem_KerConstStatic_rotlist_const);
	kernel_conform.setArg(narg++, mem_KerConstStatic_ref_coords_const);
	kernel_conform.setArg(narg++, mem_KerConstStatic_rotbonds_moving_vectors_const);
	kernel_conform.setArg(narg++, mem_KerConstStatic_rotbonds_unit_vectors_const);
	kernel_conform.setArg(narg++, dockpars.rotbondlist_length);
	kernel_conform.setArg(narg++, dockpars.num_of_atoms);
	kernel_conform.setArg(narg++, dockpars.num_of_genes);
	kernel_conform.setArg(narg++, mem_KerConstStatic_ref_orientation_quats_const);
540
	// Other kernel args are configured at every docking run
lvs's avatar
lvs committed
541
542
543
544
	#endif

	unsigned char gridsizex_minus1 = dockpars.gridsize_x - 1;
	unsigned char gridsizey_minus1 = dockpars.gridsize_y - 1;
lvs's avatar
lvs committed
545
	unsigned char gridsizez_minus1 = dockpars.gridsize_z - 1;
lvs's avatar
lvs committed
546
547
	float fgridsizex_minus1 = (float) gridsizex_minus1;
	float fgridsizey_minus1 = (float) gridsizey_minus1;
lvs's avatar
lvs committed
548
549
	float fgridsizez_minus1 = (float) gridsizez_minus1;

550
551
552
553
554
555
556
557
558
559
560
561
562
563
	#ifdef ENABLE_KRNL_INTERE
	narg = 0;
	kernel_intere.setArg(narg++, mem_dockpars_fgrids);
	kernel_intere.setArg(narg++, mem_KerConstStatic_InterE_atom_charges_const);
	kernel_intere.setArg(narg++, mem_KerConstStatic_InterE_atom_types_const);
	kernel_intere.setArg(narg++, dockpars.g1);
	kernel_intere.setArg(narg++, dockpars.g2);
	kernel_intere.setArg(narg++, dockpars.g3);
	kernel_intere.setArg(narg++, dockpars.num_of_atoms);
	kernel_intere.setArg(narg++, fgridsizex_minus1);
	kernel_intere.setArg(narg++, fgridsizey_minus1);
	kernel_intere.setArg(narg++, fgridsizez_minus1);
	kernel_intere.setArg(narg++, mul_tmp2);
	kernel_intere.setArg(narg++, mul_tmp3);
lvs's avatar
lvs committed
564
	#endif
lvs's avatar
lvs committed
565

566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
	#ifdef ENABLE_KRNL_INTRAE
	narg = 0;
	kernel_intrae.setArg(narg++, mem_KerConstStatic_IntraE_atom_charges_const);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_IntraE_atom_types_const);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_intraE_contributors_const);
	kernel_intrae.setArg(narg++, dockpars.smooth);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_reqm_const);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_reqm_hbond_const);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_atom1_types_reqm_const);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_atom2_types_reqm_const);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_VWpars_AC_const);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_VWpars_BD_const);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_dspars_S_const);
	kernel_intrae.setArg(narg++, mem_KerConstStatic_dspars_V_const);
	kernel_intrae.setArg(narg++, dockpars.num_of_atoms);
	kernel_intrae.setArg(narg++, dockpars.num_of_intraE_contributors);
	kernel_intrae.setArg(narg++, dockpars.grid_spacing);
	kernel_intrae.setArg(narg++, dockpars.num_of_atypes);
	kernel_intrae.setArg(narg++, dockpars.coeff_elec);
	kernel_intrae.setArg(narg++, dockpars.qasp);
	kernel_intrae.setArg(narg++, dockpars.coeff_desolv);
	#endif

589
590
591
	#ifdef ENABLE_KRNL_PRNG_BT_USHORT_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_bt_ushort_float.setArg(2, dockpars.pop_size);
592
	#if !defined(SW_EMU)
593
	kernel_prng_bt_ushort_float.setArg(3, mem_dummy);
594
	#endif
595
	#endif
596

597
598
599
	#ifdef ENABLE_KRNL_PRNG_GG_UCHAR
	// Other kernel args are configured at every docking run
	kernel_prng_gg_uchar.setArg(1, dockpars.num_of_genes);
600
	#if !defined(SW_EMU)
601
	kernel_prng_gg_uchar.setArg(2, mem_dummy);
602
	#endif
603
	#endif
lvs's avatar
lvs committed
604

605
606
607
	#ifdef ENABLE_KRNL_PRNG_GG_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_gg_float.setArg(1, dockpars.num_of_genes);
Leonardo Solis's avatar
Leonardo Solis committed
608
	#if !defined(SW_EMU)
609
	kernel_prng_gg_float.setArg(2, mem_dummy);
Leonardo Solis's avatar
Leonardo Solis committed
610
	#endif
611
	#endif
612

613
614
615
	#ifdef ENABLE_KRNL_PRNG_LS123_USHORT
	// Other kernel args are configured at every docking run
	kernel_prng_ls123_ushort.setArg(9, dockpars.pop_size);
616
	#if !defined(SW_EMU)
617
	kernel_prng_ls123_ushort.setArg(10, mem_dummy);
618
	#endif
619
	#endif
620

621
622
623
	#ifdef ENABLE_KRNL_PRNG_LS_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_ls_float.setArg(1, dockpars.num_of_genes);
624
	#if !defined(SW_EMU)
625
626
	kernel_prng_ls_float.setArg(2, mem_dummy);
	#endif
627
628
	#endif

629
630
631
	#ifdef ENABLE_KRNL_PRNG_LS2_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_ls2_float.setArg(1, dockpars.num_of_genes);
632
	#if !defined(SW_EMU)
633
	kernel_prng_ls2_float.setArg(2, mem_dummy);
634
	#endif
635
	#endif
636

637
638
639
	#ifdef ENABLE_KRNL_PRNG_LS3_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_ls3_float.setArg(1, dockpars.num_of_genes);
640
	#if !defined(SW_EMU)
641
	kernel_prng_ls3_float.setArg(2, mem_dummy);
642
	#endif
643
	#endif
644

645
646
647
	#ifdef ENABLE_KRNL_PRNG_LS4_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_ls4_float.setArg(1, dockpars.num_of_genes);
648
	#if !defined(SW_EMU)
649
	kernel_prng_ls4_float.setArg(2, mem_dummy);
650
	#endif
651
	#endif
652

653
654
655
	#ifdef ENABLE_KRNL_PRNG_LS5_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_ls5_float.setArg(1, dockpars.num_of_genes);
656
	#if !defined(SW_EMU)
657
	kernel_prng_ls5_float.setArg(2, mem_dummy);
658
	#endif
659
	#endif
660

661
662
663
	#ifdef ENABLE_KRNL_PRNG_LS6_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_ls6_float.setArg(1, dockpars.num_of_genes);
664
	#if !defined(SW_EMU)
665
	kernel_prng_ls6_float.setArg(2, mem_dummy);
666
	#endif
667
	#endif
668

669
670
671
	#ifdef ENABLE_KRNL_PRNG_LS7_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_ls7_float.setArg(1, dockpars.num_of_genes);
672
	#if !defined(SW_EMU)
673
	kernel_prng_ls7_float.setArg(2, mem_dummy);
674
	#endif
675
	#endif
676

677
678
679
	#ifdef ENABLE_KRNL_PRNG_LS8_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_ls8_float.setArg(1, dockpars.num_of_genes);
680
	#if !defined(SW_EMU)
681
	kernel_prng_ls8_float.setArg(2, mem_dummy);
682
	#endif
683
	#endif
684

685
686
687
	#ifdef ENABLE_KRNL_PRNG_LS9_FLOAT
	// Other kernel args are configured at every docking run
	kernel_prng_ls9_float.setArg(1, dockpars.num_of_genes);
688
	#if !defined(SW_EMU)
689
	kernel_prng_ls9_float.setArg(2, mem_dummy);
690
	#endif
691
	#endif
692

693
694
	unsigned short Host_max_num_of_iters = (unsigned short)dockpars.max_num_of_iters;
	unsigned char  Host_cons_limit       = (unsigned char) dockpars.cons_limit;
695

696
697
698
699
700
701
702
703
	#ifdef ENABLE_KRNL_LS
	narg = 0;
	kernel_ls.setArg(narg++, Host_max_num_of_iters);
	kernel_ls.setArg(narg++, dockpars.rho_lower_bound);
	kernel_ls.setArg(narg++, dockpars.base_dmov_mul_sqrt3);
	kernel_ls.setArg(narg++, dockpars.num_of_genes);
	kernel_ls.setArg(narg++, dockpars.base_dang_mul_sqrt3);
	kernel_ls.setArg(narg++, Host_cons_limit);
Leonardo Solis's avatar
Leonardo Solis committed
704
	#if !defined(SW_EMU)
705
	kernel_ls.setArg(narg++, mem_dummy);
lvs's avatar
lvs committed
706
	#endif
707
	#endif
708

709
710
711
712
713
714
715
716
	#ifdef ENABLE_KRNL_LS2
	narg = 0;
	kernel_ls2.setArg(narg++, Host_max_num_of_iters);
	kernel_ls2.setArg(narg++, dockpars.rho_lower_bound);
	kernel_ls2.setArg(narg++, dockpars.base_dmov_mul_sqrt3);
	kernel_ls2.setArg(narg++, dockpars.num_of_genes);
	kernel_ls2.setArg(narg++, dockpars.base_dang_mul_sqrt3);
	kernel_ls2.setArg(narg++, Host_cons_limit);
Leonardo Solis's avatar
Leonardo Solis committed
717
	#if !defined(SW_EMU)
718
719
	kernel_ls2.setArg(narg++, mem_dummy);
	#endif
Leonardo Solis's avatar
Leonardo Solis committed
720
	#endif
721

722
723
724
725
726
727
728
729
	#ifdef ENABLE_KRNL_LS3
	narg = 0;
	kernel_ls3.setArg(narg++, Host_max_num_of_iters);
	kernel_ls3.setArg(narg++, dockpars.rho_lower_bound);
	kernel_ls3.setArg(narg++, dockpars.base_dmov_mul_sqrt3);
	kernel_ls3.setArg(narg++, dockpars.num_of_genes);
	kernel_ls3.setArg(narg++, dockpars.base_dang_mul_sqrt3);
	kernel_ls3.setArg(narg++, Host_cons_limit);
Leonardo Solis's avatar
Leonardo Solis committed
730
	#if !defined(SW_EMU)
731
	kernel_ls3.setArg(narg++, mem_dummy);
lvs's avatar
lvs committed
732
	#endif
733
	#endif
734

735
736
737
738
739
740
741
742
	#ifdef ENABLE_KRNL_LS4
	narg = 0;
	kernel_ls4.setArg(narg++, Host_max_num_of_iters);
	kernel_ls4.setArg(narg++, dockpars.rho_lower_bound);
	kernel_ls4.setArg(narg++, dockpars.base_dmov_mul_sqrt3);
	kernel_ls4.setArg(narg++, dockpars.num_of_genes);
	kernel_ls4.setArg(narg++, dockpars.base_dang_mul_sqrt3);
	kernel_ls4.setArg(narg++, Host_cons_limit);
Leonardo Solis's avatar
Leonardo Solis committed
743
	#if !defined(SW_EMU)
744
	kernel_ls4.setArg(narg++, mem_dummy);
Leonardo Solis's avatar
Leonardo Solis committed
745
	#endif
746
	#endif
747

748
749
750
751
752
753
754
755
	#ifdef ENABLE_KRNL_LS5
	narg = 0;
	kernel_ls5.setArg(narg++, Host_max_num_of_iters);
	kernel_ls5.setArg(narg++, dockpars.rho_lower_bound);
	kernel_ls5.setArg(narg++, dockpars.base_dmov_mul_sqrt3);
	kernel_ls5.setArg(narg++, dockpars.num_of_genes);
	kernel_ls5.setArg(narg++, dockpars.base_dang_mul_sqrt3);
	kernel_ls5.setArg(narg++, Host_cons_limit);
Leonardo Solis's avatar
Leonardo Solis committed
756
	#if !defined(SW_EMU)
757
	kernel_ls5.setArg(narg++, mem_dummy);
Leonardo Solis's avatar
Leonardo Solis committed
758
	#endif
759
	#endif
760

761
762
763
764
765
766
767
768
	#ifdef ENABLE_KRNL_LS6
	narg = 0;
	kernel_ls6.setArg(narg++, Host_max_num_of_iters);
	kernel_ls6.setArg(narg++, dockpars.rho_lower_bound);
	kernel_ls6.setArg(narg++, dockpars.base_dmov_mul_sqrt3);
	kernel_ls6.setArg(narg++, dockpars.num_of_genes);
	kernel_ls6.setArg(narg++, dockpars.base_dang_mul_sqrt3);
	kernel_ls6.setArg(narg++, Host_cons_limit);
Leonardo Solis's avatar
Leonardo Solis committed
769
	#if !defined(SW_EMU)
770
	kernel_ls6.setArg(narg++, mem_dummy);
Leonardo Solis's avatar
Leonardo Solis committed
771
	#endif
772
	#endif
773

774
775
776
777
778
779
780
781
	#ifdef ENABLE_KRNL_LS7
	narg = 0;
	kernel_ls7.setArg(narg++, Host_max_num_of_iters);
	kernel_ls7.setArg(narg++, dockpars.rho_lower_bound);
	kernel_ls7.setArg(narg++, dockpars.base_dmov_mul_sqrt3);
	kernel_ls7.setArg(narg++, dockpars.num_of_genes);
	kernel_ls7.setArg(narg++, dockpars.base_dang_mul_sqrt3);
	kernel_ls7.setArg(narg++, Host_cons_limit);
Leonardo Solis's avatar
Leonardo Solis committed
782
	#if !defined(SW_EMU)
783
	kernel_ls7.setArg(narg++, mem_dummy);
Leonardo Solis's avatar
Leonardo Solis committed
784
	#endif
785
	#endif
786

787
788
789
790
791
792
793
794
	#ifdef ENABLE_KRNL_LS8
	narg = 0;
	kernel_ls8.setArg(narg++, Host_max_num_of_iters);
	kernel_ls8.setArg(narg++, dockpars.rho_lower_bound);
	kernel_ls8.setArg(narg++, dockpars.base_dmov_mul_sqrt3);
	kernel_ls8.setArg(narg++, dockpars.num_of_genes);
	kernel_ls8.setArg(narg++, dockpars.base_dang_mul_sqrt3);
	kernel_ls8.setArg(narg++, Host_cons_limit);
Leonardo Solis's avatar
Leonardo Solis committed
795
	#if !defined(SW_EMU)
796
	kernel_ls8.setArg(narg++, mem_dummy);
Leonardo Solis's avatar
Leonardo Solis committed
797
	#endif
798
	#endif
799

800
801
802
803
804
805
806
807
	#ifdef ENABLE_KRNL_LS9
	narg = 0;
	kernel_ls9.setArg(narg++, Host_max_num_of_iters);
	kernel_ls9.setArg(narg++, dockpars.rho_lower_bound);
	kernel_ls9.setArg(narg++, dockpars.base_dmov_mul_sqrt3);
	kernel_ls9.setArg(narg++, dockpars.num_of_genes);
	kernel_ls9.setArg(narg++, dockpars.base_dang_mul_sqrt3);
	kernel_ls9.setArg(narg++, Host_cons_limit);
Leonardo Solis's avatar
Leonardo Solis committed
808
	#if !defined(SW_EMU)
809
	kernel_ls9.setArg(narg++, mem_dummy);
Leonardo Solis's avatar
Leonardo Solis committed
810
	#endif
811
	#endif
812

813
	#ifdef ENABLE_KRNL_IGL_ARBITER
814
	#if !defined(SW_EMU)
815
816
	kernel_igl_arbiter.setArg(0, mem_dummy);
	#endif
817
	#endif
lvs's avatar
lvs committed
818
819
820

	printf("Docking runs to be executed: %u\n", mypars->num_of_runs); 
	printf("Execution run: ");
lvs's avatar
lvs committed
821

822
	for (unsigned int run_cnt = 0; run_cnt < mypars->num_of_runs; run_cnt++) {
lvs's avatar
lvs committed
823
		printf(" %u", run_cnt+1); 
lvs's avatar
lvs committed
824
		fflush(stdout);
lvs's avatar
lvs committed
825

826
		#ifdef ENABLE_KRNL_GA
lvs's avatar
lvs committed
827
		unsigned int Host_Offset_Pop = run_cnt * dockpars.pop_size * ACTUAL_GENOTYPE_LENGTH;
828
829
830
831
		unsigned int Host_Offset_Ene = run_cnt * dockpars.pop_size;
		kernel_ga.setArg(16, run_cnt);
		kernel_ga.setArg(17, Host_Offset_Pop);
		kernel_ga.setArg(18, Host_Offset_Ene);
lvs's avatar
lvs committed
832
		#endif
833

834
835
836
		#ifdef ENABLE_KRNL_CONFORM
		kernel_conform.setArg(8, run_cnt);
		#endif
837

838
839
840
841
		#ifdef ENABLE_KRNL_PRNG_BT_USHORT_FLOAT
		kernel_prng_bt_ushort_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 14]);
		kernel_prng_bt_ushort_float.setArg(1, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 15]);
		#endif
842

843
844
845
		#ifdef ENABLE_KRNL_PRNG_GG_UCHAR
		kernel_prng_gg_uchar.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 2]);
		#endif
846

847
848
849
		#ifdef ENABLE_KRNL_PRNG_GG_FLOAT
		kernel_prng_gg_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt]);
		#endif
850

851
852
853
854
855
856
857
858
859
860
861
		#ifdef ENABLE_KRNL_PRNG_LS123_USHORT
		kernel_prng_ls123_ushort.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 5]);
		kernel_prng_ls123_ushort.setArg(1, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 6]);
		kernel_prng_ls123_ushort.setArg(2, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 7]);
		kernel_prng_ls123_ushort.setArg(3, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 8]);
		kernel_prng_ls123_ushort.setArg(4, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 9]);
		kernel_prng_ls123_ushort.setArg(5, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 10]);
		kernel_prng_ls123_ushort.setArg(6, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 11]);
		kernel_prng_ls123_ushort.setArg(7, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 12]);
		kernel_prng_ls123_ushort.setArg(8, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 13]);
		#endif
862

863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
		#ifdef ENABLE_KRNL_PRNG_LS_FLOAT
		kernel_prng_ls_float.setArg (0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 1]);
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS2_FLOAT
		kernel_prng_ls2_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 3]);
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS3_FLOAT
		kernel_prng_ls3_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 4]);
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS4_FLOAT
		kernel_prng_ls4_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 16]);
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS5_FLOAT
		kernel_prng_ls5_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 17]);
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS6_FLOAT
		kernel_prng_ls6_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 18]);
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS7_FLOAT
		kernel_prng_ls7_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 19]);
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS8_FLOAT
		kernel_prng_ls8_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 20]);
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS9_FLOAT
		kernel_prng_ls9_float.setArg(0, cpu_prng_seeds[num_of_prng_blocks * run_cnt + 21]);
		#endif
lvs's avatar
lvs committed
890

891
		#ifdef ENABLE_KRNL_GA
892
		command_queue_ga.enqueueTask(kernel_ga);
893
		#endif
894
895
		#ifdef ENABLE_KRNL_CONFORM
		command_queue_conform.enqueueTask(kernel_conform);
896
		#endif
897
898
		#ifdef ENABLE_KRNL_INTERE
		command_queue_intere.enqueueTask(kernel_intere);
899
		#endif
900
901
		#ifdef ENABLE_KRNL_INTRAE
		command_queue_intrae.enqueueTask(kernel_intrae);
902
		#endif
903
904
		#ifdef ENABLE_KRNL_PRNG_BT_USHORT_FLOAT
		command_queue_prng_bt_ushort_float.enqueueTask(kernel_prng_bt_ushort_float);
905
		#endif
906
907
908
909
910
		#ifdef ENABLE_KRNL_PRNG_GG_UCHAR
		command_queue_prng_gg_uchar.enqueueTask(kernel_prng_gg_uchar);
		#endif
		#ifdef ENABLE_KRNL_PRNG_GG_FLOAT
		command_queue_prng_gg_float.enqueueTask(kernel_prng_gg_float);
911
		#endif
912
913
914
915
916
		#ifdef ENABLE_KRNL_PRNG_LS123_USHORT
		command_queue_prng_ls123_ushort.enqueueTask(kernel_prng_ls123_ushort);
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS_FLOAT
		command_queue_prng_ls_float.enqueueTask(kernel_prng_ls_float);
917
		#endif
918
919
		#ifdef ENABLE_KRNL_PRNG_LS2_FLOAT
		command_queue_prng_ls2_float.enqueueTask(kernel_prng_ls2_float);
920
		#endif
921
922
		#ifdef ENABLE_KRNL_PRNG_LS3_FLOAT
		command_queue_prng_ls3_float.enqueueTask(kernel_prng_ls3_float);
923
		#endif
924
925
		#ifdef ENABLE_KRNL_PRNG_LS4_FLOAT
		command_queue_prng_ls4_float.enqueueTask(kernel_prng_ls4_float);
926
		#endif
927
928
		#ifdef ENABLE_KRNL_PRNG_LS5_FLOAT
		command_queue_prng_ls5_float.enqueueTask(kernel_prng_ls5_float);
929
		#endif
930
931
		#ifdef ENABLE_KRNL_PRNG_LS6_FLOAT
		command_queue_prng_ls6_float.enqueueTask(kernel_prng_ls6_float);
932
		#endif
933
934
		#ifdef ENABLE_KRNL_PRNG_LS7_FLOAT
		command_queue_prng_ls7_float.enqueueTask(kernel_prng_ls7_float);
935
		#endif
936
937
		#ifdef ENABLE_KRNL_PRNG_LS8_FLOAT
		command_queue_prng_ls8_float.enqueueTask(kernel_prng_ls8_float);
938
		#endif
939
940
		#ifdef ENABLE_KRNL_PRNG_LS9_FLOAT
		command_queue_prng_ls9_float.enqueueTask(kernel_prng_ls9_float);
941
		#endif
942
943
		#ifdef ENABLE_KRNL_LS
		command_queue_ls.enqueueTask(kernel_ls);
944
		#endif
945
946
947
948
949
		#ifdef ENABLE_KRNL_LS2
		command_queue_ls2.enqueueTask(kernel_ls2);
		#endif
		#ifdef ENABLE_KRNL_LS3
		command_queue_ls3.enqueueTask(kernel_ls3);
950
		#endif
951
952
		#ifdef ENABLE_KRNL_LS4
		command_queue_ls4.enqueueTask(kernel_ls4);
953
		#endif
954
955
		#ifdef ENABLE_KRNL_LS5
		command_queue_ls5.enqueueTask(kernel_ls5);
956
		#endif
957
958
		#ifdef ENABLE_KRNL_LS6
		command_queue_ls6.enqueueTask(kernel_ls6);
959
		#endif
960
961
		#ifdef ENABLE_KRNL_LS7
		command_queue_ls7.enqueueTask(kernel_ls7);
962
		#endif
963
964
		#ifdef ENABLE_KRNL_LS8
		command_queue_ls8.enqueueTask(kernel_ls8);
965
		#endif
966
967
		#ifdef ENABLE_KRNL_LS9
		command_queue_ls9.enqueueTask(kernel_ls9);
968
		#endif
969
970
		#ifdef ENABLE_KRNL_IGL_ARBITER
		command_queue_igl_arbiter.enqueueTask(kernel_igl_arbiter);
971
		#endif
lvs's avatar
lvs committed
972

973
		#ifdef ENABLE_KRNL_GA	
974
		command_queue_ga.finish(); 
lvs's avatar
lvs committed
975
		#endif
976
		#ifdef ENABLE_KRNL_CONFORM	
977
		command_queue_conform.finish(); 
lvs's avatar
lvs committed
978
		#endif
979
		#ifdef ENABLE_KRNL_INTERE	
980
		command_queue_intere.finish(); 
lvs's avatar
lvs committed
981
		#endif
982
		#ifdef ENABLE_KRNL_INTRAE	
983
		command_queue_intrae.finish(); 
984
985
		#endif
		#ifdef ENABLE_KRNL_PRNG_BT_USHORT_FLOAT
986
		command_queue_prng_bt_ushort_float.finish();
987
988
		#endif
		#ifdef ENABLE_KRNL_PRNG_GG_UCHAR
989
		command_queue_prng_gg_uchar.finish();
lvs's avatar
lvs committed
990
		#endif
991
		#ifdef ENABLE_KRNL_PRNG_GG_FLOAT
992
		command_queue_prng_gg_float.finish(); 
993
		#endif
994
		#ifdef ENABLE_KRNL_PRNG_LS123_USHORT
995
		command_queue_prng_ls123_ushort.finish();
996
997
		#endif
		#ifdef ENABLE_KRNL_PRNG_LS_FLOAT
998
999
		command_queue_prng_ls_float.finish();
		#endif
1000
		#ifdef ENABLE_KRNL_PRNG_LS2_FLOAT
For faster browsing, not all history is shown. View entire blame