Skip to content

I
imagej-elphel
Commit a744e6be authored by Andrey Filippov's avatar Andrey Filippov
Browse files
Options

Added debug output for LPF filter to compare with GPU

parent ddc33b02
Hide whitespace changes
Inline Side-by-side
Showing with 79 additions and 6 deletions
src/main/java/ImageDtt.java
View file @ a744e6be
......@@ -4216,7 +4216,20 @@ public class ImageDtt {
final double [] dbg_filter= dtt.dttt_ii(filter);
for (int i=0; i < filter.length;i++) filter[i] *= 2*dct_size;
if (globalDebugLevel > 1) {
if (globalDebugLevel > 2) {
System.out.print("__constant__ float lpf_data[64]={");
for (int i=0; i<filter.length;i++){
System.out.print(String.format("%5.8ff", filter[i]));
if (i == 63) {
System.out.println("};");
} else {
System.out.print(", ");
if ((i % 8) == 7) {
System.out.print("\n ");
}
}
}
} else if (globalDebugLevel > 1) {
for (int i=0; i<filter.length;i++){
System.out.println("dct_lpf_psf() "+i+": "+filter[i]);
}
......
src/main/resources/TileProcessor.cuh
View file @ a744e6be
......@@ -254,7 +254,36 @@ __constant__ float idct_signs[4][4][4] ={
{ 1, 1, 1,-1},
{-1,-1,-1, 1}
}};
// LPF for sigma 0.9 each color (modify through cudaMemcpyToSymbol() or similar in Driver API
__constant__ float lpf_data[3][64]={
{
1.00000000f, 0.87041007f, 0.65943687f, 0.43487258f, 0.24970076f, 0.12518080f, 0.05616371f, 0.02728573f,
0.87041007f, 0.75761368f, 0.57398049f, 0.37851747f, 0.21734206f, 0.10895863f, 0.04888546f, 0.02374977f,
0.65943687f, 0.57398049f, 0.43485698f, 0.28677101f, 0.16466189f, 0.08254883f, 0.03703642f, 0.01799322f,
0.43487258f, 0.37851747f, 0.28677101f, 0.18911416f, 0.10858801f, 0.05443770f, 0.02442406f, 0.01186582f,
0.24970076f, 0.21734206f, 0.16466189f, 0.10858801f, 0.06235047f, 0.03125774f, 0.01402412f, 0.00681327f,
0.12518080f, 0.10895863f, 0.08254883f, 0.05443770f, 0.03125774f, 0.01567023f, 0.00703062f, 0.00341565f,
0.05616371f, 0.04888546f, 0.03703642f, 0.02442406f, 0.01402412f, 0.00703062f, 0.00315436f, 0.00153247f,
0.02728573f, 0.02374977f, 0.01799322f, 0.01186582f, 0.00681327f, 0.00341565f, 0.00153247f, 0.00074451f
},{
1.00000000f, 0.87041007f, 0.65943687f, 0.43487258f, 0.24970076f, 0.12518080f, 0.05616371f, 0.02728573f,
0.87041007f, 0.75761368f, 0.57398049f, 0.37851747f, 0.21734206f, 0.10895863f, 0.04888546f, 0.02374977f,
0.65943687f, 0.57398049f, 0.43485698f, 0.28677101f, 0.16466189f, 0.08254883f, 0.03703642f, 0.01799322f,
0.43487258f, 0.37851747f, 0.28677101f, 0.18911416f, 0.10858801f, 0.05443770f, 0.02442406f, 0.01186582f,
0.24970076f, 0.21734206f, 0.16466189f, 0.10858801f, 0.06235047f, 0.03125774f, 0.01402412f, 0.00681327f,
0.12518080f, 0.10895863f, 0.08254883f, 0.05443770f, 0.03125774f, 0.01567023f, 0.00703062f, 0.00341565f,
0.05616371f, 0.04888546f, 0.03703642f, 0.02442406f, 0.01402412f, 0.00703062f, 0.00315436f, 0.00153247f,
0.02728573f, 0.02374977f, 0.01799322f, 0.01186582f, 0.00681327f, 0.00341565f, 0.00153247f, 0.00074451f
},{
1.00000000f, 0.87041007f, 0.65943687f, 0.43487258f, 0.24970076f, 0.12518080f, 0.05616371f, 0.02728573f,
0.87041007f, 0.75761368f, 0.57398049f, 0.37851747f, 0.21734206f, 0.10895863f, 0.04888546f, 0.02374977f,
0.65943687f, 0.57398049f, 0.43485698f, 0.28677101f, 0.16466189f, 0.08254883f, 0.03703642f, 0.01799322f,
0.43487258f, 0.37851747f, 0.28677101f, 0.18911416f, 0.10858801f, 0.05443770f, 0.02442406f, 0.01186582f,
0.24970076f, 0.21734206f, 0.16466189f, 0.10858801f, 0.06235047f, 0.03125774f, 0.01402412f, 0.00681327f,
0.12518080f, 0.10895863f, 0.08254883f, 0.05443770f, 0.03125774f, 0.01567023f, 0.00703062f, 0.00341565f,
0.05616371f, 0.04888546f, 0.03703642f, 0.02442406f, 0.01402412f, 0.00703062f, 0.00315436f, 0.00153247f,
0.02728573f, 0.02374977f, 0.01799322f, 0.01186582f, 0.00681327f, 0.00341565f, 0.00153247f, 0.00074451f
}};
__device__ void convertCorrectTile(
struct CltExtra * gpu_kernel_offsets, // [tileY][tileX][color]
......@@ -262,6 +291,7 @@ __device__ void convertCorrectTile(
float * gpu_images,
float * gpu_clt,
const int color,
const int lpf_mask,
const float centerX,
const float centerY,
const short tx,
......@@ -300,8 +330,9 @@ __global__ void tileProcessor(
float ** gpu_images, // [NUM_CAMS],
struct tp_task * gpu_tasks,
float ** gpu_clt, // [NUM_CAMS][TILESY][TILESX][NUM_COLORS][DTT_SIZE*DTT_SIZE]
size_t dstride, // // in floats (pixels)
int num_tiles) // number of tiles in task
size_t dstride, // in floats (pixels)
int num_tiles, // number of tiles in task
int lpf_mask) // apply lpf to colors : bit 0 - red, bit 1 - blue, bit2 - green
{
dim3 t = threadIdx;
......@@ -356,6 +387,7 @@ __global__ void tileProcessor(
gpu_images[ncam], // float * gpu_images,
gpu_clt[ncam], // float * gpu_clt,
color, // const int color,
lpf_mask, // const int lpf_mask,
tt[tile_in_block].xy[ncam][0], // const float centerX,
tt[tile_in_block].xy[ncam][1], // const float centerY,
tt[tile_in_block].tx, // const short tx,
......@@ -529,6 +561,7 @@ __device__ void convertCorrectTile(
float * gpu_images,
float * gpu_clt,
const int color,
const int lpf_mask,
const float centerX,
const float centerY,
const short tx,
......@@ -902,7 +935,7 @@ __device__ void convertCorrectTile(
#ifdef DBG_TILE
#ifdef DEBUG1
#ifdef DEBUG3
if ((threadIdx.x) == 0){
printf("\nDTT Tiles after vertical shift, color = %d\n",color);
debug_print_clt1(clt_tile, color, 0xf); // only 1 quadrant for R,B and 2 - for G
......@@ -911,6 +944,33 @@ __device__ void convertCorrectTile(
__syncthreads();// __syncwarp();
#endif
#endif
// optionally apply LF
if ((lpf_mask >> color) & 1){
float * clt = clt_tile + threadIdx.x;
#pragma unroll
for (int q = 0; q < 4; q++) {
float *lpf = lpf_data[color] + threadIdx.x;
#pragma unroll
for (int i = 0; i <8; i++){
(*clt) *= (*lpf);
clt += DTT_SIZE1;
lpf += DTT_SIZE;
}
}
__syncthreads();// __syncwarp();
#ifdef DBG_TILE
#ifdef DEBUG3
if ((threadIdx.x) == 0){
printf("\nDTT Tiles after LPF, color = %d\n",color);
debug_print_clt1(clt_tile, color, 0xf); // only 1 quadrant for R,B and 2 - for G
printf("\nDTT All done\n");
}
__syncthreads();// __syncwarp();
#endif
#endif
}
int offset_src = threadIdx.x;
......@@ -919,7 +979,7 @@ __device__ void convertCorrectTile(
float * clt_dst = gpu_clt + offset_dst; // ((ty * TILESX + tx)*NUM_COLORS + color)* ( 4 * DTT_SIZE * DTT_SIZE1) + threadIdx.x; // gpu_kernels + kernel_full_index* (DTT_SIZE * DTT_SIZE * 4);
#ifdef DBG_TILE
#ifdef DEBUG1
#ifdef DEBUG3
if ((threadIdx.x) == 0){
printf("clt_src = 0x%lx\n",clt_src);
printf("clt_dst = 0x%lx\n",clt_dst);
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment