mutiple fixes: parallel printf, fixed cycle in cache, opencl refactored vecadd and sgemm, regen opencl kernels with hard-float, fixed vortex io bus interface, fixed dpi floats APi to support multicore mode, make vlsim multicore default, make rtlsim multi-core default, removed POCL binaries from repository, updated Makefiles to use external POCL

This commit is contained in:
Blaise Tine
2020-09-19 14:45:42 -04:00
parent 80f929eb61
commit f6f95e0c46
146 changed files with 116779 additions and 194258 deletions

View File

@@ -1,13 +1,11 @@
#include "tests.h"
#include <stdbool.h>
#include <vx_intrinsics.h>
#include <vx_print.h>
#include <vx_spawn.h>
#include <VX_config.h>
typedef struct
{
typedef struct {
unsigned * x;
unsigned * y;
unsigned * z;
@@ -15,7 +13,6 @@ typedef struct
unsigned numRows;
} mat_add_args_t;
unsigned x[] = {5, 5, 5, 5,
6, 6, 6, 6,
7, 7, 7, 7,
@@ -31,8 +28,7 @@ unsigned z[] = {0, 0, 0, 0,
0, 0, 0, 0,
0, 0, 0, 0};
void mat_add_kernel(void * void_arguments)
{
void mat_add_kernel(void * void_arguments) {
mat_add_args_t * arguments = (mat_add_args_t *) void_arguments;
unsigned wid = vx_warp_id();
@@ -49,73 +45,60 @@ void mat_add_kernel(void * void_arguments)
// __endif
}
int main()
{
// ensure single thread
vx_tmc(1);
int main() {
vx_printf("Let's start... (This might take a while)\n");
vx_print_str("Let's start... (This might take a while)\n");
unsigned what[36];
bool passed = true;
for (int i = 0; i < 36; i++)
{
for (int i = 0; i < 36; i++) {
what[i] = i;
// vx_print_hex(i);
// vx_printf(": ", what[i]);
if (what[i] != i)
{
if (what[i] != i) {
passed = false;
vx_printf("T1 Fail On ", i);
vx_printf("T1 Fail On %d", i);
}
}
for (int i = 0; i < 36; i++)
{
// vx_print_hex(i);
// vx_printf(": ", what[i]);
if (what[i] != i)
{
for (int i = 0; i < 36; i++) {
if (what[i] != i) {
passed = false;
vx_printf("T2 Fail on ", i);
vx_printf("T2 Fail on %d", i);
}
}
if (passed)
{
vx_print_str("Wr->read and repeat(Wr) tests passed!\n");
if (passed) {
vx_printf("Wr->read and repeat(Wr) tests passed!\n");
}
vx_print_str("Simple Main\n");
vx_printf("Simple Main\n");
// TMC test
test_tmc();
// Control Divergence Test
vx_print_str("test_divergence\n");
vx_tmc(4);
vx_printf("test_divergence\n");
test_divergence();
vx_tmc(1);
// Test wspawn
vx_print_str("test_wspawn\n");
vx_printf("test_wspawn\n");
test_wsapwn();
vx_print_str("Shared Memory test\n");
vx_printf("Shared Memory test\n");
unsigned * ptr = (unsigned *) SHARED_MEM_BASE_ADDR;
unsigned value = 0;
for (int i = 0; i < 5; i++)
{
for (int i = 0; i < 5; i++) {
*ptr = value;
unsigned read_valud = *ptr;
vx_printf("ptr: ", (unsigned) ptr);
vx_printf("Original Value: ", value);
vx_printf("Read Value: ", read_valud);
vx_print_str("-------------------\n");
vx_printf("ptr: %p\n", ptr);
vx_printf("Original Value: %x\n", value);
vx_printf("Read Value: %x\n", read_valud);
vx_printf("-------------------\n");
value++;
ptr++;
}
vx_print_str("vx_spawn_warps mat_add_kernel\n");
vx_printf("vx_spawn_warps mat_add_kernel\n");
mat_add_args_t arguments;
arguments.x = x;
@@ -124,24 +107,20 @@ int main()
arguments.numColums = 4;
arguments.numRows = 4;
int numWarps = 4;
int numThreads = 4;
vx_spawn_warps(numWarps, numThreads, mat_add_kernel, &arguments);
vx_print_str("Waiting to ensure other warps are done... (Takes a while)\n");
vx_printf("Waiting to ensure other warps are done... (Takes a while)\n");
for (int i = 0; i < 5000; i++) {}
for (int i = 0; i < numWarps; i++)
{
for (int j = 0; j < numThreads; j++)
{
for (int i = 0; i < numWarps; i++) {
for (int j = 0; j < numThreads; j++) {
unsigned index = (i * arguments.numColums) + j;
vx_print_hex(z[index]);
vx_print_str(" ");
vx_printf("0x%x ", z[index]);
}
vx_print_str("\n");
vx_printf("\n");
}
return 0;