Scruff AI
495 lines
19 KiB
Plaintext
495 lines
19 KiB
Plaintext
/* ============================================================================
|
||
* PROBE B 1024×1024 - NO FFT VERSION
|
||
* Real physics, guardian tracking, shear flow - NO FFT dependency
|
||
*
|
||
* CONSTITUTION:
|
||
* 1. NO FAKES: If step rate jumps to 300k, stop - LBM bypassed
|
||
* 2. RAW METAL: GPU fans must ramp up, or no work is being done
|
||
* ============================================================================ */
|
||
|
||
#include <cuda_runtime.h>
|
||
#include <nvml.h>
|
||
#include <cstdio>
|
||
#include <cstdlib>
|
||
#include <cstdint>
|
||
#include <cmath>
|
||
#include <chrono>
|
||
#include <vector>
|
||
|
||
#ifndef M_PI
|
||
#define M_PI 3.14159265358979323846
|
||
#endif
|
||
|
||
/* ---- Grid ---------------------------------------------------------------- */
|
||
#define NX 1024
|
||
#define NY 1024
|
||
#define NN (NX * NY)
|
||
#define Q 9
|
||
#define BLOCK 256
|
||
#define GBLK(n) (((n) + BLOCK - 1) / BLOCK)
|
||
|
||
/* ---- Test Protocol ------------------------------------------------------- */
|
||
#define TOTAL_STEPS 2000000 // ~1 hour at 5.5k steps/sec
|
||
#define STEPS_PER_BATCH 500
|
||
#define SAMPLE_INTERVAL 50000 // Sample every 50k steps
|
||
#define TOTAL_BATCHES (TOTAL_STEPS / STEPS_PER_BATCH)
|
||
#define SAMPLE_BATCHES (SAMPLE_INTERVAL / STEPS_PER_BATCH)
|
||
|
||
/* ---- LBM ---------------------------------------------------------------- */
|
||
#define OMEGA 1.0f // tau=1.0, nu=1/6 — "clear water"
|
||
|
||
/* ---- Guardian tracking ------------------------------------------------- */
|
||
#define MAX_GUARDIANS 200
|
||
#define GUARDIAN_THRESHOLD 1.01f // rho > 1.01 forms guardian
|
||
|
||
typedef struct {
|
||
float x, y; // position (grid coordinates)
|
||
float vx, vy; // velocity
|
||
float mass; // accumulated mass
|
||
int alive; // 1 if active
|
||
uint64_t born_step; // step when formed
|
||
} Guardian;
|
||
|
||
Guardian guardians[MAX_GUARDIANS];
|
||
int n_guardians = 0;
|
||
|
||
/* ---- D2Q9 --------------------------------------------------------------- */
|
||
__constant__ int d_ex[Q] = { 0, 1, 0,-1, 0, 1,-1,-1, 1 };
|
||
__constant__ int d_ey[Q] = { 0, 0, 1, 0,-1, 1, 1,-1,-1 };
|
||
__constant__ float d_w[Q] = { 4.f/9, 1.f/9, 1.f/9, 1.f/9, 1.f/9,
|
||
1.f/36,1.f/36,1.f/36,1.f/36 };
|
||
|
||
/* ======================================================================== */
|
||
/* K E R N E L S */
|
||
/* ======================================================================== */
|
||
|
||
/* ---- LBM collide & stream ---------------------------------------------- */
|
||
__global__ void lbm_collide_stream(const float* __restrict__ f_src,
|
||
float* __restrict__ f_dst,
|
||
float* __restrict__ rho,
|
||
float* __restrict__ ux,
|
||
float* __restrict__ uy,
|
||
float omega, int nx, int ny) {
|
||
const int idx = blockIdx.x * blockDim.x + threadIdx.x;
|
||
const int N = nx * ny;
|
||
if (idx >= N) return;
|
||
const int x = idx % nx, y = idx / nx;
|
||
|
||
float fl[Q];
|
||
for (int i = 0; i < Q; i++) {
|
||
int sx = (x - d_ex[i] + nx) % nx;
|
||
int sy = (y - d_ey[i] + ny) % ny;
|
||
fl[i] = f_src[i * N + sy * nx + sx];
|
||
}
|
||
|
||
float rho_val = 0.f, ux_val = 0.f, uy_val = 0.f;
|
||
for (int i = 0; i < Q; i++) {
|
||
rho_val += fl[i];
|
||
ux_val += (float)d_ex[i] * fl[i];
|
||
uy_val += (float)d_ey[i] * fl[i];
|
||
}
|
||
float inv = 1.f / fmaxf(rho_val, 1e-10f);
|
||
ux_val *= inv; uy_val *= inv;
|
||
rho[idx] = rho_val; ux[idx] = ux_val; uy[idx] = uy_val;
|
||
|
||
const float u2 = ux_val * ux_val + uy_val * uy_val;
|
||
for (int i = 0; i < Q; i++) {
|
||
float eu = (float)d_ex[i] * ux_val + (float)d_ey[i] * uy_val;
|
||
float feq = d_w[i] * rho_val * (1.f + 3.f*eu + 4.5f*eu*eu - 1.5f*u2);
|
||
f_dst[i * N + idx] = fl[i] - omega * (fl[i] - feq);
|
||
}
|
||
}
|
||
|
||
/* ---- PROBE B: Lattice shear — rotate velocity in top 25% by 90° -------- */
|
||
__global__ void probe_rotate_top(float* f, float* rho, float* ux, float* uy,
|
||
int nx, int ny) {
|
||
int idx = blockIdx.x * blockDim.x + threadIdx.x;
|
||
int N = nx * ny;
|
||
if (idx >= N) return;
|
||
int y = idx / nx;
|
||
|
||
/* Only affect top 25% */
|
||
if (y < ny * 3 / 4) return;
|
||
|
||
float r = rho[idx];
|
||
float old_ux = ux[idx];
|
||
float old_uy = uy[idx];
|
||
|
||
/* 90° rotation: (ux, uy) → (-uy, ux) */
|
||
float new_ux = -old_uy;
|
||
float new_uy = old_ux;
|
||
|
||
float u2_new = new_ux * new_ux + new_uy * new_uy;
|
||
|
||
/* Reconstruct equilibrium with rotated velocity */
|
||
for (int i = 0; i < Q; i++) {
|
||
float eu = (float)d_ex[i] * new_ux + (float)d_ey[i] * new_uy;
|
||
float feq_new = d_w[i] * r * (1.f + 3.f*eu + 4.5f*eu*eu - 1.5f*u2_new);
|
||
/* Hard set to new equilibrium — maximum disruption */
|
||
f[i * N + idx] = feq_new;
|
||
}
|
||
}
|
||
|
||
/* ======================================================================== */
|
||
/* G U A R D I A N T R A C K I N G */
|
||
/* ======================================================================== */
|
||
|
||
void update_guardians(const float* rho, const float* ux, const float* uy,
|
||
uint64_t current_step) {
|
||
// Simple guardian detection: local maxima of density
|
||
for (int y = 1; y < NY - 1; y++) {
|
||
for (int x = 1; x < NX - 1; x++) {
|
||
int idx = y * NX + x;
|
||
float rho_val = rho[idx];
|
||
|
||
// Check if this is a local maximum and above threshold
|
||
if (rho_val > GUARDIAN_THRESHOLD &&
|
||
rho_val > rho[idx - 1] && rho_val > rho[idx + 1] &&
|
||
rho_val > rho[idx - NX] && rho_val > rho[idx + NX]) {
|
||
|
||
// Check if guardian already exists nearby
|
||
int existing = -1;
|
||
for (int g = 0; g < n_guardians; g++) {
|
||
if (guardians[g].alive) {
|
||
float dx = guardians[g].x - x;
|
||
float dy = guardians[g].y - y;
|
||
if (dx*dx + dy*dy < 25.0f) { // Within 5 cells
|
||
existing = g;
|
||
break;
|
||
}
|
||
}
|
||
}
|
||
|
||
if (existing >= 0) {
|
||
// Update existing guardian
|
||
guardians[existing].x = x;
|
||
guardians[existing].y = y;
|
||
guardians[existing].vx = ux[idx];
|
||
guardians[existing].vy = uy[idx];
|
||
guardians[existing].mass += rho_val - 1.0f;
|
||
} else if (n_guardians < MAX_GUARDIANS) {
|
||
// Create new guardian
|
||
guardians[n_guardians].x = x;
|
||
guardians[n_guardians].y = y;
|
||
guardians[n_guardians].vx = ux[idx];
|
||
guardians[n_guardians].vy = uy[idx];
|
||
guardians[n_guardians].mass = rho_val - 1.0f;
|
||
guardians[n_guardians].alive = 1;
|
||
guardians[n_guardians].born_step = current_step;
|
||
n_guardians++;
|
||
}
|
||
}
|
||
}
|
||
}
|
||
}
|
||
|
||
void save_guardian_census(uint64_t current_step) {
|
||
FILE* csv = fopen("guardian_census.csv", "w");
|
||
if (!csv) return;
|
||
|
||
fprintf(csv, "id,x,y,vx,vy,mass,alive,born_step\n");
|
||
|
||
int alive_count = 0;
|
||
for (int g = 0; g < n_guardians; g++) {
|
||
if (guardians[g].alive) {
|
||
fprintf(csv, "%d,%.2f,%.2f,%.6f,%.6f,%.6f,%d,%llu\n",
|
||
g, guardians[g].x, guardians[g].y,
|
||
guardians[g].vx, guardians[g].vy,
|
||
guardians[g].mass, guardians[g].alive,
|
||
guardians[g].born_step);
|
||
alive_count++;
|
||
}
|
||
}
|
||
|
||
fclose(csv);
|
||
|
||
// Also save JSON for compatibility
|
||
FILE* json = fopen("guardian_census.json", "w");
|
||
if (json) {
|
||
fprintf(json, "{\n");
|
||
fprintf(json, " \"total_guardians\": %d,\n", alive_count);
|
||
fprintf(json, " \"current_step\": %llu,\n", current_step);
|
||
fprintf(json, " \"guardians\": [\n");
|
||
|
||
int first = 1;
|
||
for (int g = 0; g < n_guardians; g++) {
|
||
if (guardians[g].alive) {
|
||
if (!first) fprintf(json, ",\n");
|
||
first = 0;
|
||
|
||
fprintf(json, " {\n");
|
||
fprintf(json, " \"id\": %d,\n", g);
|
||
fprintf(json, " \"x\": %.2f,\n", guardians[g].x);
|
||
fprintf(json, " \"y\": %.2f,\n", guardians[g].y);
|
||
fprintf(json, " \"vx\": %.6f,\n", guardians[g].vx);
|
||
fprintf(json, " \"vy\": %.6f,\n", guardians[g].vy);
|
||
fprintf(json, " \"mass\": %.6f,\n", guardians[g].mass);
|
||
fprintf(json, " \"alive\": %d,\n", guardians[g].alive);
|
||
fprintf(json, " \"born_step\": %llu\n", guardians[g].born_step);
|
||
fprintf(json, " }");
|
||
}
|
||
}
|
||
|
||
fprintf(json, "\n ]\n");
|
||
fprintf(json, "}\n");
|
||
fclose(json);
|
||
}
|
||
}
|
||
|
||
/* ======================================================================== */
|
||
/* M A I N T E S T */
|
||
/* ======================================================================== */
|
||
|
||
int main() {
|
||
printf("=======================================================================\n");
|
||
printf(" PROBE B 1024×1024 - NO FFT VERSION\n");
|
||
printf(" Beast: RTX 4090, 1024x1024 grid\n");
|
||
printf(" Target: 2M steps (~1 hour at 5.5k steps/sec)\n");
|
||
printf("=======================================================================\n\n");
|
||
|
||
printf("CONSTITUTION:\n");
|
||
printf(" 1. NO FAKES: If step rate jumps to 300k, stop - LBM bypassed\n");
|
||
printf(" 2. RAW METAL: GPU fans must ramp up, or no work is being done\n\n");
|
||
|
||
/* ---- CUDA setup ----------------------------------------------------- */
|
||
cudaDeviceProp prop;
|
||
cudaGetDeviceProperties(&prop, 0);
|
||
printf("[CUDA] %s SM %d.%d SMs: %d\n",
|
||
prop.name, prop.major, prop.minor, prop.multiProcessorCount);
|
||
|
||
/* ---- NVML power monitoring ----------------------------------------- */
|
||
nvmlInit();
|
||
nvmlDevice_t nvml_dev;
|
||
nvmlDeviceGetHandleByIndex(0, &nvml_dev);
|
||
unsigned int power_mW;
|
||
nvmlDeviceGetPowerUsage(nvml_dev, &power_mW);
|
||
printf("[NVML] Idle power: %.1f W\n", power_mW / 1000.0f);
|
||
|
||
/* ---- Allocate memory ----------------------------------------------- */
|
||
float *f0, *f1, *rho, *ux, *uy;
|
||
cudaMallocManaged(&f0, Q * NN * sizeof(float));
|
||
cudaMallocManaged(&f1, Q * NN * sizeof(float));
|
||
cudaMallocManaged(&rho, NN * sizeof(float));
|
||
cudaMallocManaged(&ux, NN * sizeof(float));
|
||
cudaMallocManaged(&uy, NN * sizeof(float));
|
||
|
||
/* ---- Initialize equilibrium ---------------------------------------- */
|
||
printf("\n[INIT] Setting up equilibrium state (rho=1.0, u=0)...\n");
|
||
for (int i = 0; i < Q * NN; i++) {
|
||
f0[i] = 1.0f + 0.01f * (rand() / (float)RAND_MAX - 0.5f);
|
||
}
|
||
|
||
/* ---- Prepare output files ------------------------------------------ */
|
||
FILE* telemetry_csv = fopen("probeB_telemetry.csv", "w");
|
||
fprintf(telemetry_csv, "step,power_w,n_guardians,phase\n");
|
||
|
||
/* ---- Test phases --------------------------------------------------- */
|
||
enum { PHASE_BASELINE, PHASE_SHEAR, PHASE_RECOVERY } current_phase = PHASE_BASELINE;
|
||
uint64_t shear_trigger_step = 800000; // Apply shear at 800k steps
|
||
int shear_applied = 0;
|
||
|
||
auto t0 = std::chrono::steady_clock::now();
|
||
uint64_t total_steps = 0;
|
||
|
||
printf("\n[PHASE 1: BASELINE & BIRTH] Starting...\n");
|
||
printf(" Batch | Steps | Power | Guardians | Phase\n");
|
||
printf(" ------|---------|-------|-----------|--------\n");
|
||
|
||
int cur = 0;
|
||
int guardian_check_interval = 10000; // Check for guardians every 10k steps
|
||
|
||
for (int batch = 0; batch < TOTAL_BATCHES; batch++) {
|
||
// Run LBM steps
|
||
for (int s = 0; s < STEPS_PER_BATCH; s++) {
|
||
lbm_collide_stream<<<GBLK(NN), BLOCK>>>(
|
||
(cur == 0) ? f0 : f1,
|
||
(cur == 0) ? f1 : f0,
|
||
rho, ux, uy, OMEGA, NX, NY);
|
||
cudaDeviceSynchronize();
|
||
cur = 1 - cur;
|
||
}
|
||
total_steps += STEPS_PER_BATCH;
|
||
|
||
// Check for guardian formation
|
||
if (total_steps % guardian_check_interval == 0) {
|
||
update_guardians(rho, ux, uy, total_steps);
|
||
}
|
||
|
||
// Apply shear flow at trigger step (Probe B)
|
||
if (total_steps >= shear_trigger_step && !shear_applied) {
|
||
printf("\n[PHASE 3: SHEAR PUNCH] Applying Probe B at step %llu\n", total_steps);
|
||
printf(" Rotating top 25%% velocity by 90°...\n");
|
||
|
||
// Apply shear
|
||
probe_rotate_top<<<GBLK(NN), BLOCK>>>(f0, rho, ux, uy, NX, NY);
|
||
cudaDeviceSynchronize();
|
||
probe_rotate_top<<<GBLK(NN), BLOCK>>>(f1, rho, ux, uy, NX, NY);
|
||
cudaDeviceSynchronize();
|
||
|
||
shear_applied = 1;
|
||
current_phase = PHASE_SHEAR;
|
||
}
|
||
|
||
// Sample every SAMPLE_INTERVAL steps
|
||
if ((batch + 1) % SAMPLE_BATCHES == 0) {
|
||
// Get power usage
|
||
nvmlDeviceGetPowerUsage(nvml_dev, &power_mW);
|
||
float power_W = power_mW / 1000.0f;
|
||
|
||
// Log to CSV
|
||
const char* phase_str = "baseline";
|
||
if (current_phase == PHASE_SHEAR) phase_str = "shear";
|
||
else if (current_phase == PHASE_RECOVERY && shear_applied) phase_str = "recovery";
|
||
|
||
fprintf(telemetry_csv, "%llu,%.1f,%d,%s\n",
|
||
total_steps, power_W, n_guardians, phase_str);
|
||
|
||
// Print progress
|
||
printf(" %5d | %7llu | %5.0f | %9d | %s\n",
|
||
batch + 1, total_steps, power_W, n_guardians, phase_str);
|
||
|
||
// CONSTITUTION CHECK 1: Step rate
|
||
auto t_now = std::chrono::steady_clock::now();
|
||
double elapsed = std::chrono::duration<double>(t_now - t0).count();
|
||
float steps_per_sec = total_steps / elapsed;
|
||
|
||
if (steps_per_sec > 10000.0f) {
|
||
printf("\n🚨 CONSTITUTION VIOLATION: Step rate = %.0f (>10k)\n", steps_per_sec);
|
||
printf(" LBM may be bypassed. Stopping test.\n");
|
||
break;
|
||
}
|
||
|
||
// CONSTITUTION CHECK 2: Power scaling
|
||
if (power_W < 50.0f && elapsed > 60.0f) {
|
||
printf("\n🚨 CONSTITUTION VIOLATION: Power = %.1f W (<50W)\n", power_W);
|
||
printf(" GPU not under load. Stopping test.\n");
|
||
break;
|
||
}
|
||
|
||
// Phase transition: After shear, move to recovery
|
||
if (shear_applied && current_phase == PHASE_SHEAR &&
|
||
total_steps > shear_trigger_step + 100000) {
|
||
printf("\n[PHASE 4: RECOVERY] Monitoring reorganization...\n");
|
||
current_phase = PHASE_RECOVERY;
|
||
}
|
||
|
||
// Save guardian census periodically
|
||
if (n_guardians > 0 && total_steps % 100000 == 0) {
|
||
save_guardian_census(total_steps);
|
||
}
|
||
}
|
||
|
||
// Check if we've reached time limit (~1 hour)
|
||
auto t_now = std::chrono::steady_clock::now();
|
||
double elapsed = std::chrono::duration<double>(t_now - t0).count();
|
||
if (elapsed > 3600.0) { // 1 hour
|
||
printf("\n[TIME] 1 hour reached at step %llu\n", total_steps);
|
||
break;
|
||
}
|
||
}
|
||
|
||
auto t_end = std::chrono::steady_clock::now();
|
||
double runtime = std::chrono::duration<double>(t_end - t0).count();
|
||
|
||
/* ---- Final analysis ------------------------------------------------ */
|
||
printf("\n=======================================================================\n");
|
||
printf(" PROBE B TEST - FINAL RESULTS\n");
|
||
printf("=======================================================================\n");
|
||
|
||
printf("\nPERFORMANCE:\n");
|
||
printf(" Total steps: %llu\n", total_steps);
|
||
printf(" Runtime: %.1f seconds (%.2f hours)\n", runtime, runtime / 3600.0);
|
||
printf(" Steps/sec: %.0f\n", total_steps / runtime);
|
||
printf(" Expected: ~5,500 steps/sec\n");
|
||
|
||
printf("\nPOWER USAGE:\n");
|
||
nvmlDeviceGetPowerUsage(nvml_dev, &power_mW);
|
||
printf(" Final power: %.1f W\n", power_mW / 1000.0f);
|
||
printf(" Idle power: ~37 W\n");
|
||
printf(" Load power: ~290 W\n");
|
||
|
||
printf("\nGUARDIAN FORMATION:\n");
|
||
int alive_guardians = 0;
|
||
for (int g = 0; g < n_guardians; g++) {
|
||
if (guardians[g].alive) alive_guardians++;
|
||
}
|
||
printf(" Total guardians: %d\n", alive_guardians);
|
||
printf(" Expected (March 7): 194 guardians\n");
|
||
|
||
// Save final guardian census
|
||
save_guardian_census(total_steps);
|
||
|
||
printf("\nSHEAR FLOW:\n");
|
||
if (shear_applied) {
|
||
printf(" ✅ Applied at step %llu\n", shear_trigger_step);
|
||
} else {
|
||
printf(" ❌ NOT applied\n");
|
||
}
|
||
|
||
printf("\n=======================================================================\n");
|
||
printf(" V E R D I C T\n");
|
||
printf("=======================================================================\n");
|
||
|
||
int passes = 0;
|
||
int total_tests = 4;
|
||
|
||
// Test 1: Performance reality
|
||
float steps_per_sec = total_steps / runtime;
|
||
if (steps_per_sec > 4000 && steps_per_sec < 7000) {
|
||
printf("✅ PERFORMANCE: %.0f steps/sec (within 5.5k ± 25%%)\n", steps_per_sec);
|
||
passes++;
|
||
} else {
|
||
printf("❌ PERFORMANCE: %.0f steps/sec (expected ~5.5k)\n", steps_per_sec);
|
||
}
|
||
|
||
// Test 2: Power scaling
|
||
float final_power = power_mW / 1000.0f;
|
||
if (final_power > 100.0f) {
|
||
printf("✅ POWER: %.1f W (above idle, real work)\n", final_power);
|
||
passes++;
|
||
} else {
|
||
printf("❌ POWER: %.1f W (not scaling with load)\n", final_power);
|
||
}
|
||
|
||
// Test 3: Guardian formation
|
||
if (alive_guardians > 0) {
|
||
printf("✅ GUARDIANS: %d formed (real structure)\n", alive_guardians);
|
||
passes++;
|
||
} else {
|
||
printf("❌ GUARDIANS: None formed (no structure)\n");
|
||
}
|
||
|
||
// Test 4: Shear applied
|
||
if (shear_applied) {
|
||
printf("✅ SHEAR: Probe B applied at step %llu\n", shear_trigger_step);
|
||
passes++;
|
||
} else {
|
||
printf("❌ SHEAR: Not applied\n");
|
||
}
|
||
|
||
printf("\nSCORE: %d/%d tests passed\n", passes, total_tests);
|
||
|
||
if (passes == total_tests) {
|
||
printf("\n🎯 PROBE B TEST PASSED: System shows real physics\n");
|
||
printf(" The 4090 remembers how to be a brain.\n");
|
||
} else if (passes >= 3) {
|
||
printf("\n⚠️ PARTIAL SUCCESS: %d/4 tests passed\n", passes);
|
||
printf(" Some physics working, needs investigation.\n");
|
||
} else {
|
||
printf("\n🚨 TEST FAILED: Only %d/4 tests passed\n", passes);
|
||
printf(" System not exhibiting real behavior.\n");
|
||
}
|
||
|
||
printf("\nOutput files:\n");
|
||
printf(" probeB_telemetry.csv - Step-by-step telemetry\n");
|
||
printf(" guardian_census.csv - Guardian positions/mass/velocity\n");
|
||
printf(" guardian_census.json - JSON format for compatibility\n");
|
||
|
||
/* ---- Cleanup ------------------------------------------------------- */
|
||
fclose(telemetry_csv);
|
||
cudaFree(f0); cudaFree(f1);
|
||
cudaFree(rho); cudaFree(ux); cudaFree(uy);
|
||
nvmlShutdown();
|
||
|
||
return (passes >= 3) ? 0 : 1;
|
||
} |