File size: 4,181 Bytes
35aaa09 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 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 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 |
# GPU Configuration
# Docking@HOME - CUDA/CUDPP Settings
# === GPU Selection ===
use_gpu = true # Enable GPU acceleration
gpu_device_id = 0 # GPU device ID to use (-1 = auto-detect best)
use_multiple_gpus = false # Use multiple GPUs if available
gpu_ids = [0] # List of GPU IDs to use (if use_multiple_gpus = true)
# === CUDA Settings ===
cuda_device_name = "auto" # CUDA device name (auto = auto-detect)
cuda_compute_capability = "auto" # Compute capability (auto, 5.0, 6.1, 7.0, 7.5, 8.0, 8.6, 9.0)
cuda_threads_per_block = 256 # Threads per block (64, 128, 256, 512, 1024)
cuda_blocks_per_grid = 128 # Blocks per grid (32, 64, 128, 256)
# === Memory Settings ===
gpu_memory_limit_mb = 0 # GPU memory limit in MB (0 = auto, use available)
host_memory_pinned = true # Use pinned host memory for faster transfers
cache_maps_on_gpu = true # Cache grid maps on GPU memory
# === Performance Tuning ===
# AutoDock-GPU specific settings
energy_eval_per_gpu_call = 1024 # Energy evaluations per GPU kernel call
ls_method = "sw" # Local search method: sw (Solis-Wets), sd (Steepest Descent), fire
population_size = 150 # Population size for genetic algorithm
num_generations = 27000 # Number of generations
# === CUDPP Settings ===
use_cudpp = true # Use CUDPP for GPU primitives
cudpp_sort_algorithm = "radix" # Sort algorithm: radix, merge, quick
cudpp_scan_algorithm = "efficient" # Scan algorithm: efficient, work-efficient
# === Optimization Flags ===
optimize_for_speed = true # Optimize for speed vs accuracy
use_fast_math = true # Use fast math operations (less precise)
use_texture_memory = true # Use texture memory for grid maps
async_execution = true # Asynchronous kernel execution
# === Multi-GPU Load Balancing ===
load_balance_strategy = "dynamic" # static, dynamic, round-robin
tasks_per_gpu_min = 10 # Minimum tasks per GPU
tasks_per_gpu_max = 100 # Maximum tasks per GPU
# === Error Handling ===
retry_on_gpu_error = true # Retry on GPU errors
max_gpu_retries = 3 # Maximum retry attempts
fallback_to_cpu = true # Fallback to CPU on GPU failure
# === Thermal Management ===
enable_thermal_monitoring = true # Monitor GPU temperature
max_gpu_temperature = 85 # Maximum GPU temperature (°C)
throttle_at_temperature = 80 # Start throttling at this temperature (°C)
shutdown_at_temperature = 90 # Emergency shutdown temperature (°C)
check_temperature_interval = 10 # Seconds between temperature checks
# === Power Management ===
gpu_power_limit_watts = 0 # Power limit in watts (0 = default)
enable_power_monitoring = true # Monitor power consumption
# === Debugging ===
verbose_gpu_output = false # Enable verbose GPU output
profile_gpu_kernels = false # Profile GPU kernel execution times
save_gpu_debug_info = false # Save debug information
cuda_error_checking = true # Enable CUDA error checking (slower)
# === Compatibility ===
force_cpu_mode = false # Force CPU mode even if GPU available
gpu_driver_version_min = "450.0" # Minimum GPU driver version
cuda_runtime_version_min = "11.0" # Minimum CUDA runtime version
# === Specific GPU Optimizations ===
# NVIDIA RTX 30xx Series
rtx30xx_optimized = false
# NVIDIA RTX 40xx Series
rtx40xx_optimized = false
# AMD RDNA2/3
amd_rdna_optimized = false
# === Benchmark Settings ===
run_benchmark_on_startup = false # Run benchmark on startup
benchmark_duration_seconds = 30 # Benchmark duration
save_benchmark_results = true # Save benchmark results
# === Advanced CUDA Settings ===
cuda_stream_count = 2 # Number of CUDA streams
cuda_graph_enabled = false # Use CUDA graphs (CUDA 10+)
cooperative_groups = false # Use cooperative groups
unified_memory = false # Use CUDA unified memory
|