Rodrigo Rodriguez (Pragmatismo)
45e4a5e735
- Update prompt formatting in BotOrchestrator to use clearer labels (SYSTEM/CONTEXT) with emphasis markers - Remove unused token_ratio field from SystemMetrics struct - Increase default context size (2048->4096) and prediction length (512->1024) in config - Clean up metrics calculation by removing redundant token ratio computation The changes improve readability of system prompts and simplify metrics collection while increasing default model capacity.
78 lines
2 KiB
Rust
78 lines
2 KiB
Rust
use anyhow::Result;
|
|
use log::warn;
|
|
use std::collections::HashMap;
|
|
use sysinfo::{System};
|
|
|
|
/// System monitoring data
|
|
pub struct SystemMetrics {
|
|
pub gpu_usage: Option<f32>,
|
|
pub cpu_usage: f32,
|
|
}
|
|
|
|
/// Gets current system metrics
|
|
pub fn get_system_metrics(current_tokens: usize, max_tokens: usize) -> Result<SystemMetrics> {
|
|
let mut sys = System::new();
|
|
sys.refresh_cpu_usage();
|
|
|
|
// Get CPU usage (average across all cores)
|
|
let cpu_usage = sys.global_cpu_usage();
|
|
|
|
// Get GPU usage if available
|
|
let gpu_usage = if has_nvidia_gpu() {
|
|
get_gpu_utilization()?.get("gpu").copied()
|
|
} else {
|
|
None
|
|
};
|
|
|
|
|
|
Ok(SystemMetrics {
|
|
gpu_usage,
|
|
cpu_usage,
|
|
})
|
|
}
|
|
|
|
/// Checks if NVIDIA GPU is available
|
|
pub fn has_nvidia_gpu() -> bool {
|
|
match std::process::Command::new("nvidia-smi")
|
|
.arg("--query-gpu=utilization.gpu")
|
|
.arg("--format=csv,noheader,nounits")
|
|
.output()
|
|
{
|
|
Ok(output) => output.status.success(),
|
|
Err(_) => {
|
|
warn!("No NVIDIA GPU detected or nvidia-smi not available");
|
|
false
|
|
}
|
|
}
|
|
}
|
|
|
|
/// Gets current GPU utilization percentages
|
|
pub fn get_gpu_utilization() -> Result<HashMap<String, f32>> {
|
|
let output = std::process::Command::new("nvidia-smi")
|
|
.arg("--query-gpu=utilization.gpu,utilization.memory")
|
|
.arg("--format=csv,noheader,nounits")
|
|
.output()?;
|
|
|
|
if !output.status.success() {
|
|
return Err(anyhow::anyhow!("Failed to query GPU utilization"));
|
|
}
|
|
|
|
let output_str = String::from_utf8(output.stdout)?;
|
|
let mut util = HashMap::new();
|
|
|
|
for line in output_str.lines() {
|
|
let parts: Vec<&str> = line.split(',').collect();
|
|
if parts.len() >= 2 {
|
|
util.insert(
|
|
"gpu".to_string(),
|
|
parts[0].trim().parse::<f32>().unwrap_or(0.0),
|
|
);
|
|
util.insert(
|
|
"memory".to_string(),
|
|
parts[1].trim().parse::<f32>().unwrap_or(0.0),
|
|
);
|
|
}
|
|
}
|
|
|
|
Ok(util)
|
|
}
|