Module memory_estimation 

Memory estimation for tensor expressions.

This module provides tools for estimating the memory footprint of compiled tensor expressions. This is useful for:

• Planning batch sizes
• Deciding on execution strategies
• GPU memory allocation
• Optimization decisions

Usage

use tensorlogic_compiler::optimize::{estimate_memory, MemoryEstimate};
use tensorlogic_compiler::CompilerContext;
use tensorlogic_ir::{TLExpr, Term};

let mut ctx = CompilerContext::new();
ctx.add_domain("Person", 1000);
ctx.add_domain("Thing", 500);

let expr = TLExpr::pred("knows", vec![Term::var("x"), Term::var("y")]);
let estimate = estimate_memory(&expr, &ctx);

println!("Estimated memory: {} bytes", estimate.total_bytes);
println!("Peak memory: {} bytes", estimate.peak_bytes);

Structs

MemoryEstimate

Detailed memory estimate for an expression.

Functions

estimate_batch_memory

Estimate memory for a batch of similar expressions.

estimate_memory

Estimate memory usage for an expression.