Provides rough estimates of peak memory usage for different synthdid operations.
Useful for assessing feasibility of large-scale computations before running them.
Usage
synthdid_memory_estimate(
N,
T,
K = 0,
replications = SYNTHDID_SE_REPLICATIONS_DEFAULT,
include_se = TRUE
)
Arguments
- N
Total number of units (control + treated)
- T
Total number of time periods (pre + post)
- K
Number of covariates (default: 0)
- replications
Number of bootstrap/placebo replications for SE (default: 200)
- include_se
Logical. Include memory for SE computation (default: TRUE)
Value
List with memory estimates in GB for different components
Examples
if (FALSE) { # \dontrun{
# Estimate memory for California Prop 99 scale problem
synthdid_memory_estimate(N = 39, T = 31, K = 0, replications = 1000)
# Large problem with covariates
synthdid_memory_estimate(N = 1000, T = 100, K = 50, replications = 1000)
} # }