speed up model fitting in foreach loop in R - r

Question

I have a very large dataframe with > 5 millions obs. and 7 var.
I have to do thousands of bootstrap resampling and create a model with each sample, make predictions and save the results in a dataframe where each column is the prediction of one of the models
here is my code:
library(caret)
library(randomForest)
library(foreach)
library(doParallel)
#very small database example
db <-
data.frame(Y = as.factor(sample(c(0, 1), 1000, replace = T)),
X1 = sample(c(100:2000), 1000, replace = T),
X2=sample(c(100:2000), 1000, replace = T),
X3 = sample(c(100:2000), 1000, replace = T),
X4 = sample(c(100:2000), 1000, replace = T))
#bootstrap 100 times
times <- 100
set.seed(5)
#create 1000 resample indexes
res <- createResample(db$Y,times = times)
db_list <- list()
#list of database resampled
for (i in 1:times) {
db_list[[i]] <- db[res[[i]],]
}
#create 100 models from the db_list and make prediction
model_list <- list()
pred <- data.frame(Y=rep(NA,nrow(db)))
#on my machin this loop take 28.10 sec elapsed
system.time(for (l in seq_len(times)) {
set.seed(5)
model_list[[l]] <- randomForest(Y~.,data=db_list[[l]])
pred[,l] <- predict(model_list[[l]],newdata = db[,-1])#predict from the original Xn variables
db_pred <- db
db_pred <- cbind(db_pred,pred)
})
#make cluster with 11 core in my machine
cl <- makeCluster(detectCores()-1)
registerDoParallel(cl)
#on my machin this loop take 23.25 sec elapsed
system.time(foreach(l = seq_len(times),.packages = "randomForest")%dopar% {
set.seed(5)
model_list[[l]] <- randomForest(Y~.,data=db_list[[l]])
pred[,l] <- predict(model_list[[l]],newdata = db[,-1])#predict from the original Xn variables
db_pred <- db
db_pred <- cbind(db_pred,pred)
})
stopCluster(cl)
the code work well with a small dataframe but take forever with my data, despite the foreach loop.
I've tried with caret::train() but I have not found the way to store the predictions of all the models in a dataframe.
The question are:
how can I speed up the process?
Are there classifiers faster than randomforest that could be used with this data?
Thanks in advance for any help