[Exim] SpamAssassin + Exim Question

I'm curious about how SA integrates with Exim, because I'm concerned about the amount of cpu-time SA is taking up when it scans e-mails.

What I'm wondering is that, if the $spam_ related functions are called from different warn and deny sections, is the SA daemon being called to scan an e-mail each time the $spam_ is used?

Or, to say it a different way... for each incoming e-mail, regardless of how many times $spam_score is used in various ACLs, is the e-mail only scanned and scored by SA once, or is SA scanning and re-scoring the e-mail each time to determine the $spam_score?

Here's the ACL i'm using:

check_data:

  # Scan the msg and add X-Spam-Score header:
  warn  spam = nobody:true
        message = X-Spam-Score:  $spam_score ($spam_bar)

  # Add "X-Spam-Flag" header if its probably spam and write a warning
  # to the log.
  warn  condition = ${if > {$spam_score_int}{30}{yes}{no}}
        message = X-Spam-Flag: YES
        message = X-Spam-Report: $spam_report
        log_message = spam score $spam_score for $recipients

# Deny, if the message is over the reject threshold

  deny condition = ${if > {$spam_score_int}{60}{yes}{no}}
       log_message = spam score $spam_score for $recipients
       message = This message scored $spam_score spam points \
                 and has been rejected.\n

accept



---Brett.