The most interesting question is where these code increases are situated. For example, if the majority of the increase is in the testsuite, that's good news. If they are in the compiler code, probably not so good. If they are in architecture specific code, then that's probably okay.
Before 1997, the GCC release manager (RMS up through about version 2.5.5, Richard Kenner from then until the EGCS revolution) did all of the commits. This means that counting commits as their contributions does not work.
Both RMS and Kenner were big contributors; RMS did the first few GCC releases entirely on his own. But Phoronix has made a mistake in combining together pre-EGCS and post-EGCS commits.