the parsing process is not inherently slow

[Andy Armstrong]

On 2 Oct 2007, at 17:49, Eric Wilhelm wrote:
> Even if an improved parser speed becomes a bottleneck (at around 900
> cores), we can deal with that in better ways than adding conditionals
> to the harness.

For that to be true the parser would have to be capable of consuming  
TAP 900 times faster than the tests produced it. Do you believe  
that's possible?

Insisting that all the parsing happen on a single core is always  
going to hit a brick wall at some point. Right now on dual core  
systems we've already hit that wall for tests that spit out a lot of  
TAP quickly. Those tests (Regexp::Common is an example) run slower  
with muxed parsers than they do with a single parser - because we're  
already maxxing out our core and the extra housekeeping of muxed  
parsers just adds additional overhead.

Multiplexed is currently a win only in cases where we currently have  
a lot of headroom on the CPU that's running the parser - which right  
now means tests that output TAP pretty slowly. So, yes, a faster  
parser would give us more headroom and make multiplexing effective in  
more cases.

-- 
Andy Armstrong, Hexten