Re:Increasing the D Compiler Speed by Over 75%
by acqq on Thursday July 25, 2013 @08:21PM
(#6105867)
In every line there are more consecutive whitespaces at the begining unless the code isn't indented.
Re:Increasing the D Compiler Speed by Over 75%
by nkurz on Thursday July 25, 2013 @09:37PM
(#6106117)
Great, that sounds like a fine opportunity for a "best guess": if newline, expect whitespace. Currently, I'd guess that 'whitespace except newline' is the default prediction for the switch() at line 451. I'd also guess that if not followed by a space or tab, a newline is frequently followed by another newline.
Maybe you could combine the case statements for space and newline, and do a branchless 'cmov' to increment loc.linenum if the match was a newline? This could be combined with loop to grab all the whitespace/newlines in one if you think whitespace is occurs in clumps.
FWIW I'd never do a CPU dependent asm code in the lex.
I played with how to phrase that. You don't need to actually use a CMOV, just write code that allows your compiler to use one if supported.
int tmp = real;
if (foo == bar) {
tmp = newval;
}
real = tmp;
I've seen it referred to as a 'hammock'[1], and at least for GCC and ICC it usually is a strong enough hint.[1] http://people.engr.ncsu.edu/ericro/publications/conference_M...
Other tokens probably have similar but less strong "preferences". It's possible that a smart enough branch predictor will learn these on its own, but changing things so that each token ends with its own predicted branch would likely be a win.
Since the branch predictor uses the address of the branch as the 'key', the general goal is to have a separate branch instruction (if, switch) in each place where the probabilities are different enough to favor a different prediction. A wrong prediction costs about 15 cycles, but a correctly predicted branch costs less than 1, so you can put in quite a lot of these and still come out ahead.
Perhaps just ending every token handler with a best guess at what comes next?
I don't have the syntax in my head, by I think you can use 'perf' to record mispredicted indirect branches and then display that sorted by caller.