Displaying 3 results from an estimated 3 matches for "rateformula".
2019 Jun 06
2
Strange behaviour of post-legalising optimisations(?)
...look.
I have been following the code in the debugger, specially the LSRInstance::SolveRecurse function. This function traverses recursively all possible ‘Formulae’, and determines the best instruction combination for the loop generation, based on minimal cost. The SolveRecurse function uses Cost::RateFormula as the basis for the cost computation.
I found that the RateFormula function does not return the what-would-be best values for my architecture. In particular, my understanding is that the whole implementation kind of assumes that all architectures will have post-increment addressing modes, even i...
2019 Jun 05
2
Strange behaviour of post-legalising optimisations(?)
I come across a situation that I am having a hard time to understand.
When I compile the following code :
char *tst( char *dest, const char *src, unsigned int len )
{
for (int i=0 ; i<len ; i++) {
dest[i] = src[i];
}
return dest;
}
Clang generates this for the ‘for’ body:
for.body: ; preds = %for.cond
%arrayidx = getelementptr inbounds i8,
2018 May 01
0
LSR formula rating
...ssingMode
There are many addressing modes that are legal on Hexagon, yet given a
choice, some are preferable to others. LSR treats all of them as
equivalent as long as a formula can be folded completely into it. Has
anyone thought about making the cost of an addressing mode be more detailed?
4. RateFormula
A lot of that code looks rather mysterious. Is there a document that was
the original inspiration?
-Krzysztof
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
hosted by The Linux Foundation