[PATCH] D86295: [analyzer] Reorder the layout of MemRegion and cache by hand for optimal size

[Gabor Marton via Phabricator via cfe-commits]

martong added a comment.

In D86295#2233068 <https://reviews.llvm.org/D86295#2233068>, @martong wrote:

> In D86295#2232005 <https://reviews.llvm.org/D86295#2232005>, @steakhal wrote:
>
>> In D86295#2231760 <https://reviews.llvm.org/D86295#2231760>, @NoQ wrote:
>>
>>> I mean, like, you can measure the entire process with `time` or something like that. I believe @vsavchenko's docker thingy already knows how to do that.
>>
>> I tried to measure this with `time`, without much luck:
>> F12735724: time-verbose-results.txt <https://reviews.llvm.org/F12735724>
>>
>>> What i'm trying to say is that for almost every region //R// it's a trivial O(1) operation that yields `RegionOffset` (//R//, +0). It is only non-trivial for non-base regions like `FieldRegion` or `ElementRegion` or, well, `CXXBaseObjectRegion`.
>>
>> What should be my next step?
>
> Can we see any improvements in CPU counters (Linux)?
>
>   # Detailed CPU counter statistics (includes extras) for the specified command:
>   perf stat -d command

I am hoping to see a decrease in L1 <https://reviews.llvm.org/L1>-dcache-load-misses ...


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D86295/new/

https://reviews.llvm.org/D86295