AaronBallman wrote:

> @cor3ntin @AaronBallman that doesn't look overwhelmingly bad to me (e.g. 
> before and after look the same?)

I may be misinterpreting the data, but doesn't this show a significant 
regression (numbers taken from the stress test):
```
--------------------------------------------------------------------------------
  n        time(i)         total(B)   useful-heap(B) extra-heap(B)    stacks(B)
--------------------------------------------------------------------------------
  0              0                0                0             0            0
  1    186,864,557        3,485,296        3,336,946        95,382       52,968
  2    477,143,082        6,980,720        6,693,115       219,853       67,752
```
and
```
--------------------------------------------------------------------------------
  n        time(i)         total(B)   useful-heap(B) extra-heap(B)    stacks(B)
--------------------------------------------------------------------------------
  0              0                0                0             0            0
  1    286,291,056        4,757,792        4,573,558       138,186       46,048
  2    591,222,234        8,392,712        8,064,059       267,949       60,704
  3    822,224,951       10,865,504       10,439,613       365,299       60,592
```
It seems like we're taking significantly longer and using more heap space?

https://github.com/llvm/llvm-project/pull/192080
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to