diogo...@gmail.com (Diogo F. S. Ramos) writes:
> David Bremner writes:
>
>>> STRIPPED
>>> size: 4408K
>>> startup time (`$ racket -e 42'): 500ms
>>> UNSTRIPPED
>>> size: 10254K
>>> startup time (`$ racket -e 42'): 298ms
>>>
>> It would be interesting to start both from a cold cache, e.g.
David Bremner writes:
>> STRIPPED
>> size: 4408K
>> startup time (`$ racket -e 42'): 500ms
>> UNSTRIPPED
>> size: 10254K
>> startup time (`$ racket -e 42'): 298ms
>>
> It would be interesting to start both from a cold cache, e.g. by
> running
>
> echo 3 | sudo tee /proc/sys/vm/drop_caches
"Diogo F. S. Ramos" writes:
> Here are the results:
>
> STRIPPED
> size: 4408K
> startup time (`$ racket -e 42'): 500ms
> UNSTRIPPED
> size: 10254K
> startup time (`$ racket -e 42'): 298ms
>
Hi Diogo;
Thanks for running those experiments. The results are pretty
surprising. It would be
David Bremner writes:
> - About the size increase, on amd64 the racket binary is currently 3.3M.
> Diogo reported unstripped size of 10M. That's a large increase, but
> currently the combination of racket-common and racket is about 190M,
> so it would be about a 5% increase overall.
>
> - I
"Diogo F. S. Ramos" writes:
> Package: racket
> Severity: wishlist
>
> Dear Maintainer,
>
> When a racket binary is stripped from its debugging info, it's not
> possible to have a proper report from the `profile' collection on
> x86_64 [1].
>
> While the binary size grows with the debug informati
Package: racket
Severity: wishlist
Dear Maintainer,
When a racket binary is stripped from its debugging info, it's not
possible to have a proper report from the `profile' collection on
x86_64 [1].
While the binary size grows with the debug information, I find the
statistical profiler invaluable.
6 matches
Mail list logo
