• KnightMare (3/3) Jun 23 2019 https://github.com/microsoft/mimalloc
• =?iso-8859-1?Q?Robert_M._M=FCnch?= (7/8) Jul 02 2019 Wouldn't such a specialized allocator help the D runtime to have a
• a11e99z (10/16) Jul 03 2019 I share link coz I saw a few videos by Andrey and Walter where

KnightMare <black80 bk.ru> writes:

https://github.com/microsoft/mimalloc
3.5 KLOC
with benchmarks

=?iso-8859-1?Q?Robert_M._M=FCnch?= <robert.muench saphirion.com> writes:

On 2019-06-23 13:16:56 +0000, KnightMare said:

 https://github.com/microsoft/mimalloc - 3.5 KLOC with benchmarks

Wouldn't such a specialized allocator help the D runtime to have a 
lower footprint and make it possible to optimize allocations patterns?

-- 
Robert M. Münch
http://www.saphirion.com
smarter | better | faster

a11e99z <black80 bk.ru> writes:

On Tuesday, 2 July 2019 at 09:28:31 UTC, Robert M. Münch wrote:
 On 2019-06-23 13:16:56 +0000, KnightMare said:

 https://github.com/microsoft/mimalloc - 3.5 KLOC with 
 benchmarks

 Wouldn't such a specialized allocator help the D runtime to 
 have a lower footprint and make it possible to optimize 
 allocations patterns?

I share link coz I saw a few videos by Andrey and Walter where 
they talks about memory allocations
and there is talk about decoupling from C-runtime
and "the core library is less than 3500 LOC, much smaller than 
the core of other industrial strength allocators like 
tcmalloc(~20k LOC) and jemalloc(~25k LOC)." from 
https://www.microsoft.com/en-us/research/uploads/prod/2019/06/mimalloc-tr-v1.pdf
so probably such implemetnation can be useful: less code = easy 
support