 |
Post AvexkK85S99XsCNn4y by [email protected] |
 |
More posts by [email protected] |
|
Post #AvexhmB8i5iJzuaztw by [email protected] |
|
0 likes, 0 repeats |
|
|
|
Great example of "when switching to a different general purpose allocator … |
|
|
|
Post #AvexhmJIDmEgPCPVbs by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh I really don't like sweeping generalizations like this.Rust and com… |
|
|
|
Post #AvexhmPJrN3YhtEK0G by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@zeux IMHO, if a C stdlib has the choice between simplicity and performance, it… |
|
|
|
Post #AvexhmVhTeA11gDPwu by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh "should always go for simplicity" is definitely not right. Sh… |
|
|
|
Post #AvexhmbN8YhJJGrwn2 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@zeux ok, but memcpy and memset are a bit special because they are essentially … |
|
|
|
Post #AvexhmxLoqHKPR9V32 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@zeux ...it's also not just malloc/free, filesystem operations drastically … |
|
|
|
Post #AvexhtEwSLultjQdIe by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh They are only builtins when inlined.I agree a choice would be nice. But… |
|
|
|
Post #AvexiXyN0s1KECzTf6 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh @zeux rust *already* does this, it's just not applied to alpine bui… |
|
|
|
Post #AvexiY5SaVgwaCJ8iG by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@dotstdy @zeux that almost sounds like a bug tbh... I wonder if jemalloc and MU… |
|
|
|
Post #AvexipH4ga5Dt8s9g0 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh @zeux no it's nothing like that, you can enable jemalloc fine on al… |
|
|
|
Post #AvexiqUEBACNeE2Aoi by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh @zeux rust works great with musl fwiw, here's the process for makin… |
|
|
|
Post #AvexjPkP50vz0Ws5OC by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh I mean, to an extent, but the real issue there is that musl's alloc… |
|
|
|
Post #AvexjPqmhI2RKJrBKq by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@dotstdy IMHO an application shouldn't expect that the default allocator is… |
|
|
|
Post #AvexjPxAJZ8te6qHHU by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh it's not assuming the allocator is fast though, it's assuming t… |
|
|
|
Post #AvexjQ3XvqFLxtpNE8 by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh like if your system puts a giant lock around malloc that's *worse* … |
|
|
|
Post #AvexjQ9ZZR4EGaeBcW by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@dotstdy I would argue that doing so many allocations across multiple threads i… |
|
|
|
Post #AvexkK0Hv8ulU0jYvI by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@floooh idk how much I believe that honestly. People underestimate how solid a … |
|
|
|
Post #AvexkK85S99XsCNn4y by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@dotstdy @floooh The point is more, if your design absolutely requires a very f… |
|
|
|
Post #AvexkKET4QG0BzMt1c by [email protected] |
|
0 likes, 0 repeats |
|
|
|
@WAHa_06x36 @floooh I mean in practice malloc performance is pretty good, outsi… |
