August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to MrSmith | On Wednesday, 26 August 2020 at 16:08:15 UTC, MrSmith wrote:
> https://github.com/MrSmith33/tiny_jit
What kinds of memory management are supported/planned and how does this interact with slices?
I can't find any code examples in Vox. Are they represented as strings inside the .d-files?
|
August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to Per Nordlöw | On Wednesday, 26 August 2020 at 16:34:39 UTC, Per Nordlöw wrote: > On Wednesday, 26 August 2020 at 16:08:15 UTC, MrSmith wrote: >> https://github.com/MrSmith33/tiny_jit > > What kinds of memory management are supported/planned and how does this interact with slices? > > I can't find any code examples in Vox. Are they represented as strings inside the .d-files? Currently only manual MM. Check source/tests for small examples and https://github.com/MrSmith33/rltut_2019 for bigger project. Also see spec/index.md for some docs. |
August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to Per Nordlöw | On Wednesday, 26 August 2020 at 16:34:39 UTC, Per Nordlöw wrote:
> On Wednesday, 26 August 2020 at 16:08:15 UTC, MrSmith wrote:
>> https://github.com/MrSmith33/tiny_jit
>
> What kinds of memory management are supported/planned and how does this interact with slices?
>
> I can't find any code examples in Vox. Are they represented as strings inside the .d-files?
I run benchmark myself and here is what Vox code should look like based on D code:
i64 add_long_n0_h0[T](i64 x) { return x + 87734; }
i64 add_long_n0[T](i64 x) { return x + add_long_n0_h0[i64](x) + 40209; }
i64 main() {
i64 long_sum = 0;
long_sum += add_long_n0[i64](0);
return long_sum;
}
|
August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to MrSmith | On Wednesday, 26 August 2020 at 17:27:48 UTC, MrSmith wrote:
> I run benchmark myself and here is what Vox code should look like based on D code:
Ok, thanks.
How should I best build vox for maximum performance?
These alternatives correctly produce a binary:
dmd -i main.d
ldmd2 -i main.d
ldmd2 -O -release -i main.d
but running the binary produced
ldmd2 -O -release -i main.d
prints
Running 218 tests
but never completes...
|
August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to Per Nordlöw | On Wednesday, 26 August 2020 at 19:26:17 UTC, Per Nordlöw wrote:
> How should I best build vox for maximum performance?
It needs cli version passed:
ldc2 -d-version=cli -m64 -O3 -release -boundscheck=off -enable-inlining -flto=full -i main.d -of=./tjc
|
August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to MrSmith | On Wednesday, 26 August 2020 at 20:10:09 UTC, MrSmith wrote:
> It needs cli version passed:
> ldc2 -d-version=cli -m64 -O3 -release -boundscheck=off -enable-inlining -flto=full -i main.d -of=./tjc
Ahh, nice.
|
August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to MrSmith | On Wednesday, 26 August 2020 at 17:01:44 UTC, MrSmith wrote:
> Currently only manual MM. Check source/tests for small examples and https://github.com/MrSmith33/rltut_2019 for bigger project. Also see spec/index.md for some docs.
Can you elaborate on what you mean manual MM? Refcounted? I'm very curious.
|
August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to Per Nordlöw | On Wednesday, 26 August 2020 at 20:13:55 UTC, Per Nordlöw wrote:
> Can you elaborate on what you mean manual MM? Refcounted? I'm very curious.
Currently you can either do static/stack allocation and use host provided functions or OS API to dynamically allocate memory (alloc/free, mmap/VirtualAlloc/HeapAlloc etc). No GC or RC is planned.
Probably I will add more first class support for allocators like in Jai/Zig/Odin.
|
August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to MrSmith | On Wednesday, 26 August 2020 at 20:10:09 UTC, MrSmith wrote:
> It needs cli version passed:
> ldc2 -d-version=cli -m64 -O3 -release -boundscheck=off -enable-inlining -flto=full -i main.d -of=./tjc
Can `vox` only output Windows binaries?
|
August 26, 2020 Re: Is there any language that native-compiles faster than D? | ||||
---|---|---|---|---|
| ||||
Posted in reply to Per Nordlöw | On Wednesday, 26 August 2020 at 22:05:51 UTC, Per Nordlöw wrote:
> Can `vox` only output Windows binaries?
Yes. ELF and SytemV ABI is WIP
|
Copyright © 1999-2021 by the D Language Foundation