7e068154c6
[artiq] WIP
e26d6eb1b1
[artiq] WIP
69d14a3656
[artiq] WIP
4b7bfbd141
[artiq] WIP
bc9a426d32
[artiq] Fix RPC of ndarrays from host
Fix several missing i1/i8 casts
All boolean operations (e.g. icmp
) return i1
, and LLVM's ABI uses i1
for boolean parameters in functions. I would prefer to follow these conventions as we need these casts either way.
7d40808c12
meta: Use mold as linker
a2922b24e0
[core] codegen/stmt: Add more casts for boolean types
be4dd27ec4
[core] codegen/expr: Fix missing cast to i1
5b2b6db7ed
core: improve error messages
15e62f467e
standalone: add tests for polymorphism
145fcceeb3
core: WIP - Add tracer runtime
c8dfdcfdea
standalone & artiq: remove class_names from resolver
600a5c8679
Revert "standalone: reformat demo.c"
22c4d25802
core/typecheck: add missing typecheck in matmul
308edb8237
standalone: reformat demo.c
Fix several missing i1/i8 casts
a2922b24e0
[core] codegen/stmt: Add more casts for boolean types
be4dd27ec4
[core] codegen/expr: Fix missing cast to i1
ndstrides: [1] Introduce models
I did some testing and this might work:
pub struct CallFunction<'ctx, 'a, 'b, G: CodeGenerator + ?Sized> {
generator: &'b mut G,
ctx: &'b CodeGenContext<'ctx, 'a>,
name: &'b…
ndstrides: [1] Introduce models
I guess you mean there could have been a single visitor that collects the fields into a
Vec<dyn Model>
and do stuff with thatVec
(like checking field types and more).The short…
ndstrides: [1] Introduce models
I wouldn't be too concerned with using PhantomData
here. I'd rather the type bounds of the struct be explicit so that users of this class do not misuse the fields.
In fact, we should also…
ndstrides: [1] Introduce models
What about two functions, one that creates a void func and another that creates a value-returning func?