# Debugging float SIMD Intrinsics via llvmcall

**URL:** <https://discourse.julialang.org/t/debugging-float-simd-intrinsics-via-llvmcall/27855>\
**Category:** General Usage\
**Created:** [August 22, 2019, 3:53pm UTC](https://discourse.julialang.org/t/debugging-float-simd-intrinsics-via-llvmcall/27855 "2019-08-22T15:53:36Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![smallnamespace](https://avatars.discourse-cdn.com/v4/letter/s/a9a28c/32.png) [@smallnamespace](https://discourse.julialang.org/u/smallnamespace)\
**Post date:** [August 22, 2019, 3:53pm UTC](https://discourse.julialang.org/t/debugging-float-simd-intrinsics-via-llvmcall/27855/1 "2019-08-22T15:53:36Z")

</div>

I’m trying to set up a SIMD intrinsic for `rsqrt`:

```julia
function rsqrt(f::Vec{4, Float32})
    llvmcall(("declare <4 x float> @llvm.x86.sse.rsqrt.ps(<4 x float>) nounwind readnone", " 
              %2 = call <4 x float> @llvm.x86.sse.rsqrt.ps(<4 x float> %0)
              ret <4 x float> %2"), Vec{4, Float32}, (Vec{4, Float32},), f)
end

```

I’m fairly sure this has the correct mapping from Julia’s `Vec{4, Float32}` to LLVM’s `<4 x float>` vector type, but I’m still seeing:

```julia
error compiling rsqrt: error statically evaluating llvmcall argument tuple

```

What’s would be the best way to debug the LLVM error? Godbolt seems to think the LLVM instructions are fine: [Compiler Explorer](https://gcc.godbolt.org/z/4uVJE-)

---

<div class="post-metadata">

**Author:** ![kristoffer.carlsson](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/kristoffer.carlsson/32/22_2.png) [@kristoffer.carlsson](https://discourse.julialang.org/u/kristoffer.carlsson)\
**Post date:** [August 22, 2019, 4:03pm UTC](https://discourse.julialang.org/t/debugging-float-simd-intrinsics-via-llvmcall/27855/2 "2019-08-22T16:03:32Z")

</div>

```julia
function rsqrt(f::Vec{4, Float32})
    v = ccall("llvm.x86.sse.rsqrt.ps", llvmcall, NTuple{4, VecElement{Float32}}, (NTuple{4, VecElement{Float32}},), f.elts);
    Vec{4, Float32}(v)
end

```

might be a bit easier than text `llvmcall`:

```julia
julia> v = Vec{4, Float32}((1.0, 2.0, 3.0, 4.0))
<4 x Float32>[1.0, 2.0, 3.0, 4.0]
julia> rsqrt(v)
<4 x Float32>[0.99975586, 0.7069092, 0.5772705, 0.49987793]

```

The point is that you cannot use `Vec` directly for the argument and argument types. It needs to be the native type `NTuple{4, VecElement{Float32}}`

[SIMD and SIMD-intrinsics in Julia | Kristoffer Carlsson](http://kristofferc.github.io/post/intrinsics/) might have something useful.

---

<div class="post-metadata">

**Author:** ![smallnamespace](https://avatars.discourse-cdn.com/v4/letter/s/a9a28c/32.png) [@smallnamespace](https://discourse.julialang.org/u/smallnamespace)\
**Post date:** [August 22, 2019, 4:20pm UTC](https://discourse.julialang.org/t/debugging-float-simd-intrinsics-via-llvmcall/27855/3 "2019-08-22T16:20:29Z")

</div>

Thank you! I agree that `ccall` is probably better for a single intrinsic. For more context, I’m actually trying to write:

`_mm_cvtps_pd(_mm_rsqrt_ps(_mm_cvtpd_ps(m128d)))`

Can this work with nested `ccall` or does the fact that we’re converting contiguous 4 bytes from a pair of doubles to a pair of floats and back make types tricky?

Also, I didn’t have any luck substituting out `Vec`, still getting the same error with this:

```julia
function rsqrt(f::NTuple{4, VecElement{Float32}})
    llvmcall(("declare <4 x float> @llvm.x86.sse.rsqrt.ps(<4 x float>) nounwind readnone", " 
              %2 = call <4 x float> @llvm.x86.sse.rsqrt.ps(<4 x float> %0)
              ret <4 x float> %2"),
        NTuple{4, VecElement{Float32}},
        (NTuple{4, VecElement{Float32}},), f)
end

```

---

<div class="post-metadata">

**Author:** ![kristoffer.carlsson](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/kristoffer.carlsson/32/22_2.png) [@kristoffer.carlsson](https://discourse.julialang.org/u/kristoffer.carlsson)\
**Post date:** [August 22, 2019, 4:24pm UTC](https://discourse.julialang.org/t/debugging-float-simd-intrinsics-via-llvmcall/27855/4 "2019-08-22T16:24:10Z")

</div>

It needs to be `Tuple{NTuple{4, VecElement{Float32}},}` for the input argument type. Things are a bit inconsistent when you can use a tuple and when you need a tuple type.

---

<div class="post-metadata">

**Author:** ![smallnamespace](https://avatars.discourse-cdn.com/v4/letter/s/a9a28c/32.png) [@smallnamespace](https://discourse.julialang.org/u/smallnamespace)\
**Post date:** [August 22, 2019, 4:56pm UTC](https://discourse.julialang.org/t/debugging-float-simd-intrinsics-via-llvmcall/27855/5 "2019-08-22T16:56:36Z")

</div>

Perfect, that’s exactly what I missed.

For anyone else trying this, using a final tuple value works seems to work with other argument types, but not a vector of floats.