[BUG] Segfault when accessing result of tensor.__add__

[ss32]

Bug description

When accessing the result of tensor.__mul__ the program crashes. Computing the result does not elicit the crash, but attempting to print it or store it to a list does.

Stack trace:

Stack dump:
0.	Program arguments: mojo run grav.mojo
 #0 0x000056834b6f3407 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) (/home/user/.modular/pkg/packages.modular.com_mojo/bin/mojo+0x128c407)
 #1 0x000056834b6f125e llvm::sys::RunSignalHandlers() (/home/user/.modular/pkg/packages.modular.com_mojo/bin/mojo+0x128a25e)
 #2 0x000056834b6f3a9f SignalHandler(int) (/home/user/.modular/pkg/packages.modular.com_mojo/bin/mojo+0x128ca9f)
 #3 0x0000708112642520 (/lib/x86_64-linux-gnu/libc.so.6+0x42520)
 #4 0x000056834cec03d6 tcmalloc::SLL_Next(void*) (/home/user/.modular/pkg/packages.modular.com_mojo/bin/mojo+0x2a593d6)
 #5 0x000056834cec0473 tcmalloc::SLL_TryPop(void**, void**) (/home/user/.modular/pkg/packages.modular.com_mojo/bin/mojo+0x2a59473)
 #6 0x000056834cec0877 tcmalloc::ThreadCache::FreeList::TryPop(void**) (/home/user/.modular/pkg/packages.modular.com_mojo/bin/mojo+0x2a59877)
 #7 0x000056834db24350 tc_newarray_aligned (/home/user/.modular/pkg/packages.modular.com_mojo/bin/mojo+0x36bd350)
 #8 0x000070810863b7cc KGEN_CompilerRT_AlignedAlloc (/home/user/.modular/pkg/packages.modular.com_mojo/lib/libKGENCompilerRTShared.so.19.0git+0x3b7cc)
 #9 0x0000708098005185 
mojo crashed!
Please file a bug report.
Segmentation fault (core dumped)

Steps to reproduce

• Include relevant code snippet or link to code that did not work as expected.

I'm unable to reproduce in a smaller toy example but will update this if I can. Commenting out solutions.append(xx) yields no errors. The same crash can be seen if the call to append is replaced with print(xx).

from tensor import Tensor
import math
 
 
alias G = 1.0
alias M = 1.0
alias m = 1.0
alias dt = 0.1
alias tEnd = 100
alias nSteps = int(tEnd / dt)
alias float_type = DType.float64
alias simd_width = 2 * simdwidthof[float_type]()
 
 
fn norm(v: Tensor[float_type]) -> SIMD[float_type, 1]:
    return math.sqrt(v[0] * v[0] + v[1] * v[1] + v[2] * v[2])
 
fn two_body_first_order_ode(
    x: Tensor[float_type],
    t: Float64,
    masses: Tensor[float_type],
) -> Tensor[float_type]:
    var len = x.shape()[0]
    var dxdt = Tensor[float_type](len,1)
    var r1 = Tensor[float_type](3)
    var r2 = Tensor[float_type](3)
    for i in range(3):
        r1[i] = x[i]
        r2[i] = x[i + 3]
    try:
        var rr = r2 - r1
 
        var r3 = norm(rr)
        for i in range(3):
            dxdt[i] = x[i + 5]
            dxdt[i + 3] = x[i + 8]
            dxdt[i + 5] = -G * masses[0] * (r2[i] - r1[i]) / r3
            dxdt[i + 8] = -G * masses[1] * (r1[i] - r2[i]) / r3
        return dxdt
    except:
        print("Error in rr")
    return Tensor[float_type](len,1)
 
 
def compute_orbits():
    # Initial conditions
    r_sun = Tensor[float_type](
        (3, 1),
        -1,
        0,
        1,
    )
    r_earth = Tensor[float_type](
        (3, 1),
        1,
        0,
        0,
    )
    v_sun = Tensor[float_type](
        (3, 1),
        0,
        -0.5,
        0,
    )
    v_earth = Tensor[float_type](
        (3, 1),
        0,
        0.5,
        0,
    )
    earth_positions = Tensor[float_type](
        (3, nSteps),
    )
    sun_positions = Tensor[float_type](
        (3, nSteps),
    )
    x = Tensor[float_type](
        (9, 1),
    )
    for i in range(3):
        x[i] = r_earth[i]
        x[i + 3] = r_sun[i]
        x[i + 5] = v_earth[i]
        x[i + 8] = v_sun[i]
    masses = Tensor[float_type](
        (2, 1),
        m,
        M,
    )
    solutions = List[Tensor[DType.float64]]()
    for t in range(nSteps):
        dxdt = two_body_first_order_ode(x, t, masses)
        var m = (dxdt.__mul__(dt))
        var xx = x.__add__(m)
        solutions.append(xx)
 
fn main():
    try:
        compute_orbits()
    except Exception:
        print("Error in main")
        return

• If applicable, add screenshots to help explain the problem.

n/a

• If using the Playground, name the pre-existing notebook that failed and the steps that led to failure.

n/a

• Include anything else that might help us debug the issue.

System information

• What OS did you do install Mojo on ?

OS: Ubuntu Cinnamon 22.04.4 LTS x86_64
Host: MS-7C37 1.0
Kernel: 6.5.0-28-generic 
CPU: AMD Ryzen 7 3700X (16) @ 3.600GHz

• Provide version information for Mojo by pasting the output of mojo -v

$ mojo -v 
mojo 24.3.0 (9882e19d)

• Provide Modular CLI version by pasting the output of modular -v

$ modular -v
modular 0.7.4 (df7a9e8b)