UX: Avoid segfault when passed `Array`

[jariji]

If I forget to wrap an array in ROCArray, it segfaults the process, which can be annoying. Would it be possible to throw a normal error instead?

(tmp.JVnIyVIlEw) pkg> st 
Status `/tmp/tmp.JVnIyVIlEw/Project.toml`
  [21141c5a] AMDGPU v1.0.1

julia> using AMDGPU
┌ Warning: rocSPARSE is unavailable, functionality will be disabled.
└ @ AMDGPU ~/.julia/packages/AMDGPU/a1v0k/src/AMDGPU.jl:257
┌ Warning: rocSOLVER is unavailable, functionality will be disabled.
└ @ AMDGPU ~/.julia/packages/AMDGPU/a1v0k/src/AMDGPU.jl:257
┌ Warning: rocALUTION is unavailable, functionality will be disabled.
└ @ AMDGPU ~/.julia/packages/AMDGPU/a1v0k/src/AMDGPU.jl:257
┌ Warning: rocRAND is unavailable, functionality will be disabled.
└ @ AMDGPU ~/.julia/packages/AMDGPU/a1v0k/src/AMDGPU.jl:257
┌ Warning: rocFFT is unavailable, functionality will be disabled.
└ @ AMDGPU ~/.julia/packages/AMDGPU/a1v0k/src/AMDGPU.jl:257
┌ Warning: MIOpen is unavailable, functionality will be disabled.
└ @ AMDGPU ~/.julia/packages/AMDGPU/a1v0k/src/AMDGPU.jl:257

julia> let n = 10, k = 5, p = 3
           Z = (rand(n,k))
           β = -10 .+ rand(k, p)
           Z * β
       end
10×3 Matrix{Float64}:
 -20.3914  -20.394   -19.8545
 -17.1147  -17.5106  -17.2733
 -22.5435  -22.658   -22.2022
 -21.3768  -21.4355  -21.1224
 -13.8111  -13.7183  -13.4202
 -31.5885  -31.4556  -31.1918
 -27.5436  -27.8007  -27.0116
 -17.9013  -18.0811  -17.2966
 -37.8112  -37.7759  -37.1176
 -23.2461  -23.2959  -23.2583

julia> let n = 10, k = 5, p = 3
           Z = ROCArray(rand(n,k))
           β = -10 .+ rand(k, p)
           Z * β
       end

[784824] signal (11.2): Segmentation fault
in expression starting at REPL[5]:1
dgemm_itcopy_ZEN at /nix/store/2j9mrc4cv7ry3581wxy3nbw6qb215dwd-julia-bin-1.10.3/lib/julia/libopenblas64_.so (unknown line)
dgemm_nn at /nix/store/2j9mrc4cv7ry3581wxy3nbw6qb215dwd-julia-bin-1.10.3/lib/julia/libopenblas64_.so (unknown line)
dgemm_64_ at /nix/store/2j9mrc4cv7ry3581wxy3nbw6qb215dwd-julia-bin-1.10.3/lib/julia/libopenblas64_.so (unknown line)
gemm! at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/blas.jl:1524
gemm_wrapper! at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/matmul.jl:605
generic_matmatmul! at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/matmul.jl:352 [inlined]
mul! at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/matmul.jl:263 [inlined]
mul! at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/matmul.jl:237 [inlined]
* at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/matmul.jl:113
unknown function (ip: 0x7f8f1f5edf39)
_jl_invoke at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:2895 [inlined]
ijl_apply_generic at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:3077
jl_apply at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/julia.h:1982 [inlined]
do_call at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/interpreter.c:126
eval_value at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/interpreter.c:223
eval_stmt_value at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/interpreter.c:174 [inlined]
eval_body at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/interpreter.c:617
jl_interpret_toplevel_thunk at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/interpreter.c:775
jl_toplevel_eval_flex at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/toplevel.c:934
jl_toplevel_eval_flex at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/toplevel.c:877
jl_toplevel_eval_flex at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/toplevel.c:877
jl_toplevel_eval_flex at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/toplevel.c:877
ijl_toplevel_eval_in at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/toplevel.c:985
eval at ./boot.jl:385 [inlined]
eval_user_input at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/REPL/src/REPL.jl:150
repl_backend_loop at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/REPL/src/REPL.jl:246
#start_repl_backend#46 at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/REPL/src/REPL.jl:231
start_repl_backend at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/REPL/src/REPL.jl:228
_jl_invoke at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:2895 [inlined]
ijl_apply_generic at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:3077
#run_repl#59 at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/REPL/src/REPL.jl:389
run_repl at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/usr/share/julia/stdlib/v1.10/REPL/src/REPL.jl:375
jfptr_run_repl_91734.1 at /nix/store/2j9mrc4cv7ry3581wxy3nbw6qb215dwd-julia-bin-1.10.3/lib/julia/sys.so (unknown line)
_jl_invoke at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:2895 [inlined]
ijl_apply_generic at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:3077
#1013 at ./client.jl:432
jfptr_YY.1013_82700.1 at /nix/store/2j9mrc4cv7ry3581wxy3nbw6qb215dwd-julia-bin-1.10.3/lib/julia/sys.so (unknown line)
_jl_invoke at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:2895 [inlined]
ijl_apply_generic at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:3077
jl_apply at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/julia.h:1982 [inlined]
jl_f__call_latest at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/builtins.c:812
#invokelatest#2 at ./essentials.jl:892 [inlined]
invokelatest at ./essentials.jl:889 [inlined]
run_main_repl at ./client.jl:416
exec_options at ./client.jl:333
_start at ./client.jl:552
jfptr__start_82726.1 at /nix/store/2j9mrc4cv7ry3581wxy3nbw6qb215dwd-julia-bin-1.10.3/lib/julia/sys.so (unknown line)
_jl_invoke at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:2895 [inlined]
ijl_apply_generic at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/gf.c:3077
jl_apply at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/julia.h:1982 [inlined]
true_main at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/jlapi.c:582
jl_repl_entrypoint at /cache/build/builder-amdci4-2/julialang/julia-release-1-dot-10/src/jlapi.c:731
main at /etc/profiles/per-user/user/bin/julia (unknown line)
__libc_start_call_main at /nix/store/dbcw19dshdwnxdv5q2g6wldj6syyvq7l-glibc-2.39-52/lib/libc.so.6 (unknown line)
__libc_start_main at /nix/store/dbcw19dshdwnxdv5q2g6wldj6syyvq7l-glibc-2.39-52/lib/libc.so.6 (unknown line)
unknown function (ip: 0x4010b8)
Allocations: 23104858 (Pool: 23073400; Big: 31458); GC: 18

[gbaraldi]

This is probably missing something like CUDA's memory API and is just using the normal pointer code.

ERROR: ArgumentError: Illegal conversion of a CUDA.DeviceMemory to a Ptr{Float64}
Stacktrace:
  [1] convert(T::Type{Ptr{Float64}}, mem::CUDA.DeviceMemory)
    @ CUDA ~/.julia/packages/CUDA/FJf6p/lib/cudadrv/memory.jl:16
  [2] convert(::Type{Ptr{Float64}}, managed::CUDA.Managed{CUDA.DeviceMemory})
    @ CUDA ~/.julia/packages/CUDA/FJf6p/src/memory.jl:583
  [3] unsafe_convert(typ::Type{Ptr{Float64}}, x::CuArray{Float64, 2, CUDA.DeviceMemory})
    @ CUDA ~/.julia/packages/CUDA/FJf6p/src/array.jl:464
  [4] gemm!(transA::Char, transB::Char, alpha::Float64, A::CuArray{Float64, 2, CUDA.DeviceMemory}, B::Matrix{Float64}, beta::Float64, C::Matrix{Float64})
    @ LinearAlgebra.BLAS ~/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/share/julia/stdlib/v1.12/LinearAlgebra/src/blas.jl:1642
  [5] gemm_wrapper!(C::Matrix{Float64}, tA::Char, tB::Char, A::CuArray{Float64, 2, CUDA.DeviceMemory}, B::Matrix{Float64}, α::Bool, β::Bool)
    @ LinearAlgebra ~/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/share/julia/stdlib/v1.12/LinearAlgebra/src/matmul.jl:806
  [6] _syrk_herk_gemm_wrapper!(C::Matrix{…}, tA::Char, tB::Char, A::CuArray{…}, B::Matrix{…}, α::Bool, β::Bool, ::Val{…})
    @ LinearAlgebra ~/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/share/julia/stdlib/v1.12/LinearAlgebra/src/matmul.jl:533
  [7] generic_matmatmul_wrapper!(C::Matrix{…}, tA::Char, tB::Char, A::CuArray{…}, B::Matrix{…}, α::Bool, β::Bool, val::Val{…})
    @ LinearAlgebra ~/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/share/julia/stdlib/v1.12/LinearAlgebra/src/matmul.jl:507
  [8] _mul!
    @ ~/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/share/julia/stdlib/v1.12/LinearAlgebra/src/matmul.jl:328 [inlined]
  [9] mul!
    @ ~/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/share/julia/stdlib/v1.12/LinearAlgebra/src/matmul.jl:297 [inlined]
 [10] mul!
    @ ~/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/share/julia/stdlib/v1.12/LinearAlgebra/src/matmul.jl:265 [inlined]
 [11] *(A::CuArray{Float64, 2, CUDA.DeviceMemory}, B::Matrix{Float64})
    @ LinearAlgebra ~/.julia/juliaup/julia-1.12.4+0.x64.linux.gnu/share/julia/stdlib/v1.12/LinearAlgebra/src/matmul.jl:136
 [12] top-level scope
    @ REPL[10]:1
Some type information was truncated. Use `show(err)` to see complete types.

this is how this fails for CUDA