[BUG] ptxas fatal error, sm_89 not defined for gpu-name

[handdara]

Description Getting error when running the default project. Very similar to #516 , but in the reverse. Essentially, it seems my version of the cuda toolkit doesn't support sm_89. I use pop_os and have it installed as in the system76 docs. I'm not sure what to do because it doesn't seem like I can upgrade to 11.8 which would support sm_89? based on the sys76 docs that is.

Here is the error readout:

$ cabal exec accelerate-simple-exe
Resolving dependencies...
input data:
xs = Vector (Z :. 10) [0.0,1.0,2.0,3.0,4.0,5.0,6.0,7.0,8.0,9.0]
ys = Vector (Z :. 10) [1.0,3.0,5.0,7.0,9.0,11.0,13.0,15.0,17.0,19.0]

the function to execute:
\a0 a1 ->
  fold
    (\x0 x1 -> x0 + x1)
    0.0
    (delayed
      (let T1 x0 = shape a1 T1 x1 = shape a0 in (min x1 x0))
      (\(T1 x0) -> a0 ! x0 * a1 ! x0))

accelerate-simple-exe:
*** Internal error in package accelerate ***
*** Please submit a bug report at https://github.com/AccelerateHS/accelerate/issues

ptxas - -o /home/handdara/.cache/accelerate/accelerate-llvm-1.3.0.0/accelerate-llvm-ptx-1.3.0.0/llvm-hs-9.0.1/nvptx64-nvidia-cuda/sm89/rel/morp8b775bb9a1b1744ae822fcf2124b54a95ffd13d51046ea5149bb0f726c40267c.sass -arch=sm_89 (exit 255)
ptxas fatal   : Value 'sm_89' is not defined for option 'gpu-name'


CallStack (from HasCallStack):
  internalError: Data.Array.Accelerate.LLVM.PTX.Compile:185:24
  compileCUBIN: Data.Array.Accelerate.LLVM.PTX.Compile:123:20
  compile: Data.Array.Accelerate.LLVM.PTX.Compile:88:22

Steps to reproduce

Steps to reproduce the behaviour:

1. Make template project with stack new PROJECT_NAME https://github.com/AccelerateHS/accelerate/raw/stable/accelerate.hsfiles
2. Build withcabal build
3. Run with cabal exec PROJECT_NAME-exe

Expected behaviour To run the template project as planned, compute show dot product.

Your environment

• Accelerate: 1.3.0.0 (>=1.0 in cabal)
• Accelerate backend(s): accelerate-llvm-ptx
• GHC: 8.8.4
• OS: POP_OS 22.04
• LLVM: 9.0.1