Revert sequential GPU jobs: other runner jobs make it unnecessary

The `needs` dependency doesn't help since other jobs on the shared
runner contribute to GPU memory pressure regardless. The CUDA.reclaim()
fix in the test file is the actual mitigation.

Co-Authored-By: Chris Rackauckas <accounts@chrisrackauckas.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: ChrisRackauckas

.github/workflows/GPU.yml

@@ -43,7 +43,6 @@ jobs:
 
   simplenonlinearsolve-cuda:
     name: "SimpleNonlinearSolve CUDA"
-    needs: nonlinearsolve-cuda
     runs-on: [self-hosted, gpu-v100]
     timeout-minutes: 60
     steps: