[CUDAX] rename device to device_ref, add immovable device as a place to cache properties

[ericniebler]

Description

#2100 will add a global cuda::devices array. @pciolkosz mentioned that this would be a good place to keep a primary context. it would also be an excellent place to cache device properties.

to that end, we need a "heavyweight" object to represent a device. given our naming scheme, the logical name for that is device, and the current device should be renamed to device_ref.

refs #2081

[github-actions]

🟩 CI finished in 4h 33m: Pass: 100%/56 | Total: 2h 37m | Avg: 2m 48s | Max: 12m 05s | Hits: 81%/2353

• 🟩 cudax: Pass: 100%/55 | Total: 2h 25m | Avg: 2m 38s | Max: 9m 09s | Hits: 81%/2353

  🟩 cpu
    🟩 amd64              Pass: 100%/51  | Total:  2h 15m | Avg:  2m 39s | Max:  9m 09s | Hits:  81%/2181  
    🟩 arm64              Pass: 100%/4   | Total:  9m 10s | Avg:  2m 17s | Max:  2m 23s | Hits:  79%/172   
  🟩 ctk
    🟩 12.0               Pass: 100%/23  | Total:  1h 00m | Avg:  2m 38s | Max:  9m 09s | Hits:  81%/983   
    🟩 12.5               Pass: 100%/32  | Total:  1h 24m | Avg:  2m 38s | Max:  7m 21s | Hits:  80%/1370  
  🟩 cudacxx
    🟩 nvcc12.0           Pass: 100%/23  | Total:  1h 00m | Avg:  2m 38s | Max:  9m 09s | Hits:  81%/983   
    🟩 nvcc12.5           Pass: 100%/32  | Total:  1h 24m | Avg:  2m 38s | Max:  7m 21s | Hits:  80%/1370  
  🟩 cudacxx_family
    🟩 nvcc               Pass: 100%/55  | Total:  2h 25m | Avg:  2m 38s | Max:  9m 09s | Hits:  81%/2353  
  🟩 cxx
    🟩 Clang9             Pass: 100%/2   | Total:  4m 26s | Avg:  2m 13s | Max:  2m 25s | Hits:  81%/86    
    🟩 Clang10            Pass: 100%/2   | Total:  4m 20s | Avg:  2m 10s | Max:  2m 15s | Hits:  81%/86    
    🟩 Clang11            Pass: 100%/4   | Total:  8m 52s | Avg:  2m 13s | Max:  2m 18s | Hits:  81%/172   
    🟩 Clang12            Pass: 100%/4   | Total:  9m 16s | Avg:  2m 19s | Max:  2m 47s | Hits:  81%/172   
    🟩 Clang13            Pass: 100%/4   | Total:  9m 02s | Avg:  2m 15s | Max:  2m 28s | Hits:  81%/172   
    🟩 Clang14            Pass: 100%/6   | Total: 15m 35s | Avg:  2m 35s | Max:  3m 40s | Hits:  87%/258   
    🟩 Clang15            Pass: 100%/2   | Total:  4m 25s | Avg:  2m 12s | Max:  2m 22s | Hits:  81%/86    
    🟩 Clang16            Pass: 100%/6   | Total: 18m 44s | Avg:  3m 07s | Max:  5m 32s | Hits:  87%/258   
    🟩 GCC9               Pass: 100%/2   | Total:  4m 01s | Avg:  2m 00s | Max:  2m 03s | Hits:  76%/86    
    🟩 GCC10              Pass: 100%/4   | Total:  8m 12s | Avg:  2m 03s | Max:  2m 09s | Hits:  76%/172   
    🟩 GCC11              Pass: 100%/4   | Total:  7m 42s | Avg:  1m 55s | Max:  2m 06s | Hits:  76%/172   
    🟩 GCC12              Pass: 100%/12  | Total: 31m 01s | Avg:  2m 35s | Max:  3m 59s | Hits:  82%/516   
    🟩 Intel2023.2.0      Pass: 100%/1   | Total:  3m 01s | Avg:  3m 01s | Max:  3m 01s | Hits:  81%/43    
    🟩 MSVC14.36          Pass: 100%/1   | Total:  9m 09s | Avg:  9m 09s | Max:  9m 09s | Hits:  43%/37    
    🟩 MSVC14.39          Pass: 100%/1   | Total:  7m 21s | Avg:  7m 21s | Max:  7m 21s | Hits:  40%/37    
  🟩 cxx_family
    🟩 Clang              Pass: 100%/30  | Total:  1h 14m | Avg:  2m 29s | Max:  5m 32s | Hits:  83%/1290  
    🟩 GCC                Pass: 100%/22  | Total: 50m 56s | Avg:  2m 18s | Max:  3m 59s | Hits:  80%/946   
    🟩 Intel              Pass: 100%/1   | Total:  3m 01s | Avg:  3m 01s | Max:  3m 01s | Hits:  81%/43    
    🟩 MSVC               Pass: 100%/2   | Total: 16m 30s | Avg:  8m 15s | Max:  9m 09s | Hits:  41%/74    
  🟩 gpu
    🟩 v100               Pass: 100%/55  | Total:  2h 25m | Avg:  2m 38s | Max:  9m 09s | Hits:  81%/2353  
  🟩 jobs
    🟩 Build              Pass: 100%/47  | Total:  1h 53m | Avg:  2m 25s | Max:  9m 09s | Hits:  78%/2009  
    🟩 Test               Pass: 100%/8   | Total: 31m 24s | Avg:  3m 55s | Max:  5m 32s | Hits:  97%/344   
  🟩 sm
    🟩 90                 Pass: 100%/1   | Total:  1m 55s | Avg:  1m 55s | Max:  1m 55s | Hits:  76%/43    
    🟩 90a                Pass: 100%/1   | Total:  2m 00s | Avg:  2m 00s | Max:  2m 00s | Hits:  76%/43    
  🟩 std
    🟩 17                 Pass: 100%/31  | Total:  1h 12m | Avg:  2m 20s | Max:  4m 13s | Hits:  81%/1333  
    🟩 20                 Pass: 100%/24  | Total:  1h 12m | Avg:  3m 01s | Max:  9m 09s | Hits:  79%/1020  
  

• 🟩 pycuda: Pass: 100%/1 | Total: 12m 05s | Avg: 12m 05s | Max: 12m 05s

  🟩 cpu
    🟩 amd64              Pass: 100%/1   | Total: 12m 05s | Avg: 12m 05s | Max: 12m 05s
  🟩 ctk
    🟩 12.5               Pass: 100%/1   | Total: 12m 05s | Avg: 12m 05s | Max: 12m 05s
  🟩 cudacxx
    🟩 nvcc12.5           Pass: 100%/1   | Total: 12m 05s | Avg: 12m 05s | Max: 12m 05s
  🟩 cudacxx_family
    🟩 nvcc               Pass: 100%/1   | Total: 12m 05s | Avg: 12m 05s | Max: 12m 05s
  🟩 cxx
    🟩 GCC13              Pass: 100%/1   | Total: 12m 05s | Avg: 12m 05s | Max: 12m 05s
  🟩 cxx_family
    🟩 GCC                Pass: 100%/1   | Total: 12m 05s | Avg: 12m 05s | Max: 12m 05s
  🟩 gpu
    🟩 v100               Pass: 100%/1   | Total: 12m 05s | Avg: 12m 05s | Max: 12m 05s
  🟩 jobs
    🟩 Test               Pass: 100%/1   | Total: 12m 05s | Avg: 12m 05s | Max: 12m 05s
  

👃 Inspect Changes

Modifications in project?

	Project
	CCCL Infrastructure
	libcu++
	CUB
	Thrust
+/-	CUDA Experimental
	pycuda

Modifications in project or dependencies?

	Project
	CCCL Infrastructure
	libcu++
	CUB
	Thrust
+/-	CUDA Experimental
+/-	pycuda

🏃‍ Runner counts (total jobs: 56)

#	Runner
41	linux-amd64-cpu16
9	linux-amd64-gpu-v100-latest-1
4	linux-arm64-cpu16
2	windows-amd64-cpu16