client: set totalcompute from usable cores

[schmichael]

Alternative to #17

---

Linux arm64 w/ 18 cores, 9 reservable

After

Allocated Resources
CPU          Memory      Disk        Alloc Count
0/18000 MHz  0 B/16 GiB  0 B/12 GiB  0

    ^-- total & usable compute match! --v

Attributes
cpu.arch                                 = arm64
cpu.frequency                            = 2000
cpu.numcores                             = 18
cpu.reservablecores                      = 9
cpu.totalcompute                         = 18000
cpu.usablecompute                        = 18000

cpu=1000 job:

Allocated Resources
CPU             Memory          Disk            Alloc Count
1000/18000 MHz  256 MiB/16 GiB  300 MiB/12 GiB  1

cores=2 job added:

Allocated Resources
CPU             Memory          Disk            Alloc Count
5000/18000 MHz  512 MiB/16 GiB  600 MiB/12 GiB  2

Saturate with cores=1 jobs:

Allocated Resources
CPU              Memory          Disk            Alloc Count
18000/18000 MHz  900 MiB/16 GiB  2.6 GiB/12 GiB  9

---

So with 1 cores=8 job the utilization looks right:

Allocated Resources
CPU              Memory          Disk            Alloc Count
16000/18000 MHz  256 MiB/16 GiB  300 MiB/12 GiB  1

And then trying to schedule a cpu=4000 job the rejection looks right:

2026-10-16T12:31:10-07:00: Task Group "examplecpu" (failed to place 1 allocation):
 * Resources exhausted on 1 nodes
 * Dimension "cpu" exhausted on 1 nodes

---

Reversing the submission order works as expected.

cpu=4000 job results in:

Allocated Resources
CPU             Memory          Disk            Alloc Count
4000/18000 MHz  256 MiB/16 GiB  300 MiB/12 GiB  1

cores=8 job results in:

    2026-10-16T12:33:37-07:00: Task Group "batchcore" (failed to place 1 allocation):
      * Resources exhausted on 1 nodes
      * Dimension "cpu" exhausted on 1 nodes

With cpu_total_compute=20000

Allocated Resources
CPU          Memory      Disk        Alloc Count
0/20000 MHz  0 B/16 GiB  0 B/12 GiB  0

Allocation Resource Utilization
CPU          Memory
0/20000 MHz  0 B/16 GiB

Host Resource Utilization
CPU           Memory          Disk
34/20000 MHz  804 MiB/16 GiB  (tmpfs)

Allocations
No allocations placed

Attributes
cpu.arch                                 = arm64
cpu.frequency                            = 2000
cpu.numcores                             = 18
cpu.reservablecores                      = 9
cpu.totalcompute                         = 20000
cpu.usablecompute                        = 20000

After a cores=8 job:

Allocated Resources
CPU              Memory          Disk            Alloc Count
16000/20000 MHz  256 MiB/16 GiB  300 MiB/12 GiB  1

After a cpu=4000 job is added:

Allocated Resources
CPU              Memory          Disk            Alloc Count
20000/20000 MHz  512 MiB/16 GiB  600 MiB/12 GiB  2

With cpu_total_compute=50000

Trying to run multiple jobs with cores=8:

...
      * Resources exhausted on 1 nodes
      * Dimension "cores" exhausted on 1 nodes

$ nomad node status -self
...
Allocated Resources
CPU              Memory          Disk            Alloc Count
16000/50000 MHz  256 MiB/16 GiB  300 MiB/12 GiB  1
...

But cpu=4000 jobs still fit:

Allocated Resources
CPU              Memory          Disk            Alloc Count
20000/50000 MHz  512 MiB/16 GiB  600 MiB/12 GiB  2