Actually if you feel like experimenting for us, try these combos and report back on the ppd numbers.
- 4x single core regular console clients
- 3x single core regular console clients + GPU client
- SMP (Windows) client
- GPU client
- SMP (Windows) + GPU client
- 2x Linux SMP clients under VMware
- 2x LInux SMP clients under VMware + GPU client (arguably one of the Linux VM will be suffering, and that's the point on the experiment)
- 1 Linux SMP client under VMware + GPU client + single core regular client
- 1 Linux SMP client under VMware + GPU client + one idle core to do your own stuff (this is to compare the single core client to see if it is worth it in the face of already big ppd)
The Model M is not for the faint of heart. You either like them or hate them.
Gerbils unite! Fold for UnitedGerbilNation, team 2630.