No i have never put a turbo in the oven. I concede the point. My post should have read " the temperature differential across the blades"
It seems like feeding each head off of its own exhaust gas could set you up for some major complications. In the event of a partial system failure, half your cylinders would quit receiving boost.. Would a boost controller be able to open the gate on the other turbo fast enough to prevent catastrophic damage to internal engine parts? For race applications it might be worth it, but for just a sick car it seems kind of risky.
Something to consider -and im not sure how it applies in this situation, but worth mentioning- Dynos have repeatedly shown that in race motors with 3" pipes running from each bank straight to a muffler on either side, adding a crossover H or X pipe reliably adds 15-25HP by stabilizing the exhaust pressure across all cylinders.