At least when the 2023 Mac Pro M2 Ultra is not in its degraded-requires-reboot state.
See also my recommendations on which model to buy (hint: it’s not a Mac Pro for most people).
UPDATED with corrected graphs; see comments from reader Adam S at end.
Reader Adam S writes:
The "MacBook Pro 16 M1 Max" vs "MacBook Pro 16 M2 Max" numbers are crazily different.
M2 Max is faster than M1 Max, but not that much faster. Any idea what’s up?
MPG: Adam is right. I now suspect that some of the M1 Max tests were affected by high ambient temperatures; I had noticed the fans running full-out during the tests. I re-ran the 2021 MacBook Pro M1 Max tests and got much better numbers.
I also observed (for example) the PhaseOne IQ4 test take 153 seconds, then take 123 seconds after a reboot, but I am unable to detect the cause (it's obvious with M2 Ultra via hashing speed). I then observed the M1 Max take 206 seconds instead of 133 seconds (!) with Fujifilm GFX100S files, only 5 minutes apart but after a prior test. No background apps using any CPU/GPU when the test started. What is going on?!!!
Either it’s temperature downclocking as the machine gets hot, or the M1 Max has a similar Slow Mode issue as with the M2 Ultra, but so far I have not seen that behavior, so I am now on the lookout now just in case, but seems highly unlikely. In any case, the M1 Max does not perform reliably.
The 2022 MacBook Pro M2 Max could not be tested under the same conditions (loaner long gone); it was tested in early spring when temperatures were 25°F cooler in my office eg 65°F or so instead of high 85°F.