I've been running my own python3 single-threaded benchmark on my 2018 MPB 15" i9 (32MB) and my Dell XPS 9570 i9 (32MB). I'm interested in longer runtimes, this benchmark takes an 1 hour or 2. (I have a multithreaded benchmark too but it takes about 24 hours to run on 16 cores, so this is just a tentative run.) The MBP is a lot slower than the Dell despite 10.13.6 + Supplemental Update. I turn off Time Machine, bluetooth and make sure the battery is at 100% and plugged in. The Dell is running Ubuntu 18.04 LTS though.
Anyway, I decided to cool the MPB by placing it on top of an LG room air conditioner (set to max cool 60F) blowing on the underside and side vents of the case to see if it made any difference. Surprisingly, the MPB doesn't speed up much. And it never seems to make the 30W mark on Intel Power Gadget.
First, the comparison chart with my own collection of machines. (I no longer have the 2017 XPS 13" 9360 as I upgraded it with the 2018 XPS 15" 9570. The 2017 MBP 15" is maxed out, top of the line, but didn't belong to me - I ran the benchmark remotely using a friend's work machine.) I'm annoyed the MBP i9 is slower than the obsolete (by two generations) XPS 13 running Ubuntu 16.04 LTS.
Looking at Intel Power Gadget (latest version), the comparison between the two MBP i9 runs, is quite different in terms of idle temperature (e.g. after the run is done), e.g. 40C vs 18C. But during the run they're very similar, and so the difference in speed is small.
Here are sample snapshots. On top of desk with small air gap between surface and bottom of laptop:
On top of airconditioner:
[doublepost=1536720873][/doublepost]The test cooling setup. I normally use this air conditioner as a supplementary airconditioner to chill a room for treadmill running. It's vented to another room obviously.
Anyway, I decided to cool the MPB by placing it on top of an LG room air conditioner (set to max cool 60F) blowing on the underside and side vents of the case to see if it made any difference. Surprisingly, the MPB doesn't speed up much. And it never seems to make the 30W mark on Intel Power Gadget.
First, the comparison chart with my own collection of machines. (I no longer have the 2017 XPS 13" 9360 as I upgraded it with the 2018 XPS 15" 9570. The 2017 MBP 15" is maxed out, top of the line, but didn't belong to me - I ran the benchmark remotely using a friend's work machine.) I'm annoyed the MBP i9 is slower than the obsolete (by two generations) XPS 13 running Ubuntu 16.04 LTS.
Looking at Intel Power Gadget (latest version), the comparison between the two MBP i9 runs, is quite different in terms of idle temperature (e.g. after the run is done), e.g. 40C vs 18C. But during the run they're very similar, and so the difference in speed is small.
Here are sample snapshots. On top of desk with small air gap between surface and bottom of laptop:
On top of airconditioner:
[doublepost=1536720873][/doublepost]The test cooling setup. I normally use this air conditioner as a supplementary airconditioner to chill a room for treadmill running. It's vented to another room obviously.