Better (Geekbench) Performance via Battery for MacBook Air 2020 i7

[massey]

I know Geekbench isn't the end all be all but has anyone else noticed their scores higher when powering via the battery versus the power adapter? Also, as this makes complete sense, the normal CPU temperature is generally lower by 10-15 Celsius on battery mode.

Geekbench MacBook Pro 2020 i7, 16 GB Ram

• (Battery) Single Core = 1238, Multi-Core = 3225
• (Power Adapter) Single Core = 1151, Multi-Core = 2649

 

[IngerMan]

I think that has more to do with CPU usage and or temperature but more CPU usage in your case what you posted for a i7 MBA. I was getting over 1250 and over 3400 on power before any mods. I have gotten over 1300 SC and 3600 MC with power. Ive ran both with power and battery and would say my best scores have been with power.

if this seems to be true with your machine under same conditions then you should call Apple support. You have a problem with your machine.

When running GB you will get your best scores if you are not running any other programs. Your CPU usage should be at or above 98% free.

When on battery and I am not sure at what point but lower battery it seems not to be checking stuff in the background like Photo, Calendar, Accounts, System Update, Windowserver, etc. Seeing I was 98% CPU usage on battery I just ran the GB test while on battery and only 2 hours left of charge 1293SC 4109MC. Then I plugged it in and all of a sudden I am down to less than 65% CPU because I am on power and its checking everything in the background. If I ran the test now it would be total crap like you posted.

Now I wait till its done checking all the background stuff and I am on power with 98% CPU. I run the test and get about the same results. You can see from my 2 screen shots one is on battery and one is on power.

so Generally speaking you are incorrect, and if you are running 10c warmer with power vs battery and can’t determine why, call Apple. I run the same temps either way.

 

[amartinez1660]

I think that has more to do with CPU usage and or temperature but more CPU usage in your case what you posted for a i7 MBA. I was getting over 1250 and over 3400 on power before any mods. I have gotten over 1300 SC and 3600 MC with power. Ive ran both with power and battery and would say my best scores have been with power.

if this seems to be true with your machine under same conditions then you should call Apple support. You have a problem with your machine.

When running GB you will get your best scores if you are not running any other programs. Your CPU usage should be at or above 98% free.

When on battery and I am not sure at what point but lower battery it seems not to be checking stuff in the background like Photo, Calendar, Accounts, System Update, Windowserver, etc. Seeing I was 98% CPU usage on battery I just ran the GB test while on battery and only 2 hours left of charge 1293SC 4109MC. Then I plugged it in and all of a sudden I am down to less than 65% CPU because I am on power and its checking everything in the background. If I ran the test now it would be total crap like you posted.

Now I wait till its done checking all the background stuff and I am on power with 98% CPU. I run the test and get about the same results. You can see from my 2 screen shots one is on battery and one is on power.

so Generally speaking you are incorrect, and if you are running 10c warmer with power vs battery and can’t determine why, call Apple. I run the same temps either way.

View attachment 930412 View attachment 930411

MacBook Airs 2020 don’t they have the less than optimal cooling solution? Don’t know, to add to your list, might also be a case of while charging and turbo boosting through the tests it just gets hotter and throttles more than just normal.

To the original user, do you have a sort of aluminum stand or similar? (that could act as a heat dissipator help)

 

[IngerMan]

MacBook Airs 2020 don’t they have the less than optimal cooling solution? Don’t know, to add to your list, might also be a case of while charging and turbo boosting through the tests it just gets hotter and throttles more than just normal.

I think you missed my point. There is no difference!! All things being equal for both. If the poster is running 5c hotter on power and all other things are equal, he has a problem with his MBA. It’s not normal.

 

[massey]

I'll run the test again but Geekbench is the only program running and I normally wait till CPU is done doing it's background functions so usage is around 1-2%.

 

[massey]

Running a more apples to apples comparison, same room (may have a difference in the room temperature) and no additional monitor plugged in (as it was before with the power adapter benchmark), rookie mistakes.

Geekbench MacBook Pro 2020 i7, 16 GB Ram - Round 2

• (Battery) Single Core = 1245, Multi-Core = 3189
• (Power Adapter) Single Core = 1228, Multi-Core = 3201

So this makes more sense. Not sure why all the previous benchmarks were running so low with the power adapter and monitor setup (might be the cause). The single core was around 975-1154 with the multi-core at 2438-2553. Feel a bit better now but maybe some other background functions were limiting the benchmark. I did make sure the indexing was done.

 

[Erehy Dobon]

macOS is probably renice-ing background index processes while on battery, dropping their priority level so they are not chewing up CPU cycles and battery.

If you have an external monitor connected, you may be pushing more pixels into the graphics pipeline which will naturally put more load on the GPU.

 

[Tenkaykev]

Since doing the shim mod on my i5 the performance has been very good completely silent and fans running quietly in Cinebench and GB5.
Yesterday I was having a bit of a clear out and found some bathroom scales, the ones with the clear tempered glass platform and a LCD display with the mounting feet containing the strain gauges.
After removing the display it is the perfect size for a laptop stand and the glass is a better conductor of heat than the wooden desk that the Air usually sits on.
Switched the Air on this morning and went to make coffee. Came back and fired up GB5 with the results 1166/3986
I've been repeating the runs from time to time this morning, letting the temperature settle back down to a typical 35-45C.
I've had 5 runs all in the 3900's with a high of 1182/3994

That's quite respectable for an i5, and best of all is the complete lack of fan noise.

View attachment 930859

 

[Tenkaykev]

Having previously done the copper shim upgrade, this morning I fitted a thermal pad to the heatsink, sized to fit between the inner of the screw holes on the heatsink.
Now getting GB5 scores of 4K. Base isn't getting hot and running Cinebench the processor never reaches 100C.
GB5:
2837342