How does the Air manage throttling and performance loss in games?!

[macddy]

Of all the people who bought air and all the hundreds of reviewers on youtube NO ONE, and I mean NO ONE talks about or tests the macbook air gaming or heavy application use for longer periods of time. They just start a game, play for 5 min and say it runs great! Thats a BAD test. You need to test it for at least 1-4 hours and THEN say whether it runs great or not. Why arent there anyone reviewing the sustained performance!?

I'm honestly tempted to buy both just to review that, im honestly shocked no one else has yet.

 

[deeddawg]

Just how long are you thinking it takes the system to reach thermal equilibrium, where the cooling system is fully heat soaked and the CPU/GPU are operating at a level at which thermal output equals the thermal dissipation capacity?

Seems to me on other systems that takes no more than a few minutes. Do you suspect something different on the M1 MBA/

 

[macnmac]

TIL: people buy a mba to do hardcore gaming

 

[macddy]

TIL: people buy a mba to do hardcore gaming

What is hardcore gaming for you?

 

[macnmac]

What is hardcore gaming for you?

NSFW

 

[gautampw]

TIL: people buy a mba to do hardcore gaming

His money, his system, let him do what he wants. Why judge anyone?

 

[revs]

It's a good question - even for simpler games, it would be nice to know what the actual performance is after 10-20min and the machine throttles

 

[macnmac]

His money, his system, let him do what he wants. Why judge anyone?

where is there any judgement in my comment?

 

[Ocnetgeek]

Of all the people who bought air and all the hundreds of reviewers on youtube NO ONE, and I mean NO ONE talks about or tests the macbook air gaming or heavy application use for longer periods of time. They just start a game, play for 5 min and say it runs great! Thats a BAD test. You need to test it for at least 1-4 hours and THEN say whether it runs great or not. Why arent there anyone reviewing the sustained performance!?

I'm honestly tempted to buy both just to review that, im honestly shocked no one else has yet.

One of the reviewers was doing the Cinebench test for 30 minutes and noted that the Air seemed to be doing some minor throttling after about 8 minutes. He did not observe any throttling for the Pro or the Mini

 

[wytwolf]

Maxtech and Rene Ritche have done cinebench tests and/or game tests showing throttling occurs after about 8-10 minutes. Cinebench went from about 7400 done to about 6300 as the M1 went from about 3.2Ghz down to about 2.4Ghz.

 

[dboris]

I also wonder the same.
Yes benchmarks makes the system throttle by 10%, but we have yet no idea for games.
Possibly it doesn't throttle, or does throttle more since you may want to play for >1H.
Maybe heat slowly builds up as it's all passive.

 

[deeddawg]

I also wonder the same.
Yes benchmarks makes the system throttle by 10%, but we have yet no idea for games.
Possibly it doesn't throttle, or does throttle more since you may want to play for >1H.

Which goes back to my question above -- just how long do folks think it takes for the heat dissipation system to fully saturate such that we reach steady-state where the system slows no more / gets no hotter?

I'd expect that to be just a few minutes - any reason for thinking it would take longer?

 

[eyetic]

It would be nice to understand if its better or just the same as my laptop with a GTX1060TI with i7...cant figure out yet this thing to ditch my laptop and buy this m1 air or pro..

 

[4sallypat]

Macs I use for work and daily tasks but never for gaming - gave up on that when MSFS 2020 came out.
Bootcamp for Windows failed miserably on the 2019 MBP.

For gaming, I just switch platforms and use my decked out i7 gaming PC w/ 32GB RAM, NVIDIA 8GB GTX1660 Super OC card, 1ms Dell Alienware IPS 27" display + dual 27" Dell displays (triple display), flight yoke, Airbus thrust controller, quadrant controls and rudder/brake pedals.

You have use the correct tools for the mission.

 

[dboris]

Lol there's no need to lay down pc specs.
I have enough as a workstation, but I'd like to know if I'll occasionally be able to put some headshots when I'm on the go with the air.

 

[eyetic]

I like to keep my life as simple as possible, and carrying around two laptops bothers me a lot..thats why I want to replace my gaming and office laptop with this one.

 

[dboris]

I like to keep my life as simple as possible, and carrying around two laptops bothers me a lot..thats why I want to replace my gaming and office laptop with this one.

I think you'd better off waiting high end 13" pro or 16", if you plan any kind of gaming. I'll take the air, but only as a chill device, to be used on the go or in bed.

 

[archer75]

games don't peg the cpu/gpu 100% non stop. They spike up and down depending on what's going on on the screen at the time. Seeing as it takes roughly 8 minutes to throttle at 100% workload i'd say you have nothing to worry about.

 

[dboris]

games don't peg the cpu/gpu 100% non stop. They spike up and down depending on what's going on on the screen at the time. Seeing as it takes roughly 8 minutes to throttle at 100% workload i'd say you have nothing to worry about.

I'd bet on that too and I hope you are right.
But still voting up that topic as it's worth investing for people who can

 

[wyatterp]

I have the base MacBook Air and I have a huge steam library. It's blowing me away on performance, but there will be periods that crop up randomly where it will throttle down - so I would say for non FPS sensitive titles, it's more than great, but I'm still testing. Last night it throttled, but I had been in the middle of downloading over 300GB of games and kept launching them in sequence after a download to test framerate. Right now I'm playing CIV VI at medium settings for an extended period on battery just to see how it holds up.

I will say this, outside of one old title, this M1 BASE MBA I have absolutely annihilates my Dell 9310 Iris Xe i5-1135G7 iGPU for every title I throw at it.

Here are some approx benchmark impressions:

Most all at medium/900P
Rise of the Tomb Raider (last game I tested last night) - 44.9FPS!
Civ VI - Gathering Storm Benchmarks at 25FPS
Total War: Warhammer II - battle benchmark was 31FPS, and about the same for campaign overworld map
XCOM 2 - no benchmark, but looks to be running at 30FPS
Cities Skylines seemed to run a huge city (over 100K) at over 30FPS at 1050P rez
Batman Arkham City GOTY ran at 45-60FPS - some frame drops, but played really well
Mad Max - not sure why, but this game runs great on my Dell, but could not get over 9FPS on this at 900P

These are all running through STEAM, not Mac App Store versions.

 

[leman]

I also wonder the same.
Yes benchmarks makes the system throttle by 10%, but we have yet no idea for games.
Possibly it doesn't throttle, or does throttle more since you may want to play for >1H.
Maybe heat slowly builds up as it's all passive.

Andrei Frumusanu from Anandtech twitted that the GPU on his Mac mini is using 10 watts at most... it seems to be incredibly energy-efficient. Depending on the game, one might not even see much throttling. Only very few games would load up all the CPU cores to their limits, and M1 seems to achieve incredible performance with just few watts of power per core.

 

[wyatterp]

60 minutes into battery session of Civ VI - 900P/Medium (don't know how to increase rez in this game beyond this for some reason) - still running at 60FPS with little or no noticeable throttling to actual FPS.

 

[moep]

I'm also curious about this. The air would suit 98% of my needs perfectly, with the last 2% being the odd game. I prefer the wedged case of the MBA and not having a touchbar seems more of a feature than a downside to me, at least until it gets some sort of haptic feedback.

Normally a completely passive system would be a no-go for any extended gaming workload, but in this case the SoC might just be efficient enough to get away with only minor throttling once the heatspreader has reached a temperature equilibrium. I believe one of the reviews reported that the GPU alone under full load will use as little as 7W of power.

Unfortunately nobody seems to have done further testing on this.
All it really takes is one initial (Unigine Heaven?) baseline benchmark, then a one hour loop of it and then another Heaven benchmark.

I'm also wondering if one of those rather hideous laptop cooler contraptions (usually a perforated stand with a USB-connected 5V fan in it) could theoretically be used to mitigate or eliminate throttling for the occasional extended workload on a hot day. Yes, yes, I am fully aware the dual port MBP is meant for these applications.

 

[wyatterp]

i'm downloading uningine valley - i saw in maxtech review he got no image rendering on screen for heaven

 

[wyatterp]

First Valley run is scoring 41.1 FPS and 1721 points at 900P, High, 4x MSAA

Will let it run in the background for 30 minutes and retest

Ok after looping for 90+ mins here is the Valley score

27.9 FPS and 1167 points