WEBVTT 00:00:00.640 --> 00:00:06.080 maxed out max or max max with m1 max to 00:00:03.760 --> 00:00:10.559 the max the 14 inch max books pro represent 00:00:08.400 --> 00:00:15.599 double the GPU and media horsepower the m1 pro models but 00:00:12.480 --> 00:00:17.199 this raises a very interesting question 00:00:15.599 --> 00:00:21.680 how can apple fit that much more power on this same 00:00:19.359 --> 00:00:25.439 architecture no less into the same chassis 00:00:23.359 --> 00:00:29.679 for that matter is there any other advantage to getting out a macbook pro 00:00:27.359 --> 00:00:33.920 with an m1 max or is it only really useful for video editors thanks to its 00:00:31.439 --> 00:00:37.680 expanded media engine the answer may surprise you i know it surprised me 00:00:36.399 --> 00:00:41.520 but it's no surprise that today's sponsor is jackery thanks to jackery for 00:00:39.680 --> 00:00:45.920 sponsoring this video their explorer 1500 power station provides a huge 1500 00:00:44.239 --> 00:00:50.640 watt hour capacity with up to seven devices charging simultaneously and it 00:00:48.480 --> 00:00:54.719 only takes four hours to recharge from zero to eighty percent get ten percent 00:00:53.199 --> 00:00:57.960 off with code Linus tech tips at the link below 00:01:04.640 --> 00:01:11.520 from the outside there's very little to distinguish the m1 pro from m1 max so we 00:01:09.680 --> 00:01:15.439 went with space gray for the max books and silver for the pro books to tell 00:01:13.200 --> 00:01:18.720 them apart jonathan horst over at Mac Address commented that he liked silver 00:01:17.119 --> 00:01:24.400 better than space gray this time around and honestly i kinda agree it's not that 00:01:21.520 --> 00:01:27.680 the space grey looks bad but there's too little contrast for me particularly 00:01:26.159 --> 00:01:30.720 considering the blacked out keyboard that apple chose for the new generation 00:01:29.360 --> 00:01:34.240 macbooks but this is a matter of taste and i'm sure 00:01:33.040 --> 00:01:38.880 there are plenty of folks who'd like to see this thing totally stealth instead 00:01:36.479 --> 00:01:43.759 there is one major external difference between the m1 pro and m1 max though and 00:01:41.759 --> 00:01:48.240 that's in the display support where the m1 pro can handle up to two external pro 00:01:45.680 --> 00:01:52.720 display xdrs via Thunderbolt the m1 max can handle up to three while also using 00:01:49.840 --> 00:01:56.079 the HDMI 2.0 port for up to 4k 60 output that extra flexibility could lead to 00:01:54.399 --> 00:02:00.079 some impressive battle station setups for professional workloads you've got 00:01:57.840 --> 00:02:04.960 the integrated mini led display a span of three big displays worth of workspace 00:02:02.399 --> 00:02:08.640 and a large tv or color calibrated 4k display for output preview 00:02:06.880 --> 00:02:13.440 one question that's been asked a lot is whether the Thunderbolt ports could handle an HDMI 2.1 adapter and 00:02:12.160 --> 00:02:18.959 unfortunately hours didn't arrive in time for testing these laptops i'll test them all when we 00:02:17.280 --> 00:02:23.800 review the 16-inch model so make sure you get subscribed for that it's coming 00:02:20.800 --> 00:02:23.800 soon 00:02:23.920 --> 00:02:30.640 how about on the inside with these machines being otherwise so similar 00:02:28.160 --> 00:02:34.800 could apple also have given them the exact same cooling setup remember the m1 00:02:32.720 --> 00:02:38.720 pro caps out at 16 GPU cores while the m1 max can double that number 00:02:37.040 --> 00:02:42.319 and doubles the memory bandwidth which together are sure to increase the heat 00:02:40.319 --> 00:02:46.959 output let's pop the hood and compare the two oh hey apple color matched the screws 00:02:45.680 --> 00:02:51.840 how thoughtful and they're the same picture if i hadn't 00:02:50.000 --> 00:02:55.040 told you that the space gray was the m1 max you wouldn't have been able to tell 00:02:53.599 --> 00:02:58.959 the difference just by looking at them unless you looked really closely at that 00:02:57.440 --> 00:03:03.680 heat spreader it's just a little bit wider on the m1 max to cover that chip's 00:03:01.120 --> 00:03:09.040 larger footprint other than that same fans same heat pipe same everything 00:03:06.480 --> 00:03:13.280 given how hot the m1 pro got this makes me worry that we'll see some significant 00:03:10.879 --> 00:03:16.720 throttling with the m1 max we'll button everything back up and give it a go this 00:03:14.879 --> 00:03:21.840 time with the zephyrus m16 repping the pc with its cry9 and rtx 3060 mobile 00:03:23.840 --> 00:03:30.640 first we'll see if anything changed for the CPU cores on the m1 max and 00:03:28.239 --> 00:03:35.519 predictably it absolutely did not none of the m1 max's CPU scores are 00:03:32.959 --> 00:03:38.879 appreciably different in cinebench r23 and this translates over to geekbench 00:03:37.360 --> 00:03:42.400 where we're effectively looking at run to run variants same deal with blender 00:03:40.959 --> 00:03:47.360 where we're stuck with CPU rendering until blender 3.1 releases with metal 00:03:44.879 --> 00:03:50.480 GPU rendering support sometime in 2022 you don't have to wait that long for our 00:03:48.720 --> 00:03:54.480 new shirt designs and a new water bottle from ltdstore.com though they're 00:03:52.239 --> 00:03:57.840 available right now you may be surprised by these results 00:03:56.159 --> 00:04:02.400 though given how much faster the system memory is i mean ryzen would sure love 00:04:00.239 --> 00:04:06.000 to have memory that fast but it seems apple's CPU cores just 00:04:04.560 --> 00:04:10.000 aren't taking full advantage of it and based on these results alone you might 00:04:08.159 --> 00:04:16.720 not be too impressed by the m1 max at all but the CPU is only half the picture no 00:04:14.319 --> 00:04:22.639 we don't have more CPU cores or even faster CPU cores but what we do get are 00:04:19.919 --> 00:04:26.479 more GPU cores and an expanded media engine with double the encoder blocks 00:04:24.400 --> 00:04:31.120 and double the prores blocks nowhere would this be more apparent than final 00:04:28.160 --> 00:04:35.360 cut pro for this test we set up a 5x5 grid of 4k plus prores raw files and set 00:04:34.000 --> 00:04:40.400 them to appear one by one in the timeline then we set final cut to stop when it 00:04:38.240 --> 00:04:45.840 detects a drop frame with the original m1 macbook pro stalled out at just four 00:04:43.360 --> 00:04:49.759 streams and the m1 pro dropped at seven the m1 maxes manage 10 and 12 suggesting 00:04:48.320 --> 00:04:54.000 that the cut down model either has a slightly less capable prores engine 00:04:52.000 --> 00:04:57.520 or the extra GPU cores on the high end model are picking up the slack whatever 00:04:55.520 --> 00:05:02.240 the case this is a truly ridiculous number of pixels being processed all at 00:04:59.520 --> 00:05:06.639 once with no stuttering whatsoever like four or seven simultaneous streams is 00:05:04.080 --> 00:05:11.520 already a lot but let's say you don't run final cut you're a resolve shop well 00:05:09.360 --> 00:05:16.000 good news for you too when rendering the prores 422 hq to take advantage of the 00:05:13.680 --> 00:05:20.800 pro res blocks the higher end model gets scary close to a sub 10 minute mark on a 00:05:18.479 --> 00:05:25.039 timeline that is stacked with denoising chroma keying and filtering 00:05:22.960 --> 00:05:29.199 not a bad result and our timeline's frame rate keeps up too where we 00:05:27.039 --> 00:05:33.520 observed as high as 16 FPS for the high end model scaled by about 2 FPS for each 00:05:31.680 --> 00:05:37.600 step down you take from there the zephyrus meanwhile can only roughly 00:05:35.199 --> 00:05:42.240 match the lower end m1 pro that's a great result but what if you need to 00:05:39.759 --> 00:05:45.440 render some 3d assets right now the best solution on mac is cinema 4d with the 00:05:44.080 --> 00:05:49.680 redshift renderer surprisingly the jump in performance 00:05:47.280 --> 00:05:53.120 from the m1 pro to the lower end m1 max is substantial likely thanks to the 00:05:51.680 --> 00:05:57.520 doubled memory bandwidth to go along with the extra GPU cores the higher end 00:05:55.120 --> 00:06:01.440 max gives us nearly a 10 minute render which seems impressive up until the 00:05:59.840 --> 00:06:06.319 zephyrus comes along and smacks it down with a 539 00:06:03.280 --> 00:06:08.160 of course that is a 16 inch pc laptop 00:06:06.319 --> 00:06:13.120 but um spoiler for the 16 inch review 00:06:10.400 --> 00:06:18.000 the 16 m1 max doesn't do that much better we were told that the m1 max 00:06:15.440 --> 00:06:24.479 could outperform the rtx 3060 or even match the rtx 3080 and apple calls out 00:06:21.520 --> 00:06:28.160 redshift on their product page so m1 00:06:25.600 --> 00:06:32.960 gotta take the l on this one geekbench 2 shows the zephyrus and its rtx 3060 00:06:30.400 --> 00:06:38.080 absolutely pummeling the m1 max in both render modes which places the m1 max's 00:06:35.919 --> 00:06:41.280 GPU horsepower further behind than apple had led us to believe maybe there's more 00:06:39.600 --> 00:06:47.199 optimizations that still need to be done for non-video tasks on these GPU cores 00:06:43.919 --> 00:06:48.800 or maybe it's just the media engine was 00:06:47.199 --> 00:06:53.680 doing all the heavy lifting in our video editing tests to be clear it is scaling 00:06:51.600 --> 00:06:57.039 relative to the m1 and m1 pro but it's just not reaching those dedicated GPU 00:06:55.600 --> 00:07:01.199 levels of performance we kept hearing about hmm 00:06:58.960 --> 00:07:05.120 maybe they meant games yeah no again we're seeing mostly linear 00:07:03.759 --> 00:07:11.520 performance improvements across the lineup in shadow of the tomb raider where feral interactive's port makes use 00:07:08.720 --> 00:07:15.199 of apple's metal API but it's not enough to catch up to the zephyrus 00:07:13.360 --> 00:07:18.960 at least not at 14 inches few other games use metal so it's all downhill 00:07:17.120 --> 00:07:23.919 from here when we switch over to cs go where there's no difference whatsoever between 00:07:21.599 --> 00:07:27.520 the m1 pro and max ironically this time not because it's using opengl but 00:07:25.759 --> 00:07:32.240 because we've hit a CPU bottleneck at 100fps thanks to rosetta since we're 00:07:30.400 --> 00:07:37.840 looking for more ways the m1 max can differentiate itself there's one final 00:07:35.440 --> 00:07:41.520 ray of hope the venerable dolphin emulator which is native apple silicon 00:07:39.759 --> 00:07:45.120 support via universal binary unfortunately 00:07:42.800 --> 00:07:48.800 dolphin doesn't yet support the metal API but in testing we did discover a 00:07:47.280 --> 00:07:53.680 number of interesting points to cover the first head scratcher is that the 00:07:51.199 --> 00:07:56.800 high level opengl renderer is more performant than the closer to metal 00:07:55.280 --> 00:08:00.639 vulcan API this may be because apple never embraced 00:07:58.800 --> 00:08:04.800 vulcan necessitating the use of a third-party library called molten vk if 00:08:02.800 --> 00:08:09.520 developers want to support it on macOS because apple deprecated opengl several 00:08:06.960 --> 00:08:13.520 macOS revisions ago the implementation while performant is missing some new 00:08:11.520 --> 00:08:18.720 features that makes dolphin think that the GPU might be underpowered 00:08:16.080 --> 00:08:22.479 second multi-sample anti-lacing that can be enabled in most games 00:08:20.720 --> 00:08:27.039 is a huge performance killer on apple silicon it easily takes rendering speed 00:08:25.280 --> 00:08:32.080 from an impressive double of native speed to just half third some titles 00:08:30.000 --> 00:08:36.399 like rogue squadron 3 depend so heavily on the CPU that they never run at full 00:08:34.320 --> 00:08:40.320 speed even at native resolution just like the original m1 this makes sense 00:08:38.959 --> 00:08:45.040 though because the main differentiator between the m1 m1 pro and m1 max's CPU 00:08:43.039 --> 00:08:49.040 cores is that there are more performance cores available and dolphin doesn't 00:08:47.120 --> 00:08:52.800 utilize more than a couple at a time we did get performance data however by 00:08:50.880 --> 00:08:56.160 firing up super mario sunshine at 8x native resolution and an uncapped frame 00:08:54.399 --> 00:09:01.360 rate to place the burden squarely on the GPU bearing in mind that the game is 00:08:58.640 --> 00:09:06.640 native to 30fps the starting area runs at about 67 FPS on the lower end m1 max 00:09:04.080 --> 00:09:11.120 and 77 on the high end that's a pretty good improvement from the m1 pro that 00:09:08.399 --> 00:09:16.399 topped out at 47 with the original m1 taking up the rear at the sub native 21 00:09:14.240 --> 00:09:20.640 except then we look at the zephyrus again and it's a slaughter 00:09:18.720 --> 00:09:24.399 that right there is over four times the native frame rate for this game and 00:09:22.000 --> 00:09:28.960 nearly double what the best m1 max can put out in a laptop that costs a little 00:09:26.800 --> 00:09:33.139 over half as much at retail 00:09:34.880 --> 00:09:41.279 so again we're seeing differences between the chips and the performance is 00:09:39.519 --> 00:09:45.600 objectively good but we're just not coming up to the performance claims apple made during 00:09:44.000 --> 00:09:50.000 their unleashed event no matter what we do we're legitimately not sure if this is 00:09:48.240 --> 00:09:53.760 an optimization problem or if apple over-promised with cherry-picked data 00:09:51.839 --> 00:09:57.600 because as we've seen it's quite difficult to find anything that runs 00:09:55.600 --> 00:10:01.279 metal on apple silicon to take full advantage of the m1 max 00:09:59.360 --> 00:10:05.920 now let's take a look at thermals for a test we could run in our m1 pro video we 00:10:04.160 --> 00:10:10.240 focus on the CPU so we'll be looking at the GPU this time around for the m1 max 00:10:08.000 --> 00:10:14.720 by running redshift the fact that it hovers around 90 degrees isn't too 00:10:12.240 --> 00:10:18.399 surprising but given how much surface area the GPU occupies the rest of the 00:10:16.720 --> 00:10:22.079 soc is also brought up to those temperatures which indicates that we're 00:10:20.560 --> 00:10:25.040 probably at the threshold of what the cooler can manage without maxing out the 00:10:23.839 --> 00:10:29.040 fans those only ever got to around 50 speed 00:10:27.680 --> 00:10:35.360 so there's definitely still some headroom if you want to max the fans out manually we'll be doing a more strenuous 00:10:33.040 --> 00:10:39.200 combined CPU plus GPU torture test when we test out the 16-inch form factor in 00:10:37.120 --> 00:10:43.839 our next review so stay tuned to find out by how much when we look at the 00:10:41.120 --> 00:10:47.519 flare it becomes clear that the m1 max is really pushing the upper limits of 00:10:45.519 --> 00:10:51.600 the 14-inch chassis cooling capacity easily hotter than even the CPU test we 00:10:49.680 --> 00:10:55.920 did on the m1 pro chips in our previous review we're seeing over 50 degrees in the 00:10:54.160 --> 00:10:59.440 hotspot and well above that exhausting out of the display the bottom is still 00:10:57.920 --> 00:11:05.440 within the realm of comfort and if it's sitting in your lap chances are the underside's hot spot won't be touching 00:11:02.399 --> 00:11:07.200 your skin so points to apple there now 00:11:05.440 --> 00:11:11.519 let's talk battery last time we showed the m1 pro's 00:11:08.959 --> 00:11:14.959 endurance was nothing short of amazing while doing youtube video playback 00:11:13.440 --> 00:11:18.800 easily beating out contemporaries like the surface laptop studio with a literal 00:11:16.880 --> 00:11:22.399 all-day battery life result for the m1 max we matched all the 00:11:20.720 --> 00:11:26.880 screens at 4 ticks of brightness and let them go this time around both of the m1 00:11:24.880 --> 00:11:31.519 pro models lasted a little over 2 hours longer than both of the m1 max models so 00:11:29.760 --> 00:11:35.839 while battery life is still phenomenal at about 20 hours in this test you do 00:11:33.920 --> 00:11:39.600 give up a little bit of endurance for the extra horsepower even when you're 00:11:37.360 --> 00:11:45.040 not making use of it depending on your priorities though this could be a side 00:11:41.680 --> 00:11:47.360 grade a net benefit or a net loss a fun 00:11:45.040 --> 00:11:50.800 fact this test failed once and these endurance figures being what they are 00:11:48.800 --> 00:11:56.720 made that a huge pain overall for an extra 200 for the 24 core 00:11:53.440 --> 00:11:58.240 variant m1 max is not a bad upgrade 00:11:56.720 --> 00:12:02.240 it's just difficult to recommend in general unless you've got a need for 00:11:59.920 --> 00:12:06.399 more than 32 gigs of memory the extra GPU horsepower and memory bandwidth seem 00:12:04.480 --> 00:12:10.639 very situational particularly in light of the dearth of apple silicon native 00:12:08.320 --> 00:12:13.839 applications with metal support this may change moving forward but it's a 00:12:12.399 --> 00:12:17.120 teething pain that's going to hamper these higher end socs for the coming 00:12:16.000 --> 00:12:23.440 years rosetta can only do so much as we've 00:12:19.440 --> 00:12:25.040 seen as such m1 max only really shines 00:12:23.440 --> 00:12:30.480 in high-end video editing as of right now for everything else we're looking at 00:12:27.600 --> 00:12:35.279 something much more akin to the rtx 3050 ti much less the rtx 3080 we were 00:12:33.200 --> 00:12:39.200 promised apple's leaning heavily on special purpose silicon here and while 00:12:37.440 --> 00:12:43.600 that's an approach that may work out for them they run a very real risk of having 00:12:41.600 --> 00:12:49.120 purpose-built hardware that becomes dead weight once consumers inevitably move on 00:12:46.800 --> 00:12:53.680 just like you should move on from untidy nether regions 00:12:51.360 --> 00:12:56.800 with manscaped manscapes provides an all-in-one 00:12:55.279 --> 00:13:00.399 grooming kit that has you covered from head to toe 00:12:58.560 --> 00:13:04.959 their performance package 4.0 features their awesome lawnmower 4.0 waterproof 00:13:02.800 --> 00:13:08.959 body trimmer their weed whacker ear and nose trimmer plus a whole lot of other 00:13:06.959 --> 00:13:13.440 goodies as well for a limited time you get all of this plus two free gifts the 00:13:11.440 --> 00:13:17.760 shed travel bag and the manscaped anti-chafing boxer briefs visit 00:13:15.440 --> 00:13:22.320 manscape.com tech or click the link below for 20 off and free shipping 00:13:20.480 --> 00:13:25.760 thanks for watching guys the 16 inch review is going to include both 00:13:23.760 --> 00:13:31.440 processors so go check out our review of the m1 pro macbooks for a little more on 00:13:27.600 --> 00:13:31.440 the whole platform before the finale