👇 Recommended Zen 5 Laptops: - Vivobook S 14 (Refreshed 2024): go.magik.ly/ml/265v4/ - Zenbook S 16 (365): go.magik.ly/ml/265ux/ - Zenbook S 16 (HX 370): go.magik.ly/ml/265uz/ - Pro Art PX 13: go.magik.ly/ml/265zm/ - Pro Art P 16: go.magik.ly/ml/265zg/ 👉 All Our Recommended Laptops: www.justjosh.tech/best-tech/ Corrections: 1) Our Geekbench graphs reflect the performance of the MacBook Pro 14 with M3 Max instead of the M3 Pro. The corrected Geekbench score for CPU should be listed as 3,083 for single core, 15,345 for multi-core, and 78,380 for GPU Compute (Metal). These are still better than the Zen 5 chips but by a much smaller margin for Multi-core CPU. 2) The "c" cores are also on 4nm. 3nm is only used on EPYC servers.
try running the proart 13 in eco mode it turns of the dedicated gpu am run my test on the proart p16 in eco mode as well to see how the 890m runs alone never owned an amd laptop and now am lovin the power and efficiency of these new chips yes arc graphic is good but intel needs more battery life out of them good reviews just josh glad to see i choose a good laptop
Fantastic video but can I please make a request? Please don't flash slides on the screen for 1 or 2 seconds like at 5:06. The problem is also in some of your earlier videos! On mobile devices it often takes 2 seconds to pause. It literally took me > 90 secs and THREE attempts to freeze that slide! Please try to show each slide for 5 seconds minimum!
8:50 if the fan was running that's probably why battery life is not very good on the ProArt 13". Looking at the internals, there are 3 fans, not. 2, and they are much larger. They are probably very inefficient at low cooling loads like in a battery lifetime test, eating too much power... Look at how lack of fans doubles macbook air efficiency - this has to be the reason!
For some weird reason, being upfront about your findings makes me want to watch till the end to enjoy the entire process of arriving at your conclusions. This channel is literally my go to for all things laptop. Big ups Josh and team❤❤
Great video, it's nice to see that efficiency is becoming more and more important. One thing though; I think that the graph at 9:44 is a bit misleading, running Cinebench for 30 minutes doesn't say how much progress the laptops had on the "task" - The M3 Macbook might have had 89% battery left after 30 minutes but we don't know how much of the job was finished in those 30 minutes. So the graph won't tell us if the battery will last to finish the task. Maybe the Macbook air dies before the job is finished, whereas the px13 finishes on one charge. We'd need a standardised task that could run for 30 minutes, and if the task finishes before 30 minutes the time it took and how much charge was left could be noted down. I do like that you are upfront with the conclusion in the beginning of the video!
I hate his battery comparisons and charts...show me which laptop lasted longest with web browsing or some kind of light work...not this efficiency bullshit
@@EmilPettersson1 it shows that when you ate constantly loading the CPU, like working with heavy Applications, how far it will bring you. This could include rendering live previews on CPUs in editing software. The different performance only changes the quality of the preview and not the stress on the system. However, I understand your concern. But the efficiency can easly be derived by dividing the benchmark score by the battery % used. With that number you now that when your workload is simmilar to "100k cinebench points" you know how much power will be used.
@@1world0 but most people don't use theirs laptops like that on battery. They're mostly browsing, copying files, watching videos and working in Office apps.
@@1world0 But the graph only shows the battery remaining, ignoring how much "work" Cinebench did during that time. We can't derive anything other than: "The computer will have this much battery after 30 minutes of max power usage", not how much work will be done after 30 minutes of max power usage.
Important correction: Zen5c is *not* build on 3nm as claimed at 2:32 Zen 5 mobile is 100% 4 nm. Only the server Zen5c will be built on 3-nm and maybe future Mobile parts, but not strix point. Another thing, I do not know for sure but appears to be the case is, that the GB6 multicore results of the MBP are actually with an M3 Max and not from an 12C M3 pro
Thank you very much. You are correct. The 3nm we read was related to EPIC server chips. We have posted a correction in the description and pinned comment
Thanks for the comparisons, I don't think anybody does a more thorough job. We have several ASUS products in our house (laptop, desktop, even ProArt monitors) so I can't say I'm surprised by your pick.
Yes, please keep each chart up for 5s minimum, not 1.5s as the price/performance chart was flashed! On a phone there is a 2 seconds delay for TH-cam to respond to pauses, and often it takes THREE OR FOUR replays to be able to see the chart-flashes on-screen, please avoid doing this !!
Maybe they should add timestamps for the chart as well. Yes, it will make following the topics a bit messy but will be able to pause the video & go to the chart directly.
I've been watching laptop youtubers for years and it seemed like you came out of nowhere but it feels like you're the only one who truly understands laptops as devices, the marketing behind them, and how to actually test and review results!! you and your team are doing great work. thanks for being objective!
something important i've noted on the NPU chart: the M4's 38 TOPS was calculated at INT4 precision while all the others use INT8, at INT8 the M4 falls down to 19 TOPS
This is false! The M4 is measured at int8(=38TOPS), whereas the M3 and previous M series chips were measured at FP16(=18TOPS). For M3 at int8=(36TOPS).
"Keep in mind that the M4's 38 TOPS is based on INT8, while the 45 TOPS number from Qualcomm is based on INT4. According to NVIDIA Benchmark Data, INT4 precision can potentially bring an additional 59% speedup compared to INT8, which means the 45 TOPS (INT4) benchmark number could be around 28 TOPS (INT8). The Apple M4 38 TOPS (INT8) can possibly be 35% better than the Qualcomm Snapdragon X Elite 28 TOPS (INT8) in ML performance." Btw the Snapdragon Hexagon still the 2013 project what was just 32bit (int4 - 32-bit values in range), and Apple completly eliminated the 32bit compability so they are only measuring in 64bit (int8 - 64-bit values in range).
Because I got the most important info on the beginning of the video, I'm definitely watching the whole video, since I feel my time appreciated. Thanks for the amazing work as always!
Can you conduct comprehensive and serious battery tests, rather than the incomplete ones you ve always shown? Marc Andrews' tests of the S16's battery show improvement, but it still falls significantly short compared to Qualcomm Snapdragon (Surface laptops etc). The Samsung snapdragon laptop with the X84 beats the S16 with its Zen 5 in performance.
@@alexs.5107 been saying this for a while. Case in point : streaming videos Will always favor OLED screens vs web browsing where IPS screens excel. They should test The battery with mixed uses of browsing, video playback and office productivity apps
The snapdragon chips still have better efficiency than the Zen 5 at low power tasks. This channel has a slight anti-qualcomm bias in all their videos. Probably because of how badly the snapdragon was overhyped. It rubbed some people the wrong way, including this reviewer.
Snapdragon is twice as efficient in single-core tasks, look up the Notebookcheck review (Cinebench 2024 single-thread perf/W). Somehow Josh doesn't feel like mentioning this fact, despite single-core efficiency strongly correlating to real-world battery life. If you fire the laptop on all cylinders, sure, Zen 5 shows similar efficiency to Snapdragon. But that's not reflective at all of day-to-day usage. Also Snapdragon doesn't have efficiency cores, if the next generation will also have efficiency cores they will easily crush Zen 5 (or 6).
Thanks for the comprehensive analysis in your video. Still not convinced that the Zen 5 chips are more efficient than ARM. I have looked at many other reviews and none seem to say that the Zen 5 chip is as power efficient as the Snapdragon chip.
Power efficiency varies with load. The takeaway is that Snapdragon chips have diminishing returns as you increase load. When you do medium workloads, it is close to AMD efficiency and, in heavy loads, as bad as Intel efficiency. The XElite are good at low-power tasks like office, browsing and watching media, where they are second to Apple.
@@andyH_England yes, which, in real life terms, translates to more battery life as those are the tasks people spend more time on (generally) and at the end of the day that means less energy usage.
Snapdragon x elite is the first attempt from the nuvia team and it's not very good! It's very muscular but it wastes power. According to Jim Keller both Apple arm designer and AMD ryzen designer, ARM is not implicitly more efficient. Intel just makes power hungry shit, thats all (the latter is my own assertion).
So the best Zen 5 laptop has the SAME peak CPU performance as the slowest X Elite, whilst running louder with worse battery life? And we should be impressed by this? And where are the tests for standby time? Non intensive tasks like Web browsing or using productivity tools? How about some tests running on battery vs AC? How about testing the general smoothness of the OS, opening apps, using application features?? Like you know, what most people do on their laptops? People do not buy laptops to run benchmarks lol. All this time spent testing theses laptops yet there's little to know useful information as to how these will perform for normal everyday users.
The X elite used in this video was the fastest in multicore in their testing. Take a look at the recent dedicated x elite video. They only have compared it to to the X1E80 midel though, but the X1E84 is inonly very few and very expensive laptops and without data it isn't clear if that chip would be much faster as the multicore score might not be significantly impacted by the CPU designation. And as you can see in the video in the netflix playback test and in the full load test the HX 370 without dGPU used less battery % than the x elite.
@@1world0 the point is he is not representing all of SD x elite laptops, while he got all 3 ryzen laptops and each one is tuned for either battery life or performance. You can clearly see how pro art throws watts to get performance but for battery life he takes the other 2 laptops. So why not the same for SD x elite cos they will look better than amd and goes against his bias
@@nanjappa, The X Elites cannot live with the AMD in GPU tasks, where the battery efficiency is lost on the ProArt (the S16 matches X Elite). 100% of people needing GPU power will buy ProArt over X Elite; the Snapdragon is not in the game. So your point is moot, as the S16 AMD is as good as the X Elite in efficiency for casual workloads.
@@nanjappa, As for my previous comment, I was saying that if you need a good GPU, you should not buy the X Elite, so they would not be on the radar. So why mention the ProART? Snapdragon does not compete in that GPU league.
Fantastic work, as always! I'm confused by the discussion of efficiency and battery life, though - there's always variation between different reviewers' battery life tests, but I can't find anyone else who's found the S 16 to last anywhere near as long as a Snapdragon or M3 laptop. Everyone else seems to have found the battery life to be no more than mediocre in comparison to those alternatives, but in your test it's as good or better. I'd love to know what explains the discrepancy, here. For example, on Laptop Mag 's tests the S16 lasted for 11:35, and the Air 15 lasted for 15:03; on Andrew Marc David's test the S16 lasted for 15:37 and the Air 15 for 18:33; on NotebookCheck the S16 lasted for 10:40 and the Air l5 lasted for nearly 17:00. These times suggest that Zen 5 isn't as efficient as Snapdragon X, as with similar sized batteries those laptops match or surpass the Air. Why aren't other reviewers finding Zen 5 as efficient as Snapdragon in practice?
Wonderful video buddy. I ordered the SL 7 15 inch version (16gb ram), it should arrive sometime next week. I am wondering how snappy are the new AMD laptops? One main reason for ARM is the snappiness, for example you open any window, it just feels like zero lag or zero stutter (I tried out the SL7 at stores before ordering and I was blown by the snappiness of it). Are the new AMD chips the same or there is a minute lag (compared to ARM chips)? It would be great if you can elaborate more on that. Great content, and looking forward for more :)
As someone who currently has both the 15" SL7 and the Zenbook S16, I'm happy to share my thoughts. They are both snappy for everyday tasks (browsing, email, office apps, etc...) with SL7 having a slight upper hand but only kind of noticeable if you have them side by side and barely. Personally, I would not let that be the determining factor for which one to go with since they are both exceptional in that dept yet vastly different in other areas. I am still undecided in which one to keep since each one has something I like better than the other. I should point out that I'm not a gamer and I don't do anything heavy. Technical aspects of either of these are not of great concern to me since both are more than enough for what I will use them for and for my use case there are no compatibility issue with ARM. My area of focus is on form factor, build, look and feel as well as well as battery life. My ideal form factor is a 16" screen, OLED, no numpad. If the 16" Samsung GB4P or Edge didn't come with that horrendous touchpad, I would have been satisfied.... I really love the 15" SL7. It's almost perfect for me. Incredible build quality, phenomenal battery life, awesome touchpad and keyboard. The only reason why I'm not fully in love is because I really prefer a 16" screen size as well as OLED, hence the Zenbook S16.... I'm trying to love the S16 but been having issues, mostly with build. Comparing to a Mac, SL7 and GB4P, the S16 feels on a cheaper side, not quite premium. I'm currently on my third unit. The first two were 24gb "white" ones (light gray really) and this third one is the 32gb dark gray one. The bottom panel/cover is quite thin and if you squeeze the laptop at the very front, it can cause a popping sound. That was my very first initial experience as I was taking the first unit out of the box (Maybe a bit of exaggeration but think a sound that a thick empty can would make if squeezed). As I squeezed, there would be a gap or opening created wide enough to put my fingernail through along with a popping sound. Not good. Fast fwd to the third unit which just arrived today, the dark gray 32gb model, the touchpad rattles when I tap on it, like it's loose. It's quite annoying. This was not the case with the first two units... So long story short, there seem to be some major qc issues with Asus. I absolutely love this form factor, it's almost perfect for me, but when I go and pick up my SL7, I can't help but asking myself why I'm willing to go in circles with the S16 especially since the battery life is not even close to the 15" SL7 and the SL7 is nearly perfect for my use case. And also, the screen on SL7, even though not OLED is definitely brighter than the S16 screen which can be beneficial in the dark mode. Also, the keyboard, on the Asus it light up like a Christmas tree which can be distracting to some of us where on the SL7 light only comes through the key symbols. So that's just a matter of personal preference. So that's my experience. I hope this somehow helps you or anyone faced with the same choice.
@@VladdyDaddy369 Thank you for your reply. Can you say something to these concerns, how they are different on each of the devices? Restart / Boot up time Wake from standby battery lost in standby (e. g. overnight) fan noise trackpad (does it make a big difference?) is the reflection on the screen noticeably worse on SL7?
@@AllTernative90 Restart / Boot up time Both fast but S16 is faster by 2-4 seconds from both the shutdown and restart boot (I disabled all startup apps options shown on both and tried 3 times) Wake from standby Both nearly identical, pretty much instant. Open the lid and the ir camera recognizes immediately. Battery lost in standby (e. g. overnight) SL7 about 2-3% depending on duration of "overnight". I never checked it on the first two S16 units and won't know on this third one until tmrw obviously Fan noise Virtually non-existent with SL7. I got to hear the fans when running Cinebench 2024 and the top of the laptop above the Fn keys by the bottom of the screen was super hot when running multi-core. Otherwise they never kick on for my normal daily use and the laptop stays very cool.... S16 gets a little warm even when not doing much. How much fans kick on depends on the fan profile selected but they kick in a bit when I don't expect while browsing on Chrome. trackpad (does it make a big difference?) OMG absolutely yes! That haptic touchpad on the SL7 is worth every penny. It's so smooth with no palm rejection issues. It's simply phenomenal. Is the reflection on the screen noticeably worse on SL7? With both laptops side by side and turned off, yes for sure. However, I never really notice it while using because I keep screen brightness at 100% in dark mode which obviously makes a big difference. Speaking of, just did both laptops side by side, dark mode and screen at 100%, the SL7 screen actually looks better to me because of brightness. The S16 is almost like it's missing that one-more-click. I feel like I can see and read better and easier on the SL7. Hmm, interesting, didn't notice that before. I don't think I have found my "the-one" yet, I'm curious what the next couple of months will bring with both Apple and Intel but for now I know which one I'm keeping. Chatting with you actually really helped me decide lol. Thanks!
@@VladdyDaddy369 Thank you so much for your time, that was super helpful! I'm leaning heavily on the Snapdragon side (especially on the SL7) but with all the reviews of the Zen5 chips I'm getting second thoughts. The SL7 from your description sounds very nice but I hoped it would loose less battery in standby. The trackpad does sound amazing and it's afaik the only laptop with it. I worried about the screen if it's good enough, but your explanation helps. Especially 3:2 aspect ratio is such a great thing to have, and there is no other laptop ticking the boxes like the SL7. I hope there are discounts now with zen5 chips out, but I really doubt it.
Watching this video again. It was that good. BTW, Josh, I enjoyed your recent cameo on HC! They produce some great content. Also, I think it would be cool if you collab'd with a few smaller channels whose content is top notch, including Kyle Erickson, Mark Linsangan and Notebookcheck Reviews (I understand notebookcheckreviews was in the HC video as well). I like when larger channels throw some shine on the quality up-and-comers (maybe there are some you know that I don't even watch yet). Of course, when you collab with JT - we can't get enough of those videos. JT has gotta be one of the best in the business of reviewing laptops.
I finally subscribed. I love your personality, the willingness to help people, and the hard work. The data presented is surely top notch, and I am sure it takes a lot of work and time. Hello from Indonesia
at 11:10 I think the M3 Pro 12C results are wrong, according to geekbench website, should be around 68000 points instead of 157000 ... I think that result is from the M3 Max variant with the 40 gpu cores
You are SPOT ON, we have checked all other graphs and it is only the 2 Geekbench ones that have this incorrect. We are going to post the correction in the description and pinned comment
Man, looking those power efficiency charts and graph(s) and all I can think about is a new Razer Blade 14 with these Zen 5 chips inside. The 2024 model (RTX 4070 + 8845HS etc.) is already amazing, now with this proper CPU upgrade... 😳
@@jackofthecoke Indeed! Exactly my thoughts, especially with a 240 Hz display so there's no need to carry an external one (although it's recommended, obviously).
The amount of work team Just Josh puts in for these evaluations is insane. Thank you so much. This channel is my reference for computer reviews. Once you discover this channel, the majority of those other highly popular "reviews" channels are like infomercials*. (* Incidentally: "A Pixel 9 review agreement required influencers to showcase the Pixel over competitors or have their relationship terminated." --The Verge)
What people don't understand is the original Commentor might have meant Ryzen 7000series CPU plus RTX 4070 GPU hence the name Ryzen 4070! Instead of considering this angle, people preferred to make this a meme(ZTT)!
In high-performance mode running video it will use the same wattage as in balanced. The difference is that in high-performance mode you do not get power throttled, which you would in balanced.
I got the Vivobook S 14 and am extremely happy with the Machine. The HX370 in the Vivobook actually runs at 45W in Performance Mode (at around 80°C) without getting too loud and is fast enough for most games I play (with AI scaling from 1200p to 1880p). A 75Wh battery in a 14" chassis is actually impressive (for 16" on the other hand I'd like to see around 90Wh instead of just the same 75Wh/78Wh ones in the larger Vivobook/Zenbook Variants). Display and Battery Life are also great, my only caveats are: Only one of the two Type-C Ports is 40Gbit (the Zenbook has two full speed ones) and it has a WiFi6 Chipset (performant, but not amazingly futureproof).
Anything else you can tell me about that I should be aware of before purchasing the Vivobook. No one on the internet is reviewing the Vivobook. Only the Zenbook and ProArt reviews are available and I think that's largely due to their better build quality, battery and efficiency
@@jatingumbhir8954 Not a lot in terms of Hardware (I've only had it for just under a week by now) but found some Software Tweaks. 1. If you want to game on it you need to set the reserved GPU memory in the "MyASUS" App to something appropriate (like 4GB). The default "Auto" setting just uses 512 MB and is relying on Windows to virtually swap VRAM over PCIe to/from the iGPU instead of having it be in direct control. 2. The Keyboard RGB is controlled in the Windows Personalization Settings Panel. As for the Review Samples: The Vivobook Chassis Design is older and less impressive (though imo actually more functional). The same physical components and Layout were used in late 2023 / early 2024 which is also the Reason for just WiFi6 and only one 40Gbit port. It's technicall not a new Laptop and instead "just" a CPU upgrade (but a very significant one, especially for efficiency)
@@jatingumbhir8954 I've been hearing many complaining about how hot the laptops get, even during low stress loads. Other than that, it seems like a good value for the money
I sent mine back the 60hz and lack of gsync means is rubbish for gaming. Plus it has a massively power constrained 4070 in it. I went for the zephyrus g16 with the refreshed amd processor in it
These Pro Art laptops would have been pretty fantastic if the eco mode worked properly and they actually stopped feeding power to the Nvidia dGPU. I refuse to buy a laptop with a dGPU if it's not optimized properly and can't be fully turned off when not needed, as i value battery life, especially when mobile, otherwise, these would have been pretty fantastic. I don't even mind the 60hz screen which is fine for on the go computing and better battery life. I use external monitors with much higher refresh rates when i'm stationary.
@@alexanderruchkov7629 honestly, i'm leaning towards entering the apple ecosystem due to the inefficiencies and bugs of the Windows and Linux ecosystems.
Hey Josh, could you please test if these Ryzen AI laptops support HW acceleration for decoding HEVC 4:2:2 10bit video? I’m finding it difficult to find that info and apparently only Apple silicon or Intel 11th gen and newer support it, which is relevant for any user filming video with cameras like Sony Alphas
To my knowledge, Zen 5 does not feature any updates to AMD's hardware video encoder, which means that it still doesn't support HEVC 4:2:2. And yes, that's a feature reserved to the latest version of Quick Sync, and I couldn't find anything except a random Reddit comment to confirm it works on Macs.
Awesome news, loved your livestream and this video is a very impressive "summary" of all your findings. I bet you and your team spent a lot of time just for this video
as per usual, regarding AI tasks, I'm kindly requesting an Ollama local test with 8b and 70b models such as llama 3 and seeing how these laptops perform on it. As most likely this is what most local AI is going to do, generate text and use it inside various programs
Hi Josh and team, Seriously guys, I don't know anyone on TH-cam or anywhere on the internet for that matter who does this as professionally and tastefully as you guys. Really appreciate the awesome, and meticulously detailed content. How u don't have at least 10 million subs by now is mind boggling. Wish u all the best. Thanks again🙏
Thanks for the absolute great video. I just received my PX13 and can confirm there must be something wrong with its dGPU's ECO mode: by its description it should turn off the dGPU completely, and as you mentioned in the video and I also confrimed, the dGPU disappeared from both device manager and task manager. However, on my PX13, the proart creator hub (ASUS's laptop control center) shows the GPU fan is still running with 2000+ RPM, and its number is different from the CPU fan's RPM, which makes me believe the dGPU is not turned off at all, or the laptop thinks the heat from CPU somehow needs 2 fans both run 2000+ RPM to exhuast (I am at wisper mode, CPU runs at 2.57Ghz, with less than 50 C temperature reading).
Zen 5 and X Elite are showing that the node is the most important factor, they are built on the same 4nm node, while Apple have an exclusivity deal with TSMC and are on 3nm
It's not just the 3nm.. there are a few test where even the 4 years old M1 can outperform the X Elite... so the 5nm M1 perform better in a few cases like the 4nm X Elite (and that M1 chip also had the same Watt limitations.. so it's still 20W against 40-50W.. and the 20W 5nm beat the 50W 4nm..
It's actually not the process node that is everything. Note that the Ryzen AI HX 370 is on the same 4nm process node as the 8945 HX, but its both faster and uses about half the power. RDNA 3.5 brought all the power saving ideas from RDNA 4! That's from AMD brainwork, not the VLSI process node. ARM holdings CPUs are slow as dog doodoo thats why Apple and now Qualcomm are designing far more powerful ones to go into laptops with batteries 12x -14x larger than the phones ...
Since 14nm the new process nodes have been saving power but there has been almost ZERO speedup from Moore's law. All the speedup comes from tuning the circuits or using combinatorially more complex ones, e.g. multipliers can consume an almost infinite number of gates if you let them ..
That's because he's only comparing multi-core efficiency. Snapdragon doesn't have low-power cores, so Zen 5 has an edge there. For single-core workloads Snapdragon is twice as efficient, which would translate into increased battery life with light usage. Now idea why that's not mentioned or measured, single-core efficiency is very important for long battery life.
That goes against his conclusions, so he won't say that. Users do bursty workloads on laptops not cine bench. His battery methodology is flawed and he won't change it
Nice video Josh (and the team). Now im actually considering the upgrade from my i5-8350u laptop. AMD did an amazing job on these mobile chips lately, however i see a different story with their new desktop 9000 series.
Ooo. If I'm reading this right, and I'd like to think that I am (because it was very well explained!), these new chips will be very good performers, across a variety of tasks. I'll look forward to seeing all the Zen 5 laptops reviewed!
@@iz723 i bought the 4060 ultra 9 yoga pro and no regrets, this thing feels like a macbook pro but for windows, but i can run games like val at 350fps with a 160hz monitor LOL i would've went for AMD's too because it's clearly better but no complaints here!!
Hi Josh, when you are showing the charts why not show the bars in descending order? It helps a lot to compare with other contestants really quick. Thanks, Love your content.
I am loving my ProArt P16 it really is the Windows MacBook Pro, with some advantages. I noticed on your battery test you shut off the Nvidia graphics on the PZ13. However, did you put it on quiet mode that will up the battery life with acceptable performance for general. These AMD chips, along with the Asus ProArt series, are among the best Windows laptops ever made.
Maybe also include the MacBook with Crossover (Wine) with D3DMetal support. Just for comparison even though it takes a performance hit through the emulations layers.
@@franzpleurmann2585 right...Im planning to do so such reviews on my own actually. Almost no youtuber does true comparisons like this. Waiting for my galaxy book to be delivered in order to compare the ultra 7 and m3
out of curiosity, why not unplug the battery before measuring power draw from the wall? That allows you to isolate another variable so that you can more accurately show power consumption. Just my 2 cents. :)
Please, Josh, when you edit your audio, use a rebreather from Adobe or iZotope to remove it. It's a bit distracting when you focus on it. Hehe. Oh, and thank you for calling out TH-camrs focused on sponsorship and not being frank about poor devices. That was IMPORTANT!
Good stuff. A lot of nuances that the "casual user of a desktop' will miss but vital if you use laptops long term. Gamers probably don't care about power efficiency.
Josh its an obvious design mistake to put a 60Hz OLED panel in a dGPU Laptop. Nobody wants to blow money on an OLED panel thats 60hz. There are no such TVs of this type that I am aware of. I say this as a ProArt monitor buyer (love the brand.) The S16 is really a highly superior 32Gb macbook air alternative and it will run on 17w for most of its life, except for rare times gaming away from a desktop. I feel like its closer to perfect ...
I know you really did not like the 16s 370 but it was an instant buy from me. I needed something thin and light that would augment more powerful systems. To me it strike the right balance giving more performance than older chips while keeping power draw low. If I am traveling and need to get real work done I still have the 4080 razerblade that I can break out at my destination but this new zenbook should hold me over between. Hopefully I still agree with this statement in a week when mine arrives.
Would you still recommend the HP Spectre 14? It's already out for a few months and it seems like a laptop with a newer processor (like the AMD Zen 5) would be a better deal. But your review back then totally convinced me to get the Spectre. Also, it ticks all of my boxes (2in1, good display, battery life, keyboard, size, design and even little things like a privacy shutter for the webcam)
Please also compare vivobook S 14 with these laptops. I don't know why nobody is talking about this 14" laptop. Also why is this processor Asus exclusive?
@@IvoPavlik Don't agree. I have a qualcolmm for my personal use. I even use for develop PLC code on automation. That software is only x86 and it runs better on QUALCOMM for some reason. The only really reason I see bad on SnapDragon is drivers for weird devices like DAQ cards. Everything else can be emulated on PRISM and work pretty well. I think Josh has a very bad test for Qualcolmm, Intel and APPLE , and he loves AMD since many years he is being very bias in favor of them. Every one has a favorite Brand, I like AMD but I recognize in every built I have done with AMD I got more issues than Intel. You need to be a power user on AMD to actually know which setting to apply while on Intel it just works!.
Josh, take it from a person who setup LLM gor developers over the last year: nothing comes close to a full sized GPU with at least 16GB video memory. So Nvidia 3090, 4090 or AMD 7800xt or 7900xtx ... The LLMs xan be run on a separate workstation and users can connect over the network to chat or to perform tasks. That is yhe only practical way to use LLMs from a laptop today
@@alexanderruchkov7629Anything that supports HDMI 2.1 or DP 2.0 (DP1.4 works if you're not using HDR) can provide 4K@120Hz on an external display. The AI HX processors support both of these standards.
I am sorry for the folowing question that may be a stupid one... Nowadays, videos are made in two formats: 1080p and 4k (except for youtube of course). Why does a customer want to buy a laptop with 3K monitor? Technincally there should be worse video quality and worse battery life, right? Thanks for the explanation. Love your videos
OLED 1080p looks worse than LCD 1080p due to the pixel structure (arrangement), so you really need a minimum QHD for OLED for a retina display. Most OEMs vary between 2.5K, 2.8K and 3K, which can depend on the aspect ratio and supply (prices). Some use higher, like 4K for professional users.
@@andyH_England ok, that's clear. So better avoiding OLED 1080p. However, if we consider OLED vs LCD, for entertainment purpose, you will probably see 1080p video worse on a OLED 3k than on a LCD 1080p, right?
@@davidemora8762 I understand that OEMs use upscaling to fill in pixels on a 4K screen if the video is 1080 P. I assume it will be slightly better than 1080 P video playback on a native 1080 P display.
Great detail's as always. A question on how battery performance to "percentage left" is measured how? Is it calculated from the watt hours drawn over the period and calculated from the battery capacity, which would have had a prior draw down test to turn off? If it's measured only using the battery percentage provided by the system, this is likely misleading as it's not linear in most devices and varies.
I’m really sensitive to screen quality when it comes to laptops. I don’t know what OEM Lenovo sourced their LCD from but my Slim 7 Pro X (2022) has one of the best screens yet it’s a 14.5” 3072x1920 IPS panel with 120Hz refresh and 100% sRGB coverage. It’s pretty comparable to the MacBook Pro 14 screen just without the notch and the mini led backlight.
I honestly want a Flow series laptop, preferably x16 with this cpu & 4070, srgb 240hz screen & upgradeable ram with more external io. Basically an upgraded proart with more features.
Could you look at the macbook air 13 performance per watt statistic again? This statistic seems like an extreme outlier in this video. In another video of yours it was also an outlier even compared to other apple devices. Are you sure there is not something wrong with the macbook air measurement?
I do like the benchmark Vs wattage graphs however I think the Geekbench Vs. Wattage graph does probably tell you something...specifically that X Elite is more efficient when it comes to single core usage (almost double compared to AMD or Intel according to Notebookcheck in cinebench 24)...and as you say, geekbench mixes a lot of bursty activity with full cpu usage....explaining why X elite sits closer to the top here. Maybe worth doing your Cinebench Vs Wattage for single core scores too?
Looks to me the only thing saving apple here is the node level that keeping just above the others. AMD is a hybrid design between the two - thus will be drawing a little more power and using a bit more space per chiplet. It also doesnt have as much cache it looks like - and we know at least in gaming, cache can play a roll in some applications. Drawing a bit futher than they are now, looks like AMD is pretty much on the toes of Apple SoC and that pretty much gives us a picture on how the all 3nm SoC will look like follow by whatever M4 looks like a few years later down the line as they move to 2 or 1nm.
The M4 is Apple's 2024 chip as this is AMD, so for the M3 Max to still be on top of the best from AMD is remarkable, bearing in mind the M4 is a massive generational leap which will put the M4 single-core about 35% above AMD! So in 2-3 months, when most AMDs arrive and the M4 is released in the Pro models, there will be a no-contest, and the 2-3 gen CPU single-core gap will return.
👇 Recommended Zen 5 Laptops:
- Vivobook S 14 (Refreshed 2024): go.magik.ly/ml/265v4/
- Zenbook S 16 (365): go.magik.ly/ml/265ux/
- Zenbook S 16 (HX 370): go.magik.ly/ml/265uz/
- Pro Art PX 13: go.magik.ly/ml/265zm/
- Pro Art P 16: go.magik.ly/ml/265zg/
👉 All Our Recommended Laptops: www.justjosh.tech/best-tech/
Corrections:
1) Our Geekbench graphs reflect the performance of the MacBook Pro 14 with M3 Max instead of the M3 Pro. The corrected Geekbench score for CPU should be listed as 3,083 for single core, 15,345 for multi-core, and 78,380 for GPU Compute (Metal). These are still better than the Zen 5 chips but by a much smaller margin for Multi-core CPU.
2) The "c" cores are also on 4nm. 3nm is only used on EPYC servers.
Hello, just want to ask, not expecting a reply but can you add the “Lenovo loq 15 or 16” plz - to your website?
try running the proart 13 in eco mode it turns of the dedicated gpu am run my test on the proart p16 in eco mode as well to see how the 890m runs alone never owned an amd laptop and now am lovin the power and efficiency of these new chips yes arc graphic is good but intel needs more battery life out of them good reviews just josh glad to see i choose a good laptop
@arnulforamirez4836 We did that btw. Thank you
Fantastic video but can I please make a request? Please don't flash slides on the screen for 1 or 2 seconds like at 5:06. The problem is also in some of your earlier videos! On mobile devices it often takes 2 seconds to pause. It literally took me > 90 secs and THREE attempts to freeze that slide! Please try to show each slide for 5 seconds minimum!
8:50 if the fan was running that's probably why battery life is not very good on the ProArt 13". Looking at the internals, there are 3 fans, not. 2, and they are much larger. They are probably very inefficient at low cooling loads like in a battery lifetime test, eating too much power... Look at how lack of fans doubles macbook air efficiency - this has to be the reason!
For some weird reason, being upfront about your findings makes me want to watch till the end to enjoy the entire process of arriving at your conclusions.
This channel is literally my go to for all things laptop. Big ups Josh and team❤❤
Great video, it's nice to see that efficiency is becoming more and more important. One thing though; I think that the graph at 9:44 is a bit misleading, running Cinebench for 30 minutes doesn't say how much progress the laptops had on the "task" - The M3 Macbook might have had 89% battery left after 30 minutes but we don't know how much of the job was finished in those 30 minutes.
So the graph won't tell us if the battery will last to finish the task. Maybe the Macbook air dies before the job is finished, whereas the px13 finishes on one charge. We'd need a standardised task that could run for 30 minutes, and if the task finishes before 30 minutes the time it took and how much charge was left could be noted down.
I do like that you are upfront with the conclusion in the beginning of the video!
I hate his battery comparisons and charts...show me which laptop lasted longest with web browsing or some kind of light work...not this efficiency bullshit
@@EmilPettersson1 it shows that when you ate constantly loading the CPU, like working with heavy Applications, how far it will bring you. This could include rendering live previews on CPUs in editing software. The different performance only changes the quality of the preview and not the stress on the system.
However, I understand your concern. But the efficiency can easly be derived by dividing the benchmark score by the battery % used. With that number you now that when your workload is simmilar to "100k cinebench points" you know how much power will be used.
@@1world0 but most people don't use theirs laptops like that on battery. They're mostly browsing, copying files, watching videos and working in Office apps.
@@1world0 But the graph only shows the battery remaining, ignoring how much "work" Cinebench did during that time. We can't derive anything other than: "The computer will have this much battery after 30 minutes of max power usage", not how much work will be done after 30 minutes of max power usage.
Yeah, using a remaining battery % at end of render task test would be much better 👍
Speaking of power efficiency, your data output per staff member is bonkers. Thank you so much for all the valuable work you do!
think it hasnt been displayed, its not that important but a data anyways indeed, could be informative
Important correction: Zen5c is *not* build on 3nm as claimed at 2:32 Zen 5 mobile is 100% 4 nm. Only the server Zen5c will be built on 3-nm and maybe future Mobile parts, but not strix point.
Another thing, I do not know for sure but appears to be the case is, that the GB6 multicore results of the MBP are actually with an M3 Max and not from an 12C M3 pro
Thank you very much. You are correct. The 3nm we read was related to EPIC server chips. We have posted a correction in the description and pinned comment
The M3 Pro results do seem much higher than I'd expect, so I also wonder if that's the case.
@@Slavolko The Geekbench graphs were off. We have posted a correction in the videos description and pinned comment btw
@@JustJoshTech Thanks for the correction! Keep up the good work.
What a great video structure, including a generous up-front summary. No tosh from Josh! 🤣
Thanks for the comparisons, I don't think anybody does a more thorough job. We have several ASUS products in our house (laptop, desktop, even ProArt monitors) so I can't say I'm surprised by your pick.
I'm sorry, Josh, but I truly can't handle these rapid fire charts. I'm even really struggling to pause the video....
Change playback speed.
Same for me. If he slows down a little he has more content (longer videos) that are better to follow.
Yes, please keep each chart up for 5s minimum, not 1.5s as the price/performance chart was flashed! On a phone there is a 2 seconds delay for TH-cam to respond to pauses, and often it takes THREE OR FOUR replays to be able to see the chart-flashes on-screen, please avoid doing this !!
Not saying you are wrong, but fyi if you are on desktop you can go frame by frame with angle bracket keys
Maybe they should add timestamps for the chart as well. Yes, it will make following the topics a bit messy but will be able to pause the video & go to the chart directly.
I've been watching laptop youtubers for years and it seemed like you came out of nowhere but it feels like you're the only one who truly understands laptops as devices, the marketing behind them, and how to actually test and review results!!
you and your team are doing great work. thanks for being objective!
For some reason your channel flew under my radar.
I was not expecting someone bringing in a fresh take on tech at this point.
A joy to watch!
something important i've noted on the NPU chart: the M4's 38 TOPS was calculated at INT4 precision while all the others use INT8, at INT8 the M4 falls down to 19 TOPS
Wow! wow wow wow. Where did you source this?
Yeah we need a source. I heard that M3 & below was INT16, while M4 was INT8
This is false! The M4 is measured at int8(=38TOPS), whereas the M3 and previous M series chips were measured at FP16(=18TOPS). For M3 at int8=(36TOPS).
"Keep in mind that the M4's 38 TOPS is based on INT8, while the 45 TOPS number from Qualcomm is based on INT4. According to NVIDIA Benchmark Data, INT4 precision can potentially bring an additional 59% speedup compared to INT8, which means the 45 TOPS (INT4) benchmark number could be around 28 TOPS (INT8). The Apple M4 38 TOPS (INT8) can possibly be 35% better than the Qualcomm Snapdragon X Elite 28 TOPS (INT8) in ML performance."
Btw the Snapdragon Hexagon still the 2013 project what was just 32bit (int4 - 32-bit values in range), and Apple completly eliminated the 32bit compability so they are only measuring in 64bit (int8 - 64-bit values in range).
@@TamasKiss-yk4st They're all INT8 except the M3 lol.
Stop spreading misinformation.
The Asus ProArt Px13 is such a beast, if only it wasn't 60hz
fr, i really don't understand why they decided to so heavily bottleneck an otherwise amazing laptop
That laptop was making me regret getting a Surface Pro 11 until I saw that and the price.
true, nothing pro about a 60hz screen
Exactly. What were they thinking 🤦🏻
🔥
Because I got the most important info on the beginning of the video, I'm definitely watching the whole video, since I feel my time appreciated. Thanks for the amazing work as always!
Thanks mate!
Can you conduct comprehensive and serious battery tests, rather than the incomplete ones you ve always shown? Marc Andrews' tests of the S16's battery show improvement, but it still falls significantly short compared to Qualcomm Snapdragon (Surface laptops etc). The Samsung snapdragon laptop with the X84 beats the S16 with its Zen 5 in performance.
Exactly my words...I hate his battery and efficiency benchmarks...tells me nothing
@@alexs.5107 been saying this for a while. Case in point : streaming videos Will always favor OLED screens vs web browsing where IPS screens excel. They should test The battery with mixed uses of browsing, video playback and office productivity apps
@Pepo5288 It s amost like he has beef with Qualcomm. We just want fair ,objective reviews. Period!
The snapdragon chips still have better efficiency than the Zen 5 at low power tasks. This channel has a slight anti-qualcomm bias in all their videos. Probably because of how badly the snapdragon was overhyped. It rubbed some people the wrong way, including this reviewer.
Snapdragon is twice as efficient in single-core tasks, look up the Notebookcheck review (Cinebench 2024 single-thread perf/W). Somehow Josh doesn't feel like mentioning this fact, despite single-core efficiency strongly correlating to real-world battery life. If you fire the laptop on all cylinders, sure, Zen 5 shows similar efficiency to Snapdragon. But that's not reflective at all of day-to-day usage. Also Snapdragon doesn't have efficiency cores, if the next generation will also have efficiency cores they will easily crush Zen 5 (or 6).
I appreciate your work in putting together detailed efficiency graphs and really explaining how the new tech compares to other offerings.
Thanks for the comprehensive analysis in your video. Still not convinced that the Zen 5 chips are more efficient than ARM. I have looked at many other reviews and none seem to say that the Zen 5 chip is as power efficient as the Snapdragon chip.
Because it isn't.
Power efficiency varies with load. The takeaway is that Snapdragon chips have diminishing returns as you increase load. When you do medium workloads, it is close to AMD efficiency and, in heavy loads, as bad as Intel efficiency. The XElite are good at low-power tasks like office, browsing and watching media, where they are second to Apple.
@@andyH_England yes, which, in real life terms, translates to more battery life as those are the tasks people spend more time on (generally) and at the end of the day that means less energy usage.
Snapdragon x elite is the first attempt from the nuvia team and it's not very good! It's very muscular but it wastes power. According to Jim Keller both Apple arm designer and AMD ryzen designer, ARM is not implicitly more efficient. Intel just makes power hungry shit, thats all (the latter is my own assertion).
Very comprehensive and thorough review! Thanks!
So the best Zen 5 laptop has the SAME peak CPU performance as the slowest X Elite, whilst running louder with worse battery life? And we should be impressed by this?
And where are the tests for standby time? Non intensive tasks like Web browsing or using productivity tools? How about some tests running on battery vs AC? How about testing the general smoothness of the OS, opening apps, using application features?? Like you know, what most people do on their laptops? People do not buy laptops to run benchmarks lol.
All this time spent testing theses laptops yet there's little to know useful information as to how these will perform for normal everyday users.
The X elite used in this video was the fastest in multicore in their testing. Take a look at the recent dedicated x elite video. They only have compared it to to the X1E80 midel though, but the X1E84 is inonly very few and very expensive laptops and without data it isn't clear if that chip would be much faster as the multicore score might not be significantly impacted by the CPU designation. And as you can see in the video in the netflix playback test and in the full load test the HX 370 without dGPU used less battery % than the x elite.
@@1world0 the point is he is not representing all of SD x elite laptops, while he got all 3 ryzen laptops and each one is tuned for either battery life or performance. You can clearly see how pro art throws watts to get performance but for battery life he takes the other 2 laptops. So why not the same for SD x elite cos they will look better than amd and goes against his bias
@@nanjappa, The X Elites cannot live with the AMD in GPU tasks, where the battery efficiency is lost on the ProArt (the S16 matches X Elite). 100% of people needing GPU power will buy ProArt over X Elite; the Snapdragon is not in the game. So your point is moot, as the S16 AMD is as good as the X Elite in efficiency for casual workloads.
@@andyH_England lol, do you know the price difference why would anyone looking for gaming laptop be lookin at x elite
@@nanjappa,
As for my previous comment, I was saying that if you need a good GPU, you should not buy the X Elite, so they would not be on the radar. So why mention the ProART? Snapdragon does not compete in that GPU league.
Fantastic work, as always! I'm confused by the discussion of efficiency and battery life, though - there's always variation between different reviewers' battery life tests, but I can't find anyone else who's found the S 16 to last anywhere near as long as a Snapdragon or M3 laptop. Everyone else seems to have found the battery life to be no more than mediocre in comparison to those alternatives, but in your test it's as good or better. I'd love to know what explains the discrepancy, here.
For example, on Laptop Mag 's tests the S16 lasted for 11:35, and the Air 15 lasted for 15:03; on Andrew Marc David's test the S16 lasted for 15:37 and the Air 15 for 18:33; on NotebookCheck the S16 lasted for 10:40 and the Air l5 lasted for nearly 17:00. These times suggest that Zen 5 isn't as efficient as Snapdragon X, as with similar sized batteries those laptops match or surpass the Air.
Why aren't other reviewers finding Zen 5 as efficient as Snapdragon in practice?
It seems as though he is comparing battery life in performance tasks or video watching instead of things like standby or light web browsing.
Wonderful video buddy. I ordered the SL 7 15 inch version (16gb ram), it should arrive sometime next week. I am wondering how snappy are the new AMD laptops? One main reason for ARM is the snappiness, for example you open any window, it just feels like zero lag or zero stutter (I tried out the SL7 at stores before ordering and I was blown by the snappiness of it). Are the new AMD chips the same or there is a minute lag (compared to ARM chips)? It would be great if you can elaborate more on that.
Great content, and looking forward for more :)
I'd like to know as well
As someone who currently has both the 15" SL7 and the Zenbook S16, I'm happy to share my thoughts. They are both snappy for everyday tasks (browsing, email, office apps, etc...) with SL7 having a slight upper hand but only kind of noticeable if you have them side by side and barely. Personally, I would not let that be the determining factor for which one to go with since they are both exceptional in that dept yet vastly different in other areas. I am still undecided in which one to keep since each one has something I like better than the other.
I should point out that I'm not a gamer and I don't do anything heavy. Technical aspects of either of these are not of great concern to me since both are more than enough for what I will use them for and for my use case there are no compatibility issue with ARM. My area of focus is on form factor, build, look and feel as well as well as battery life.
My ideal form factor is a 16" screen, OLED, no numpad. If the 16" Samsung GB4P or Edge didn't come with that horrendous touchpad, I would have been satisfied.... I really love the 15" SL7. It's almost perfect for me. Incredible build quality, phenomenal battery life, awesome touchpad and keyboard. The only reason why I'm not fully in love is because I really prefer a 16" screen size as well as OLED, hence the Zenbook S16....
I'm trying to love the S16 but been having issues, mostly with build. Comparing to a Mac, SL7 and GB4P, the S16 feels on a cheaper side, not quite premium. I'm currently on my third unit. The first two were 24gb "white" ones (light gray really) and this third one is the 32gb dark gray one. The bottom panel/cover is quite thin and if you squeeze the laptop at the very front, it can cause a popping sound. That was my very first initial experience as I was taking the first unit out of the box (Maybe a bit of exaggeration but think a sound that a thick empty can would make if squeezed). As I squeezed, there would be a gap or opening created wide enough to put my fingernail through along with a popping sound. Not good. Fast fwd to the third unit which just arrived today, the dark gray 32gb model, the touchpad rattles when I tap on it, like it's loose. It's quite annoying. This was not the case with the first two units... So long story short, there seem to be some major qc issues with Asus. I absolutely love this form factor, it's almost perfect for me, but when I go and pick up my SL7, I can't help but asking myself why I'm willing to go in circles with the S16 especially since the battery life is not even close to the 15" SL7 and the SL7 is nearly perfect for my use case. And also, the screen on SL7, even though not OLED is definitely brighter than the S16 screen which can be beneficial in the dark mode. Also, the keyboard, on the Asus it light up like a Christmas tree which can be distracting to some of us where on the SL7 light only comes through the key symbols. So that's just a matter of personal preference.
So that's my experience. I hope this somehow helps you or anyone faced with the same choice.
@@VladdyDaddy369 Thank you for your reply. Can you say something to these concerns, how they are different on each of the devices?
Restart / Boot up time
Wake from standby
battery lost in standby (e. g. overnight)
fan noise
trackpad (does it make a big difference?)
is the reflection on the screen noticeably worse on SL7?
@@AllTernative90
Restart / Boot up time
Both fast but S16 is faster by 2-4 seconds from both the shutdown and restart boot (I disabled all startup apps options shown on both and tried 3 times)
Wake from standby
Both nearly identical, pretty much instant. Open the lid and the ir camera recognizes immediately.
Battery lost in standby (e. g. overnight)
SL7 about 2-3% depending on duration of "overnight". I never checked it on the first two S16 units and won't know on this third one until tmrw obviously
Fan noise
Virtually non-existent with SL7. I got to hear the fans when running Cinebench 2024 and the top of the laptop above the Fn keys by the bottom of the screen was super hot when running multi-core. Otherwise they never kick on for my normal daily use and the laptop stays very cool.... S16 gets a little warm even when not doing much. How much fans kick on depends on the fan profile selected but they kick in a bit when I don't expect while browsing on Chrome.
trackpad (does it make a big difference?)
OMG absolutely yes! That haptic touchpad on the SL7 is worth every penny. It's so smooth with no palm rejection issues. It's simply phenomenal.
Is the reflection on the screen noticeably worse on SL7?
With both laptops side by side and turned off, yes for sure. However, I never really notice it while using because I keep screen brightness at 100% in dark mode which obviously makes a big difference. Speaking of, just did both laptops side by side, dark mode and screen at 100%, the SL7 screen actually looks better to me because of brightness. The S16 is almost like it's missing that one-more-click. I feel like I can see and read better and easier on the SL7. Hmm, interesting, didn't notice that before.
I don't think I have found my "the-one" yet, I'm curious what the next couple of months will bring with both Apple and Intel but for now I know which one I'm keeping. Chatting with you actually really helped me decide lol. Thanks!
@@VladdyDaddy369 Thank you so much for your time, that was super helpful! I'm leaning heavily on the Snapdragon side (especially on the SL7) but with all the reviews of the Zen5 chips I'm getting second thoughts. The SL7 from your description sounds very nice but I hoped it would loose less battery in standby. The trackpad does sound amazing and it's afaik the only laptop with it. I worried about the screen if it's good enough, but your explanation helps. Especially 3:2 aspect ratio is such a great thing to have, and there is no other laptop ticking the boxes like the SL7.
I hope there are discounts now with zen5 chips out, but I really doubt it.
Watching this video again. It was that good. BTW, Josh, I enjoyed your recent cameo on HC! They produce some great content. Also, I think it would be cool if you collab'd with a few smaller channels whose content is top notch, including Kyle Erickson, Mark Linsangan and Notebookcheck Reviews (I understand notebookcheckreviews was in the HC video as well). I like when larger channels throw some shine on the quality up-and-comers (maybe there are some you know that I don't even watch yet). Of course, when you collab with JT - we can't get enough of those videos. JT has gotta be one of the best in the business of reviewing laptops.
I finally subscribed. I love your personality, the willingness to help people, and the hard work. The data presented is surely top notch, and I am sure it takes a lot of work and time. Hello from Indonesia
at 11:10 I think the M3 Pro 12C results are wrong, according to geekbench website, should be around 68000 points instead of 157000 ... I think that result is from the M3 Max variant with the 40 gpu cores
You are SPOT ON, we have checked all other graphs and it is only the 2 Geekbench ones that have this incorrect. We are going to post the correction in the description and pinned comment
Man, looking those power efficiency charts and graph(s) and all I can think about is a new Razer Blade 14 with these Zen 5 chips inside. The 2024 model (RTX 4070 + 8845HS etc.) is already amazing, now with this proper CPU upgrade... 😳
Would be a dream LAN party machine.
@@jackofthecoke Indeed! Exactly my thoughts, especially with a 240 Hz display so there's no need to carry an external one (although it's recommended, obviously).
@JustJoshTech 3:53 You used the M3 Max nT score. Your, 'MacBook Pro M3 (14 & 16): You're Being Misled' video reflects this
The amount of work team Just Josh puts in for these evaluations is insane. Thank you so much. This channel is my reference for computer reviews. Once you discover this channel, the majority of those other highly popular "reviews" channels are like infomercials*.
(* Incidentally: "A Pixel 9 review agreement required influencers to showcase the Pixel over competitors or have their relationship terminated." --The Verge)
Ryzen 4070 dominates better
Fr fr 🗣️🗣️🗣️🗣️
What people don't understand is the original Commentor might have meant Ryzen 7000series CPU plus RTX 4070 GPU hence the name Ryzen 4070!
Instead of considering this angle, people preferred to make this a meme(ZTT)!
PC building brainrot is crazy
brain not found
How did it make it here
But why test battery life on the highest performance mode? That's not how people use their laptops.
In high-performance mode running video it will use the same wattage as in balanced. The difference is that in high-performance mode you do not get power throttled, which you would in balanced.
I appreciate all the work put into this. Thanks!
I got the Vivobook S 14 and am extremely happy with the Machine. The HX370 in the Vivobook actually runs at 45W in Performance Mode (at around 80°C) without getting too loud and is fast enough for most games I play (with AI scaling from 1200p to 1880p). A 75Wh battery in a 14" chassis is actually impressive (for 16" on the other hand I'd like to see around 90Wh instead of just the same 75Wh/78Wh ones in the larger Vivobook/Zenbook Variants). Display and Battery Life are also great, my only caveats are: Only one of the two Type-C Ports is 40Gbit (the Zenbook has two full speed ones) and it has a WiFi6 Chipset (performant, but not amazingly futureproof).
Anything else you can tell me about that I should be aware of before purchasing the Vivobook. No one on the internet is reviewing the Vivobook. Only the Zenbook and ProArt reviews are available and I think that's largely due to their better build quality, battery and efficiency
@@jatingumbhir8954 Not a lot in terms of Hardware (I've only had it for just under a week by now) but found some Software Tweaks.
1. If you want to game on it you need to set the reserved GPU memory in the "MyASUS" App to something appropriate (like 4GB). The default "Auto" setting just uses 512 MB and is relying on Windows to virtually swap VRAM over PCIe to/from the iGPU instead of having it be in direct control.
2. The Keyboard RGB is controlled in the Windows Personalization Settings Panel.
As for the Review Samples:
The Vivobook Chassis Design is older and less impressive (though imo actually more functional). The same physical components and Layout were used in late 2023 / early 2024 which is also the Reason for just WiFi6 and only one 40Gbit port. It's technicall not a new Laptop and instead "just" a CPU upgrade (but a very significant one, especially for efficiency)
@@jatingumbhir8954 I've been hearing many complaining about how hot the laptops get, even during low stress loads. Other than that, it seems like a good value for the money
px13 looks so good, compact with enough power even for good gaming
isn't it with just a 60Hz screen?
I sent mine back the 60hz and lack of gsync means is rubbish for gaming. Plus it has a massively power constrained 4070 in it.
I went for the zephyrus g16 with the refreshed amd processor in it
Giving your findings and results in the beginning is amazing. Love the videos, appreciate the honest content. Thanks!
I absolutely appreciate what Josh and his team is doing in laptop reviews these are proper reviews
Great work, guys.
These Pro Art laptops would have been pretty fantastic if the eco mode worked properly and they actually stopped feeding power to the Nvidia dGPU. I refuse to buy a laptop with a dGPU if it's not optimized properly and can't be fully turned off when not needed, as i value battery life, especially when mobile, otherwise, these would have been pretty fantastic. I don't even mind the 60hz screen which is fine for on the go computing and better battery life. I use external monitors with much higher refresh rates when i'm stationary.
This here is kinda a valid argument. I would love to see a laptop with dgpu than has an outstanding battery life, otherwise it's macbook max
@@alexanderruchkov7629 honestly, i'm leaning towards entering the apple ecosystem due to the inefficiencies and bugs of the Windows and Linux ecosystems.
Hey Josh, could you please test if these Ryzen AI laptops support HW acceleration for decoding HEVC 4:2:2 10bit video? I’m finding it difficult to find that info and apparently only Apple silicon or Intel 11th gen and newer support it, which is relevant for any user filming video with cameras like Sony Alphas
To my knowledge, Zen 5 does not feature any updates to AMD's hardware video encoder, which means that it still doesn't support HEVC 4:2:2. And yes, that's a feature reserved to the latest version of Quick Sync, and I couldn't find anything except a random Reddit comment to confirm it works on Macs.
Maybe Yes? Have you try to search RDNA 3.5 feature?
Really enjoyed the structure of this video Josh 👌🏻
Nice channel
Keep going mate
Awesome news, loved your livestream and this video is a very impressive "summary" of all your findings. I bet you and your team spent a lot of time just for this video
I can't wait to see the ThinkPads with these. Hopefully they come out soon, so I can see your review on it.
as per usual, regarding AI tasks, I'm kindly requesting an Ollama local test with 8b and 70b models such as llama 3 and seeing how these laptops perform on it. As most likely this is what most local AI is going to do, generate text and use it inside various programs
Yes, please. There is even a ollama-benchmark available which can be installed with `pip install llm-benchmark`.
Interesting review. Thanks for the time you put into these tests.
Thanks for making this, really helps in decision.
Hi Josh and team,
Seriously guys, I don't know anyone on TH-cam or anywhere on the internet for that matter who does this as professionally and tastefully as you guys.
Really appreciate the awesome, and meticulously detailed content.
How u don't have at least 10 million subs by now is mind boggling.
Wish u all the best.
Thanks again🙏
Thanks for the absolute great video. I just received my PX13 and can confirm there must be something wrong with its dGPU's ECO mode: by its description it should turn off the dGPU completely, and as you mentioned in the video and I also confrimed, the dGPU disappeared from both device manager and task manager. However, on my PX13, the proart creator hub (ASUS's laptop control center) shows the GPU fan is still running with 2000+ RPM, and its number is different from the CPU fan's RPM, which makes me believe the dGPU is not turned off at all, or the laptop thinks the heat from CPU somehow needs 2 fans both run 2000+ RPM to exhuast (I am at wisper mode, CPU runs at 2.57Ghz, with less than 50 C temperature reading).
Zen 5 and X Elite are showing that the node is the most important factor, they are built on the same 4nm node, while Apple have an exclusivity deal with TSMC and are on 3nm
It's not just the 3nm.. there are a few test where even the 4 years old M1 can outperform the X Elite... so the 5nm M1 perform better in a few cases like the 4nm X Elite (and that M1 chip also had the same Watt limitations.. so it's still 20W against 40-50W.. and the 20W 5nm beat the 50W 4nm..
It's actually not the process node that is everything. Note that the Ryzen AI HX 370 is on the same 4nm process node as the 8945 HX, but its both faster and uses about half the power. RDNA 3.5 brought all the power saving ideas from RDNA 4! That's from AMD brainwork, not the VLSI process node. ARM holdings CPUs are slow as dog doodoo thats why Apple and now Qualcomm are designing far more powerful ones to go into laptops with batteries 12x -14x larger than the phones ...
Since 14nm the new process nodes have been saving power but there has been almost ZERO speedup from Moore's law. All the speedup comes from tuning the circuits or using combinatorially more complex ones, e.g. multipliers can consume an almost infinite number of gates if you let them ..
Let me just ask you this. Why is the battery 🔋 last longer on SnapDragon if the new Zen 5 has better efficiency?!
That's because he's only comparing multi-core efficiency. Snapdragon doesn't have low-power cores, so Zen 5 has an edge there. For single-core workloads Snapdragon is twice as efficient, which would translate into increased battery life with light usage. Now idea why that's not mentioned or measured, single-core efficiency is very important for long battery life.
im guessing its because snapdragon is good at sipping power, but once the power increases, the snapdragon becomes pretty bad
That goes against his conclusions, so he won't say that. Users do bursty workloads on laptops not cine bench. His battery methodology is flawed and he won't change it
Because this methodology isn't true to real life scenarios, which is what matters.
@@RequiemAeternumI mean, in real life scenarios people don't tend to use their laptops while on battery for upwards of 10 hours, either.
Tip for the team, disable the discrete GPU in device manager when you want to run benchmarks for iGPU. 11:24
Yeah these software solutions provided don't always work properly.
Good work! I like it details and precision. Thanks!
Best video for comparison cpu i saw..
Thank you!
Superb Review.Thanks a lot for such good attention to detail and effort invested.
Again, super great work! Thanks for publishing!
Really looking forward to individual reviews for all these new Zen 5 laptops
Thanks for measuring the battery consumption properly!
Thank you so much
Just picked up an ally x and wow the performance from the z1 extreme makes me really appreciate how powerful AMD makes there chips
Clever testing > big labs built for testing, just awesome work Josh very awesome.
Nice video Josh (and the team). Now im actually considering the upgrade from my i5-8350u laptop. AMD did an amazing job on these mobile chips lately, however i see a different story with their new desktop 9000 series.
Ooo. If I'm reading this right, and I'd like to think that I am (because it was very well explained!), these new chips will be very good performers, across a variety of tasks. I'll look forward to seeing all the Zen 5 laptops reviewed!
A Legion or Yoga Pro 9i laptop with these chips would be god tier.
aren't 9i's an intel variant?
@@alexanderruchkov7629 True, it would be Yoga Pro 9. But they currently don't exist
@@iz723 i bought the 4060 ultra 9 yoga pro and no regrets, this thing feels like a macbook pro but for windows, but i can run games like val at 350fps with a 160hz monitor LOL
i would've went for AMD's too because it's clearly better but no complaints here!!
Hi Josh, when you are showing the charts why not show the bars in descending order? It helps a lot to compare with other contestants really quick. Thanks, Love your content.
I am loving my ProArt P16 it really is the Windows MacBook Pro, with some advantages. I noticed on your battery test you shut off the Nvidia graphics on the PZ13. However, did you put it on quiet mode that will up the battery life with acceptable performance for general. These AMD chips, along with the Asus ProArt series, are among the best Windows laptops ever made.
Can't wait for the ProArt PX13 review, that looks like a really interesting laptop!
Any guess how Lunar Lake will compare to Zen5 in terms of performance per watt?
We'll see soon, hopefully.
Really liked the cyberpunk tests, hope u add more games like this. Thanks a lot
Maybe also include the MacBook with Crossover (Wine) with D3DMetal support. Just for comparison even though it takes a performance hit through the emulations layers.
@@franzpleurmann2585 right...Im planning to do so such reviews on my own actually. Almost no youtuber does true comparisons like this. Waiting for my galaxy book to be delivered in order to compare the ultra 7 and m3
out of curiosity, why not unplug the battery before measuring power draw from the wall? That allows you to isolate another variable so that you can more accurately show power consumption. Just my 2 cents. :)
Please, Josh, when you edit your audio, use a rebreather from Adobe or iZotope to remove it. It's a bit distracting when you focus on it. Hehe. Oh, and thank you for calling out TH-camrs focused on sponsorship and not being frank about poor devices. That was IMPORTANT!
Really well made video. Thank you very much for your amazing work.
Thank you!
Good stuff. A lot of nuances that the "casual user of a desktop' will miss but vital if you use laptops long term. Gamers probably don't care about power efficiency.
Josh its an obvious design mistake to put a 60Hz OLED panel in a dGPU Laptop. Nobody wants to blow money on an OLED panel thats 60hz. There are no such TVs of this type that I am aware of. I say this as a ProArt monitor buyer (love the brand.) The S16 is really a highly superior 32Gb macbook air alternative and it will run on 17w for most of its life, except for rare times gaming away from a desktop. I feel like its closer to perfect ...
I know you really did not like the 16s 370 but it was an instant buy from me. I needed something thin and light that would augment more powerful systems. To me it strike the right balance giving more performance than older chips while keeping power draw low. If I am traveling and need to get real work done I still have the 4080 razerblade that I can break out at my destination but this new zenbook should hold me over between. Hopefully I still agree with this statement in a week when mine arrives.
We actually have changed our tune on them. Video out soon
How is it so far?
Would you still recommend the HP Spectre 14? It's already out for a few months and it seems like a laptop with a newer processor (like the AMD Zen 5) would be a better deal. But your review back then totally convinced me to get the Spectre. Also, it ticks all of my boxes (2in1, good display, battery life, keyboard, size, design and even little things like a privacy shutter for the webcam)
Please also compare vivobook S 14 with these laptops. I don't know why nobody is talking about this 14" laptop.
Also why is this processor Asus exclusive?
asus gets a head start. laptops with these CPUs will start appearing at the end of the month and especially next month
Would love to see both Zenbook S 16 compared. The Lunar Lake and the Strix Point ones. This will be an unique opportunity for a direct face off.
Literally every other video I've seen has said that Qualcomm is still more power efficient?
Me too, JustJosh seems to have some bad blood with Qualcomm
@@alexs.5107Josh seems to be testing his hardware thoroughly unlike other "reviewers" (looking at you, Dave2D).
@@IvoPavlik Don't agree. I have a qualcolmm for my personal use. I even use for develop PLC code on automation. That software is only x86 and it runs better on QUALCOMM for some reason. The only really reason I see bad on SnapDragon is drivers for weird devices like DAQ cards. Everything else can be emulated on PRISM and work pretty well.
I think Josh has a very bad test for Qualcolmm, Intel and APPLE , and he loves AMD since many years he is being very bias in favor of them. Every one has a favorite Brand, I like AMD but I recognize in every built I have done with AMD I got more issues than Intel. You need to be a power user on AMD to actually know which setting to apply while on Intel it just works!.
What a helpful and well done video. Keep up the great reviews and analysis.
Josh, take it from a person who setup LLM gor developers over the last year: nothing comes close to a full sized GPU with at least 16GB video memory. So Nvidia 3090, 4090 or AMD 7800xt or 7900xtx ... The LLMs xan be run on a separate workstation and users can connect over the network to chat or to perform tasks. That is yhe only practical way to use LLMs from a laptop today
Why does the asus zenbook 14 2024 woth the ryzen 7 costs in the US 800$ and in germany 1200€? This is annoying
So... Radeon 880M is the same as 780M, or maybe even worse depending on the laptop
no it's 20% faster in games and sometimes even more.
@@coffee7180 still no 4k@120Hz desktop on external display, meh
I'd expect LNL BTM iGP to beat this by a healthy margin if the drivers work
@@alexanderruchkov7629Anything that supports HDMI 2.1 or DP 2.0 (DP1.4 works if you're not using HDR) can provide 4K@120Hz on an external display. The AI HX processors support both of these standards.
I am sorry for the folowing question that may be a stupid one... Nowadays, videos are made in two formats: 1080p and 4k (except for youtube of course). Why does a customer want to buy a laptop with 3K monitor? Technincally there should be worse video quality and worse battery life, right? Thanks for the explanation. Love your videos
OLED 1080p looks worse than LCD 1080p due to the pixel structure (arrangement), so you really need a minimum QHD for OLED for a retina display. Most OEMs vary between 2.5K, 2.8K and 3K, which can depend on the aspect ratio and supply (prices). Some use higher, like 4K for professional users.
@@andyH_England ok, that's clear. So better avoiding OLED 1080p. However, if we consider OLED vs LCD, for entertainment purpose, you will probably see 1080p video worse on a OLED 3k than on a LCD 1080p, right?
@@davidemora8762 I understand that OEMs use upscaling to fill in pixels on a 4K screen if the video is 1080 P. I assume it will be slightly better than 1080 P video playback on a native 1080 P display.
Excellent video Josh. Really good data.
Honestly I don’t think u gave the vivobool s 14 enough of a shout out here, that thing was killing in the bench marks
Awesome data and graphs
Great detail's as always. A question on how battery performance to "percentage left" is measured how? Is it calculated from the watt hours drawn over the period and calculated from the battery capacity, which would have had a prior draw down test to turn off?
If it's measured only using the battery percentage provided by the system, this is likely misleading as it's not linear in most devices and varies.
I got the ProArt 13, it's a little beast. Battery life is annoyingly brief though, hopefully a BIOS update will stop the power leak to the 4050.
I’m really sensitive to screen quality when it comes to laptops.
I don’t know what OEM Lenovo sourced their LCD from but my Slim 7 Pro X (2022) has one of the best screens yet it’s a 14.5” 3072x1920 IPS panel with 120Hz refresh and 100% sRGB coverage. It’s pretty comparable to the MacBook Pro 14 screen just without the notch and the mini led backlight.
Awesome information!
Waiting for a ThinkPad T series release.
You're amazing Josh thanks a lot :)
I honestly want a Flow series laptop, preferably x16 with this cpu & 4070, srgb 240hz screen & upgradeable ram with more external io. Basically an upgraded proart with more features.
If only these could come out a few months ago for students
good review, I appreciate the effort.
Games should always be a option, you never know when a game shows up that you like to play or even just to play it with friends.
Could you look at the macbook air 13 performance per watt statistic again? This statistic seems like an extreme outlier in this video. In another video of yours it was also an outlier even compared to other apple devices. Are you sure there is not something wrong with the macbook air measurement?
2:19 "These processors are certainly no skynet!" That's gotta hurt!
I do like the benchmark Vs wattage graphs however I think the Geekbench Vs. Wattage graph does probably tell you something...specifically that X Elite is more efficient when it comes to single core usage (almost double compared to AMD or Intel according to Notebookcheck in cinebench 24)...and as you say, geekbench mixes a lot of bursty activity with full cpu usage....explaining why X elite sits closer to the top here. Maybe worth doing your Cinebench Vs Wattage for single core scores too?
Very interesting idea, we will try this
@@JustJoshTechInstead of constantly praising yourself, learn from other reviewers too.
I know most people probably won’t see this but as a mechanical engineering student would you recommend a 4070 g14 or a 4060 px13 (both from 2024)
Review out tomorrow
Tx Josh, perfect review
When are we going to see Zen 5 laptops from manufacturers other than Asus?
September 4
It’s interesting that in some other reviews the zenbook s 16 outperforms m3 due to it being set to full speed mode instead of performance mode
Looks to me the only thing saving apple here is the node level that keeping just above the others. AMD is a hybrid design between the two - thus will be drawing a little more power and using a bit more space per chiplet. It also doesnt have as much cache it looks like - and we know at least in gaming, cache can play a roll in some applications. Drawing a bit futher than they are now, looks like AMD is pretty much on the toes of Apple SoC and that pretty much gives us a picture on how the all 3nm SoC will look like follow by whatever M4 looks like a few years later down the line as they move to 2 or 1nm.
The M4 is Apple's 2024 chip as this is AMD, so for the M3 Max to still be on top of the best from AMD is remarkable, bearing in mind the M4 is a massive generational leap which will put the M4 single-core about 35% above AMD! So in 2-3 months, when most AMDs arrive and the M4 is released in the Pro models, there will be a no-contest, and the 2-3 gen CPU single-core gap will return.