Review Asus ROG Zephyrus G14 Review: AMD Ryzen 9 4900HS Tested

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

csbin

Senior member
Feb 4, 2013
838
351
136

On the Shadow of the Tomb Raider benchmark (highest, 1920 x 1080), the Zephyrus ran at 49 fps, tying it with both the Acer Predator Triton 500 (i7-8750H, RTX 2060) and Dell G7 15 (i7-9750H, RTX 2060).


Asus’ laptop ran Hitman (ultra, 1920 x 1080) at 89 fps, two frames ahead of the Predator and one frame ahead of the Dell.


The Zephyrus outperformed on Grand Theft Auto V’s benchmark (very high, 1920 x 1080) at 115 fps, losing by two frames to the Dell but easily beating the Predator with 87 fps.

We also ran our gaming stress test, in which we ran Metro Exodus 15 times on a loop to simulate half an hour of gaming. In this case, we ran the game at the ultra preset at 1080p. The game ran at an average of 40.5 fps, and with RTX on it dropped to 37.8 fps. The average CPU clock speed was 3.1 GHz, and it had an average temperature of 78.4 degrees Celsius (173.1 degrees Fahrenheit). The GPU ran at an average of 425.1 MHz and a temperature of 64.8 degrees Celsius (148.6 degrees Fahrenheit).


pnYUIDFMoEKSgBH.png



HOjb3NPAuiLXrkt.png


Oz5wAilKQET6Fxj.png






It’s truly incredible how long the Zephyrus lasted on our battery test. Generally, only the best ultraportables last between eight and 10 hours on our test, which continuously browses the web, streams video and runs OpenGL tests, all while connected to Wi-Fi with the display at 150 nits brightness.

The Zephyrus endured for 11 hours and 32 minutes. That’s incredible for a gaming notebook and even for some ultrabooks. The premium gaming average is just under 4 hours. This means the Zephyrus is suitable to use as your everyday system in addition to being your gaming machine.

For comparison, the Acer Predator Triton 500, with an i7-8750H and RTX 2060, ran for 4:24 and the Dell G7 15, with an i7-9750H and RTX 2060, died after 3:12. The Razer Blade Stealth 13, with a 25W Ice Lake processor ( i7-1065G7) lasted 8:53. The Dell XPS 13, with a 6-core/12-thread i7-10710U Comet Lake CPU ran for 7:56, albeit with a more taxing 4K display.

And while it’s not quite the best comparison, the MSI Alpha 15, a budget all-AMD gaming laptop with an AMD Ryzen 7 3750H and a Radeon RX 5500M graphics lasted only 3:53.



oZBe6xafzRNPyku.png
 

Gideon

Golden Member
Nov 27, 2007
1,598
3,520
136
So for most benches, it's running @ 54W.

Yet the 45W 4800H still performed better (possibly boosting at even higher wattage for a short while?)

Marginaal 2020-03-31 170500.jpg

, forget about 35W no way in hell that the CPU is running at 35W while the GPU is not in use, thats why it performs so close to a 3700X
Anyway, that behaviour is quite similar to how Intel CPUs operate and it's not like the 15W 4800U running in lenovo 25W chassis is performing that much worse:

ES-aiPwWAAA5ETT
 

mopardude87

Diamond Member
Oct 22, 2018
3,348
1,575
96
Boosting algo for the G14 is as follows:

65W for 10 seconds.
54W for 150 seconds.
35W indefinitely.

Check Hardware's Unboxed's review for this:
c12211fb39340787cef1c30894b6c7ff.jpg

Yeah i love how that functions cause with luck most things you do will certainly be done in 150 seconds. Not a bad way honestly to handle things. Good job AMD.
 
  • Like
Reactions: DarthKyrie

Shivansps

Diamond Member
Sep 11, 2013
3,828
1,505
136
And it dosent drop to 35W after 150 seconds, it goes down gradually over the course of another 50 seconds.

Well this explains why it benches so closely to a 3700X.
 

Hitman928

Diamond Member
Apr 15, 2012
5,159
7,595
136
And it dosent drop to 35W after 150 seconds, it goes down gradually over the course of another 50 seconds.

Well this explains why it benches so closely to a 3700X.

EDIT: That Handbrake test image seems to show 80W for 10 seconds 72-75W for 150 seconds and around 65W after that, what im missing?

Blue is clock speed, yellow/brownish is power.
 

moinmoin

Diamond Member
Jun 1, 2017
4,926
7,608
136
The meaning of TDP gets tweaked constantly, lol. Yeah, it's essentially the way Intel does it already so OEMs should feel right at home with it. And by offering a way to connect external sensors to keep track of the chassis temperature AMD may even stay in more control of not artificially limiting the CPU too much due to conservative configuration by the OEM like happened in the last several gens.

Computerbase had a table of the different modes offered by Asus' software:

ModusCPU Package Power
(permanent)
CPU Package Power
(peak)
Frequency
(permanent, ca.)
Noise
Asus Zephyrus G14 „Silent“25 Watt50 Watt2,95 GHz31–38 dB
Asus Zephyrus G14 „Silent“ (on battery)15 Watt20 Watt2,40 GHz31–33 dB
Asus Zephyrus G14 „Performance“35 Watt65 Watt3,25 GHz31–44 dB
Asus Zephyrus G14 „Performance“ (on battery)20 Watt25 Watt2,90 GHz31–37 dB
Asus Zephyrus G14 „Turbo“35 Watt65 Watt3,25 GHz37–45 dB


I don't know if you remember that fiasco but yeah they went out of their ways to call HWUB trolls when the dude over there asked for a meeting, that is beyond disrespectful. He simply wanted to meetup and compare data cause obviously something was way off given Toms was the outlier while the rest of the people found similar issues with cheap motherboards.As someone about to buy a B450 and prob 3900x/4900x before end of year, if i listened to Toms i would be on one heck of a flakey motherboard instead of settling on a B450 Tomahawk Max. That one has been proven at least.

I will buy whatever HWUB sells before Toms, they did out of the case and in the case testing, they went up and above the call of duty with their testing. Heck i will buy TPU before Toms.

Edit: Yeah here is the video where he calls Toms out. Yeah NEVER again. :)

I remember the fiasco well. My two points were: It's easy to ignore Toms since there are plenty reviews to choose from anyway. And the blame for the unprofessional handling of said fiasco is down to its author Luke Hill. Toms review of the ROG Zephyrus G14 is by Andrew Freedman, after a quick look across his articles he seemed fine.
 

amrnuke

Golden Member
Apr 24, 2019
1,181
1,772
136
Well, I may have finally found a replacement for my 6700HQ laptop.
Dell XPS 15?
I have the same CPU. Honestly surprised it's feeling so old already. I got it in late 2017 and it just feels sluggish at times, despite 4C/8T, a decent SSD, and 16GB RAM. It does have a 4K touchscreen which I suspect causes some of the lag.
 

DrMrLordX

Lifer
Apr 27, 2000
21,570
10,762
136
Yet the 45W 4800H still performed better (possibly boosting at even higher wattage for a short while?)

Depends, I'm not sure how they have Renoir set up. On desktop, all the 105W TDP CPUs can boost up to 142W, and I've seen my 3900x sustain that power draw easily. It really never goes any higher than that. Mobile is going to be different, and it'll be limited by cooling solution. That Asus lappie is not going to offer thermal limits when the GPU is out of use.

Really, each individual implementation would have to be reviewed for power usage.

That's impressive. It just takes 35W to maintain ~3.3 GHz indefinitely for 8 cores.

That IS pretty damn good. Realistically-speaking, most benches/workloads are going to be finished in under 150s (especially on a laptop) so I don't expect it to sit there too often.
 

uzzi38

Platinum Member
Oct 16, 2019
2,556
5,531
146
-U series starts off with a similar hyper-agressive boost.

All the way up to 80W perhaps? Would only be that initial 10s boost though.

There is at least one 1 4800U that scores around 1800CB in CB15, which is short enough for them to boost so hard.
 

Shivansps

Diamond Member
Sep 11, 2013
3,828
1,505
136
The channel Tech Epiphany is posting a lot of videos of games running from the Vega 7 and the performance is very impressive.
AMD already have a killer APU for next year just switching Vega for Navi.

The 4800HS can mantain 1600mhz for Vega7 that puts it between margin of error of a desktop 3200G at 1600mhz, witch is impressive to say at least.
 

Arkaign

Lifer
Oct 27, 2006
20,736
1,377
126
An absolutely stellar SKU. When paired with a dGPU, these mobile Zen2 CPUs bring just absolutely monumental performance. I feel they should have even pushed this lineup first over the desktop variants to be honest.
 

DrMrLordX

Lifer
Apr 27, 2000
21,570
10,762
136
An absolutely stellar SKU. When paired with a dGPU, these mobile Zen2 CPUs bring just absolutely monumental performance. I feel they should have even pushed this lineup first over the desktop variants to be honest.

Margins are probably lower on the mobile parts. Also, Matisse is like Rome's toenail clippings. Once you have Rome, it's easy enough to cobble together Matisse for the desktop crowd. Renoir obviously re-uses the same IP, but design-wise it's much more involved.
 
  • Like
Reactions: Thunder 57 and lobz

Gideon

Golden Member
Nov 27, 2007
1,598
3,520
136
and here is the TDP of Comet lake cpus for comparison. suddenly brief 65w bursts on renoir don't seem so bad?

Intel's recommended PL2 (turbo power) for the Core i9 is 135 W, and Intel says the recommended 'Tau' is set to 56 seconds for the i9, and 28 seconds for all the other CPUs. OEMs don't often adhere to these values for notebooks, but they are provided as a guide. It does mean that in order to hit 5.3 GHz, the Core i9 is by default allowed to take 135 W across two cores, or 67.5 W per core. Even at 60W per core, you're looking at 50A of current per core... in a laptop.

from Anandtech article:
 

uzzi38

Platinum Member
Oct 16, 2019
2,556
5,531
146
Please don't tell me anyone thought CML-H was going to be any different to this?

It's more 14nm, a node they've already pushed to the very limit. To be honest, getting another +100mhz year on year was already impressive, getting any more than that was simply not happening.

CML-H is a housefire, and we all knew and saw it coming.

(Note: I'm not including TVB, but just TB2.0 rated frequencies)