Monday, March 30th 2020

AMD Ryzen 9 4900HS Torpedoes Intel's Core i9 Mobile Lineup, Fastest Mobile Processor

Reviews of AMD's flagship mobile processor, the Ryzen 9 4900HS went live today, and the verdict is clear. Intel has lost both performance and battery-efficiency leadership over its most lucrative computing segment: mobile client computing. In a Hardware Unboxed review comparing the 4900H to Intel's current Core i9 flagship, the i9-9880H, the AMD chip at its stock 45 W TDP beats the Intel one even with the Intel chip configured to 90 W cTDP.

The 4900HS posts 11.9% higher CineBench R20 score (both chips are 8-core/16-thread) when the Intel chip is bolstered with 90 W cTDP, and a whopping 33% faster when the i9-9980H is at its stock settings, and 54% faster when its capped at 35 W cTDP. It also ends up over 150% faster than AMD's last fastest mobile processor, the 12 nm "Picasso" based Ryzen 7 3750H. The story repeats with CineBench R15 (4900H being 34% faster than stock i9-9880H), 18% faster at Handbrake HEVC, 25% faster at Blender "Classroom," and 35% faster at 7-Zip benchmark. The AMD chip lags behind by 12% in the less-parallelized Photoshop. On creativity apps that do scale with cores, such as Premiere "Warp Stabilizer 4K," the 4900HS is 12.6% faster. Gaming performance remains an even split between the two chips. Find several more interesting test results and commentary in the Hardware Unboxed presentation here. Intel has already announced a response to the 4900HS in the form of the i9-10980HK.
Add your own comment

211 Comments on AMD Ryzen 9 4900HS Torpedoes Intel's Core i9 Mobile Lineup, Fastest Mobile Processor

#51
R0H1T
Huh, what a scam probably costs much as like their overpriced routers :nutkick:
Posted on Reply
#52
Vya Domus
londisteNot instruction sets. 4 additional cores, AVX2 and 7nm.
The number of cores and instructions are the same. Zen 2 simply has twice the FP throughput because the vector FPUs have been doubled in width.
Posted on Reply
#53
QUANTUMPHYSICS
The vast majority of buyers will still aim for intel and Nvidia.
Posted on Reply
#54
Imsochobo
R0H1THuh, what a scam probably costs much as like their overpriced routers :nutkick:
somehow 1400 $, 1920x1080 120 HZ, 4900HS, RTX 2060 max-q, 1tb 660p, 16gb ram, wifi6.. (not sure on the battery)
but.. yeah... you probably pay a bit more with the best battery but it's nearly the same as a router.


Also: repeating summarized things as we're on a new page so people can read this hopefully..
It is zen2,
it has 8mb total L3 vs 32mb on desktop
no pci-e 4.0
it has no chipset, it is a SOC (ALL INCLUDED!!!!)
it is 35 W not 45W.
it is using turbo of 60 W~ before going to 35 W for steady state. (intel uses turbo like this too)
Posted on Reply
#56
ARF
QUANTUMPHYSICSThe vast majority of buyers will still aim for intel and Nvidia.
There is really little to no sense whatsoever to do so.

The right path is AMD Ryzen + AMD Radeon..
TheLostSwedeIt has 8 or 16GB soldered down, plus a slot, so it's not single channel.
As for the camera, it's an "up-sell" to this:
www.asus.com/ROG-Republic-Of-Gamers/ROG-Eye/
Configuration 8 GB soldered + 16 GB on the slot, 24 GB - will it be single or dual channel because of the different capacities?
Posted on Reply
#57
Rahnak
Imsochobosomehow 1400 $, 1920x1080 120 HZ, 4900HS, RTX 2060 max-q, 1tb 660p, 16gb ram, wifi6.. (not sure on the battery)
but.. yeah... you probably pay a bit more with the best battery but it's nearly the same as a router.
R0H1T was talking about the ROG Eye webcam, not the laptop.
Posted on Reply
#58
ARF
AMD Ryzen 9 4900HS vs Intel 5.3 GHz Core i9-10980HK:

Posted on Reply
#59
SL2
TheLostSwedeIt has 8 or 16GB soldered down, plus a slot, so it's not single channel.
I didn't say it was. The reviews clearly shows that it's only one slot.

Posted on Reply
#61
R0H1T
RahnakR0H1T was talking about the ROG Eye webcam, not the laptop.
Yes, but I guess not everyone needs a webcam on their laptop like ~ well me. I barely use it though that's largely due to the sub 720p resolution & basically no practical use.
Posted on Reply
#62
TheLostSwede
News Editor
MatsI didn't say it was. The reviews clearly shows that it's only one slot.

Those timings though... :(:wtf:
Posted on Reply
#63
ARF
TheLostSwedeThose timings though... :(:wtf:
Low-end, cheap memory modules.
Posted on Reply
#64
IceShroom
TheLostSwedeThose timings though... :(:wtf:
Standard JEDEC timing.
Posted on Reply
#65
EarthDog
IceShroomStandard JEDEC timing.
Yep... that new 3200 MHz spec that came out last year... ick...
Posted on Reply
#66
ARF
Normally, the modules are CAS16 DDR4-3200, and most of the available sticks on the market are such..

Posted on Reply
#67
John Naylor
Won't pay attention to performance benchmarks that have no relevance to application performance. As with 3900x vs 9900KF, far more often that not, application performance on apps most folks use every day see no benefit from these "press release" performance benchmarks.

Does anyone sit around waiting to choose a CPU until 7-Zip performance benchmarks come out ? Why No Premier, PhotoShop, AutoCAD, office Suoites , gaming ? 3900x reportedly kicked 9900k to the cub and yet the 9900KF won in gaming, traded insignificant wins in Office apps, won in Photoshop and Premier and most important, won in our primary app AutoCAD. Now certainly if our big apps were the ones that the 3900X excelled in like brain simulation, 7-Zip, virtual machines and game develpment, rendering / animation.... the 3900X is the obvious
of course better choice ... but anyone who chooses a CPU on the bases of performance benchmarks in apps they don't use on a frequent basis isn't doing it right.
Posted on Reply
#68
SL2
TheLostSwedeThose timings though... :(:wtf:
Yeah, call me old school but I prefer dual slots. ;)
Posted on Reply
#69
ARF
John NaylorWon't pay attention to performance benchmarks that have no relevance to application performance. As with 3900x vs 9900KF, far more often that not, application performance on apps most folks use every day see no benefit from these "press release" performance benchmarks.

Does anyone sit around waiting to choose a CPU until 7-Zip performance benchmarks come out ? Why No Premier, PhotoShop, AutoCAD, office Suoites , gaming ? 3900x reportedly kicked 9900k to the cub and yet the 9900KF won in gaming, traded insignificant wins in Office apps, won in Photoshop and Premier and most important, won in our primary app AutoCAD. Now certainly if our big apps were the ones that the 3900X excelled in like brain simulation, 7-Zip, virtual machines and game develpment, rendering / animation.... the 3900X is the obvious
of course better choice ... but anyone who chooses a CPU on the bases of performance benchmarks in apps they don't use on a frequent basis isn't doing it right.
Here:



Here:


translate.google.com/translate?sl=auto&tl=en&u=https%3A%2F%2Fwww.sweclockers.com%2Ftest%2F29309-asus-rog-zephyrus-g14-ryzen-9-4900hs-och-geforce-rtx-2060-max-q%2F5%23content
Posted on Reply
#70
Darmok N Jalad
MatsI didn't say it was. The reviews clearly shows that it's only one slot.

What an odd configuration. I’d think they’d either go all soldered or all slot. Usually one slot sits under the keyboard, but not this one.
Posted on Reply
#71
SL2
John NaylorWon't pay attention to performance benchmarks that have no relevance to application performance. As with 3900x vs 9900KF, far more often that not, application performance on apps most folks use every day see no benefit from these "press release" performance benchmarks.

Does anyone sit around waiting to choose a CPU until 7-Zip performance benchmarks come out ? Why No Premier, PhotoShop, AutoCAD, office Suoites , gaming ? 3900x reportedly kicked 9900k to the cub and yet the 9900KF won in gaming, traded insignificant wins in Office apps, won in Photoshop and Premier and most important, won in our primary app AutoCAD. Now certainly if our big apps were the ones that the 3900X excelled in like brain simulation, 7-Zip, virtual machines and game develpment, rendering / animation.... the 3900X is the obvious
of course better choice ... but anyone who chooses a CPU on the bases of performance benchmarks in apps they don't use on a frequent basis isn't doing it right.
That was a bit different situation. The 3900X is faster thanks to having 8 more threads, and the 9900KF is faster because of higher clock speeds, lower latency.

This time around they have the same amount of threads, and Intel can't use their hot running Mhz advantage because it's mobile. At 35 W, that is.

The GE65 and the Aero 15 are both >2.2 kg, the G14 is 1.6 kg.




ARFEveryone excited but MSI will come and say :D

Excuse us, guys, but we don't work with AMD APUs.
I'm not so sure about that.
www.msi.com/Laptop/Alpha-15-A3DX/Specification

Posted on Reply
#72
Totally
Imagine when this chip makes it into a less constrained chassis, 'tiger with wings' anyone?
Posted on Reply
#73
B-Real
XuperBattery life among all reviews :

www.kitguru.net/wp-content/uploads/2020/03/Battery-Page.png
images.hothardware.com/contentimages/article/2954/content/battery-life-asus-rog-zephyrus-g14.png
cdn.mos.cms.futurecdn.net/6hPiit6d3p4iC6LXTaxxDU-650-80.png
hexus.net/media/uploaded/2020/3/8ec7191f-c434-4869-85ac-3dc3e82b8969.png
hexus.net/media/uploaded/2020/3/d0bc753d-ec5d-43ab-a3a5-b3132f9d588b.png
hexus.net/media/uploaded/2020/3/44303c92-ad06-460c-ba71-d3bf80773e23.png
hexus.net/media/uploaded/2020/3/e65b0ac4-0567-422d-b19c-fea57d21156a.png
images.idgesg.net/images/article/2020/03/asus_rog_zephyrus_g14_video_run_down-100836710-orig.jpg
This with the given performance. It is incredible, really.
John NaylorWon't pay attention to performance benchmarks that have no relevance to application performance. As with 3900x vs 9900KF, far more often that not, application performance on apps most folks use every day see no benefit from these "press release" performance benchmarks.

Does anyone sit around waiting to choose a CPU until 7-Zip performance benchmarks come out ? Why No Premier, PhotoShop, AutoCAD, office Suoites , gaming ? 3900x reportedly kicked 9900k to the cub and yet the 9900KF won in gaming, traded insignificant wins in Office apps, won in Photoshop and Premier and most important, won in our primary app AutoCAD. Now certainly if our big apps were the ones that the 3900X excelled in like brain simulation, 7-Zip, virtual machines and game develpment, rendering / animation.... the 3900X is the obvious
of course better choice ... but anyone who chooses a CPU on the bases of performance benchmarks in apps they don't use on a frequent basis isn't doing it right.
This CPU just crushes the top Intel mobile CPUs and you are still ranting. :D
Posted on Reply
#74
Cheeseball
Not a Potato
ARFNo webcam? The CPU can become hot?!

Why did they make this 14" and how will you game on it? Using a proper monitor?

Just wow!
An integrated webcam would be useless when a streamer/collaborator would just use a higher-quality USB 3.0 device that you can freely position.

14" is a good portable size. I've been waiting for more powerful, compact solutions, similar to the Razer Blade Stealth with a GTX 1650.

The panel in this is a 144 Hz IPS, albeit slower response time. May not be used for competition, but it's still manageable for on-the-road gaming.
MatsThat was a bit different situation. The 3900X is faster thanks to having 8 more threads, and the 9900KF is faster because of higher clock speeds, lower latency.

This time around they have the same amount of threads, and Intel can't use their hot running Mhz advantage because it's mobile. At 35 W, that is.

The GE65 and the Aero 15 are both >2.2 kg, the G14 is 1.6 kg.
Looks like that Gigabyte Aero 15 is throttled badly. The GE65 does have good thermals (similar to the Acer Predator Helios 300 with the same specs) from what I've read about it.
Posted on Reply
#75
TheLostSwede
News Editor
Sadly the Zephyrus G14 is WAY too expensive in Sweden once you add the local markup and VAT. Was going to get one, but I can only afford the base config and I'm not interested in the GTX 1650 Max-Q. For some reason the Ryzen 5 4600H is not available with a better GPU option. Thanks Asus...
Posted on Reply
Add your own comment
Apr 24th, 2024 19:51 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts