Only 28% difference on Graphics scale? Damn Intel caught up nicely on GPU performance.

sverek:

Only 28% difference on Graphics scale? Damn Intel caught up nicely on GPU performance.

Well, it is the 15 watt part. Can't be too big of a GPU.

sverek:

Only 28% difference on Graphics scale? Damn Intel caught up nicely on GPU performance.

Is this a joke? A 30% difference for a 15W TDP part is massive.

Toss3, as you said, in a very limited TDP as 15W for both CPU and GPU, 30% above an already good performance GPU on the Intel, is massive. Intel had bad drivers, AMD and nvidia have excellent drivers, so games should run at a i7 + MX250 level but at 40% of the TDP and less size.

Hilbert Hagedoorn:

At geekbench a Lenovo notebook embedded with Ryzen 7 4800U 8-core/16-thread processor has been spotted. That's renoir based on ZEN2 cores.... AMD Renoir Ryzen 7 4800U Gets Geekbenched

In the screencap, it states 4700U. Are we sure its 4800U and not 4700U ? edit : Also, the memory bandwidth makes it look very likely that it was in Single channel mode, with only one Memory stick. (~20GB/s) Single channel DDR4-2400/2666

rmcrys:

Toss3, as you said, in a very limited TDP as 15W for both CPU and GPU, 30% above an already good performance GPU on the Intel, is massive. Intel had bad drivers, AMD and nvidia have excellent drivers, so games should run at a i7 + MX250 level but at 40% of the TDP and less size.

They only tested 3DMark Timespy which isn't the best case scenario for the 4700u. Also not sure why you would assume Intel has bad drivers?

Now if only they'd stop using the next gen model number range for current gen APUs.

Toss3:

They only tested 3DMark Timespy which isn't the best case scenario for the 4700u. Also not sure why you would assume Intel has bad drivers?

AMD and nvidia drivers are very frequently optimised for released games, not to mention those who are optimised for AMD/ nvidia; also AMD GPU are the base from the present and future PS/XBOX which means developers know the hardware. Intel just happens to be a CPU maker who has a generic GPU which happens to be rather ok on the 10th (!!!) Gen CPU and only on the G4/G7 versions. Everything else (G1, UHD, HD ...) are just crap, as a lot of games just don't recognize the GPU as a supported one. So, until Intel GPUs can compete for more generations at least with crappiest AMD/ nvidia chips and get developers to optimize for their hardware, they will be always the 3rd and last option. And worse: the G4 GPU version should be standard but somehow Intel figured that the 14nm "10th gen" and most 10nm should have G1/UHD graphics, so perhaps 95% of their 10th gen chips won't have decent GPU. Will a developer optimize games for ... 5% of Intel sold chips? Which accounts perhaps for 0.01% of all GPUs in 2020-2021?

sverek:

Only 28% difference on Graphics scale? Damn Intel caught up nicely on GPU performance.

You do realise that we're talking about 28-30% GPU difference in a 15W TDP for one thing, on a APU which offers TWICE the amount of CPU cores (8 cores and 16 threads in 15W TDP, vs Intel's 4 cores and 8 threads). Honestly, this is a huge win for AMD. And the fact that its using Vega iGP with improved core performance (and LESS cores than previous APU's says a lot).

Kool64:

Now if only they'd stop using the next gen model number range for current gen APUs.

Yes, I was a bit disappointed with that myself, however, those Vega iGP's did receive pretty much a similar treatment in various engines, etc. as Navi did. Plus those Vega cores are MUCH faster than the previous generation (while at the same time, these new APU's have less CU's). Zen 3 however should be using Navi for iGP in APU's... I'm just not sure which Navi (RDNA 1 or RDNA 2).

Evildead666:

In the screencap, it states 4700U. Are we sure its 4800U and not 4700U ? edit : Also, the memory bandwidth makes it look very likely that it was in Single channel mode, with only one Memory stick. (~20GB/s) Single channel DDR4-2400/2666

Good catch. I noticed the Geekbench screen saying 4700U too. So, either Geekbench is reading the APU wrong, or this article made a mistake - its possible the article made a mistake indeed - but we'll need to see.

Tomshardware is also posting the same results and using 4700U... so I think this article is wrong (that is, if Geekbench is reporting the correct APU). I suggest a wait and see approach to determine whether this is indeed the 4700U APU. If it is, then the multithreaded score would need to be much higher (and single threaded too would be at least same or slightly higher).