Also don't forget that it includes an AI chip, which might come in handy sooner or later (who knows).
Personally waiting for Minisforum to use this chip and buy one as soon as it's out.

QuoteOn the GPU side, the Radeon 780M inside the APU scored a rating of 44.3% which is roughly on par with a desktop RX 570 and slightly behind the GTX 1060 Mobile (46.7%).

Sounds too good to be true. Especially since the recent Timespy Leak showed 3000 points (780m) vs 3700 points (gtx1060).

I have a feeling that the user-benchmark-score profits from dual-issue, and thus it can calculate some theoretical s*** (such as FP32 performance) faster and come close to the gtx1060. However it's a whole other story when it comes to gaming. We don't have any comparison to that yet, so I'm still not sure how it will turn out...

for crying out loud, stop giving that website free advertising. it's just stupid to make such articles based on results taken from there.

why are you dragging your reputation through the mud? for a few extra clicks?

The more outlandish userbenchmark's bias becomes can be used as a reliable indicator of how well AMD is performing vs Intel, i.e., userbenchmark's bias is by itself a kind of benchmark.

Quoteit is important to mention that UserBenchmark is a contentious platform and accused by many of favoring Intel parts over AMD.

The linked 7940HS result literally calls it "AMD Advanced Marketing Devices 7940HS" and "AMD Advanced Marketing Devices 780M". So it's a safe bet that UserCringemark is not to be taken seriously.

I'm a tad confused.
How can it be rated higher than 5950x but 6% behind desktop 7600?

Me thinks, there's a bit of a loss of information/context.

7940HS is an 8c/16th CPU.
There's no way it can be more powerful than 5950x in real world software under multi-threaded operations (especially with 55W TDP).
I think the score is skewed because of AVX-512 in that regard.
6% lower single threaded score than desktop 7600 makes sense though.

The iGP being on par with RX 570... that might be possible.

But, UserBenchmark is a joke and not something that should EVER be used to compare hardware.

I'd rather wait for third party reviews where this APU is put through its paces under real life conditions and software.