Yeah yeah the M1 dominate...

Actually, I prefer less closed ecosystem and good performances all-round than "merciless crushing" in selective bench and scenari... (Just like when the M1 was shown against willow cove... But Intel also has that behaviour  ::) aka: twisting words and meaning and using specific cases)

Lol, citing a Youtube channel none other than the Apple-clown Max Tech was the most hilarious thing I've ever seen🤣 Is this the same joke of an author that wrote about the BS article about iOS being superior to Android?

My swiftX 5800u at 15w can score better than his "65w" 5900hx.

Abnormally low results for the 5900HX in CPU tests.
This is why no one takes YouTube videos seriously.

Quote from: hachi on September 04, 2021, 17:21:08

Quote from: Anonymousgg on September 04, 2021, 14:12:15
Those GPU results in particular are embarrassing. AMD has a chance to rectify this next year with Rembrandt, 12 CUs of RDNA2 instead of 8 Vega.

For what ? This cpu will pair with 3070 or 3070 in laptop. Increase graphic processing power with the expense of cpu power ? No thanks.

But they used this CPU in a mini PC. No discrete graphics. The 5800U which also has Vega 8 and is less likely to be paired with discrete graphics would do even worse against the Apple M1.

AMD obviously doesn't agree with you because we already know they are putting 12 RDNA2 CUs in top Rembrandt, and aren't adding any more CPU cores.

Quote from: Anonymousgg on September 04, 2021, 14:12:15
Those GPU results in particular are embarrassing. AMD has a chance to rectify this next year with Rembrandt, 12 CUs of RDNA2 instead of 8 Vega.

For what ? This cpu will pair with 3070 or 3070 in laptop. Increase graphic processing power with the expense of cpu power ? No thanks.

don't misunderstand my point bro. Even if my 5800H is beaten by M1 in some workload in term of finish time. My 5800H score 13k Cinebench R23

My point is x86's strength is about multitasking.
Need to run multi workload to stress out 5800H 100% & then you'll see the difference, M1 performance will suffer when it's being pushed hard.

@FlatWhite
I've done quite a lots of programming software benchmark between Ryzen 7 5800H vs M1,

Really? i made Cinebench R23 on 5800H and i can get more than 13000 points.

So why a stronger CPU like 5900 are only 10000 here?

lol, it's not about hardware, it's software issue.
Those benchmark doesn't fully utilize 100% CPU, I've done quite a lots of programming software benchmark between Ryzen 7 5800H vs M1, M1 is maxed out & finish faster while 580HH is only 40-60% Util.
The strength of x86 CPU is multitasking.

He's super biased on Apple. I mentioned it on his video, he even puts a picture at some point of a mac mini using 22 watts of power (from an anandtech post), but he says the mac mini uses 15 watts.

It's Max Tech.

He's very biased against non apple product. He'd use clickbait title and thumbnail to showcase how superior apple product is when it's known he put apple rivals in disadvantage situation.

Anyone shouldn't take anything he said seriously.

Something is definitely off with the CB23 score, this other review
www.youtube.com/watch?v=fHpBsPxixkg
gives it 12035; on the other hand the Geekbench V5 score is higher in the comparison video shown in this article.

The review is very biased. Test results for 5900HX are much lower than we have in other independent studies. For instance, the median CB R23 score for 5900HX is nearly 13000, but not 10700. The same story takes place with the Geekbench scores. The average scores for 5900HX are nearly 1500 and 8400-8600.

The other manipulation is power consumption. M1 chip consumes nearly 25W under full load, but not 15W.

It is very easy to manipulate the data if you want to show that one platform is better.

Those GPU results in particular are embarrassing. AMD has a chance to rectify this next year with Rembrandt, 12 CUs of RDNA2 instead of 8 Vega.

Cinebench score seems very weak for the 5900. Notebookcheck net have much higher values. Even a Ryzen 7 5800H a 45w piece is much faster than 10000.