M4-powered MacBook Pro flexes on Cinebench, crushing the Core Ultra 9 288V and Ryzen 9 AI 370 HX

[Redaktion]

Apple's M4 SoC is no slouch - we already knew that. However, Apple's 14-inch MacBook Pro allows the M4 to truly flex its muscles with active cooling, and the results are outrageous, crushing its Intel, AMD, and Qualcomm counterparts by as much as 62%.

https://www.notebookcheck.net/M4-powered-MacBook-Pro-flexes-on-Cinebench-crushing-the-Core-Ultra-9-288V-and-Ryzen-9-AI-370-HX.899722.0.html

[Johan]

From Notebook Check's own review you show the 35W configuration of the HX 370 exceed this multi-core score.

Truly bipolar reporting here...

[dada_dave]

From Notebook Check's own review you show the 35W configuration of the HX 370 exceed this multi-core score.

Truly bipolar reporting here...

They wrote:

AMD's Ryzen 9 AI 370 HX, despite being more of a competitor for the M4 Pro, falls behind the M4 by a whopping 33% in single-core performance, while being almost neck and neck in multi-core despite its much more generous power envelope.

From the HX 370 article you reference, in multi-core the 35W TDP HX 370 actually draws 57W during the Cinebench run and scores 1022. This is only 5% more than the score the M4 is reported to get, hence the term "neck and neck" to describe their scores. The M3 Air only drew about 21W and while we don't know the power draw of the M4 in the Macbook Pro it is unlikely to be significantly higher, at least relative to 57W, hence why they noted the HX 370's "more generous power envelope" to achieve that score. While occasionally notebookcheck, like every site, is prone to fits of exaggeration, in this instance they are, if anything, understating.

[Drew9]

CPU's are boring.

[Johan]

From Notebook Check's own review you show the 35W configuration of the HX 370 exceed this multi-core score.

Truly bipolar reporting here...

They wrote:

AMD's Ryzen 9 AI 370 HX, despite being more of a competitor for the M4 Pro, falls behind the M4 by a whopping 33% in single-core performance, while being almost neck and neck in multi-core despite its much more generous power envelope.

From the HX 370 article you reference, in multi-core the 35W TDP HX 370 actually draws 57W during the Cinebench run and scores 1022. This is only 5% more than the score the M4 is reported to get, hence the term "neck and neck" to describe their scores. The M3 Air only drew about 21W and while we don't know the power draw of the M4 in the Macbook Pro it is unlikely to be significantly higher, at least relative to 57W, hence why they noted the HX 370's "more generous power envelope" to achieve that score. While occasionally notebookcheck, like every site, is prone to fits of exaggeration, in this instance they are, if anything, understating.

No, they're quite wrong. M4 scores lower than the HX 370 in Cinebench 2024. Their own testing shows this. Yes, M4 scores better for its power draw but that isn't the point here. The author repeated a claim from a tweet on X without checking their own results. The author didn't even order the name of the HX 370 correctly in the title. There's no need to defend sloppy reporting with inaccurate headlines simply because the M4 is good. Lying about results isn't good regardless.

[missingxtension]

I thought I was on a mac news site, I didn't realize it was notebook check until I read the comments. This reporting is right from one of those 9-5 apple sites. The first I saw was that we don't know if this is actually retail silicon or cherry ⛏️'d batches running pre release firmware. Either way, its just a bad article. Take it with a grain of salt

[dada_dave]

From Notebook Check's own review you show the 35W configuration of the HX 370 exceed this multi-core score.

Truly bipolar reporting here...

They wrote:

AMD's Ryzen 9 AI 370 HX, despite being more of a competitor for the M4 Pro, falls behind the M4 by a whopping 33% in single-core performance, while being almost neck and neck in multi-core despite its much more generous power envelope.

From the HX 370 article you reference, in multi-core the 35W TDP HX 370 actually draws 57W during the Cinebench run and scores 1022. This is only 5% more than the score the M4 is reported to get, hence the term "neck and neck" to describe their scores. The M3 Air only drew about 21W and while we don't know the power draw of the M4 in the Macbook Pro it is unlikely to be significantly higher, at least relative to 57W, hence why they noted the HX 370's "more generous power envelope" to achieve that score. While occasionally notebookcheck, like every site, is prone to fits of exaggeration, in this instance they are, if anything, understating.

No, they're quite wrong. M4 scores lower than the HX 370 in Cinebench 2024. Their own testing shows this. Yes, M4 scores better for its power draw but that isn't the point here. The author repeated a claim from a tweet on X without checking their own results. The author didn't even order the name of the HX 370 correctly in the title. There's no need to defend sloppy reporting with inaccurate headlines simply because the M4 is good. Lying about results isn't good regardless.

Typos notwithstanding, they aren't lying and aren't wrong. I quoted you the results from the very article you referenced and explained why they said what they said even including your own target of the "35W TDP" (57W) score. Yes, the HX 370 can be pushed far past that and score even higher at the expense of even more power ... hence why they also said in the article that the HX 370 is truly "an M4 Pro competitor" rather than a competitor for the base M4. This is also why the author exudes excitement for what the M4 Pro and Max might bring to the table.

The reason for the slightly clickbait headline isn't that the M4 is merely scoring better at its power draw of ~20W, but rather that it takes the HX 370 until 57W to match a likely 20W processor meant primarily for fan-less devices and with a smaller core count, most of which are E cores - tiny by comparison to even the "Zen5c" cores (less than half the size). YMMV but I think calling that a crushing result in the context is quite defensible, so did the author apparently, not only because of the result itself but what it means for the M4 family of chips.

And that's not even accounting for the single thread results which, while again, we don't know power draw here (Geerkerwan's testing obviously didn't include CB R24 and primarily used software-measured power rather than wall power so not directly comparable to Notebook Check's results), Apple cores typically use 10-14W of wall power in ST CB R24. Maybe the M4 Macs will use more, who knows? But the AMD chip is using at least double that, around 30W in Notebook Check's testing and getting a 33% lower score. What word would you describe for that? If not "crushing", then at least "impressive" perhaps?

Now, CB R24 being primarily an endurance test as well as primarily a FP test means that Apple does rather better in terms of CB R24 scores than in other benchmarks. Also, from what I can tell, AMD does rather worse in CB R24 compared to other benchmarks, so from purely a performance standpoint this a worst case scenario for the AMD chip vs Apple.

I thought I was on a mac news site, I didn't realize it was notebook check until I read the comments. This reporting is right from one of those 9-5 apple sites. The first I saw was that we don't know if this is actually retail silicon or cherry ⛏️'d batches running pre release firmware. Either way, its just a bad article. Take it with a grain of salt

If real and they do appear to be real, these are almost certainly stolen devices - so definitely not cherry picked. There's reportedly about 200 being sold on social media sites in Russia. That said, given the provenance, or lack thereof, yes the results should be suspect and of course even if accurate Notebook Check's own results with power draw will be much more informative. And of course there will be variance between devices and even run-to-run.

[Sambit Saha]

I understand this might ruffle a few feathers, but the M4 is indeed running circles around the competition and I do not feel the need to say otherwise.

In single-core, the M4 is ahead of the HX 370 and 288V by 51% and 42% respectively.

In multi-core, the M4 outpaces the 288V by almost 70%. The considerably more power-hungry HX 370 (remember that it can be found with a wide variety of power configs, some north of 50 W) is almost neck and neck, as our average score for the HX 370 is 994 (M4 gets 971).

Therefore, I stand by what I wrote.

[Worgarthe]

I understand this might ruffle a few feathers, but the M4 is indeed running circles around the competition...

...if you need MacOS. If you need Windows though...

[debatable]

In multi-core, the M4 outpaces the 288V by almost 70%.

you're comparing a higher end cpu with intel's thin and light offering, that isn't a fair comparison. The M4 Pro should be compared with Arrow Lake.

[debatable]

Let's take time to remember that M-series MacBook Pro's consume up to to 76 watts when under sustained load which includes benchmarks, gaming, and high-end applications, so M4 in the MBP will hardly be different. From NBC's review of the M3P in the MBP:

During our stress test, we did note over 70 watts, meaning the supplied 70-watt power supply can't deliver the necessary power. If you put both components under load permanently, then it will slowly drain the battery. During our one-hour stress test, we noted 3 per cent. Using a more powerful power supply, we noted a maximum consumption of 80 watts which then balanced out at around 76 watts.

[Sambit Saha]

@debatable

you're comparing a higher end cpu with intel's thin and light offering, that isn't a fair comparison. The M4 Pro should be compared with Arrow Lake.

Let's take time to remember that M-series MacBook Pro's consume up to to 76 watts when under sustained load which includes benchmarks, gaming, and high-end applications, so M4 in the MBP will hardly be different. From NBC's review of the M3P in the MBP:

This article is about the M4. Not the M4 Pro.

G'day :)

[Exotic]

@debatable

Let's take time to remember that M-series MacBook Pro's consume up to to 76 watts when under sustained load which includes benchmarks, gaming, and high-end applications, so M4 in the MBP will hardly be different. From NBC's review of the M3P in the MBP:

This score is from the M4, not M4 Pro. The base M4 replaces the base M3 which has max power load of 37 watts and the M3 Pro has a max load of 73 watts as per NBC review. There is a big difference between the base and Pro SKU
The M4 Pro was not tested here, it was the base M4. The 14" MBP currently comes with 3 SKUs, base M3, M3 Pro and M3 Max later and this month there will be three SKUs of based on M4 architecture: base M4, M4 Pro and M4 Max,  all three will go into Macbook Pros.

In multi-core, the M4 outpaces the 288V by almost 70%.

you're comparing a higher end cpu with intel's thin and light offering, that isn't a fair comparison. The M4 Pro should be compared with Arrow Lake.

The article is talking the base M4 scores which will compete with Lunar Lake. The M4 Pro chip will compete with Arrow Lake H and M4 Max will compete with Arrow lake HX.

So it is a fair comparison as the base M4 is also Apple's thin and light offering. The base M4 also goes in iPad Pros.

[David Alsh]

Big for nothing because it's nerfed by MacOS.

Really wish Apple would offer first class support for Linux.