Ryzen 7 5800U thin and light laptops should offer stellar single-core performance as AMD's Zen 3 APU scores higher than an Intel Core i9-10900K in CPU-Z

[Redaktion]

An AMD Ryzen 7 5800U has been delivering some curious results thanks to recently leaked CPU-Z and Cinebench benchmarks. Single-core and single-thread scores were incredible, to the point that the laptop Ryzen 7 5800U could even outscore a desktop Intel Core i9-10900K. However, there seemed to be an issue with the Cezanne APU's multi-core performances.

https://www.notebookcheck.net/Ryzen-7-5800U-thin-and-light-laptops-should-offer-stellar-single-core-performance-as-AMD-s-Zen-3-APU-scores-higher-than-an-Intel-Core-i9-10900K-in-CPU-Z.514047.0.html

[Hoisin]

Intel needs more Netburst and lower security standards to compete.

[Make it Stop]

If only there were a class of mobile processors from Intel that you could have used for comparison against these upcoming AMD mobile processors... hell, there are even leaked Cinebench scores for the i9-11900 you could have used if you wanted to compare same-generation laptop and desktop parts.

[smh]

Intel can actually make a competing chip. They only have to do one thing.
Add freaking 4 more cores to their laptop chips. They already have the best single threaded performance in their laptop chips, and their iGPUs are the best as well. Now, they need to just add 4 more cores. smh.

[havefun]

The Ryzen 9 5900H amassed a single-core score of 584 points and a strong multi-core score of 5,264 points.

i dont believe 5800U could be better.

[Irish_Adam]

Intel can actually make a competing chip. They only have to do one thing.
Add freaking 4 more cores to their laptop chips. They already have the best single threaded performance in their laptop chips, and their iGPUs are the best as well. Now, they need to just add 4 more cores. smh.

That's not really true, their 4 core 10nm uses twice the power per core that AMD uses in their CPU's. Intel can't bolt on 4 more cores and clock it the same while using the same power.

This along with horrible yields are the reasons why we have not seen an 8 core 10nm CPU from Intel, even though they have been promising one for ages.

[_MT_]

Intel can actually make a competing chip. They only have to do one thing.
Add freaking 4 more cores to their laptop chips. They already have the best single threaded performance in their laptop chips, and their iGPUs are the best as well. Now, they need to just add 4 more cores. smh.

While it's certainly possible, who would it be for? Who exactly can utilize eight cores? U series, that's primarily ultrabooks. Managers running Outlook and Excel probably won't utilize eight cores. The biggest benefit is for oddballs like Lenovo using ULV chips where they really shouldn't.

I think improving efficiency (and therefore real world performance) of higher end H series processors is much more important than introduction of eight core U series. Also, fitting eight cores plus a powerful iGPU will take quite a bit of space (H series will ship with a small iGPU). They've got yield issues, they've got capacity issues. At this point, most of their mobile processors are still not on a 10 nm node. And according to reports, it will never have the same productivity as their 14 nm nodes did. Throwing silicon out of the window where the market might not need it wouldn't be a very good strategy.

[_MT_]

That's not really true, their 4 core 10nm uses twice the power per core that AMD uses in their CPU's. Intel can't bolt on 4 more cores and clock it the same while using the same power.

The trick is that the relationship between frequency (and therefore performance) and power is not linear. In multi-core loads that scale well, an eight core processor should be significantly more powerful than a four core processor (same architecture, same node) within the same power envelope. Yes, the cores would be running at lower frequencies. But when you cut power in half, frequency doesn't go down to a half (not to mention that at lower frequencies, you waste fewer cycles by waiting). Above about 2 GHz, the higher the frequency, the lower the efficiency. The four core processors consume so much power mainly because they can. Not because it's smart. They're operating far from their sweet spot. If you actually wanted to compare efficiencies of the respective cores, you'd have to control frequency and map the curve.

The practical problem is that most workloads don't scale well across eight cores. While it's possible to do and we can use benchmarks like Cinebench to compare, the benefit is limited. I belong to the few that actually benefit from it. I would take even 16 cores in a laptop. But most people are not like me. And if you need that many cores, you can probably do with an H series processor. What AMD has done, has been nice because now, ultrabooks are powerful enough for me to do some work on them without having to offload everything to a server. I might be able to ditch one laptop. But I have those facilities in place. I have been using them for over ten years. Another problem is that the selection of laptops with AMD processors is quite limited and even if you manage to find something you like, availability is terrible, at least in my part of the world.

[john mon]

The Ryzen 9 5900H amassed a single-core score of 584 points and a strong multi-core score of 5,264 points.

i dont believe 5800U could be better.

it could, depending how much juice and cooling the OEM gives it.

I played with a lenovo 4500u laptop, and it had high performance mode where it would supply 25w, and it would hold turbo all cores indefinitely.
4ghz all 6 cores

[smeegle]

apple m1 should be part of any discussion about relative laptop cpu performance at this point.  does this even come close?  if you really want apples to apples (and to stick it to apple), you could even compare the m1 running in a windows vm or emulating x86, but it probably still comes out on top.

[Astar]

Intel can actually make a competing chip. They only have to do one thing.
Add freaking 4 more cores to their laptop chips. They already have the best single threaded performance in their laptop chips, and their iGPUs are the best as well. Now, they need to just add 4 more cores. smh.

That's complete idiocy. What is wrong with you Intel & CrApple fangirls. Learn some basics of semiconductor design & performance. You don't just add more cores like a Lego toy.

Higher heat reduces silicon performance. You think adding 4 more cores has no heat penalty on the entire SOC and motherboard? There is a reason why EPYC cores are clocked lower than Threadripper cores which are in turn clocked lower than Zen desktop ones.

Like what some others have already said. Anybody can make very low power consumption chips that seem to offer great performance per watt numbers. But its AMD's ability to do that as well as scale up the power consumption to EPYC's high core count TDPs that made it such a media darling. What AMD has managed to do for the entire x86/x64 stack from 4 to 128 cores is what has been revolutionary. There is a reason why AMD, with Ryzen has gone from $9 to $90 in share price. Tsk tsk...

[Astar]

Intel can actually make a competing chip. They only have to do one thing.
Add freaking 4 more cores to their laptop chips. They already have the best single threaded performance in their laptop chips, and their iGPUs are the best as well. Now, they need to just add 4 more cores. smh.

And to add to that, Intel has 8 core chips as well but their current state of the art consumes so much more power per core in order to pretend to try to stay competitive on AMD's coat tails, that it is simply not viable anymore. All they do is overclock and overvolt.

And get your damn facts right. Intel has lost the single-thread crown long ago to AMD.

I've said elsewhere that they should just use Intel chips in the next KFC-console Gen 2 that can COOK, instead of just warm fried chicken.

[Astar]

apple m1 should be part of any discussion about relative laptop cpu performance at this point.  does this even come close?  if you really want apples to apples (and to stick it to apple), you could even compare the m1 running in a windows vm or emulating x86, but it probably still comes out on top.

Why do CrApple fangirls keep wanting to interject in every AMD thread?
CrApple whatever M1 chip is only strong in synthetic cherry picked benchmarking software tests.

The so called benchmarks are all cherry picked by CrApple. Even Anandtech is an Apple lackey that will pander to anybody with a big marketing budget.

Also a lot of real world laptop performance reviews have already shown that AMD Zen 2 (with Zen 3 on the way) performance still beats the crap out of CrApple's M1, at a lower price point and running at cool temperatures.

See the review of the best Windows laptop featuring Zen 2 Renoir, beating the hell out of M1 and everything else out there:
https://youtu.be/-VLbbw7FuZw

I won't care to explain again to fools like you just gushing about CrApple's marketing brochure claims. You will never get understand that CrApple marketing hype uses cherry picked benchmarking scores. I'll just cut and paste my other posts for others to read regarding this point since I know its beyond your puny intellect:

-------------------
EEtimes and PC Perspectives, among many others have stated time and again how Geekbench is basically useless to a lot of well-informed folks.


As per https://www.eetimes.com/new-geekbench-5-improves-tests-but-adds-bias/ ---



"In a review of some of the results posted for the same platforms on both Geekbench 4 and Geekbench 5, we found that the change in performance numbers was higher on the Android OS and AMD processor than on Apple's operating systems or Intel processors (the table provides a sample of these results). The Apple OSs and Intel processors appear to have benefitted most from the benchmark changes, much of which is likely attributed to the elimination of the memory and battery test that benefit the AMD processors and Android OS."


"PC Perspective did a similar comparison both versions of the benchmark on a pair Intel and AMD-based PCs and the differences are rather drastic — showing a much higher relative performance of the Intel-based platforms than the AMD platforms."

--------------------

[kek]

no such thing as superior or not, any s perfect