We had to wait quite a while but we finally got our hands on AMD's latest mobile processors from the Ryzen 6000U generation. We compare the Zen3+ cores of the new Ryzen 7 6800U with Intel's current Alder Lake-P platform as well as Apple's M1 Pro SoCs. Who builds the best chip for compact laptops?
https://www.notebookcheck.net/AMD-Ryzen-7-6800U-Efficiency-Review-Zen3-beats-Intel-Alder-Lake.623763.0.html
This is bogus.. everyone knows last 5% perf takes 20-30% more power. Power efficiency tests should be done when we have tests doing exactly same work... the faster processor will idle faster and save battery... notebook check needs to send their author to some computer science 101 class
Quote from: Barebooh on May 31, 2022, 20:12:03
Absolutely f**k-ugly unit. Taiwanese skip art classes like Hitler. Ridiculous chip (needs MOAR CORES!!!!1!) ... Scorching wattage. "the power of sun, in the palm of my hand...
somebody please tell me what the hell is going on!
What's going on is you're an imbecile LOL ;D
Who is the genius that decided to compare P (Performance) segment laptops to U (Ultramobile) in efficiency tests? In completely different types of laptops. And then you just brush aside the performance differences to focus on efficiency, despite there being a 20% performance difference between the 6800u and 1260p in ST.
No wonder why notebookcheck is no longer a trusted source and has been replaced by other outlets.
to Rav4
I'm not sure why maximum performance is relevant in an efficiency comparison, especially since most of the processors mentioned found its way into small notebooks for which efficiency is important. How is that relevant?
It is specified which benchmarks are used and whether they run natively.
Efficiency can be specified as the energy used per output. In this statement, time is not a present unit at all.
Example: car: 5 liters/100 km; Sports car: 10 litres/100 km
Your argument: the sports car is there faster, so it consumes less.
Please actually read the article.
Another fantastic article.
"Intel's hybrid architecture with the 10 nm cores is not really suited for very compact systems"
Biggest reason why I can't honestly recommend Intel 12th Gen P-series laptops to anyone - terrible battery life and arbitrary fan noise, just look at Yoga 9i 14. For some reason, unlike on performance laptops OEMs refuse to put cTDP tweaking and fan control software into ultrabooks.
"AMD Ryzen 6000U chips are the best mobile processors for Windows laptops right now"
Problem is 6800U starts at ~$1300 with Asus (ZenBook S) and goes up to $2000+ in premium models from Lenovo (Thinkpad) and HP (EliteBook). Most people won't find that price justifiable when other U-series laptops usually start at $600 price point.
Quote from: SayQuantum1MoreTime on June 01, 2022, 01:07:57
to Rav4
I'm not sure why maximum performance is relevant in an efficiency comparison, especially since most of the processors mentioned found its way into small notebooks for which efficiency is important. How is that relevant?
It is specified which benchmarks are used and whether they run natively.
Efficiency can be specified as the energy used per output. In this statement, time is not a present unit at all.
Example: car: 5 liters/100 km; Sports car: 10 litres/100 km
Your argument: the sports car is there faster, so it consumes less.
Please actually read the article.
It's important because someone looking into 6800U is probably a type of person who needs to perform heavy multi-threaded CPU work for long period of time on battery power. When unplugged, chips eventually throttle during sustained load. So measuring how much maximum horsepower the chips can output at a gradually lower wattage than advertised TDP is most relevant to potential consumers.
The main problem is that the AMD completely failed the supply of chips in 2021. There were practically no laptops from 5600U/5800U in grade 15.6 (and this is the mainstream of sales). It is enough to introduce 5600U/5800U in Google and you will not find a single ponocent model (with a full -sized keyboard) except HP Probook for all 2021. In fact, AMD in 2021 deceived the expectations of consumers around the world. On the shelves for 1 laptop with Zen3, there were generally 10 models with Intel. It is not clear to me which idiots buy the AMD shares and support their price overplined once every 5. After all, in reality, almost the entire market of mobile chips control Intel and almost all the profit of the world has...
Quote from: Dylan on June 01, 2022, 00:17:17
Who is the genius that decided to compare P (Performance) segment laptops to U (Ultramobile) in efficiency tests? In completely different types of laptops. And then you just brush aside the performance differences to focus on efficiency, despite there being a 20% performance difference between the 6800u and 1260p in ST.
No wonder why notebookcheck is no longer a trusted source and has been replaced by other outlets.
Because this is how Intel is selling itsP series in a way?
Almost ALL ultrabooks that used to be equipped with 11th Intel U series have switched to the P series this year. It's also the fault of Intel who marketed the P series as a 28W TDP processor when in reality this processor can only express itself at higher wattage.
The 11th gen U was marketed at 15-28W and a lot of ultrabooks were having the 1165 or 1185 configured at 25-28W so the P series was on paper a good successor.
2Zrat:
Most people will still be forced to buy laptops with Intel, for the reasons I have voiced below. AMD does not have its own factories, and TSMC is not available for them in volumes comparable to Intel - they are "second-class people" there compared to Apple, Qualcomm/Mediatek, and now, since 2022 and Intel. All of them are in the first stage for conveyors. AMD, on the other hand, gets what remains of the industry's majors. AMD cannot get out of narrow lot without own factories and the highest priority on those of others. That is why its prospects are deplorable, and the value of shares is overestimated at times. What has become definitively clear since 2021.
The technological advantage becomes overwhelming when it is possible to supply such chips indefinitely. These are the harsh realities of the real world.
I, for once, want to say that I'm impressed with Apple's numbers. It's day and night of difference.
It's pretty sad they are keeping it locked to their systems.
As for AMD vs Intel, I will just only say that the achilles heel in Intel's cpu is that turbo boost thing: it clocks as high as it can while generating a lot of heat and consuming a lot of energy. At least in my 1185G7, things have been better since I turned it off. Sure, there must have been some performance lost, but I havent noticed, and I prefer to run on stock clocks rather than boosting for everything while on Windows.
I dont know if AMD has an equivalent to this or how do they manage it, but Intel seriously needs to keep that thing in check. It brings more negatives than positives, in laptops at least.
Quote from: Rav4 on May 31, 2022, 20:45:50
Power efficiency tests should be done when we have tests doing exactly same work... the faster processor will idle faster and save battery...
While the basic premise is true -- it is the reason why it doesn't pay to be too slow, it's better to get the job done faster and idle longer as idle allows you to shut things down to minimize leakage -- Turbo Boost is beyond diminishing returns. Meaning, the energy you save during idle will never compensate for the extra energy needed to sustain boost. If you're after efficiency, it's best to avoid boost altogether.
Quote from: Rav4 on May 31, 2022, 20:45:50
Power efficiency tests should be done when we have tests doing exactly same work... the faster processor will idle faster and save battery...
The biggest fallacy in these articles is the treatment of efficiency as if it's a single number. No, it's a function. You can attain a range of efficiencies and there are trade-offs to be made. The Whisper mode illustrates this rather nicely.
The single-threaded section could practically be called worst-case efficiency as the processor is in a "all-you-can-eat buffet mode," it's not even trying to be efficient.
Hi any chance you could set the 1260p to 28w and do the comparison that way? (I have this processor and do it through trottlestop). Would be an interesting comparison, since the 1265u is not avaible yet .
Quote from: Rav4 on May 31, 2022, 20:45:50This is bogus.. everyone knows last 5% perf takes 20-30% more power. Power efficiency tests should be done when we have tests doing exactly same work... the faster processor will idle faster and save battery... notebook check needs to send their author to some computer science 101 class
Its much worse then that... Look up the Whisper mode numbers.
The CPU loses 50% performance but it using ~233 a 250% less power. The main reason being that the CPU is forced to use reduced frequency and thus better performance.
The race for faster and faster frequencies just kills performance. Wait for Zen4 and enjoy the s*** show that 5.5Ghz will become. People expecting that TSMC 5nm its 20% power saving compared to 7nm, are in for a rude awaking when a 5.5Ghz mobile CPU will be more power hungry, despite the advanced node power saving. Sure, you will gain maybe 15% performance from the higher frequency but you will be spending easily another 300% of power draw on that 4.7 -> 5.5Ghz range.
That is the problem when CPU's are made with desktop first strategy, where power is not a issue. Need more power? Well, just increase the TDP target to 170W ( Zen 4 ) or like Intel 240W+. And then try to downscale that same technology down to mobile platforms.
Apple their M1's are this efficient because they are designed for a mobile platform first, where you need to first look at each feature, what power does it use and what benefit does it give you. Unlike our PC platform, where its first what performance does it give and that is it. Because benchmark winning is more important then power saving, as our world waste energy everywhere. 70W ~ 100W idle PC's, no problem. That is what PC's used to do, when fully loaded on ancient nodes!!! A lot of our performance has come at the cost of power. And lets not look at NVIDIA/AMD GPU's with NVIDIA wanting 600W power connectors for their GPU's. Insanity!
What is this? Did you measure the draw from the wall? The M1 perf TDP is entirely off.
Just saying... do not get me wrong.
Quote from: Sparkyte on June 04, 2022, 11:23:44What is this? Did you measure the draw from the wall? The M1 perf TDP is entirely off.
Since the article talks of package and core power, it's self-reported. There is no way you could measure that at a wall.
Quote from: Ben29 on June 03, 2022, 14:58:28The CPU loses 50% performance but it using ~233 a 250% less power. The main reason being that the CPU is forced to use reduced frequency and thus better performance.
You mean better efficiency. It actually stems from the fact that lower frequencies allow you to use lower voltage which modern CPUs can exploit. And saying things like 250% less power is just ridiculous, no offence. What in the world that means? That the processor is making electricity? You mean 64 % of performance for 43 % of power which improves efficiency by 50 %.
Ummm... you compared Zen3+ to just the P-cores in the Intel CPU for the efficiency test?? Did I get that right? So the E-cores, you know the efficiency cores, are irrelevant for an efficiency comparison?
Why do you only emphasize the difference between 6000 series and Intel and not also the fact that 6000 series is worse in efficiency than the 5000 series zen 3 from AMD???
Quote from: litetaker on June 08, 2022, 22:33:40Ummm... you compared Zen3+ to just the P-cores in the Intel CPU for the efficiency test?? Did I get that right? So the E-cores, you know the efficiency cores, are irrelevant for an efficiency comparison?
Hi, that's not quite right. For the single—core tests, applications will automatically use the faster P-cores.
All cores are active in multi-core tests anyway, so the efficiency cores are included there.
Are you kidding me? Is it AMD payed bullshit? Or ya just f... amateurs?
You should compare 6800U w-mode vs 1260p running on E-cores only w/o turbo vs m1(m2) E-cores only. It's easy as f... Just google howto cinebench e-cores only. also 1260p zenbook has "whisper" mode too!
And show us "whisper" mode frequency, please.
i guess m2 will be 10 times more efficient than 6800u and 3x 1260p
Still no review after all this time ?...