i5-2520m vs i5-4200U/5200U/6200U vs Core M 6y30?

[zijin_cheng]

I'm been looking at benchmarks for the old i5-2520m vs the newer i5-X200U chips. Based on some benchmarks, the performance of the 2520m vs the 4200U seems very similar, so why go with the new U series of chips instead of the old one?

Also, how does the Core M perform compared to both the 2520m and the 4200U, about 20-30% slower?

 

[jhu]

I'm been looking at benchmarks for the old i5-2520m vs the newer i5-X200U chips. Based on some benchmarks, the performance of the 2520m vs the 4200U seems very similar, so why go with the new U series of chips instead of the old one?

Also, how does the Core M perform compared to both the 2520m and the 4200U, about 20-30% slower?

Better battery life maybe?

 

[Burpo]

Also, how does the Core M perform compared to both the 2520m and the 4200U, about 20-30% slower?

https://www.youtube.com/watch?v=M4EmpbCQ5cY

 

[Maxima1]

Better battery life maybe?

Right, but it's still stupid. Much of the battery life is taken by the screen, not the CPU, and it's not costly or weight burdensome to add a higher capacity battery.

 

[jpiniero]

Right, but it's still stupid. Much of the battery life is taken by the screen, not the CPU, and it's not costly or weight burdensome to add a higher capacity battery.

Maybe not with Core M given how much it draws. If you want any of the U parts you'd need a fan and that of course makes a difference.

 

[coercitiv]

Right, but it's still stupid. Much of the battery life is taken by the screen, not the CPU, and it's not costly or weight burdensome to add a higher capacity battery.

Stupid? That i5-2520m has a 35W TDP, while an entire modern ultrabook consumes 30W under full load.

 

[Zodiark1593]

I'm been looking at benchmarks for the old i5-2520m vs the newer i5-X200U chips. Based on some benchmarks, the performance of the 2520m vs the 4200U seems very similar, so why go with the new U series of chips instead of the old one?

Also, how does the Core M perform compared to both the 2520m and the 4200U, about 20-30% slower?

Under peak loads, a Core M should outpace the 2520M, though in extended workloads, the Core M will throttle back greatly depending on OEM configuration.

 

[crashtech]

Just because screens draw inordinate amounts of power doesn't mean the rest of the system shouldn't be made more efficient. Also, the essential part of putting more CPU power in a mobile form factor is to design for higher performance per watt. Although benchmarks are scarce, I'm betting the i7-6920HQ has no peer when it comes to no-holds barred mobile performance while staying within the tried and true 45W TDP maximum laptop envelope.

 

[Maxima1]

Stupid? That i5-2520m has a 35W TDP, while an entire modern ultrabook consumes 30W under full load.

If you are under full load, then it isn't likely you are using it for portable use, or if you are, the battery will tank within 1-2 hours. And as I said, it's likely that the a given ultrabook has half the battery capacity anyways.

Just because screens draw inordinate amounts of power doesn't mean the rest of the system shouldn't be made more efficient.

Well, duh, but that still happens with refinements and shrinks. Let's see. The new Dell Latitude 5000 series uses standard voltage CPUs with higher than average battery capacity. My bet is the 3000 series will use ULV. So was the ULV necessarily better?

 

[frozentundra123456]

Well, simply look at the prices of ultrabooks, and you know why Intel and OEMs are going for lower power and "thin and light" instead of higher performance. Actually, what I would like to see intel do is include some middle level of performance between the 15 watt U chips and the 45 watt quads. Maybe a higher clocked dual core or a lower clocked quad. I think there is actually supposed to be a 28 watt Skylake chip, but whether it will see any real availability is a big question.

 

[Maxima1]

Well, simply look at the prices of ultrabooks, and you know why Intel and OEMs are going for lower power and "thin and light" instead of higher performance.

Well, look at the most expensive Macbook Pro and its appearance.

Actually, what I would like to see intel do is include some middle level of performance between the 15 watt U chips and the 45 watt quads. Maybe a higher clocked dual core or a lower clocked quad. I think there is actually supposed to be a 28 watt Skylake chip, but whether it will see any real availability is a big question.

They have had 28 watt SKUs, though the OEMs haven't used them. They rather just sell the "core i7" ULV 15 watt as the flagship even if it's a 17 inch notebook.

 

[Raftina]

The 28W U processors have higher TDP to accommodate the more powerful GPU. Any benefit the CPU might derive is entirely incidental. You also need to consider the form factor built around the CPU. The 28W CPUs are almost always put into "premium" thin and light form factors that have poor cooling, with the result that they throttle much more easily under load than a 15W CPU in a business laptop.

As for the original question: The Haswell/Broadwell/Skylake U processors doing the same work under much lower power consumption means that they also produce significantly less heat, so the heatsink fan does not work nearly as much. The difference is not merely theoretical: A ~20% load on a Youtube video on my X220t (2520M) causes the fan to run noticeably, while the HSF for the 5200U on my mother's X250 barely moves.

There's also the fact that the iGP on the 4xxxU/5xxxU/6xxxxU is far more powerful than the HD 3000.

 

[Maxima1]

The 28W U processors have higher TDP to accommodate the more powerful GPU. Any benefit the CPU might derive is entirely incidental. You also need to consider the form factor built around the CPU. The 28W CPUs are almost always put into "premium" thin and light form factors that have poor cooling, with the result that they throttle much more easily under load than a 15W CPU in a business laptop.

It isn't just for the GPU. They have higher turbos/base clocks. If it meant nothing, then why do OEMs put in 15 watt i7 ULV?

 

[coercitiv]

If you are under full load, then it isn't likely you are using it for portable use, or if you are, the battery will tank within 1-2 hours.

Even when left completely idle a modern ULV notebook will consume about 50% less power than an old Sandy Bridge platform.

My Toshiba Z30-A has 15W TDP Haswell CPU, 52Wh battery and weighs 1.2kg. Care to share with us a similar Sandy Bridge system to match performance, weight and stamina of this model?

And as I said, it's likely that the a given ultrabook has half the battery capacity anyways.

You mean most normal notebooks have 90-120Wh batteries?

 

[VirtualLarry]

Right, but it's still stupid. Much of the battery life is taken by the screen, not the CPU, and it's not costly or weight burdensome to add a higher capacity battery.

That's what they say, but my 15.6" Dell with Atom N2830 CPU lasts for more than 8 hours on a charge, and it doesn't have an extraordinarily large battery.

My Lenovo 11.6" laptop with Atom Z3735F also lasts for more than 8-9 hours on a charge. The battery in that unit is non-user-removable.

My Lenovo 15.6" laptop with AMD A6-6310 lasts 4-5 hours on a charge.

 

[dark zero]

It isn't just for the GPU. They have higher turbos/base clocks. If it meant nothing, then why do OEMs put in 15 watt i7 ULV?

Simple. To save costs. And that is not good for the consumer who wants the best and now they deliver us the worst.

 

[Maxima1]

Even when left completely idle a modern ULV notebook will consume about 50% less power than an old Sandy Bridge platform.

My Toshiba Z30-A has 15W TDP Haswell CPU, 52Wh battery and weighs 1.2kg. Care to share with us a similar Sandy Bridge system to match performance, weight and stamina of this model?

I don't really disagree depending on the price of an old system and a person's needs. I buy refurbished of current models, since past generations don't go down enough to even consider it, especially if you are a user who needs graphics performance.

You mean most normal notebooks have 90-120Wh batteries?

??? Standard voltage hasn't been normal. Many of the standard voltage CPUs have been pushed up into top end models or gaming. They have higher battery capacities. I doubt a 52Wh battery is the norm for ULV.

That's what they say, but my 15.6" Dell with Atom N2830 CPU lasts for more than 8 hours on a charge, and it doesn't have an extraordinarily large battery.

Larry, I don't want to see Atom creeping up especially in >= 15"* lol

 

[Maxima1]

Some good examples of my point: the Dell XPS has an 84 whr battery as an option. The cheaper 7559 has 74 whr.

You really don't get that kind of thing in ULV unless it comes with discrete graphics. And since it last long enough for portable use, why the heck wouldn't I want standard voltage over ULV? It doesn't even increase the weight of the laptop by even a half of a pound.

 

[zijin_cheng]

I see, thanks for all the replies, very much appreciated. So performance wise, the 2520m/4200U/5200U and maybe the 6200U are all very similar in performance, the difference is how much power they draw?

And you mentioned taht the 6y30 has better bursty performance. I know what that means, but would photoshop/video editing (not the rendering, but the actual editing aspect) be considered bursty?

 

[Raftina]

And you mentioned taht the 6y30 has better bursty performance. I know what that means, but would photoshop/video editing (not the rendering, but the actual editing aspect) be considered bursty?

No. Those use sustained power draw. If this is for hobby editing, you can get away with a dual core M or U processor. If this is for professional editing, go for a workstation with a HQ processor and, just as importantly, a nice screen with good color space.

It isn't just for the GPU. They have higher turbos/base clocks. If it meant nothing, then why do OEMs put in 15 watt i7 ULV?

Price is one. Another is, of course, that putting a 28W package in the laptop means thermal design to accommodate a 28W CPU--at least theoretically. Most ultrabook manufacturers fail to do this. And then, the 28W packages are consumer models. They lack vPro--and until Haswell, VT-d.

Which reminds me, the 2520M is comparable to the x300U, not the x200U.

Some good examples of my point: the Dell XPS has an 84 whr battery as an option. The cheaper 7559 has 74 whr.

You really don't get that kind of thing in ULV unless it comes with discrete graphics.

The T450, T450s, and X250 can be configured with a 24+72 Whr configuration (96 Whr total) for all of $5 more than the base price. All 3 ship with ULV CPUs. The latter 2 are ultrabooks.

 

[sm625]

The reason they make them smaller and more power efficient is because that is the only way to squeeze the gullible consumer year after year. They cannot deliver real tangible performance increases, so they just increase the efficiency and hope the consumer laps it up anyway. For the most part they do, but in the long term it is simply a waste of resources. Most notebooks spend most of their lives plugged in to the mains anyway, so the increase in "portability" is largely moot. In every sense of the word, it is simply a scam.

 

[dark zero]

The reason they make them smaller and more power efficient is because that is the only way to squeeze the gullible consumer year after year. They cannot deliver real tangible performance increases, so they just increase the efficiency and hope the consumer laps it up anyway. For the most part they do, but in the long term it is simply a waste of resources. Most notebooks spend most of their lives plugged in to the mains anyway, so the increase in "portability" is largely moot. In every sense of the word, it is simply a scam.

Totally agree with you. Is now a mere scam. The H tier is badly treated when that tier is way superior than the U line. Maybe OEMs wants to scam everyone, even worse in the bussiness which is mandatory the performance over the energy consumption.

I have a LOT of problems with an U tier laptop (5th geb Core i5). Is so laggy in some apps, specially Virtual Machines that I am forced to use another more powerful machine (my own 4th gen Quad Core i7 MQ) in order to complete the tasks.

 

[Maxima1]

Price is one. Another is, of course, that putting a 28W package in the laptop means thermal design to accommodate a 28W CPU--at least theoretically.

Which is funny because standard voltage performs better and is cheaper.

Most ultrabook manufacturers fail to do this. And then, the 28W packages are consumer models. They lack vPro--and until Haswell, VT-d.

I don't see how that is such a huge engineering feat. There are thin laptops with i7 quads and big discrete cards. They can easily do 28 TDP. But I'm not just specifically talking about "ultrabooks" either. The trend has been to put ULV in many notebooks.

The T450, T450s, and X250 can be configured with a 24+72 Whr configuration (96 Whr total) for all of $5 more than the base price. All 3 ship with ULV CPUs. The latter 2 are ultrabooks.

Exceptional (I see many are going the integrated route, too), and you just proved my point showing how cheap the batteries really are. A big battery kills the point of the ULV.

 

[Dufus]

Although benchmarks are scarce, I'm betting the i7-6920HQ has no peer when it comes to no-holds barred mobile performance while staying within the tried and true 45W TDP maximum laptop envelope.

Isn't that a contradiction? Might as well do the Intel marketing trick and limit it to 4.5W while competing with a core m.

Have seen one of those rare 6920HQ WPrime benches at 4GHz while using a -80mV offset vs a 6820HK at 4GHz with a -60mV offset.

The 6920HQ Intel estimated pkg power was 70W vs 49W for the 6820HK

 

[hhhd1]

The iGPU in newer processors is supposed to be much better.

The main 2 differences are effeicency and iGPU, if you do not care about either, then raw cpu performance is quite close.