Yoga Pro 3 review: Broadwell is a mixed blessing

The design is ever more exotic, but the performance is worse.

Read the whole story

 

[sonicmerlin]

ichemandrew[/url]":3f2o2csf]Core M really doesn't seem that impressive for the price. It also seems strange that it needs a fan. Even at inaudible speed, that's still another watt or two thrown away, which matters for a small battery.

Intel's reference 10" Broadwell tablet was fanless and its GPU benchmarks were much better than what the Nvidia K1 and the iPad Air 2 can do.

http://www.anandtech.com/show/8515/quic ... a-mountain

Lenovo did something very wrong here. I've read speculation they've capped the TDP at 2.5 W and that's causing a ton of throttling.

 

[graigsmith]

They fail at centering the touch pad.

 

[DrPizza]

sonicmerlin[/url]":fueownyt]

ichemandrew[/url]":fueownyt]Core M really doesn't seem that impressive for the price. It also seems strange that it needs a fan. Even at inaudible speed, that's still another watt or two thrown away, which matters for a small battery.

Intel's reference 10" Broadwell tablet was fanless and its GPU benchmarks were much better than what the Nvidia K1 and the iPad Air 2 can do.

http://www.anandtech.com/show/8515/quic ... a-mountain

Lenovo did something very wrong here. I've read speculation they've capped the TDP at 2.5 W and that's causing a ton of throttling.

There were some (vague) reports that it was capped at 2.5W in early versions, but ours arrived last week I think, so I would expect it to be OK. I haven't seen any positive confirmation either way, though.

 

[Callitrax]

심돌산[/url]":2wjrblww]

GlennHowes[/url]":2wjrblww]Am I right in reading this that the graphics (GFXBench Offscreen) of the Yoga 3 are now significantly slower than an iPad Air 2? And that the GeekBench Multi-core has the iPad a bit faster? For other things like the web benchmarks, the Yoga is faster, sometimes by a lot.

The graphics don't surprise me. Apple is prioritizing graphics in their SoCs, more so than Intel does in their chips.

As for the CPU scores, for anything single threaded, the Yoga 3 wins easily in comparison with the iPad Air 2. Apple has come a long way (baby), but not that far yet.

They surprise me. Honestly something had to have gone very wrong to get that low of a score on T-Rex. My Baytrail tablet gets 15.5 FPS in T-Rex offscreen. That is on an older process (22 nm) with 4 EUs instead of 24* EUs in Core M, and probably with a lower TDP. The Core M should do better than 10% increase. With more cores you can run at lower speeds/voltages/power and still get better results - thats what Apple does, they spend on silicon to save power. That tells me that somewhere between Intel and the test being run something went sideways. Maybe Intel set the firmware wrong to maximize the performance/power curve or the CPU is using too much of the power budget or there are problem with the OpenGL drivers. Maybe Lenovo set the power settings wrong, or have a bad cooling setup in the design. Or Kishonti screwed up somewhere in the coding. Or Ars screwed up the test (I'm guessing not that).

But this chip should do better than that graphically even in a low TDP.

* Apparently the smallest GPU setup in broadwell is 24 Execution Units

 

[F16PilotJumper]

theoilman[/url]":1mxa3ttq]benchmarks aside, does the new chip actually make for a noticeable difference in day to day use?

the loss of battery life with screen brightness is still worrying anyway though, and I have to agree that keyboard looks bad (and keyboard is one of the things Lenovo is supposed to be good at too, so for shame on them).

That's....that's how screen brightness works.

I am very disappointed that Ars can't spend a small amount of $$$ on a Spyder4 and set the monitors to the same brightness level for testing. That is a major omission from the test. They talk about the Yoga 3 being "too dim" but was the Yoga 2 tested at 100%? What screen brightness was the Apple Air tested at??

Lots of other sites that care about laptop battery life do this, in my mind if Ars can't even be bothered then they shouldn't do hardware reviews.

 

[Callitrax]

Joriarty[/url]":2ttn52uz]
Also - although pricing for the Core M isn't public, I've read it's in the $300 range.

So it seems that A8X is an order of magnitude cheaper, certainly consumes less power, and might also be faster.

ARM is kicking some serious ass. It's a pity that architecture changes aren't an entirely smooth affair.

Is the A8X cheaper? The lowest price you can get one is $499. Yeah you get a lot more with that but you also get the Apple high margin markup at that point. With the Intel based systems the highest margin part is the CPU where the markup is applied earlier in the supply chain. Given that the A8X has 3 Billion transisters, and the die sizes seem to be in the same range, I'm doubting that the production cost of the A8X Chips is markedly different from that of the Core M, the markup is just applied at a different point in the chain. (Short version, analyzing cost structures is a pain, but the chips shouldn't really be materially different in production cost)

And we don't inherently know that it uses less power. I've never seen real power measurements of the IPad SOCs under load, and there are so many other power draws in an operating computer that it can be hard to compare. But Intel chips have been shown to work in the same power ranges as other ARM chips.

 

[QuikMix]

lack of dedicated F keys is an immediate dis qualifier. I guess I'll wait for the update to the Thinkpad Yoga

 

[hornetfig]

maskarova[/url]":3l99r5ya]Would really love to see a review of the yoga 11e windows version. They appear thin on the ground at the moment.

I've had demo units of both the Yoga and non-Yoga ThinkPad 11e Windows:

* build quality is fine
* the Yoga version's screen is pretty good for the category and a lot better than the cheap TN in the non-Yoga. Touch response was fine. In any case the units are 1366x768, so I'd never personally buy one
* both versions are heavy. It's weird to compare to like an X240 which is vastly lighter.
* battery life I didn't benchmark against their claims; that the battery is non-removable is concerning though for the market segment it's aimed at

 

[DrPizza]

F16PilotJumper[/url]":2zqarnm0]

theoilman[/url]":2zqarnm0]benchmarks aside, does the new chip actually make for a noticeable difference in day to day use?

the loss of battery life with screen brightness is still worrying anyway though, and I have to agree that keyboard looks bad (and keyboard is one of the things Lenovo is supposed to be good at too, so for shame on them).

That's....that's how screen brightness works.

I am very disappointed that Ars can't spend a small amount of $$$ on a Spyder4 and set the monitors to the same brightness level for testing. That is a major omission from the test. They talk about the Yoga 3 being "too dim" but was the Yoga 2 tested at 100%? What screen brightness was the Apple Air tested at??

Lots of other sites that care about laptop battery life do this, in my mind if Ars can't even be bothered then they shouldn't do hardware reviews.

One is on its way to me, it's just somewhere inside FedEx. I'm not sure where.

 

[theoilman]

F16PilotJumper[/url]":2y6qouuq]

theoilman[/url]":2y6qouuq]benchmarks aside, does the new chip actually make for a noticeable difference in day to day use?

the loss of battery life with screen brightness is still worrying anyway though, and I have to agree that keyboard looks bad (and keyboard is one of the things Lenovo is supposed to be good at too, so for shame on them).

That's....that's how screen brightness works.

I am very disappointed that Ars can't spend a small amount of $$$ on a Spyder4 and set the monitors to the same brightness level for testing. That is a major omission from the test. They talk about the Yoga 3 being "too dim" but was the Yoga 2 tested at 100%? What screen brightness was the Apple Air tested at??

Lots of other sites that care about laptop battery life do this, in my mind if Ars can't even be bothered then they shouldn't do hardware reviews.

Of course screen brightness reduces battery life but 4 hours is insane, and the fact that the yoga pro 2 is far less affected highlights a problem with the screen and battery life in the new one.

 

[The Real Blastdoor]

심돌산[/url]":40kv7rzr]If you look at these benchmarks side by side with the benchmarks from Ars' iPad Air 2 review, you will see that the Yoga 3 wins everything single-threaded, usually by a large margin.

If find it very hard to interpret benchmarks comparing Core-M to the A8X. Core-M will turbo for brief periods of time up to 2.6 GHz while the A8X stays at 1.5 GHz all the time. For the web benchmarks, which take milliseconds to complete, it may very well be that Core-M is benefiting from a short clock speed spike. But is that really informative about real world performance?

I'd love to see a detailed comparison of these two processors, comparing sustained performance/watt. Alas, I don't think we'll be seeing that kind of careful analysis anytime soon, here or anywhere else. It just doesn't seem to be something reviewers are interested (or able?) to take on.

 

[name99]

It's interesting, as an intellectual exercise, to compare these Broadwell results with the A8X.

The short summary is:
Geekbench single core is about 40% faster
Geekbench multicore is a few to ten percent slower
Graphics has A8X wiping the floor, at around 50% faster or more, on or off screen
For browser benchmarks (which always have the problem that they also reflect software differences which can make a severe difference --- so good as a platform test, not so much as a CPU test) Broadwell+IE beats Android+Safari by 20 to 150 or so percent.

So two interesting takeaways.
- The story has been that Intel's iGPUs are way ahead of the ARM world. Doesn't seem to be the case here...
- Broadwell costs 5x or more what Apple pays TSMC for an A8X, and has the advantage of all that 14nm FF goodness, but it doesn't do it THAT MUCH good. Yes the single-threaded performance is superior, but not ridiculously so. It's not a stretch to imagine that, if Apple cared, they could bump the A9X by 40% over the A8X. They got 25% with the A8 over A7 without trying that hard (IMHO because they simply didn't need to --- A57's still haven't shipped --- so they devoted all their efforts to the watch chip and the A9, and phoned in the A8 as a set of minor tweaks).

 

[F16PilotJumper]

theoilman[/url]":307loafv]

F16PilotJumper[/url]":307loafv]

theoilman[/url]":307loafv]benchmarks aside, does the new chip actually make for a noticeable difference in day to day use?

the loss of battery life with screen brightness is still worrying anyway though, and I have to agree that keyboard looks bad (and keyboard is one of the things Lenovo is supposed to be good at too, so for shame on them).

That's....that's how screen brightness works.

I am very disappointed that Ars can't spend a small amount of $$$ on a Spyder4 and set the monitors to the same brightness level for testing. That is a major omission from the test. They talk about the Yoga 3 being "too dim" but was the Yoga 2 tested at 100%? What screen brightness was the Apple Air tested at??

Lots of other sites that care about laptop battery life do this, in my mind if Ars can't even be bothered then they shouldn't do hardware reviews.

Of course screen brightness reduces battery life but 4 hours is insane, and the fact that the yoga pro 2 is far less affected highlights a problem with the screen and battery life in the new one.

There are two major sources of power draw in a modern Intel laptop - the CPU/GPU (one die now) and the screen.

The Yoga 2 Pro had a 15W TDP CPU and a a screen that draws some power X, and a 55 Whr battery

The Yoga 3 Pro has 4.5W TDP CPU and lets say the same screen with power X, and a 44 Whr battery.

With less of the system's draw coming from the CPU, and a smaller battery, the effect of changing the screen brightness will be amplified because a larger percentage of the power draw is coming from the screen. The smaller battery is sized to produce a run time at some standard brightness that Lenovo uses in their testing.

This is not something mysterious or some screw up it's just a consequence of driving the CPU draw down even lower, that any other change will be amplified. And laptop manufacturers do not size batteries for 100% brightness.

 

[steelgrass]

I looked at the Yoda Pro 3 earlier, and came away confused. This review was helped a great deal.

For me the compromises are fine with one exception - the 5 hours battery life.

The amazing thing is this wasn't caused by the CPU. Intel has apparently nailed it with Broadwell.

I guess it also explains why the Apple Air has "only" a 1440x900 screen.

 

[romkyns]

pythagoreanmetronome[/url]":1qunj2fa]EDIT: what's next "No Pause Break key for YOU!"

Next?!?! It's already a reality. Just look at Samsung series 9.

graigsmith[/url]":1qunj2fa]They fail at centering the touch pad.

Oh yeah good point, that's also a Yoga 2 Pro fail. When you're typing, the right hand is on the touch pad. Oh well, give up everything for the looks, I guess, that's the way one sells things.

 

[kliu0x52]

Joriarty[/url]":2oea2fq6]"the [function] keys may be power user features..."

Eh, really? What about as a set of media keys (with the F1 - F12 functions needing fn key)? I like having music, display brightness, and a few other shortcuts right at my fingertips. I don't think that's a power user thing! I agree that the removal of that row of keys seems pointless. I see lots of space above the keyboard where they could go.

I have a friend who's also a power user. She types faster than me (and I'm pretty fast) and navigates forms the proper way, by tabbing through fields (she even taught me that [Shift] reverses the traversal direction, which at the time I didn't know about even after a decade and a half of using computers), and she uses all kinds of keyboard shortcuts. One day, I was helping fix her netbook that, by default, had the F-keys mapped to media functions (you had to press the blue Fn modifier to get them to act as F-keys), and I got so annoyed by that behavior that I changed the default so that they acted as F-keys on normal presses and media keys when modified by Fn. To my surprise, that annoyed her because she really wanted those media key functions. Surely someone who's so at home with the keyboard would want proper F-key behavior instead. I asked, have you ever renamed a file using F2? No? Ever reloaded using F5? I didn't know I could do that.

So I guess the moral of the story is: Media keys are important too. And the problem with keyboard shortcuts is discoverability. Even power users don't know them all. Like me not knowing, for 15 years, that Shift-Tab could be used for reverse traversal.

 

[katorga]

I bought a Yoga 2 Pro and returned it within 5 days, the screen could not display yellow accurately. If it is the same screen, it will have the same problem.

 

[mrnomnoms]

It'll be interesting to see how the same hardware performs when Windows 10 is released given the much rumours going around about how the next version will be more memory and power efficient that Windows 8.1 - hopefully part of the problem seen with the Broadwell Y 'Yoga Pro 3' has been the lack of optimisation of Windows and that the gap will be filled once Windows 10 is released. I'd love to hear whether some here have tried running alternative operating systems on it such as Linux or *BSD and whether the issue is actually the hardware rather than the operating system.

 

[theoilman]

F16PilotJumper[/url]":3abfgx3r]

theoilman[/url]":3abfgx3r]

F16PilotJumper[/url]":3abfgx3r]

theoilman[/url]":3abfgx3r]benchmarks aside, does the new chip actually make for a noticeable difference in day to day use?

the loss of battery life with screen brightness is still worrying anyway though, and I have to agree that keyboard looks bad (and keyboard is one of the things Lenovo is supposed to be good at too, so for shame on them).

That's....that's how screen brightness works.

I am very disappointed that Ars can't spend a small amount of $$$ on a Spyder4 and set the monitors to the same brightness level for testing. That is a major omission from the test. They talk about the Yoga 3 being "too dim" but was the Yoga 2 tested at 100%? What screen brightness was the Apple Air tested at??

Lots of other sites that care about laptop battery life do this, in my mind if Ars can't even be bothered then they shouldn't do hardware reviews.

Of course screen brightness reduces battery life but 4 hours is insane, and the fact that the yoga pro 2 is far less affected highlights a problem with the screen and battery life in the new one.

There are two major sources of power draw in a modern Intel laptop - the CPU/GPU (one die now) and the screen.

The Yoga 2 Pro had a 15W TDP CPU and a a screen that draws some power X, and a 55 Whr battery

The Yoga 3 Pro has 4.5W TDP CPU and lets say the same screen with power X, and a 44 Whr battery.

With less of the system's draw coming from the CPU, and a smaller battery, the effect of changing the screen brightness will be amplified because a larger percentage of the power draw is coming from the screen. The smaller battery is sized to produce a run time at some standard brightness that Lenovo uses in their testing.

This is not something mysterious or some screw up it's just a consequence of driving the CPU draw down even lower, that any other change will be amplified. And laptop manufacturers do not size batteries for 100% brightness.

If the 3 starts with almost double battery life and turning up screen brightness brings it to the same battery life as the 2 then that shows the screen in the 3 is less efficient or something is being used inefficiently. Otherwise the 3 would still have more life than the 2 at high brightness.

 

[karadoc]

theoilman[/url]":1l14awj6]

F16PilotJumper[/url]":1l14awj6]

theoilman[/url]":1l14awj6]

F16PilotJumper[/url]":1l14awj6]

theoilman[/url]":1l14awj6]benchmarks aside, does the new chip actually make for a noticeable difference in day to day use?

the loss of battery life with screen brightness is still worrying anyway though, and I have to agree that keyboard looks bad (and keyboard is one of the things Lenovo is supposed to be good at too, so for shame on them).

That's....that's how screen brightness works.

I am very disappointed that Ars can't spend a small amount of $$$ on a Spyder4 and set the monitors to the same brightness level for testing. That is a major omission from the test. They talk about the Yoga 3 being "too dim" but was the Yoga 2 tested at 100%? What screen brightness was the Apple Air tested at??

Lots of other sites that care about laptop battery life do this, in my mind if Ars can't even be bothered then they shouldn't do hardware reviews.

Of course screen brightness reduces battery life but 4 hours is insane, and the fact that the yoga pro 2 is far less affected highlights a problem with the screen and battery life in the new one.

There are two major sources of power draw in a modern Intel laptop - the CPU/GPU (one die now) and the screen.

The Yoga 2 Pro had a 15W TDP CPU and a a screen that draws some power X, and a 55 Whr battery

The Yoga 3 Pro has 4.5W TDP CPU and lets say the same screen with power X, and a 44 Whr battery.

With less of the system's draw coming from the CPU, and a smaller battery, the effect of changing the screen brightness will be amplified because a larger percentage of the power draw is coming from the screen. The smaller battery is sized to produce a run time at some standard brightness that Lenovo uses in their testing.

This is not something mysterious or some screw up it's just a consequence of driving the CPU draw down even lower, that any other change will be amplified. And laptop manufacturers do not size batteries for 100% brightness.

If the 3 starts with almost double battery life and turning up screen brightness brings it to the same battery life as the 2 then that shows the screen in the 3 is less efficient or something is being used inefficiently. Otherwise the 3 would still have more life than the 2 at high brightness.

That's not necessarily the case. If the 3 has a lower-capacity battery, but a more efficient CPU, then that could result in a longer battery life for low-brightness, and shorter battery life for high-brightness - compared to the 2.

If the 2 is spending most of its energy on the CPU, and the 3 is spending most of its energy on the screen, then clearly screen brightness will have a larger impact on the 3 than it will on the 2.

 

[aleph_nought]

If they're hellbent on getting the TDP down, Lenovo should have gone with a fanless Atom SoC like what it's using on the cheap Miix tablets. These Yoga Pros are meant to be a Surface Pro competitor with notebook-class processing power and decent battery life, but the strange reduction in battery life and performance here seems very strange.

 

[F16PilotJumper]

karadoc[/url]":29joknpz]

theoilman[/url]":29joknpz]

F16PilotJumper[/url]":29joknpz]

theoilman[/url]":29joknpz]

F16PilotJumper[/url]":29joknpz]

theoilman[/url]":29joknpz]benchmarks aside, does the new chip actually make for a noticeable difference in day to day use?

the loss of battery life with screen brightness is still worrying anyway though, and I have to agree that keyboard looks bad (and keyboard is one of the things Lenovo is supposed to be good at too, so for shame on them).

That's....that's how screen brightness works.

I am very disappointed that Ars can't spend a small amount of $$$ on a Spyder4 and set the monitors to the same brightness level for testing. That is a major omission from the test. They talk about the Yoga 3 being "too dim" but was the Yoga 2 tested at 100%? What screen brightness was the Apple Air tested at??

Lots of other sites that care about laptop battery life do this, in my mind if Ars can't even be bothered then they shouldn't do hardware reviews.

Of course screen brightness reduces battery life but 4 hours is insane, and the fact that the yoga pro 2 is far less affected highlights a problem with the screen and battery life in the new one.

There are two major sources of power draw in a modern Intel laptop - the CPU/GPU (one die now) and the screen.

The Yoga 2 Pro had a 15W TDP CPU and a a screen that draws some power X, and a 55 Whr battery

The Yoga 3 Pro has 4.5W TDP CPU and lets say the same screen with power X, and a 44 Whr battery.

With less of the system's draw coming from the CPU, and a smaller battery, the effect of changing the screen brightness will be amplified because a larger percentage of the power draw is coming from the screen. The smaller battery is sized to produce a run time at some standard brightness that Lenovo uses in their testing.

This is not something mysterious or some screw up it's just a consequence of driving the CPU draw down even lower, that any other change will be amplified. And laptop manufacturers do not size batteries for 100% brightness.

If the 3 starts with almost double battery life and turning up screen brightness brings it to the same battery life as the 2 then that shows the screen in the 3 is less efficient or something is being used inefficiently. Otherwise the 3 would still have more life than the 2 at high brightness.

That's not necessarily the case. If the 3 has a lower-capacity battery, but a more efficient CPU, then that could result in a longer battery life for low-brightness, and shorter battery life for high-brightness - compared to the 2.

If the 2 is spending most of its energy on the CPU, and the 3 is spending most of its energy on the screen, then clearly screen brightness will have a larger impact on the 3 than it will on the 2.

Exactly.

I am glad I bought the original Helix with the Ivy Bridge i7 because the Helix 2 with Core M just isn't going to do it performance-wise for me.

 

[Xiao-zhi]

About that hinge

More than 600 parts?

I think Lenovo just out-gimmicked Samsung, and that takes some doing.

Well, it might be good news for iFixit because that's a lot of little spare parts to buy.

 

[foxyshadis]

pythagoreanmetronome[/url]":2u2x3j8h]No F-keys. I am sorry maybe I am a bit of a geezer but not having F-keys is ludicrous. I suppose there must be some way to enter the BIOS during boot up doing it some other way than normal is odious. This thing is going to be a $1000 coaster in two years I am sure. How the heck I am supposed to install Ubuntu on this thing? Because I have never owned a laptop that I didn't put Linux on first thing out of the box.

EDIT: what's next "No Pause Break key for YOU!"

You haven't been able to enter the "BIOS" with Fn keys since Win8 and other secure boot OSes (Ubuntu and Red Hat), at least for UEFI-based PCs, unless it was specially enabled in the system firmware (HP, Lenovo, and Dell allow this on their business lines, but not their consumer lines, like the IdeaPads). Now you have to use system menus in the OS to reboot into the EFI menu. When Lenovo does implement a boot key, it's always a separate blue ThinkVantage key.

It's often possible to reset the EFI to legacy BIOS mode, though, if you're installing an older operating system. In that case the firmware always gives you some way to get in.

Even way back in the day, the Del key was pretty common for entering BIOS, and that's probably what legacy mode uses on the Yoga 3.

 

[foxyshadis]

DrPizza[/url]":1dch2we9]

sonicmerlin[/url]":1dch2we9]

ichemandrew[/url]":1dch2we9]Core M really doesn't seem that impressive for the price. It also seems strange that it needs a fan. Even at inaudible speed, that's still another watt or two thrown away, which matters for a small battery.

Intel's reference 10" Broadwell tablet was fanless and its GPU benchmarks were much better than what the Nvidia K1 and the iPad Air 2 can do.

http://www.anandtech.com/show/8515/quic ... a-mountain

Lenovo did something very wrong here. I've read speculation they've capped the TDP at 2.5 W and that's causing a ton of throttling.

There were some (vague) reports that it was capped at 2.5W in early versions, but ours arrived last week I think, so I would expect it to be OK. I haven't seen any positive confirmation either way, though.

Do you have a Kill-A-Watt you can test with? That'd be conclusive. RMClock can show and let you test all the states the CPU should be able to hit, but I doubt it supports Broadwell yet.

 

[foxyshadis]

The Real Blastdoor[/url]":1sxbmnbp]

심돌산[/url]":1sxbmnbp]If you look at these benchmarks side by side with the benchmarks from Ars' iPad Air 2 review, you will see that the Yoga 3 wins everything single-threaded, usually by a large margin.

If find it very hard to interpret benchmarks comparing Core-M to the A8X. Core-M will turbo for brief periods of time up to 2.6 GHz while the A8X stays at 1.5 GHz all the time. For the web benchmarks, which take milliseconds to complete, it may very well be that Core-M is benefiting from a short clock speed spike. But is that really informative about real world performance?

I'd love to see a detailed comparison of these two processors, comparing sustained performance/watt. Alas, I don't think we'll be seeing that kind of careful analysis anytime soon, here or anywhere else. It just doesn't seem to be something reviewers are interested (or able?) to take on.

They're both interesting. Sub-second performance is interesting, and running sustained tasks that are going to be thermally limited is useful too. I'd imagine that some tasks in Photoshop would feel much snappier, even as others felt slower. Trade-offs are weird.

 

[pehash]

TheFerenc[/url]":2cc1g0zt]I'd have liked to see the ThinkPad Yoga tossed into the mix, as well.

Bothers me they call these things "Pro" but then have a version in their actual pro line -- in case the marketing has managed to convince you otherwise, these are still IdeaPads, not ThinkPads.

I seriously doubt a 'pro' user could be mislead by Lenovo's marketing strategy. If you don't know that ThinkPads are the real thing, you deserve your 'Pro' machine.. Maybe the real con is using a term like 'pro' to identify a slightly more knowledgeable buyer than my grandma? Tickle their vanity a bit.. make them pay for it!..

 

[Kingu]

The one standout app, for all the wrong reasons, was called OneKey Optimizer. For all the world this app feels like a throwback to those 1990s "Windows optimizers" that make a bunch of ill-advised "registry tweaks" to notionally optimize system performance but in reality break things in subtle ways.

Sounds hilariously well named to a British English speaker. ("OneKey" is homophonous with wonky).

 

[The Real Blastdoor]

DrPizza[/url]":39tm0w12]

sonicmerlin[/url]":39tm0w12]

ichemandrew[/url]":39tm0w12]Core M really doesn't seem that impressive for the price. It also seems strange that it needs a fan. Even at inaudible speed, that's still another watt or two thrown away, which matters for a small battery.

Intel's reference 10" Broadwell tablet was fanless and its GPU benchmarks were much better than what the Nvidia K1 and the iPad Air 2 can do.

http://www.anandtech.com/show/8515/quic ... a-mountain

Lenovo did something very wrong here. I've read speculation they've capped the TDP at 2.5 W and that's causing a ton of throttling.

There were some (vague) reports that it was capped at 2.5W in early versions, but ours arrived last week I think, so I would expect it to be OK. I haven't seen any positive confirmation either way, though.

Perhaps the "very wrong" thing that Lenovo did was to believe Intel's hype and put Core-M in this form factor. Maybe what Intel calls a TDP of 4.5 watts isn't what people think it means.

 

[The Real Blastdoor]

Callitrax[/url]":v0y80ctt]

심돌산[/url]":v0y80ctt]

GlennHowes[/url]":v0y80ctt]Am I right in reading this that the graphics (GFXBench Offscreen) of the Yoga 3 are now significantly slower than an iPad Air 2? And that the GeekBench Multi-core has the iPad a bit faster? For other things like the web benchmarks, the Yoga is faster, sometimes by a lot.

The graphics don't surprise me. Apple is prioritizing graphics in their SoCs, more so than Intel does in their chips.

As for the CPU scores, for anything single threaded, the Yoga 3 wins easily in comparison with the iPad Air 2. Apple has come a long way (baby), but not that far yet.

They surprise me. Honestly something had to have gone very wrong to get that low of a score on T-Rex. My Baytrail tablet gets 15.5 FPS in T-Rex offscreen. That is on an older process (22 nm) with 4 EUs instead of 24* EUs in Core M, and probably with a lower TDP. The Core M should do better than 10% increase. With more cores you can run at lower speeds/voltages/power and still get better results - thats what Apple does, they spend on silicon to save power. That tells me that somewhere between Intel and the test being run something went sideways. Maybe Intel set the firmware wrong to maximize the performance/power curve or the CPU is using too much of the power budget or there are problem with the OpenGL drivers. Maybe Lenovo set the power settings wrong, or have a bad cooling setup in the design. Or Kishonti screwed up somewhere in the coding. Or Ars screwed up the test (I'm guessing not that).

But this chip should do better than that graphically even in a low TDP.

* Apparently the smallest GPU setup in broadwell is 24 Execution Units

The big question comes at the end of your post -- "even in a low TDP"

There's no doubt that Core-M can post stellar benchmarks. It can also run at low power. What has not been fully determined is whether it can do both at the same time in real products. Lenovo may very well have screwed up here -- that kind of thing happens. But it could also be that Core-M was designed by marketers to win benchmarks at power draws that aren't feasible in real products.

 

[The Real Blastdoor]

sonicmerlin[/url]":2vb0vo0j]

ichemandrew[/url]":2vb0vo0j]Core M really doesn't seem that impressive for the price. It also seems strange that it needs a fan. Even at inaudible speed, that's still another watt or two thrown away, which matters for a small battery.

Intel's reference 10" Broadwell tablet was fanless and its GPU benchmarks were much better than what the Nvidia K1 and the iPad Air 2 can do.

http://www.anandtech.com/show/8515/quic ... a-mountain

Lenovo did something very wrong here. I've read speculation they've capped the TDP at 2.5 W and that's causing a ton of throttling.

A few notes on that anandtech article:

1. So far as I can tell, it wasn't a review -- they just put Intel marketing benchmarks in a graph next to results they achieved themselves when reviewing other products.

2. That reference tablet weighs 45% more than a 2013 iPad Air.

3. Intel actually set the TDP in this reference tablet to 6W rather than 4.5.

So.... hmmm.

 

[julienm]

심돌산[/url]":3swf4xs5]

GlennHowes[/url]":3swf4xs5]Am I right in reading this that the graphics (GFXBench Offscreen) of the Yoga 3 are now significantly slower than an iPad Air 2? And that the GeekBench Multi-core has the iPad a bit faster? For other things like the web benchmarks, the Yoga is faster, sometimes by a lot.

The graphics don't surprise me. Apple is prioritizing graphics in their SoCs, more so than Intel does in their chips.

As for the CPU scores, for anything single threaded, the Yoga 3 wins easily in comparison with the iPad Air 2. Apple has come a long way (baby), but not that far yet.

I'm not familiar with GFXBench, but if it runs at the native screen resolution (onscreen, and maybe offscreen too?), then we can't compare the results of this benchmark running on the lower resolution ipad air 2 (2048x1536) and the result of it running on the screen of the yoga pro 3 (3200x1800).

Also, if this test is using openGL, it won't perform as well as proper direct3d-based portage.

Since most apps will use direct3d on windows, it's a bit absurd to compare the performance of a chip by running an openGL-based benchmark on a Windows PC when we know that openGL support on Windows is usually slower than Direct3D support.

 

[ethd]

julienm[/url]":2ci8f0r2]

심돌산[/url]":2ci8f0r2]

GlennHowes[/url]":2ci8f0r2]Am I right in reading this that the graphics (GFXBench Offscreen) of the Yoga 3 are now significantly slower than an iPad Air 2? And that the GeekBench Multi-core has the iPad a bit faster? For other things like the web benchmarks, the Yoga is faster, sometimes by a lot.

The graphics don't surprise me. Apple is prioritizing graphics in their SoCs, more so than Intel does in their chips.

As for the CPU scores, for anything single threaded, the Yoga 3 wins easily in comparison with the iPad Air 2. Apple has come a long way (baby), but not that far yet.

I'm not familiar with GFXBench, but if it runs at the native screen resolution (onscreen, and maybe offscreen too?), then we can't compare the results of this benchmark running on the lower resolution ipad air 2 (2048x1536) and the result of it running on the screen of the yoga pro 3 (3200x1800).

Also, if this test is using openGL, it won't perform as well as proper direct3d-based portage.

Since most apps will use direct3d on windows, it's a bit absurd to compare the performance of a chip by running an openGL-based benchmark on a Windows PC when we know that openGL support on Windows is usually slower than Direct3D support.

Unless GFXBench is substantially different than when I last used it on iOS a while back, both Windows and iOS (and Android for that matter) render offscreen tests at 1080p. The Windows version is subtitled "DXBench," or in other words, it's a native test of Direct3D. (EDIT: actually, looks like even the Windows test is OpenGL now. That is really strange.) Unless it's changed since the last time I looked at it, which is likely with some new iOS graphics APIs, the iOS version uses OpenGL.
For whatever it's worth, my current phone is an HTC One and its GPU is comparable to the Intel HD Graphics 4000 in my Ivy Bridge laptop, at least when you compare it in GFLOPS. Mobile GPU's are actually getting really good.

 

[jouse16]

The Real Blastdoor[/url]":1aua2e3p]

sonicmerlin[/url]":1aua2e3p]

ichemandrew[/url]":1aua2e3p]Core M really doesn't seem that impressive for the price. It also seems strange that it needs a fan. Even at inaudible speed, that's still another watt or two thrown away, which matters for a small battery.

Intel's reference 10" Broadwell tablet was fanless and its GPU benchmarks were much better than what the Nvidia K1 and the iPad Air 2 can do.

http://www.anandtech.com/show/8515/quic ... a-mountain

Lenovo did something very wrong here. I've read speculation they've capped the TDP at 2.5 W and that's causing a ton of throttling.

A few notes on that anandtech article:

1. So far as I can tell, it wasn't a review -- they just put Intel marketing benchmarks in a graph next to results they achieved themselves when reviewing other products.

2. That reference tablet weighs 45% more than a 2013 iPad Air.

3. Intel actually set the TDP in this reference tablet to 6W rather than 4.5.

So.... hmmm.

Wow, Intel set the TDP to 6 W? That's even more impressive. That shows what a good cooling design can do. Lenovo really did some bad engineering when they designed the cooling system on the Yoga 3 Pro that's larger than Intel's reference design and still has a fan.

 

[melgross]

심돌산[/url]":rbnn2zy8]

Joriarty[/url]":rbnn2zy8]

dmsilev[/url]":rbnn2zy8]

GlennHowes[/url]":rbnn2zy8]Am I right in reading this that the graphics (GFXBench Offscreen) of the Yoga 3 are now significantly slower than an iPad Air 2? And that the GeekBench Multi-core has the iPad a bit faster? For other things like the web benchmarks, the Yoga is faster, sometimes by a lot.

If I were Intel I'd be worried.

http://meincmagazine.com/apple/2014/10/th ... y-package/

Well, the A8X has three cores vs. two on this Broadwell chip, so Intel still has a significant edge there. Especially when you consider that these chips are the bottom tier of performers in the Core line and the A8X is Apple's faster chip to-date (and is near the top of the current generation of mobile ARM chips). The graphics, on the other hand, is kind of startling; is 'Manhattan off-screen' the same test on x86 and ARM? If it is, then the iPad is out-pointing even the gruntier Haswell chip, never mind the Broadwell.

Also - although pricing for the Core M isn't public, I've read it's in the $300 range.

So it seems that A8X is an order of magnitude cheaper, certainly consumes less power, and might also be faster.

ARM is kicking some serious ass. It's a pity that architecture changes aren't an entirely smooth affair.

If you look at these benchmarks side by side with the benchmarks from Ars' iPad Air 2 review, you will see that the Yoga 3 wins everything single-threaded, usually by a large margin.

Apple has been doing great work, but so has Intel - and for a lot longer.

And you are mistaken if you think there is some inherent advantage to ARM. The instruction set really doesn't matter for two reasons. 1) x86-64 cleaned up a worst of the legacy cruft in the x86 architecture and 2) instruction decoding is the only part affected by the instruction set and that is a minuscule part of a modern Intel design.

You miss the point here. The fact that the A8x is close, which it is, is what's important. Intel is advancing performance just between 5-15% per generation, whereas Apple has been advancing performance between 49-109% per generation. Graphics performance is increasing faster. It was said, before Apple forced the hand of the ARM industry, that ARM could never, in any way, compete with x86. We can now see that that isn't true. In addition the OS and software for ARM is more efficient, because earlier ARM SoCs were so lacking in power.

While Apple may not be able to keep those rather large steps over the next few generations, it's clear that in another one or two, they will be there. Considering the difference in the cost of an Apple SoC, when compared to even a less expensive Intel oart, which still requires additional chips, ARM is much more efficient in a price performance measurement. If Apple were to use two of these chips, it would still cost considerable less than the Intel part, and would have far better graphics performance, and significantly better multi core performance.

If Apple just increases per core performance by 25% next year, it will equal an x86 part, easily. If they increase graphics by 50%, certainly the minimum we'll see, then that will be in Intel territory as well.

There's considered to still be a lot of cruft in Intel's architecture.

 

[melgross]

Callitrax[/url]":249dpnz6]

Joriarty[/url]":249dpnz6]
Also - although pricing for the Core M isn't public, I've read it's in the $300 range.

So it seems that A8X is an order of magnitude cheaper, certainly consumes less power, and might also be faster.

ARM is kicking some serious ass. It's a pity that architecture changes aren't an entirely smooth affair.

Is the A8X cheaper? The lowest price you can get one is $499. Yeah you get a lot more with that but you also get the Apple high margin markup at that point. With the Intel based systems the highest margin part is the CPU where the markup is applied earlier in the supply chain. Given that the A8X has 3 Billion transisters, and the die sizes seem to be in the same range, I'm doubting that the production cost of the A8X Chips is markedly different from that of the Core M, the markup is just applied at a different point in the chain. (Short version, analyzing cost structures is a pain, but the chips shouldn't really be materially different in production cost)

And we don't inherently know that it uses less power. I've never seen real power measurements of the IPad SOCs under load, and there are so many other power draws in an operating computer that it can be hard to compare. But Intel chips have been shown to work in the same power ranges as other ARM chips.

The iPad may cost $499, but the SoC is believed to cost about $30. Apple doesn't get a "high end" markup, they make a good profit, unlike most other manufacturers who are barely making any profit on their machines. We've known that for many years, it's why we've seen so much bloatware, which has been where most of the profit for Windows machines have come from. But this is an expensive machine.

We can look at the price of the Surface Pro 3 to see that using an x86 chip raises the price substantially. The low price Surface Pro 3, at $799, sans the required keyboard, uses an ultra low power i3, while the better performing model at $999 uses the i5 version. While they need to be spec'd up because they run Windows, which needs far more RAM and storage than do Android or iOS, a large part of that higher pricing are the x86 chips inside.

 

[CrymsonKyng]

A few years ago I told anyone who asked mobile socs would never be able to touch even laptop chips. Every year my foot is a little farther into my mouth.

 

[Mitlov]

Xiao-zhi[/url]":2t827h8f]About that hinge

More than 600 parts?

I think Lenovo just out-gimmicked Samsung, and that takes some doing.

Well, it might be good news for iFixit because that's a lot of little spare parts to buy.

Pure number of parts has very little to do with durability. A drive chain (hundreds of pieces) made out of durable materials will far out-last a drive belt (one piece) made out of flimsy materials. A well-made band to a wristwatch can last decades or longer. Some very simple laptop hinges can break quickly if they're flimsy or made with poor-quality plastics.

And Samsung's laptops have historically been very traditional and conservative in design. It's Dell, Lenovo, and Sony which have really been creative on the convertible hinge front, not Samsung. So even if parts-number-count was a useful test for reliability, your flame makes no sense.

 

[Callitrax]

melgross[/url]":3picfup9]

Callitrax[/url]":3picfup9]

Joriarty[/url]":3picfup9]
Also - although pricing for the Core M isn't public, I've read it's in the $300 range.

So it seems that A8X is an order of magnitude cheaper, certainly consumes less power, and might also be faster.

ARM is kicking some serious ass. It's a pity that architecture changes aren't an entirely smooth affair.

Is the A8X cheaper? The lowest price you can get one is $499. Yeah you get a lot more with that but you also get the Apple high margin markup at that point. With the Intel based systems the highest margin part is the CPU where the markup is applied earlier in the supply chain. Given that the A8X has 3 Billion transisters, and the die sizes seem to be in the same range, I'm doubting that the production cost of the A8X Chips is markedly different from that of the Core M, the markup is just applied at a different point in the chain. (Short version, analyzing cost structures is a pain, but the chips shouldn't really be materially different in production cost)

And we don't inherently know that it uses less power. I've never seen real power measurements of the IPad SOCs under load, and there are so many other power draws in an operating computer that it can be hard to compare. But Intel chips have been shown to work in the same power ranges as other ARM chips.

The iPad may cost $499, but the SoC is believed to cost about $30. Apple doesn't get a "high end" markup, they make a good profit, unlike most other manufacturers who are barely making any profit on their machines. We've known that for many years, it's why we've seen so much bloatware, which has been where most of the profit for Windows machines have come from. But this is an expensive machine.

We can look at the price of the Surface Pro 3 to see that using an x86 chip raises the price substantially. The low price Surface Pro 3, at $799, sans the required keyboard, uses an ultra low power i3, while the better performing model at $999 uses the i5 version. While they need to be spec'd up because they run Windows, which needs far more RAM and storage than do Android or iOS, a large part of that higher pricing are the x86 chips inside.

My point was that we don't know WHAT the A8X costs, because nobody can buy one. $30 dollars is the number people like to throw around for the high end ARM chips. And I don't know that may even be what Apple pays TSMC to fab it. Also remember its a lot bigger than the regular A8, so if the A8 is $30 this may be more like $40.

But if Apple were to sell them alone? They would charge a lot more than that. Why? Because they could. Same reason Intel charges what they do for their mobile processors. Production wise the Core M should be relatively similar to the I3 (Same CPU and GPU config) which sells for as low as $100. The Core M will be a little more because it is packaged the extra southbridge chip - which is much smaller and on a process intel is really good at by now, and more CPUs won't make the cut for power reasons (although those just get sold as higher TDP units).

But production wise the A8X and the Come M shouldn't be too dissimilar. Apple is different from the other ARM vendors, they make bigger chips to drive lower power at the same or better performance. Since die size is the main driver of cost in the actual production of a CPU trying to judge to cost of Apple chips by using those from Qualcomm or Samsung as analogues is not ideal. The A8X is about 123mm^2. The Core M, well I haven't seen a die size yet, but its probably in the same vicinity - bigger versions of Haswell were in the 177-183mm^2 range with more cores or GPU units, and on a bigger process, so Core M should be much smaller and probably fairly similar to A8X. (See the image of Haswell and Broadwell dies here - the Core M is much smaller).

Basically the cost of producing those chips should be in the same ballpark, but comparing how much they "cost" the consumer is borderline impossible. The companies are very dissimilar, Apple sell systems and not chips, Intel chips but no systems. They've both dumped huge amounts of intellectual capital into developing the chips which is of coarse hard to quantify in the cost contribution. And both companies like high profit margins so they both have big markups on what they sell. Yes tablet and variants with the Intel chips sell for more than a base IPad - but they also have other things driving up costs - large real SSDs, more RAM, etc.

Short version, the A8X and Core M are going to be similar in production costs, the Core M is probably more, but its maybe 50%, not the order of magnitude that some might suggest. After that its all business decisions and differences that influence cost to user. Yes the Intel CPU is a large fraction of the cost of a system using it. But given that Apple sells at large profit margins and their SoCs are one of the prime differentiators from other products (along with IOS) its possible to attribute some of that selling price to the CPU.

 

[Jon Ghast]

Callitrax[/url]":tlahuylh]Short version, the A8X and Core M are going to be similar in production costs, the Core M is probably more, but its maybe 50%, not the order of magnitude that some might suggest. After that its all business decisions and differences that influence cost to user. Yes the Intel CPU is a large fraction of the cost of a system using it. But given that Apple sells at large profit margins and their SoCs are one of the prime differentiators from other products (along with IOS) its possible to attribute some of that selling price to the CPU.

But if they can't compare the single unit retail/OEM price of Intel chips to the per unit cost Apple pays on multi- million piece orders, how can they make Apple look good in articles that have nothing to do with Apple?