Intel Broadwell Thread

[dahorns]

Dell Venue 11 pro with the 5y10 has appeared on geekbench. It puts up the best score yet for the 5y10.

http://browser.primatelabs.com/geekbench3/1319713

2180 Single
4506 Multi

This basically matches the very best multi core score for the Yoga 3 with 5y70.

Pretty much crushes the Air 2 in everything but SHA.

 

[dahorns]

Same system compared to the best score from the previous Venue 11 pro with the i5 4300y

http://browser.primatelabs.com/geekbench3/compare/728160?baseline=1319713

And compared to the atom version:

http://browser.primatelabs.com/geekbench3/compare/547141?baseline=1319713

 

[Sweepr]

Dell Venue 11 pro with the 5y10 has appeared on geekbench. It puts up the best score yet for the 5y10.

http://browser.primatelabs.com/geekbench3/1319713

2180 Single
4506 Multi

This basically matches the very best multi core score for the Yoga 3 with 5y70.

Nice find dahorns. Time to get past the unimpressive Yoga 3 Pro scores, I think we can expect better results from other designs.
Here's some scores of today's best 28nm/20nm ARM SoCs (GeekBench 3 ST/MT score):

Exynos 7 Octa 5433 (1.9GHz 4C + 1.3GHz 4C A57/A53 @ Android 32-bit 20nm): 1282 / 4334
Tegra K1 Denver (2.5GHz 2C @ Android 64-bit 28nm): 2216 / 3819
Apple A8 (1.4GHz 2C @ iOS 64-bit 20nm): 1617 / 2904
Apple A8X (1.5GHz 3C @ iOS 64-bit 20nm): 1812 / 4477

 

[III-V]

Unfortunately, Ars doesn't seem to standardize their screen brightness test on an absolute brightness, so it's hard to tell what's really going on here. One thing this does highlight, though, is that the display probably has a larger impact on battery life than the CPU ever did. Intel might be focusing on a problem (CPU power consumption) that doesn't really exist in this market segment, at the expense of performance.

Well, they've tried other avenues. They've been pushing PSR forever. But... the backlight thing is bizarre. Why would a backlight kill the battery so much? They don't need a stronger one for a higher pixel count display, do they? That should stay constant, and really, with LEDs improving in efficiency over time, shouldn't be regressing either.

 

[pw257008]

Well, they've tried other avenues. They've been pushing PSR forever. But... the backlight thing is bizarre. Why would a backlight kill the battery so much? They don't need a stronger one for a higher pixel count display, do they? That should stay constant, and really, with LEDs improving in efficiency over time, shouldn't be regressing either.

I don't know, both my phone and my laptop run down much faster at max brightness vs. min brightness, and power saving mode in Windows turns the backlight down.

 

[III-V]

Well, I just looked at mobile device consumption charts, broken down by component... It definitely looks like the backlight can indeed cut battery life in half if you go to max brightness. Man, LEDs need a breakthrough.

 

[witeken]

Well, I just looked at mobile device consumption charts, broken down by component... It definitely looks like the backlight can indeed cut battery life in half if you go to max brightness. Man, LEDs need a breakthrough.

We've just had (or having) IGZO, which cut power consumption in half.

 

[sefsefsefsef]

Well, they've tried other avenues. They've been pushing PSR forever. But... the backlight thing is bizarre. Why would a backlight kill the battery so much? They don't need a stronger one for a higher pixel count display, do they? That should stay constant, and really, with LEDs improving in efficiency over time, shouldn't be regressing either.

A higher resolution screen actually DOES need a brighter backlight. This is because pixel size scales down faster than ... well, how to say this ... than the space "between" the pixels does. Imagine each pixel has a bezel around it. As the PPI of a display increases, then so does the amount of light being blocked by these "bezels." Therefore, for the user to perceive the same brightness, the backlight itself needs to actually be brighter on a high PPI display.

 

[NTMBK]

No really, CPU power consumption does matter in tablets. I have a Haswell Venue 11 Pro- even with a fan, that thing gets noticeably hot when you load the APU. The top right corner gets toasty to the touch.

Really impressed that they seem to have fitted the same performance into a fanless tablet this year. It's a great device.

 

[Fjodor2001]

We've just had (or having) IGZO, which cut power consumption in half.

Not seen any nearly any laptops or tablets yet though. I wonder why, because it looks promising.

 

[III-V]

We've just had (or having) IGZO, which cut power consumption in half.

Right, but that cuts the power for the transistors, not the backlight. What we really need now is better LEDs. Wonder how LED backlight efficiency has been progressing... I know it's been improving steadily for home lighting (especially cost), and they're the same color LEDs, but they're separate markets and have been progressing at different rates, perhaps.

Not seen any nearly any laptops or tablets yet though. I wonder why, because it looks promising.

$$$

A higher resolution screen actually DOES need a brighter backlight. This is because pixel size scales down faster than ... well, how to say this ... than the space "between" the pixels does. Imagine each pixel has a bezel around it. As the PPI of a display increases, then so does the amount of light being blocked by these "bezels." Therefore, for the user to perceive the same brightness, the backlight itself needs to actually be brighter on a high PPI display.

Ah, that's interesting.

 

[Roland00Address]

Not seen any nearly any laptops or tablets yet though. I wonder why, because it looks promising.

Several dell laptops since 2013 has included them main on the xps and business laptops. Sharp started creating izgo laptop screens in 2012.

 

[Haserath]

Not seen any nearly any laptops or tablets yet though. I wonder why, because it looks promising.

Apple's iPad Air was the first IGZO screen. Sharp was having issues beforehand, which is the reason the iPad3 wasn't igzo, so Apple most likely took all the volume for awhile. And with it being a new technology and OEMs being in the 'cheap game', I'm not too surprised that many haven't been introduced yet.

Right, but that cuts the power for the transistors, not the backlight. What we really need now is better LEDs. Wonder how LED backlight efficiency has been progressing... I know it's been improving steadily for home lighting (especially cost), and they're the same color LEDs, but they're separate markets and have been progressing at different rates, perhaps.

IGZO does reduce backlight power. The parts of the display that block backlight are significantly reduced(wires and transistors). IGZO is able to use 50% less power with less flicker.

Low temperature poly silicon is actually one step further than this but is more expensive.

What interests me more is the fact that Intel has been experimenting with similar materials for 10nm.

 

[Nothingness]

Dell Venue 11 pro with the 5y10 has appeared on geekbench. It puts up the best score yet for the 5y10.

http://browser.primatelabs.com/geekbench3/1319713

2180 Single
4506 Multi

This basically matches the very best multi core score for the Yoga 3 with 5y70.

Pretty much crushes the Air 2 in everything but SHA.

Your last two sentences are in contradiction since the best Yoga 3 score is 2567/4570 vs 11 pro 2180/4506 vs iPad Air 2 1824/4575. The single thread score difference is similar between 5Y10 and 5Y70, as it is between 5Y10 and A8X

That 7140 seems to have good battery life with the keyboard, it's a pity the screen resolution isn't better (this would likely have an impact on battery life).

 

[Nothingness]

Nice find dahorns. Time to get past the unimpressive Yoga 3 Pro scores, I think we can expect better results from other designs.
Here's some scores of today's best 28nm/20nm ARM SoCs (GeekBench 3 ST/MT score):

Exynos 7 Octa 5433 (1.9GHz 4C + 1.3GHz 4C A57/A53 @ Android 32-bit 20nm): 1282 / 4334
Tegra K1 Denver (2.5GHz 2C @ Android 64-bit 28nm): 1932 / 3432
Apple A8 (1.4GHz 2C @ iOS 64-bit 20nm): 1617 / 2904
Apple A8X (1.5GHz 3C @ iOS 64-bit 20nm): 1812 / 4477

Tegra K1 Denver / Nexus 9 is getting 2216/3819. I guess your score is for 32-bit rather than 64-bit.

 

[mavere]

Tegra K1 Denver / Nexus 9 is getting 2216/3819. I guess your score is for 32-bit rather than 64-bit.

Even when looking at just 64-bit, the Tegra K1's results are all over the place. In contrast, this has reasonable, reliable clustering of results.

 

[NTMBK]

Your last two sentences are in contradiction since the best Yoga 3 score is 2567/4570 vs 11 pro 2180/4506 vs iPad Air 2 1824/4575. The single thread score difference is similar between 5Y10 and 5Y70, as it is between 5Y10 and A8X

That 7140 seems to have good battery life with the keyboard, it's a pity the screen resolution isn't better (this would likely have an impact on battery life).

Honestly, I don't see why you need more than 1080p in a 11" device. It looks great to me, and 1080p content looks lovely because it isn't scaled at all.

 

[Nothingness]

Honestly, I don't see why you need more than 1080p in a 11" device. It looks great to me, and 1080p content looks lovely because it isn't scaled at all.

Personal preference: I see extremely well at short distance (I use 8pt fonts for programming on my 27" screen) and I don't use tablets to watch videos.

@mavere: Very ugly indeed.

 

[Sweepr]

Tegra K1 Denver / Nexus 9 is getting 2216/3819. I guess your score is for 32-bit rather than 64-bit.

It was actually 64-bit, but for some reason Nexus 9 results vary a lot. Updated anyway, thanks.

 

[386DX]

It would be pretty easy for the reviewers to do a battery life test without the display (brightness, type, resolution) affect the battery life, just use an external monitor and turn off the internal LCD. Yes you won't get real life runtime but if you're testing how much power improvement the CPU/platform has over the previous gen eliminating as much variables as you can is the first step.

 

[Fjodor2001]

And with it being a new technology and OEMs being in the 'cheap game', I'm not too surprised that many haven't been introduced yet.

The extra money you spend on the IGZO-display can be saved elsewhere. Because it reduces overall power consumption, so you can get by with e.g. a smaller battery and less expensive/power-efficient CPU.

I hope we'll be seeing some more devices with IGZO display in early 2015 and onwards. If it can reduce power consumption as much as promised it is a much more important factor in improving battery life than making the CPU more power efficient.

I assume OLED would be even better though...

 

[Roland00Address]

It would be pretty easy for the reviewers to do a battery life test without the display (brightness, type, resolution) affect the battery life, just use an external monitor and turn off the internal LCD. Yes you won't get real life runtime but if you're testing how much power improvement the CPU/platform has over the previous gen eliminating as much variables as you can is the first step.

Many cpus/computers turn their gpu clocks on maximum when you connect an external display even if you turn off the primary display. The reason why they assume you want the best experience if you are connecting an external display.

This is especially obvious on some desktop GPUs which have loud fans. Just running two monitors even if the monitor is a blank background will rev up the clocks, which raise temperatures, which rev the fans.

 

[rushmore]

Your last two sentences are in contradiction since the best Yoga 3 score is 2567/4570 vs 11 pro 2180/4506 vs iPad Air 2 1824/4575. The single thread score difference is similar between 5Y10 and 5Y70, as it is between 5Y10 and A8X

That 7140 seems to have good battery life with the keyboard, it's a pity the screen resolution isn't better (this would likely have an impact on battery life).

My 7140 should arrive Tuesday and will compare to my 5130. Same keyboard dock works, so will just switch them out. Still seems the Yoga 3 anchor is the high res display. It drags the cpu down due to the balancing of power needed by the GPU and also hits battery life.

I prefer 1080p for PC devices and smaller displays. Heck, my LED desktop display is 1080p and is 27". Makes my Note 4's Qhd display seem even more ironic.

 

[rushmore]

Your last two sentences are in contradiction since the best Yoga 3 score is 2567/4570 vs 11 pro 2180/4506 vs iPad Air 2 1824/4575. The single thread score difference is similar between 5Y10 and 5Y70, as it is between 5Y10 and A8X

That 7140 seems to have good battery life with the keyboard, it's a pity the screen resolution isn't better (this would likely have an impact on battery life).

Considering I am happy with a 27" 1080p desktop display, 1080p on a 11" display is laser sharp

I prefer the 1080p since more efficient for power and though joking, I am fine with my 27" desktop display at 1080p. The Yoga 3's battery and performance anchor is the higher res display.

 

[Haserath]

The extra money you spend on the IGZO-display can be saved elsewhere. Because it reduces overall power consumption, so you can get by with e.g. a smaller battery and less expensive/power-efficient CPU.

I hope we'll be seeing some more devices with IGZO display in early 2015 and onwards. If it can reduce power consumption as much as promised it is a much more important factor in improving battery life than making the CPU more power efficient.

I assume OLED would be even better though...

Maybe if it isn't too exensive. Apple could afford a bit of a hit to reduce heat and weight in their tablet. And they went from 42Wh to 32Wh without losing battery life or even gaining a bit(can be attributed to other changes like 32->28nm and that thing they did with stunning performance). The Air 2 uses a 20nm SoC and a 28Wh battery yet it loses battery life compared to the Air, so the power reduction mostly came down to the screen.

I'm not sure how good OLED is comparatively; Samsung has been progressing for the last few years. It still uses a lot of power to display white pages but images and blacks are most likely better; Samsung even has a setting on their phones to create more black on the screen for savings. It would seem like they could dynamically adjust the brightness for white to use a bit less, who knows...