The Intel Atom Thread

Page 141 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Sweepr

Diamond Member
May 12, 2006
5,148
1,142
131
This is very, very insteresting

This company is pulling off a surprise every day. Looks like an awesome Windows-based Chromebook alternative, somehow they manage to pack a 1920x1200 IPS screen in a $190 notebook. I know I've already said this in Broadwell's thread but time to take a note from Pipo and update the 1440x900/1366/768 TN screens from the >$899 Macbook Air Apple. :p

Dell Venue 8 7000 Reviews

dellvenue87000-14.0.jpg


The Verge said:
"Intel's processor performs just as well as the best from Qualcomm"

In addition to its unique design, the Venue 8 uses an Intel processor, a relative rarity among mobile devices. It’s a 2.3GHz, quad-core Atom chip paired with 2GB of RAM and it provides a fast and responsive experience. If the Venue 8 didn’t have a big Intel logo emblazoned on the back of it, I wouldn’t have assumed it was powered by anything different than the standard Qualcomm fare. Games played smoothly, and though the tablet got warm in spots, it never got uncomfortable to hold (unlike the Nexus 9). The Atom chip is power-efficient, too: the Venue 8 lasted for 9 hours and 15 minutes on our rundown test and had no problem going for multiple days in normal usage.

www.theverge.com/2015/2/9/8004117/dell-venue-8-7000-tablet-review

Notebookcheck said:
Performance is pretty good overall: in the single-core portion of the Geekbench 3 benchmark test, our review tablet performs exactly like the Samsung Galaxy Tab S 8.4 with the Exynos 5420 Octa CPU and ARM Mali-T628 MP6 GPU; the Apple iPad Mini 3 with its 64-bit A7 CPU performs over 45 % better. The multi-core performance is once again similar to the Samsung, but this time ahead of the iPad Mini 3 (~ + 20 %), which is only equipped with a dual core processor. The results of the graphic-intensive 3DMark 2013 Ice Storm Unlimited are pretty compelling: the Venue 8 7000 manages to distance itself from the Galaxy Tab S 8.4 by a fair margin (about 50 % better performance). Apple's iPad Mini 3 generally offers less performance as well even though it is equipped with the same GPU. The most likely reason: the clock speed in Apple's implementation is a bit lower. The performance of the Nvidia Shield Tablet with Tegra K1 Kepler GPU remains untouched.

Idle
Dell Venue 8 7000: 1790 minutes
iPad Mini 3: 1215
Galaxy Tab S 8.4: 1172
Nexus 9: 1422

Load
Dell Venue 8 7000: 402 minutes
iPad Mini 3: 210
Galaxy Tab S 8.4: 250
Nexus 9: 224


The Venue 8 7000 comes with a 5900 mAh battery which provides 21 Wh. The Google Nexus 9 (25.5 Wh) and the Apple iPad Mini 3 (24 Wh) both have larger batteries; the Samsung Galaxy Tab's battery has a lower capacity of under 19 Wh.

Notebookcheck said:
The Venue 8 7000 only reached a maximum of 25 °C during idle, which is an excellent result. The Google Nexus 9, as an example, runs a good 0 °C hotter; the Galaxy Tab S 8.4 and the iPad Mini 3 reached 31 °C and 32 °C, respectively. During the stress test, the Venue topped out at just over 31 °C maximum, which is far lower than the other tablets in our comparison. Next closest is Apple's iPad 3 at 37 °C, the other two tablets reached almost 40 °C.

www.notebookcheck.net/Dell-Venue-8-7000-Tablet-Review.135578.0.html

Battery life @ idle and load destroyed the iPad mini 3 and Galaxy Tab S 8.4 while WiFi was just a bit lower. Looks like 22nm Moorefield is quite efficient compared to 28nm ARM competitors. Meanwhile performance is very respectable and despite being the thinnest tablet out there it doesn't get nearly as warm as competitors.
 
Last edited:

witeken

Diamond Member
Dec 25, 2013
3,899
193
106
Moorefield looks like a good FinFET SoC. I hope it gets some more prestigious design wins. It Intel can finally pull ahead with 14nm Broxton and <1 year afterwards the 10nm die shrink, Intel could finally gain some momentum without delays etc.
 

liahos1

Senior member
Aug 28, 2013
573
45
91
Moorefield looks like a good FinFET SoC. I hope it gets some more prestigious design wins. It Intel can finally pull ahead with 14nm Broxton and <1 year afterwards the 10nm die shrink, Intel could finally gain some momentum without delays etc.

whats crazy is that arm just put out their earnings report and gave their outlook for the rest of the year and they said they expect to lose no market share to intel in mobile haha. it is just baffling. hopefully we see some sofia based product out soon,.
 
Mar 10, 2006
11,715
2,012
126
whats crazy is that arm just put out their earnings report and gave their outlook for the rest of the year and they said they expect to lose no market share to intel in mobile haha. it is just baffling. hopefully we see some sofia based product out soon,.

This is interesting...

9 ARMv8-A processor licences signed, including an architecture licence for high performance computing
 

Nothingness

Platinum Member
Jul 3, 2013
2,371
713
136
whats crazy is that arm just put out their earnings report and gave their outlook for the rest of the year and they said they expect to lose no market share to intel in mobile haha. it is just baffling. hopefully we see some sofia based product out soon,.
That's crazy, I wonder why ARM just don't surrender to almighty Intel, after all it's not as if ARM results were good.
 
Mar 10, 2006
11,715
2,012
126
That's crazy, I wonder why ARM just don't surrender to almighty Intel, after all it's not as if ARM results were good.

Intel's mobile product portfolio isn't really good enough in 2015 to grow share meaningfully. SoFIA/Cherry Trail will primarily be used to replace the 22nm Bay Trails in designs that Intel already had at a better cost structure, at least this is the impression Hermann Eul gave at the 2014 investor meeting.

SoFIA might win some share in smartphones, but I don't think it'll be all that huge...the first SoFIA is a 3G part, so market for that is limited. Second SoFIA is coming in mid-2015 which could do better, but in the scheme of the kinds of volumes MediaTek/Qualcomm have, and in light of Intel's mobile guidance ($800M reduction in loss, much of reduction due to contra-revenue evaporating, in addition to opex custs implies not a lot of unit shipment growth), I don't think they'll be all that huge.

IMHO, ARM's assessment is a fair one.
 
Last edited:

liahos1

Senior member
Aug 28, 2013
573
45
91
That's crazy, I wonder why ARM just don't surrender to almighty Intel, after all it's not as if ARM results were good.

its not about surrendering or anything. its about being realistic. last year nobody assumed intel would take market share and they took 25% of the market in the 2h (albeit at some cost). ARM has literally 99% share in mobile phones and their going assumption is that they are going to keep that. just doesnt seem realistic.
 

III-V

Senior member
Oct 12, 2014
678
1
41
its not about surrendering or anything. its about being realistic. last year nobody assumed intel would take market share and they took 25% of the market in the 2h (albeit at some cost). ARM has literally 99% share in mobile phones and their going assumption is that they are going to keep that. just doesnt seem realistic.
I think it's a reasonable expectation for this year. Intel's probably going to be treading water this year, whereas previously they were drowning. Next year they'll probably make some more progress.
 

liahos1

Senior member
Aug 28, 2013
573
45
91
I think it's a reasonable expectation for this year. Intel's probably going to be treading water this year, whereas previously they were drowning. Next year they'll probably make some more progress.

I think it depends on how sofia 3g/lte along with jv soc's do at the low end. Intel won back Lenovo this year at the premium end. and those phones have yet to be released. I am not saying intel is going to take the type of share they took in tablets in 2014. I just would be surprised if they literally gained no share in phones this year. then again this company (at least in mobile) has had its share of fits and starts and has disappointed in the past. I guess we will wait and see.
 

witeken

Diamond Member
Dec 25, 2013
3,899
193
106
I'd love to see that, but I'm not holding my breath.

The Trail platform is dead. Even Atom for desktops is (re)named (to) Braswell. Intel's 2 platforms for the coming years have yet to launch (maybe except for Moorefield, but that doesn't have Intel graphics). Intel wanted to start the Tick-Tock pragmatism at 22nm, but failed because of 14nm and apparently things are simply taking longer than anticipated. But once Tick-Tock gets going (I'm still quite optimistic on the 2 year cadence), how could you possibly stop Intel, a company with an astonishing manufacturing skill and the omnipotent Core architecture and a healthy Atom pipeline as assets?

The process node has an SoC flavour since 22nm and a leading density since 14nm. Both architectures benefit from that and are designed anyway with an unparalleled dynamic range in mind, so no useless Big/Little approach is needed. To compare: Intel's 2014 contra-revenue budget is ARM's full year revenue or so.
 

liahos1

Senior member
Aug 28, 2013
573
45
91
Did anyone notice ARM called out Rockchip as an a72 licensee.

"Increasingly, consumers are adopting mobile devices including smartphones, tablets, phablets and other large screen devices as their primary compute platforms. These larger form factor devices demand higher performance and energy-efficiency that scales across a variety of processor configurations," said Mr. Feng Chen, chief marketing officer, Rockchip. "Rockchip is pleased to partner with ARM to introduce the Cortex-A72 to enable a wide range of premium mobile platforms that consumers can take advantage of. Rockchip is dedicated to bringing cutting-edge CPU and GPU technologies and solutions to our customers and end users to enable a compelling user experiences for personal and enterprise use."

This doesnt seem to fit within the framework of eventually transitioning rockchip to IA. Anyone have any conspiracy theories/insight/thoughts/speculation they want to share?
 

witeken

Diamond Member
Dec 25, 2013
3,899
193
106
This doesnt seem to fit within the framework of eventually transitioning rockchip to IA. Anyone have any conspiracy theories/insight/thoughts/speculation they want to share?

Sure.

Will you find A72 in low-end devices where SoFIA is meant to go? No.
This begs the question: will Intel license Broxton to other companies? I don't think so. This is probably because the low-end phone market doesn't have a good ROI, so they will leave this to other companies to exploit to gain market share for IA (because there's obviously a lot of volume in the low-end).
 

III-V

Senior member
Oct 12, 2014
678
1
41
I think it depends on how sofia 3g/lte along with jv soc's do at the low end. Intel won back Lenovo this year at the premium end. and those phones have yet to be released. I am not saying intel is going to take the type of share they took in tablets in 2014. I just would be surprised if they literally gained no share in phones this year. then again this company (at least in mobile) has had its share of fits and starts and has disappointed in the past. I guess we will wait and see.
I think it's inevitable that Intel's absolute sales numbers will go up this year, but their relative sales numbers, compared to ARM, may not. It's a growing market, and a rising tide lifts all boats. I do think Intel will gain market share from ARM this year, but it will be small enough to be of no real significance.
 

coercitiv

Diamond Member
Jan 24, 2014
6,151
11,686
136
Intel wanted to start the Tick-Tock pragmatism at 22nm, but failed because of 14nm and apparently things are simply taking longer than anticipated. But once Tick-Tock gets going (I'm still quite optimistic on the 2 year cadence), how could you possibly stop Intel, a company with an astonishing manufacturing skill and the omnipotent Core architecture and a healthy Atom pipeline as assets? [...] To compare: Intel's 2014 contra-revenue budget is ARM's full year revenue or so.
Indeed, apparently things are simply taking longer than anticipated.
 
Mar 10, 2006
11,715
2,012
126
Sure.

Will you find A72 in low-end devices where SoFIA is meant to go? No.
This begs the question: will Intel license Broxton to other companies? I don't think so. This is probably because the low-end phone market doesn't have a good ROI, so they will leave this to other companies to exploit to gain market share for IA (because there's obviously a lot of volume in the low-end).

And yet, BK has publicly stated that Rockchip/Spreadtrum will eventually transition to IA across the board ;)
 

erunion

Senior member
Jan 20, 2013
765
0
0
And yet, BK has publicly stated that Rockchip/Spreadtrum will eventually transition to IA across the board ;)

I wasn't aware of that.

Though I imagine the A72 is a very small invest from rockchip, considering ARMs business model of charging per core built.
 

liahos1

Senior member
Aug 28, 2013
573
45
91
Sure.

Will you find A72 in low-end devices where SoFIA is meant to go? No.
This begs the question: will Intel license Broxton to other companies? I don't think so. This is probably because the low-end phone market doesn't have a good ROI, so they will leave this to other companies to exploit to gain market share for IA (because there's obviously a lot of volume in the low-end).

i think this explanation makes sense. but isnt it odd that on one hand you just jv'd with a partner to push intel product while on the other hand you compete with that customer in another segment of the market

its sort of like a mini apple relationship. if i could only spend 1 hour with BK and the team that works on ATOM...id have a lot of questions
 

svarog19

Member
Feb 11, 2015
32
0
0
This company is pulling off a surprise every day. Looks like an awesome Windows-based Chromebook alternative, somehow they manage to pack a 1920x1200 IPS screen in a $190 notebook. I know I've already said this in Broadwell's thread but time to take a note from Pipo and update the 1440x900/1366/768 TN screens from the >$899 Macbook Air Apple. :p

Dell Venue 8 7000 Reviews

dellvenue87000-14.0.jpg




www.theverge.com/2015/2/9/8004117/dell-venue-8-7000-tablet-review





www.notebookcheck.net/Dell-Venue-8-7000-Tablet-Review.135578.0.html

Battery life @ idle and load destroyed the iPad mini 3 and Galaxy Tab S 8.4 while WiFi was just a bit lower. Looks like 22nm Moorefield is quite efficient compared to 28nm ARM competitors. Meanwhile performance is very respectable and despite being the thinnest tablet out there it doesn't get nearly as warm as competitors.

You forgot that those tablets have 3G or 4G which increases consumption and not to mention higher quality lower latency higher resolution screens which consume considerably more than screen used in that tablet.

If you have put screen on the same level and a 3G or 4G chip then battery life would drop like a rock.
 

Sweepr

Diamond Member
May 12, 2006
5,148
1,142
131
You forgot that those tablets have 3G or 4G which increases consumption and not to mention higher quality lower latency higher resolution screens which consume considerably more than screen used in that tablet.

If you have put screen on the same level and a 3G or 4G chip then battery life would drop like a rock.

iPad mini 3 and Nexus 9 pack lower-resolution LCD screens (Dell Venue 8 7000 has a 2560x1600 OLED panel). The Nexus 9 they tested is WiFi only and even if their iPad mini 3 and Galaxy Tab S 8.4 support LTE nowhere does it say that they were activated and using a 4G/LTE connection. And again, both Nexus 9 and iPad mini 3 pack larger batteries than the Dell Venue 8 7000. So yes, I'm impressed.
 
Last edited:

Bryf50

Golden Member
Nov 11, 2006
1,429
51
91
You forgot that those tablets have 3G or 4G which increases consumption and not to mention higher quality lower latency higher resolution screens which consume considerably more than screen used in that tablet.

If you have put screen on the same level and a 3G or 4G chip then battery life would drop like a rock.

It has literally the same screen as the galaxy tab and a higher res screen than the ipad mini. Nowhere does it say they tested against the 3G/4G models of the other tablets and even if they were they would likely have the modems off anyway.
 

erunion

Senior member
Jan 20, 2013
765
0
0
The Trail platform is dead. Even Atom for desktops is (re)named (to) Braswell. Intel's 2 platforms for the coming years have yet to launch (maybe except for Moorefield, but that doesn't have Intel graphics). Intel wanted to start the Tick-Tock pragmatism at 22nm, but failed because of 14nm and apparently things are simply taking longer than anticipated. But once Tick-Tock gets going (I'm still quite optimistic on the 2 year cadence), how could you possibly stop Intel, a company with an astonishing manufacturing skill and the omnipotent Core architecture and a healthy Atom pipeline as assets?

The process node has an SoC flavour since 22nm and a leading density since 14nm. Both architectures benefit from that and are designed anyway with an unparalleled dynamic range in mind, so no useless Big/Little approach is needed. To compare: Intel's 2014 contra-revenue budget is ARM's full year revenue or so.

Honestly, Intel already has the right node. What they need are the right products. Tick-tock has become almost meaningless in mobile.

They have been iterating Baytrail like crazy. All sorts of SKUs have been released over its lifespan, with major BOM reductions. With the 14nm delay Sofia(and porting to TSMC) is looking smarter and smarter.

Basically, i think intel has realized that tick-tock isn't fast enough.
 

jdubs03

Senior member
Oct 1, 2013
377
0
76
It is going to be really interesting to see the Apple A9 bench when the next iPhone and iPad come out this year (safe guess is mid-September). With 14nmFF we could be seeing something that is damn near close to as high performing as Core M; maybe even the 5Y71 SKU. Apple is going to get a big boost literally, enabled by faster frequencies, and maybe even basing their next custom enhancements (cyclone's successor) off the Cortex-A72 core. A lot of stars are aligning to potentially challenge the top Broadwell Core M SKU. The only saving grace for Intel will probably be Skylake, and even here I'm not as confident as I used to be. I can see why there is a fuss about Apple abandoning Intel for lower spec'd devices, though I don't necessarily agree.

The atom line doesn't stand a chance I don't think; and that's why I think we've see the Broxton delay into 2016.