Nexus 9 Ships Today With 'Denver' Processor And Android 5.0 Lollipop

Status
Not open for further replies.

anthony8989

Distinguished
Feb 2, 2013
653
2
19,165
57
"The Nexus 9 is the first Android tablet to have an Nvidia Denver CPU that has very high single-threaded performance. However, it doesn't have a very high multi-threaded performance due to the fact that it's only a dual-core CPU."

What are you basing these claims on? If it's your own speculation then you know pretty damn little about the Denver K1, or in SoC's in general...

You do understand that the Denver K1 employs a 7 way superscalar, higher core frequency, and an L1 cache that blows the Snapdragon 805's clear out of the water. Early benchmarks are showing this chip beating out the Note 4, iPhone 6 +, S5 , pretty much everything on the market to date.

Honestly why make baseless assertions like that?
 

InvalidError

Titan
Moderator
Considering how bad my luck was with getting a fully functional N7-2013, how Google's forums were flooded with the same handful of complaints for the first many months from launch and how people buying them today still encounter many of the same launch issues, I would not be too surprised if Google was limiting initial availability to field-test it before risking another embarrassing flood of complaints.
 

rishiswaz

Honorable
Mar 10, 2012
564
0
11,060
31
"The Nexus 9 is the first Android tablet to have an Nvidia Denver CPU that has very high single-threaded performance. However, it doesn't have a very high multi-threaded performance due to the fact that it's only a dual-core CPU."

What are you basing these claims on? If it's your own speculation then you know pretty damn little about the Denver K1, or in SoC's in general...

You do understand that the Denver K1 employs a 7 way superscalar, higher core frequency, and an L1 cache that blows the Snapdragon 805's clear out of the water. Early benchmarks are showing this chip beating out the Note 4, iPhone 6 +, S5 , pretty much everything on the market to date.

Honestly why make baseless assertions like that?
Your support is rather baseless as well unless you have benchmarks to prove otherwise. He is going off of what is known, it is a dual core 64 bit SoC with a Kepler GPU. We really don't know how Denver is going to be until there are hard numbers behind it. Unless you have proof besides Nvidia's spec sheet you have no argument either.
The iPhone 6 outperforms the Galaxy S 5 in a lot of benchmarks, a dual core 1.3 Ghz A8 vs a quad core 2.4GHz Qualcomm Snapdragon 801. Denver may well blow everything out of the water but it is still a big question mark.
 

anthony8989

Distinguished
Feb 2, 2013
653
2
19,165
57
Your support is rather baseless as well unless you have benchmarks to prove otherwise. He is going off of what is known, it is a dual core 64 bit SoC with a Kepler GPU. We really don't know how Denver is going to be until there are hard numbers behind it. Unless you have proof besides Nvidia's spec sheet you have no argument either.
The iPhone 6 outperforms the Galaxy S 5 in a lot of benchmarks, a dual core 1.3 Ghz A8 vs a quad core 2.4GHz Qualcomm Snapdragon 801. Denver may well blow everything out of the water but it is still a big question mark.
Why don't you try googling the things you talk about before talking out of your ass?

I recommend googling " Nexus 9 review" or "Nvidia Denver K1 benchmarks"...
 

DSpider

Distinguished
Jan 10, 2009
531
0
18,980
0
16 GB is pathetic. 16 GB had my 2009-bought iPod touch 2G (5 years ago, which is a LONG time in technology terms). 16 GB flash sticks are dirt cheapt. What the hell, Google?

You place a Full HD movie on it and that's it, the OS takes up the rest.
 

house70

Splendid
Considering how all my Nexus devices had zero problems and that they were always the first to receive updates, I have no problem getting the new tablet. The old ones will be passed down in my family.
 

house70

Splendid


That's why you can get the 32GB version. Or you could get an USB OTG stick for your media.
 

house70

Splendid


SD support has been phased out since JB. It's up to individual manufacturers to keep it or not.
 

InvalidError

Titan
Moderator

The extra cost for 16GB is grossly marked up and Google got in the habbit of ripping USB storage support out of official Nexus firmware for their newer devices so USB OTG storage does not work natively; third-party software (such as Nexus Media Importer) or third-party firmware is required.

As for SD support getting "phased out", SD support is alive and well in Android: nearly every Android device aside from Google's own support it and that includes tons of sub-$100 tablets. Google and Apple are pretty much the only mobile device vendors who shun external/removable local storage.
 

somebodyspecial

Honorable
Sep 20, 2012
1,459
0
11,310
20


You just gave an example (iphone6) of WHY the author should have never said what he said. What we know is quad doesn't mean much (yet?) on mobile as Apple usually proves. Also as Anthony said there are benchmarks out already for Denver. Not to mention as extremetech says:
http://www.extremetech.com/computing/187862-nvidia-details-64-bit-denver-tegra-k1-claims-haswell-class-performance-for-first-64-bit-android-chip
"in reality a dual-core CPU should be more than capable of handling most workloads (most apps and games are still bottlenecked by single-threaded performance)."
Considering many games don't efficiently use a quad on even PC's yet (better perf at times, but not scoring 2x a dual regularly), I don't think dual is a problem on mobile yet...LOL.

http://hothardware.com/News/Leaked-Nexus-9-Benchmark-Results-Show-DenverBased-64Bit-Tegra-K1-Performing-Very-Well/
3166 for Nexus 9 multicore, 1903 single core.

iphone 6: geekbench single core:1624 muti core:2911
3-core iPad Air 2: Geekbench single core - 1812, Multi - 4477
http://www.redmondpie.com/ipad-air-2-features-3-cores-2gb-ram-55-faster-than-iphone-6-6-plus-68-than-ipad-air/

One thing not to forget here though, is this is 28nm Denver vs. 20nm Tricore apple for ipad2air. I can't wait for NV's shrink scores ;) The A8 is ~89mm^2 (a8x obviously bigger with 3rd core etc), K1 is 123mm^2 IIRC, so a shrink will probably come in under 89 unless they up gpu or cpu (meaning change gpu out for maxwell at 20nm, no idea if K1 will be shrunk or not). I can't remember if there is a difference in size of K1 32vs.64, I don't think Denver version has been sized yet, so don't quote me. Either way K1 is impressive here toppling A8 20nm in iphone6. Granted you'd have to cut some speed to match phone's size, but it's still going to probably match apple and this is PRE-SHRINK of K1.

http://browser.primatelabs.com/android-benchmarks
Galaxy Note 4 Single core 1161 (LOL, note this is better than K1 A15 at 1079, but also note S805 loses to this), multicore for note4 8 core is 3914. But I'd say while this is a winner here, how many apps effectively use 8 cores? ;) Also note the Note4 with S805 scores 3169 (K1 32 scores 3211 here). Again, I suspect there won't be many places a Denver will be hindered for being dual in a usage scenario.

http://www.engadget.com/2014/11/03/google-nexus-9-review/
Some odd scores in there, so I'm thinking Lollipop or something here needs a bit more polish maybe. CF/Quadrant/Sunspider here sucked vs. Shield, yet we know geekbench is virtual tie in multicore. We'll find out more as we get more reviews. Also I thought Denver was supposed to be 2.5Ghz, but this runs 2.3. Maybe they had to take a cut to ensure enough chips for an early/large release? We'll see as more Denver based devices come out I guess.
 

somebodyspecial

Honorable
Sep 20, 2012
1,459
0
11,310
20
http://www.cnet.com/products/google-nexus-9/2/
Another review of nexus 9 from cnet. Odd it scores an extra 1600 or so in 3dmark IS unlimited here.

Also as noted, I think some spit and polish is needed:
"For all its speedy prowess, the Nexus 9 wasn't without its hiccups. Occasionally, apps launched to a blank, unresponsive screen or quickly loaded only to consistently crash. Once, it restarted on me without prompt after I unlocked the screen and too frequently I encountered lagging after simply swiping to navigate the user interface.

It's reasonable to believe that a fair share of the bugginess I encountered, despite the tablet's state-of-the-art specs, is because Android 5.0 is a fresh bun out of the oven and not all apps are seamlessly compatible with it."

Google should have waited another few weeks I think and fixed some stuff. They only needed to be out a week before black friday or so. That said, it still did very well against ipad air 2.

http://www.pcworld.com/article/2842245/google-nexus-9-review-a-powerful-tablet-made-for-everyone.html
PCworld, geekbench higher at 3358 multicore vs. 3166 of the leaked scores mentioned before. Who knows what revs of the device or software all these sites used, but clearly things can change here.

Note PC world Vellamo scores vs. Endgadget 2,653:
"In Vellamo’s Chrome and multi-core benchmarks, it scored 5855 and 2805, respectively, which is way higher than what the Shield turned in."

Not sure which endgadget is quoting there, but both PCworlds are different, but at endgadget shield WON Vellamo, so again weird as it loses at pcworld. Also almost 26000 in icestorm vs. engadget's 24256 (again, quite a difference). They shouldn't be all over the board like this, so hopefully google gets this straightened out shortly and tells reviewers to update and re-bench or something.
 

somebodyspecial

Honorable
Sep 20, 2012
1,459
0
11,310
20
http://gizmodo.com/nexus-9-review-googles-flagship-tablet-is-nothing-spe-1654016772
Another and he already got a graphics driver update:
"The problem was that the Nexus 9 was chewing through battery on the homescreen, and just generally swiping around pretty low-impact apps. Since then, the Nexus 9 got a graphics driver update which I think may help. At the very least the tablet doesn't get as hot anymore, which is a very good sign. I'll have to try it some more and update here afterwards, but so far it seems like the battery life has improved."

Other reviews say a driver update was sent during review also. It might be a bit before we get the real story on lollipop/denver, but a gpu driver update wouldn't do much to cpu side I'd guess but that's just an assumption. I'm wondering if this is NV or Google or both that need more work here. As the 1st 64bit OS device, no shock it's getting updated, but I think they could have gotten better reviews a few weeks from now. This looks rushed. They would have sold out 3 weeks from now anyway, so not sure why they rushed it with a new OS and chip. I would have thought they'd want to get this "RIGHT". ;)

I hope NV puts out a Denver device so we get a better idea of what happens when the chip company themselves does the device if google can't get their act together here soon. A new shield handheld? :) Then again I'd rather wait for a 20nm version of NV's next handheld. That is a device I'd buy (or a 20nm tablet 13in+ with from NV). I don't think I'll bite on any NV soc until it says maxwell+20nm inside though ;) I can wait out all the 28nm stuff. No rush. I'd rather let Lollipop/drivers/apps age a few months on 64bit anyway.
 

InvalidError

Titan
Moderator

This is easier said than done: no matter how much in-house testing Google might do before launching a new device or Android version, it will still only account for only a tiny fraction of all possible hardware and software configurations people may end up using in the wild.

Personally, I consider new Android version launches like open-betas: if you buy a device soon after launch or receive the OTA update notification early for an existing device, you just volunteered for field-testing. If you want to skip the beta phase, wait a month or two before buying/updating.
 

InvalidError

Titan
Moderator

A few years ago, Google said they removed external storage support in Nexus devices because it was "too confusing for most consumers." Supposedly, people had trouble figuring out what got stored where.

Totally lame reason but the bottom line is Google does not want people to use local storage and I doubt we will ever see a new Google-branded device with an SD-slot. In their Nexus-branded devices, Google even goes through the trouble of disabling USB storage support.
 

house70

Splendid
Last but not least, there are plenty other Android-BASED systems to choose from; that's why it dominates the market. Show me your options when it comes to it's competitor(s). That's right... <crickets>
 

anthony8989

Distinguished
Feb 2, 2013
653
2
19,165
57


At least once a day I hope to myself that Nvidia releases a smartphone with Maxwell integrated Denver SoC on 20nm or less. The day they announce that phone... :bounce:

 

tobalaz

Honorable
Jun 26, 2012
276
0
10,780
0
There's also a licensing fee that has to be paid to Microsoft close to $10 a device for FAT storage. Nexus is cheaper because they don't include the microsd, so no FAT license has to be paid.
I'm with everyone else that says 8gb and 16gb are a freaking joke and 32 needs to be the default (with 64gb and 128gb options) since many apps are hitting 500mb in size themselves and Google is trying to keep the apps off of the sd card.
Seriously, a class 10 64gb card can be had for $30, I paid a $100 for my 32gb class 10 4 years ago, look at how far down SSD prices have fallen, we need more space and its cheaper then ever and we keep getting lowballed on the amount of storage we get and overcharged for minor upgrades to it. It needs to change.
 

InvalidError

Titan
Moderator

While there is some truth to the licensing part of this, you might want to keep in mind that even $50 tablets have an SD slot so that alone does not excuse the lack of an SD slot on Google's $300+ devices.

BTW, vFAT/FAT32 is 20 years old now. This means most of its related patents are either already expired or soon will.
 
Status
Not open for further replies.

ASK THE COMMUNITY