Nokia's Present & Future

Still expecting some news from Intel re: Moorestown at MWC, will have to see if its Nokia specifc but I expect to hear news of actual handsets this time round. Whether its a small run in advance of Medfield is another matter.
 
That is exactly opposite to the point I'm making. I'm saying why would NOKIA release Cortex-A9 MeeGo devices instead of releasing Atom ones from the start, if that's their ultimate goal? That should have been clear from my post, I think you read the first sentence then hit the reply button...

Easy there. First you say it doesn't make sense to go from A9 to Atom, then you say Atom isn't a good choice for smartphones.
My assumption wasn't that far-fetched..

Anyways, the opposite still applies. MeeGo is already supporting the ARM architecture. Makes some sense for them to use the U8500 for the higher-end smartphones until Intel comes out with a 28/32nm Atom that's capable of offering higher performance than the ARM-based competitors, at the same power levels (IMHO that's where Intel may be able to win this battle: faster transitions to smaller nodes).



So everyone thinks we'll see Symbian devices with Cortex A9s?
I still don't agree at all.
I still think Nokia will keep Symbian^3 to mid-end smartphones, increasingly enhancing the user experience while optimising the UI by taking advantage of the BCM2727 to the max, and they'll keep this hardware for at least the end of 2011.



Of course x86 phones are happening. But I still don't see the practical benefit.

I always hoped to see development+OS+application unification across all devices (high-powered desktop to low-power handheld) someday, and I always thought it would come from the top to the bottom. Bringing x86 to handhelds would be a part of that.
Given Microsoft's announcement for Windows 8 with ARM, I guess it may go the opposite way after all.
 
Btw it is likely that many of you have read anandtech's N8 review. I would like to point out that I did tests on my own N8 and found that my web browser scores were much higher than theirs. I think they used Opera Mobile 10.0 as opposed to 10.1. Opera 10.1 has a new javascript engine and I got sunspider scores about 10x faster than theirs (putting it in competition with iPhone 4 on this test) and browsermark about 50-60% higher than theirs (lower than iPhone and Android phones but higher than most WP7 phones). At least one other person with an N8 has reproduced my findings so I am sure Anandtech's scores are based on older version of Opera and NOT a fair representation of N8.

I have asked them to review their web browsing data. I am also suspicious of their wifi throughput tests as in an informal test I am getting much higher results than theirs but need to test more to be sure.

I think this review was done by a relatively inexperienced writer and he accidentally did some important mistakes. I have written to them and hope they will rectify their review soon.
 
Opera Mobile 10.1 for Symbian uses JIT (hence the performance increase), but it's still just using the 680MHz ARM11 CPU.

I expect the new Symbian browser, coming out in the first Symbian^3 update, to be fully hardware accelerated and thus have a much higher performance.

The Symbian^3 devices desperately need a new default browser.
 
Opera Mobile 10.1 for Symbian uses JIT (hence the performance increase), but it's still just using the 680MHz ARM11 CPU.

I expect the new Symbian browser, coming out in the first Symbian^3 update, to be fully hardware accelerated and thus have a much higher performance.

The Symbian^3 devices desperately need a new default browser.

Doesn't opera mobile already use hardware acceleration for drawing? what else is there to accelerate? I think the browser score just reflects the limits of the ARM11 CPU.
 
Doesn't opera mobile already use hardware acceleration for drawing?

AFAIK, no. I don't remember ever seeing it in a changelog.

Opera 10.1 for Symbian was more directed to Cortex A8-equipped Symbian devices, like i8910 and Vivaz.
 
http://www.engadget.com/2011/01/25/nokias-leaked-meego-device-resembles-dual-core-st-ericsson-u850/

It seems the latest leaked image of a Nokia tablet is the same device ST Ericsson showed in the U8500 demo.
Background color and icons resemble MeeGo.
Doesn't seem like the U8500 is coming to Symbian devices (even less if the (now rumoured to be cancelled) N9 would use an OMAP3 for the top-end O.S). What's the point of launching a top-end SoC for the mid-end O.S: and a mid-end SoC for the high-end O.S.?

I think the C5-03 doesn't use the 434 mhz cpu so maybe they moved to something else if they do more s60v5 devices?
It's probably the same 600MHz CPU Nokia has been using with some keyboard-enabled devices, like X5-01 and the original C5.
It could even be the exact same CPU from TI, which could be rated at 600MHz but Nokia underclocks/undervolts it to 434MHz in order to lower power consumption and maintain battery life with the bigger screen devices.


I kinda hope the C5-03 will be their last Symbian^1 device. Coupled with Nokia's commitment with software updates and customer support, prolonguing Symbian^1's life will be quite the hassle when they further fragment their market with MeeGo handsets.
 
Last edited by a moderator:
http://www.nokiaport.de/forum/thread.php?threadid=7351
Rumoured new Nokia MeeGo smartphone.

New device's specs found in a MeeGo bug report:

CPU - Intel 1,6 GHz

RAM - 1 GB
- High Level Memory: 117MB
- Low Level Memory: 889MB

CLK
- Reference/System Clock: 19,2MHz
- DSP Clock: 440MHz
- Memory Clock: 26MHz
- Grapcics clock (LNC core): 200MHz

Video
- Graphics Memory: 512MB
- HDMI
- Display: 480 x 854 Pixel

GPS - Broadcom BCM4751
- Single chip GSM
- A-GPS

NFC - NPX Semiconductors PN544 NFC Controller
- HT80C51MX Core processor
- up 424kBit/s
- all NFC standards
- NFC-WI Secure interface
- RAM: 5K; ROM: 128K; EEPROM: 52K
- 10cm oprating distance

UART - Maxim MAX3110
- RS232
- 300baud - 230kbaud max
- Freq: 3.125MHz

Audio - Intel HD Audio - ALSA

LAN
- Intel PRO/1000 - 1 GBit Ethernet
- Intel PRO/100 - 100 MBit Ethernet
- IPv4 / IPv6

UBS
- USB 1.1 (OHCI)
- USB 2.0 (EHCI)
- USB Mass Storage
- USB On-The-Go (penwell_otg)

Mass Memory Interfaces
- MMC1
- MMC2
- MMC3
- MMC4
- SDHC Interface (ADMA)

Touch
- Cypress True Touch Controller
- Atmel maXTouch Touchscreen/Controller (MXT224?)

I/O Expander - Toshiba TC35894XBG
- up to 24 GenIO ports
- support for 26 keys

RF - Infineon XMM 6260 3G modem (???)
- HSPA+

Connectivitiy
Texas Instruments WiLink 6.0 (WL1271)
- WLAN
- Bluetooth v2.1 + EDR
- FM Radio (Transmitter/Receiver)

Bluetooth
- Profile: L2CAP, SCO, RFCOMM, BNEP, HIDP
- Core Version: 2.15

WLAN
- Standards: a (54MBit/s) / b (12MBit/s) / g (54MBit/s), n (300MBit/s)

Camera Interfaces: 2
- Camera 1 @ I2C
- Camera 2 @ I2C

Sensors

Avago APDS980x series
- Ambient Light Sensor APDS9802als (analogue)
- Proximity Sensor APDS9802ps (digital)

Accelerometer - STMicroelectronics LIS3LV02DL
- MEMS inetrial sensor, 3-axis
- Measurement range: +2g/-2g + +6g/-6g
- 16 Bit

Magnetometer - AKM Semiconductors AK8974
- 3-axis electronic ompass

Else
Thermal zone 1, 2, 3
Lunix media interface
Linux video capture interface
PCI SSP Controller

Filesystem - EXT3


- Screen resolution means it's most likely a smartphone.
- Graphics clock shows it's probably not the SGX535 in Moorestown (reported to work @ 400MHz).. which in turn means it's probably not a Moorestown. Unless it's in powersaving mode, but then again the CPU seems to be running at full speed. Plus, the highest-clocked Moorestown was Z625 @ 1.5GHz, consuming 2.2W (obviously not for a smartphone).
- The presence of a dedicated DSP is also an indicator that it's not a Moorestown.
- Memory clock is just weird.. probably some bogus clock detection?
- Having an ethernet controller is just plain weird, just like the RS232 controller, IMHO.. maybe it's just an engineering sample in an open motherboard, for debug?



This kinda goes in line with the rumours launched two weeks ago, but Arun seemed pretty sure the N9 has an OMAP3, so..
 
Last edited by a moderator:
Nothing in that rumoured bug report suggest it relates to an N9, or Nanythingelse. Just that its being developed with Meego, unless there are bits that you didn't quote.

1 -. Display Resolution is 480 x 854. Not only the resolution but also the screen orientation tells me it's definitely a smartphone-sized handheld.

2 - Nokia is the only smartphone company that has publicly stated that it will launch consumer devices with MeeGo

3 - Nokia has publicly stated that their next N-series devices will all use MeeGo.


So it's almost certain it's a Nokia device (unless it's a pure MeeGo dev board) and the N-series are the only series we're officially sure that's going to use MeeGo.


Of course, anyone could do endless argument loops to try to prove "how wrong I am", but the logic steps are obvious IMO.
 
1 -. Of course, anyone could do endless argument loops to try to prove "how wrong I am", but the logic steps are obvious IMO.

sensitive type huh. Who's trying to prove you wrong ?...is there some conspiracy that I am unaware off?
Kinda reminiscent of the top of this thread when you "went off on one" regarding people fixated on the USA, when the only person that mentioned the USA was you.

I said there was nothing in the leak to indicate it was N9. It could well be, or it could be developement for a product further down the line called N-series, or some other new series name.

Your "so...." at the end of the posting virtually agrees with the above !
 
could be the N9-01 as opposed to N9-00, sorry just trying to help ;-)

Well, there are rumours of the N9 being scratched, so maybe they scratched the original (and a bit disapointing) OMAP3 internals - N9-00, kept the chassis and changed to an Oak Trail - N9-01.

There's the X3-00 and then the X3-02. It probably means there was a X3-01 at some point but it was scratched, so it's nothing Nokia has never done.


EDIT: Not Oak Trail, it's Medfield. Medfield is Moorestown's 32nm successor and will supposedly include the southbridge functions in a single SoC.
 
Last edited by a moderator:
nokrdspend.jpg


Source: Digital Daily

I'm vexed with these numbers.. So where is all this money going to? Does launching many new products every year increase R&D costs that much for Nokia?

And how come Apple spends so little? Weren't they reported to be spending billions in new tech? Didn't they buy Intrinsity in 2010? How come the total R&D spending comes down to some 750 million $?
 
ToTTenTranz said:
And how come Apple spends so little? Weren't they reported to be spending billions in new tech? Didn't they buy Intrinsity in 2010? How come the total R&D spending comes down to some 750 million $?
They are investing and prepaying billions to secure their future supply chain (and locking out competitors at the same time). That has little to do with R&D and goes to a different item in their accounting tables.

The Intrinsity buy goes to yet another accounting item, though the expenses to pay their employees should be included after the closing of the deal.
 
They are investing and prepaying billions to secure their future supply chain (and locking out competitors at the same time). That has little to do with R&D and goes to a different item in their accounting tables.

The Intrinsity buy goes to yet another accounting item, though the expenses to pay their employees should be included after the closing of the deal.

Getting the Intrinsity thing out of the graph sounds a bit like cheating for the R&D graph.

If you pay hundreds of researchers for new hardware, it goes to the chart, but if you buy a whole research company, it stays out.
Honestly, what's the point of even making a chart like that?
 
Back
Top