Texas Instruments and Qualcomm are the two leading providers of CPUs for Android phones. Texas Instruments has the OMAP series of SoCs, while Qualcomm makes the Snapdragon platform. The OMAP is an ARM Cortex-A8 CPU, and Qualcomm's chip is an ARM chip very similar to the Cortex-A8. Both platforms are extremely robust.
In terms of speed, the 1GHz Snapdragon in the Nexus One is 400MHz faster than the Cortex-A8 in the iPhone. There are additional optimizations as well, which gives the Nexus One the ability to crunch 80 million polygons per second to the iPhone's 28 million. The Nexus One is quite a bit better, to say nothing of its <b><i>drastically</b></i> superior display (800x480 vs 480x320, and a world of difference in brightness/contrast).
To address your first question, both TI and Qualcomm have faster/better versions of these ARM chips floating around. Some are dual core, some have better GPUs, some have higher frequencies and so on. The trick is to make them suitable for a smartphone without killing battery life.
I am excited at with the future of the smartphone industry. I just wish they could find a better way to provide their equipment at a more reasonable price... While I really like my Motorola Cliq (I loved it 6 weeks ago), I am already seeing how far from cutting edge it really was. Knowing I am stuck for 22 more months and knowing that at launch, my phone was several months behind other new android phones... hard to swallow.
So... you mentioned battery life. Without holding back the CPU speed, what are their options for getting the Android platform to be more battery friendly? Bigger batteries seem to always be the first answer but is bigger really better? My Cliq barely gets a 15 hours without ever using GPS, WiFi, and not really doing much with it. I know Android v2.0 is better at battery life, is the answer all code based?
At this point, a reasonable expectation of any smartphone is between 6-7 hours of straight usage. This battery life is typical of any make or model you look at today. Standby time is about 3-4 days, and an "average use" lifetime (a little talk, a little browsing) should give you an entire day without any issue.
Efficient coding is part of it, but only a small part. Real battery life savings will come from more efficient displays (OLED) and smaller CPU process nodes, like shrinking from 45nm to 32nm, or reducing the flash memory's process node to 34 or 22nm. That will take time, but it'll happen.
0
KwitkoSheriff of Banning (Retired)By the thing near the stuffIcrontian
edited January 2010
AT&T needs to do something to offset the loss of exclusivity on the iPhone.
Thrax
🐌Austin, TX Icrontian
Comments
Also... how does the Nexus standup to the newest iPhone in terms of speed?
In terms of speed, the 1GHz Snapdragon in the Nexus One is 400MHz faster than the Cortex-A8 in the iPhone. There are additional optimizations as well, which gives the Nexus One the ability to crunch 80 million polygons per second to the iPhone's 28 million. The Nexus One is quite a bit better, to say nothing of its <b><i>drastically</b></i> superior display (800x480 vs 480x320, and a world of difference in brightness/contrast).
To address your first question, both TI and Qualcomm have faster/better versions of these ARM chips floating around. Some are dual core, some have better GPUs, some have higher frequencies and so on. The trick is to make them suitable for a smartphone without killing battery life.
The future for Android phones is bright.
So... you mentioned battery life. Without holding back the CPU speed, what are their options for getting the Android platform to be more battery friendly? Bigger batteries seem to always be the first answer but is bigger really better? My Cliq barely gets a 15 hours without ever using GPS, WiFi, and not really doing much with it. I know Android v2.0 is better at battery life, is the answer all code based?
Efficient coding is part of it, but only a small part. Real battery life savings will come from more efficient displays (OLED) and smaller CPU process nodes, like shrinking from 45nm to 32nm, or reducing the flash memory's process node to 34 or 22nm. That will take time, but it'll happen.