The Palm Pre vs. Apple's iPhone 3G: Preliminary Results and 3GS Discussion
by Anand Lal Shimpi on June 9, 2009 12:00 AM EST- Posted in
- Anand
I got up early, went down to very sketchy Sprint store and bought a Palm Pre on Saturday. I've been playing with it and testing it ever since and before I finish the full review I thought I'd share some data with you all.
First, battery life:
| Phone | Web Browsing (Cell Network) | Web Browsing (WiFi) | Talk Time |
| Apple iPhone 3G | 218 minutes | 400 minutes | 289 minutes |
| Palm Pre | 219 minutes | 351 minutes | 312 minutes |
| T-Mobile G1 | 398 minutes (on Edge) | 435 minutes | 218 minutes (on Edge) |
T-Mobile doesn't have any 3G coverage in NC yet so all of my tests for the G1 were on Edge, thus we get a much longer web browsing battery life. The thing that surprised me the most however was the eerily similar battery life results between the iPhone 3G and the Palm Pre. The Pre didn't do so well in WiFi web browsing but in the other two tests it lasted around as long as the iPhone 3G.
Performance
The Palm Pre uses TI's OMAP 3430 processor, clocked somewhere around 600MHz. The OMAP 3430 uses an ARM Cortex A8 core. The iPhone 3G uses an ARM11 based processor running at somewhere around 400MHz (thank you guys for the correction). The ARM11 in the iPhone 3G is a much older design than the Cortex A8 in the Palm Pre. Both processors are in-order architectures, but while the ARM11 was a single-issue chip the Cortex A8 is dual-issue.
The ARM11 has an 8 stage integer pipeline compared to a 13 stage integer pipeline in the Cortex A8, so the A8 loses some of its advantage there but makes up for it with its superscalar nature. There should be no contest when it comes to performance between these two chips, the Pre's Cortex A8 has the clear advantage. It's why Palm is able to enable pre-emptive multitasking while the iPhone pretty much can't.
The recently announced iPhone 3GS does address the performance issue, presumably by introducing a Cortex A8 based processor to the iPhone 3G. Apple is claiming significant improvements in battery life for everything but 3G talk time with the new iPhone. What this tells me is that Apple did a great job squeezing the most performance per watt out of its ARM11 based processor in the iPhone 3G. The new iPhone 3GS should have performance levels simliar to the Palm Pre, but if Apple's numbers are to be believed it means that battery life will go up significantly.
For Palm this means that there is a lot of room left on the table to improve battery life. In most of my interaction with the Pre I've gotten the impression that if Palm only had a few more months the Pre would be significantly more polished, I suspect that battery life falls under that observation as well.
Other Pre Notes
When I compare the T-Mobile G1 to the iPhone 3G it's no contest, Apple's smartphone takes the cake. The G1 feels more like the smartphones that existed before the iPhone rather than something competitive with it. With the Palm Pre however, it's difficult to make such an apples-to-apples comparison. In many ways the Pre falls short of the iPhone, but in others it's completely untouched by Apple's offering. I'm nowhere near my conclusion but I don't think I'll see a clear victor in this review.
The Palm Pre brings multitasking to the smartphone market better than any of its predecessors. It's almost as if Apple did it. I say almost because the implementation isn't as polished as I'd like. Despite the significant performance advantage of the Pre's CPU, the multitasking just isn't as smooth as I'd want it to be. I'm guessing battery life isn't the only thing Palm could stand to optimize on the Pre.
What the Pre lacks is what the original iPhone had going for it: mastery of key features. The Pre does many things but it does very few things well. The original iPhone on the other hand didn't do a lot, but what it did do, it did better than any other phone on the market. Palm comes very close to achieving that, but I think it needs another 6 months with the Pre to produce the level of polish I feel is necessary to pose a true threat to Apple.
What is truly striking about the Pre is how far Palm was able to take it. Going from Palm's position to truly out-innovating Apple is a serious accomplishment. There are things about the Pre that even Apple's iPhone 3GS can't touch.
More in the review to come...
Facebook
Twitter
RSS
44 Comments
View All Comments
Griswold - Wednesday, June 10, 2009 - link
"Despite the fact that the Pre can finish loading a large complex webpage a bit faster than the current iPhone, the iPhone is faster and smoother when it comes to panning and zooming loaded web pages. "No, its not. Check out the engadget video somebody linked above. My iphone 3G scrolls and pans pretty jerky, just like in the video, whereas the pre is buttesmooth.
Why do you people keep saying these things - you're the second to make a false claim as far as browsing goes...
lifeblood - Tuesday, June 9, 2009 - link
And they will get the 6 months to improve it. By releasing the Pre exclusively on an almost dead network (Sprint), they are in effect "beta testing it under fire". Sprint users get to find the bugs thus allowing Palm to iron them out while tweaking it's performance before releasing it into a truly competitive network (Verizon).Palm must have taken a page from Microsoft.
HelToupee - Tuesday, June 9, 2009 - link
"By releasing the Pre exclusively on an almost dead network (Sprint), they are in effect "beta testing it under fire"."Really? Sprint's almost dead? They've got 15 million subscribers. That's up against Verizon's what, 19 or 20 million? They've got arguably better 3G coverage (especially in the Midwest) than Verizon.
Are you trolling, or are you really that misinformed?
theslug - Tuesday, June 9, 2009 - link
Judging by the last sentence of his post, it's clearly trolling.cocoviper - Tuesday, June 9, 2009 - link
Sprint has ~50 million subscribers, Verizon has about 87 million. Both are big, neither are "dead".As with the computer hardware industry- don't believe the FUD that competitors try to feed you.
anandtech02148 - Tuesday, June 9, 2009 - link
the smartphone war is exciting,i want to read Anand's insight into this, you could also do Nokia n97, Iphone 3Gs,& Palm Pre. Web Os is very impressive base on a lot of you tube videos. Symbian Os reminds me of OpenVms but with Gui, love that Os very much, i'm hoping symbian becomes the wild Os like in the days of Windows. And Apple is just Apple, they delivere supports like HDSPA 7.2mb that will make most smartphone obsolete in about 6months.
the phone war reminds me of the old pc wars eventually its all about content and a kickass web browser. Apple is ahead with the browser performance. I want Anand to have a dedicated section to smartphone hardwares and Os.
aftlizard - Tuesday, June 9, 2009 - link
I believe Sprint's 3g is at 7.2 mbs as well and actually IIRC tested faster than any other carrier.cocoviper - Tuesday, June 9, 2009 - link
Nope EVDO RevA is 3.1Mbps vs. Att's currently rolled out 3.6Mbps 3G (what the iPhone 3G uses). Their network is slowly being upgraded to 7.2 but there are other items besides the theoretical speed to be considered.1-Spectrum. Sprint uses 1.9GHz whereas Att uses 800MHz typically (and they are moving more of their 3G coverage to 800 because it propagates further). Lower frequencies are better at penetrating walls but they also carry less data per second. For any given channel, an equivalent signal on 1.9GHz will always be faster than an 800MHz signal- it can simply carry more data.
2-EVDO (CDMA) seems to consistantly get closer to its theoretical data speeds better than HSPA (GSM). This is one of the key reasons that Sprint (as well as Verizon) are known for the speed and quality of their data network vs. Att and T-Mobile are typically considered subpar to horrible when it comes to data speeds (see all the class action lawsuits filed over the iPhone 3G being twice as fast, Gizmodo's 3G test http://gizmodo.com/5111989/the-definitive-coast+to...">http://gizmodo.com/5111989/the-definitive-coast+to... , etc...)
taywyn - Thursday, June 11, 2009 - link
I think your spectrum explanation is invalid or imprecise. Data speed is relative to the width of the signal envelope; center frequency has little to do with that except that you typically get assigned narrower bandwidth as the center frequency decreases. For any channel of a fixed width (a simplified case of 1 MHz at 800 MHz vs 1 MHz at 1.9 GHz), you have equivalent data transfer rates with the same encoding and modulation schemes. This is mathematically provable.Better propagation is a trade-off with user density; more users on a single cell (high propagation) means more bandwidth contention for all users, whereas with worse propagation, where you have coverage, you probably have decent speed. It just costs your network more (cell sites and therefore money) to get you that coverage.
aftlizard - Tuesday, June 9, 2009 - link
You're right. However I was correct in remembering that Sprint's network tested out faster, at least according to gizmodo.http://gizmodo.com/5111989/the-definitive-coast+to...">http://gizmodo.com/5111989/the-definitive-coast+to...