Take an IPhone or any other current cell phone that you know well enough. To what computer is it equivalent travelling back in time?
286? 386? Quadra? PowerPC? Pentium?
If you wanted to equate it to the middle of the pack mainstream computer of a year, what year would it be?
ETA: Is the gap closing?
The iPhone 3GS is using a 600MHz processor. With 16GB Drive and 256M of RAM, it’s roughly equivalent to 2002 vintage iMac G4 in raw specs, although the iPhone would probably beat the pants off of the iMac.
If you go by raw clock speed [MHZ], the Windows/Intel world got a 600 MHZ Pentium III on August 2, 1999.
Figure on that being a median clock speed for new computers circa March 2000.
The fastest Cellphone CPU you can actually buy is the 1GHZ Snapdragon that is in the Nexus One. There is supposed to be 1.5 GHZ dual core version out soon.
http://www.qctconnect.com/products/snapdragon.html
In terms of raw clock speed, that is about what Intel was shipping 7 or 8 years ago, but there is a lot of custom silicon on the Snapdragon for handling things like HD and 3D video that may actually make it faster than current low-end Intel chips for some purposes.
http://www.mobiletopsoft.com/board/6890/google-nexus-one-3d-benchmark-test.html
It’s not really meaningful just to look at clock rates. The amount of work a CPU can do in any given clock cycle varies dramatically depending on the design. Mobile processors can be much more specialized than general purpose CPUs and hence only good at a few things.
It’s hard to give a meaningful answer to this. A 1 GHZ ARM processor may do some stuff better than an 800 mhz pentium 3, and some stuff much worse. You’d have to figure out what sort of metrics you’d want to compare.
This website has some simple iterative benchmarks for browsers. There are 5 tests. My iPhone scores between 75% and 25% of the speed of my 2.5GHz G5 (.022 vs .009, .008 vs. .002, .013 vs. .005, .009 vs. .006, .004 vs. .003, iphone vs G5, in seconds). That implies that the iPhone processor is no worse than the G5 on a clock-for-clock basis, and sometimes much better.
Now hold on a moment - I’m using an ancient PC (over 6 years old) and the times I get are:
0.001
0.000
0.001
0.001
0.000
Surely something’s amiss - I would expect my PC to be much slower?
Thanks for all the answers so far.
True. Let’s go for a balanced benchmark of the stuff you normally do on a cellphone. Internet browsing and light gaming.
Hard to say because cell browsers have to do a lot of processing to get the page rendered for their little screen and be able to tap it to zoom in etc. I counted 18-mississippi to load this forum on my iPhone 3G with the built-in Safari browser, and 1-2-mississippi to load it on my desktop.
ArsTechnica has some benchmarks in their Nexus One Review. Where they compare it to the Droid and the iPhone.
http://arstechnica.com/gadgets/reviews/2010/01/nexus-one-review.ars/4
This article implies that the GPU on the 3GS might actually be faster than the Nexus One.
I didn’t find any benchmarks for Android running on a netbook so we could compare an Intel CPU with a Snapdragon yet.
What are your machine’s specs?
The benchmark is pure javascript - you can look at the code by viewing the page source. My guess is that your javascript interpreter is optimizing the code to the point that it does nothing.
This. The phone chips are so different in architecture vs a PC CPU that head to head metrics are close to meaningless.
Good point. Even finding meaningful benchmark tests is difficult because the platforms are so different. Benchmarks designed for desktop OSes may not be appropriate. Smartphones only have to drive low-resolution screens, for example.
It’s not just apples (heh) vs. oranges, more like apples vs. xylophones.
True, but in the end user experience is what matters. I don’t much know or care about GHz and petaflops, I care about how long it takes to download a webpage or what game I can run. This is why there are “real world usage” benchmarks. They are admittedly less precise but they do help to compare apples to oranges.
Times to download webpages are a common benchmark and they were shown in one of the links upthread. I guess one would have to rummage through old PC-World magazines to find when those times in second where the norm among desktop PCs.
Game performance is probably trickier due to the absence of equal games to compare frame rates, for example. I guess the closest we could get is looking at notable games that push the cellphones to the top and seeing when was it that a game that looked the more or less the same was pushing the PCs of the time to the top.
There are a couple of flight simulators for the iPhone, for example that look pretty amazing. I remember playing Falcon on my PC (first among my friends with a VGA) and it not looking half as good as these do today. I am sure there are plenty other examples of cellphone games that could be compared to PC games of some earlier time.
It must be. I got the same exact results as Colophon initially, but when I tried the benchmark test in Firefox instead of Safari, I got real results (.004, .001, .005, .006, .002).
The ARM chip, first developed by DEC before they collapsed, is used in many cell phones. This and probably some ASICs make up the guts. That said, I have no idea what generation of the common PC CPU chip this compares to.
Would love to see any cell phone CPU on the market today working on rendering 1024x768 video. Dollars to doughnuts the PC from 5 years ago does exponentially better.
Actually, developed by Acorn Computers Ltd. “ARM” = Acorn RISC Machine. DEC got involved later.
You’ll see this when the iPad comes out. Everything points to better-than-adequate video performance.
But that’s handled by a special-purpose chip, not the CPU. That’s why apples-to-apples comparisons between smartphones and computers are so difficult. All of the heavy lifting that a smart phone does (audio coding and decoding and video output) is done by a dedicated chip.