Slashdot Mirror
Qualcomm Ships Dual-Core Snapdragon Chipsets
rrossman2 writes "Qualcomm has issued a press release revealing it has started shipping new dual-core Snapdragon chipsets. These chipsets run each core at up to 1.2GHz, include a GPU that supports 2D/3D acceleration engines for Open GLES 2.0 and Open VG 1.1, 1080p video encode/decode, dedicated low-power audio engine, integrated low-power GPS, and support for 24-bit WXGA 1280x800 resolution displays. These chipsets come in two variants, the MSM8260 for HSPA+ and the MSM8660 for multi-mode HSPA+/CDMA2000 1xEV-DO Rev B. The press release also lists QSD8672 as a third-gen chipset like the two mentioned, but doesn't go into any detail of what its role is. With this announcement of shipping chipsets, how long until HTC makes a super smartphone?"
I'll wait until it makes a super ULTRA smart phone.
Om
Since the whole "smartbook revolution" seems to be a puff of hot air, the thing to hope for would be that some sort of "assembly kit" possibilities for computer-building hobbyists interested in RISC/ARM architecture could be available. This seems to be a market entirely owned by x86, with tons of pieces that can be stuck together like lego. I for one would love to have a full-size passively cooled laptop with low-energy processor and screen.
Great timing to reveal this just ahead of Steve Jobs iPhone 4/HD A4-processor equipped phone. I almost feel badly for Mr. Jobs getting beaten up like this. Even Dan Lyons (aka Fake Steve Jobs) is getting an HTC Incredible.
I said almost.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
One core for flash, one core for you.
Domestic spying is now "Benign Information Gathering"
Some power-draw information for H.264 decode, full tilt GPU utilisation, 25/50/100% CPU utilisation of one/both cores would be welcome.
That would be a very bad move since any x86 OS is both bloated, and not suited for a touch screen only interface. They all want keyboard/mouse inputs. Even Apple realized that OS/X was not the thing to run on a smartphone, while HP has dropped Windows 7 for their Slate, Google offers Android, not Chrome, for phones, and Microsoft Win 7 Mobile is really looking iffy to appear at all.
This is also why Microsoft Office and Open Office aren't available on the Android phone yet. They are not suited for this type of hardware, memory limitations, screen limitation, and lack of keyboard/mouse.
And most SSD drives these days are just about the size of your entire phone. Try to realize why a smartphone is a different paradigm altogether.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
Those of us old British farts who remember the BBC Micro will be celebrating. Who would have thought that, nearly thirty years on, its descendants would at last become a threat to (at least the low end of) the Intel/Microsoft domination of personal computing?
From scarped cliff or quarried stone she cries "A thousand types are gone, I care for nothing, no not one."
Good thing I don't have mod points today. I wouldn't know whether to mod you funny, insightful, or troll.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
We will not get the great dream phones we all want until the current patent mess is sorted out. As soon as HTC brings out a proper iPhone competitor, Apple will sue the crap out of them, making sure that at least they drag the new product into a mire of fud and drawn out proceedings.
Net result? The customer doesn't get a better device.
Seven Days with Ubuntu Unity
Microsoft Win 7 Mobile is really looking iffy to appear at all.
Uh, what?
Visual IRC: Fast. Powerful. Free.
Sounds good for a phone, but awesome on a tablet, where there is more room for battery. The iPad got the right form factor and weight, but I also need a SD slot, HDMI output, user freedom and uncrippled USB. That's one tablet I would buy.
I did a lot of assembly programming in the 1980s, for nearly every major processor available at the time. The 8086 rocked, in comparison to the others, at least until the 68000 came out.
The one processor that really stunk, IMHO, was the z80, and that's why its lineage died after being so popular. But the others, like the 6809 and 6502, were rather limited in comparison to the 8086.
Of course, virtual memory is a different beast and adapting x86 was a kludge. But I don't see RISC as being any improvement. If anything, they should have gone to a *more* complex instruction set, otherwise you start losing efficiency at the lowest level with all the library function calls that are needed. One example of a superb implementation of CISC for virtual memory was the VAX instruction set. The VAX was easily the winner in ease of assembly programming.
In the end, I'd rather have a good CISC implementation than RISC. For an example of how RISC sucks, take the Microchip PIC architecture. They even claim "only 35 instructions to learn" in their marketing, as if this was an advantage.
In conclusion, here's the car analogy: RISC is like a muscle car, all power but cannot make curves.
You can improve ARM code density using the Thumb extension, but it's the variable instruction cycle length that kills x86. Pipelining, branch prediction, etc, is much easier with RISC.
The ARM architecture is far superior to the x86 which is why one of the most competitive markets, mobile phones, has moved there. ARM has consolidated there as they do not have the marketing or R&D budget to take on Intel head to head. The margins have been much higher with desktop CPUs, with marketing and playing the GHz game driving sales more than processor efficiency.
Once ARM processors take over the netbook market, there will then be an incentive to increase their maximum raw performance. The server market would be the next target. However, they are unlikely to challenge the desktop market any time soon. Intel is cash-rich enough to dump processors onto the market at as loss if necessary to drive them out. Shame, as my ARM-based desktop machine was incredibly fast.
Phillip.
Property for sale in Nice, France
Mobile usages are limited by available battery technology at least as much as processing power; and the former moves forward much slower. Process lead of Intel doesn't quite work the same as before in this case...
Sure, there's one future, unreleased, next year Intel product; as you can see from the article, basically "smartphones only", no Win for you or generic Linux distro (not a big deal so far). But now it gets interesting..."southbridge" has "system controller/32 bit risc" - would be surprising if that's not some ARM (plus at least another one in radio interface; that's already probably more ARM cores than x86 ones, to keep power consumption at merely acceptable levels; Intel just couldn't do it without ARM). Less efficient and more expensive multichip solution (and of course other manufacturers are expected to make this effort, for miniscule portion of the market...while Intel doesn't risk anything; but anyway, there are no announcements - while phones would need to get certs quite some time before release; Android players have no incentive to switch; Apple has none, either, considering their inhouse ARM team; Samsung goes its own way, their own SoCs; Nokia devices with MeeGo are an uberniche product - they will certainly ride on Symbian for a long time)
Plus Intel doesn't even tell everything - they show those nice power usage numbers only in scenarios...when x86 core is idling; when the "supporting" hardware (with a great help of ARM cores :D ) does the real work. Power usage when x86 is doing something intensive (using its "impressive" speed) is strangely absent...
It will be still probably around an order of magnitude difference. Plus ARM won't stand still, look at the progress in the past decade from, say, ARM7TDMI to latest Cortex.
Again - a progress constrained by battery technology; Intel offering doesn't help that, quite the contrary - their greatest strength, process shrinking, no longer works quite the way as before.
BTW, how is the i960 or Itanium going?
One that hath name thou can not otter
Regarding Open Office on Android: Android Tablets are coming, Android phones take a BT keyboard, and some have video out...
Regarding Win7 and other regular OSes on mobiles: it may be impossible to get Win7 to be energy efficient, and keep the oodles of power-sucking services (and the basic architecture) of that server/desktop OS. Unlickily, those are probably required for compatibility.
Regarding x86 mobile: x86 was never designed as a low-power, high-efficiency CPU. Attempts to backport that are somewhat succesful, but I can't imagine x86 ever being as efficient as ARM cores that have been designed from the ground up to be precisely that. The one advantage Intel has is process technologies. See http://netbooked.net/blog/arm-vs-atom-size-vs-power-vs-performance/ for a biased source :-p
Other than that, I agree with you. Oh, wait ...
The Cloud - because you don't care if your apps and data are up in the air.
Yes, a HDMI output port would make these things much more like a small PC. Wander around with it, reading your emails, then get to the office and plug a cable into it from your TV/Monitor. Add a bluetooth keyboard and you have something every salesman, accountant, and manager dreams of.
I reckon that's the future of computing devices, not Windows anymore.
That would be a very bad move since any x86 OS is both bloated, and not suited for a touch screen only interface.
This is pure BS. Look at the upcoming Moorestown and the OSes available to run on them. MeeGo runs on it - completely touch-based OS. And, Android also runs on it. There is nothing inherent in the x86 to make it touch-averse. Where it has been lacking so far was performance for the limited power envelop. Moorestown will fix that. The next thing where it will still be lacking is not tight enough integration of communications capability - which is key to create a mobile platform that runs well with limited power consumption. Osho