http://www.linleygroup.com/Newsletters/LinleyMobile/lm090129.htmlSo I was right after all that qcom will try to put scorpion almost everywhere so that they could benefit from the hard work they done on designing itBTW, it feels pretty good being spot on (in the original news piece) - now if only someone else didn't have the scoop straight from Qualcomm even before I speculated about it, bah...
Maybe when they were designing scorpion they figured out some way to overcome this 'flaw' but even if they didn't it will still be better than current A8 implementation.
Well, these are generic designs. Someone might need DSPs, someone Neon etc. If you want something specific you have to build your own design like AppleIndeed, definitely looks like they want to amortize Scorpion in a lot of designs. They're also being a lot more aggressive with video & 3D hardware on 45nm. All good news for the industry...
But well, you know, I'm a tad subjective when it comes to NEON because I 300% agree with the point of view of a certain third-party I won't name that think it's as much of a joke as I think I am, and will stick to VFP. These companies have three zillion redundant units capable of doing the same thing, then they act all surprised when their die size is higher than the competition. Oh gosh, hoocoodanode?
Just look at Snapdragon: you could do video decoding on NEON. Or you could do it on the advanced DSP. Or you could do it in the dedicated accelerators! Errr, what exactly is the point, guys? And how many application developers do you really think are going to use SIMD intrinsics on an open platform where 90%+ of the user base doesn't have it available? The same thing is true to differing extents for other architectures, including OMAP4. It's as if all these guys were living in a world of their own where there's no competition and higher die size is merely a benefit because it allows you to justify a higher selling price... Oh well, I really should stop being so bitter!
Indeed, definitely looks like they want to amortize Scorpion in a lot of designs. They're also being a lot more aggressive with video & 3D hardware on 45nm. All good news for the industry...
But well, you know, I'm a tad subjective when it comes to NEON because I 300% agree with the point of view of a certain third-party I won't name that think it's as much of a joke as I think I am, and will stick to VFP. These companies have three zillion redundant units capable of doing the same thing, then they act all surprised when their die size is higher than the competition. Oh gosh, hoocoodanode?
Just look at Snapdragon: you could do video decoding on NEON. Or you could do it on the advanced DSP. Or you could do it in the dedicated accelerators! Errr, what exactly is the point, guys? And how many application developers do you really think are going to use SIMD intrinsics on an open platform where 90%+ of the user base doesn't have it available? The same thing is true to differing extents for other architectures, including OMAP4. It's as if all these guys were living in a world of their own where there's no competition and higher die size is merely a benefit because it allows you to justify a higher selling price... Oh well, I really should stop being so bitter!
Why would someone *need* that? I have yet to see a single smartphone manufacturer do anything really fancy with this stuff. They just use it for multimedia, often with partner-supplied libraries. Maybe slow-motion video encode but that can be done great in fixed-function hardware if thought of in advance.Well, these are generic designs. Someone might need DSPs, someone Neon etc. If you want something specific you have to build your own design like Apple
But who cares about that stuff? The entire point of those chips is the media acceleration. Those third party apps are great for older processors, but I don't really see the point now. Surely every phone with those chips will have a hardware accelerated media player, and if that thing's GUI is so bad you want to use another one then you probably shouldn't have bought the phone in the first place. Plus, even discounting that, I think there's a real willingness from some chip designers to expose the APIs to successful third party video software developers.
Whadda?
The industry standard is becoming to put a small audio processor on its own power island so you can shutdown everything else while playing music. That's how you get to numbers like 100+ hours for audio on OMAP4/Tegra. On NEON, I don't even see how you could get to 20 hours... Even with the OMAP3 DSP, Archos' MID has some pretty pitiful battery life for audio, doesn't it?Wow, poor DSP designers, I think you just insulted every single one of them that has ever lived...That's all great really but I wouldn't call those DSP's advanced if they can't process 720p h.264 high profile video with average 5 mbits
I'm certainly not aware of 720p H.264 High Profile having shipped in a commercial product using a classic DSP and no accelerator even for the tight loops... But then again I guess you meant a DSP+Accelerator approach probably.I can hear people at NVIDIA crying right now! j/k (and quite a few other companies too, I'm not sure anyone but TI & Apple is doing High Profile in their next-gen SoCs). Anyway to address your reason:
Wow, a 720p library not being usable until the ends of time... shock horror
How many people already have their library of BDrips anyway? Seems like a niche of a niche to me. And this generation nobody supports High Profile, by next generation everyone will be 1080p so you'd have to re-encode anyway. Plus 5mbps is not very high so you'd probably want to encode it again on a SoC that supports 10-20Mbps. Finally, consider that even OMAP4 doesn't support bitrates anywhere near high enough to play any Bluray stream without transcoding.Yeah, that's definitely very unfortunate, and is probably the best use-case I can think of where the user would be pissed off because he wouldn't understand why it doesn't work on an ARM netbook that is claimed to be so awesome at HD playback.Even youtube HD would be too much
It would be extremely amusing if Apple was the first to come out with that - I wouldn't be overly optimistic at this point though. Ah well, we'll see what happens.OMAP4 should be capable of playing any Blu-ray stream. Maybe not with it's DSPs alone, but the SGX540 (and Dual-Core ARM9) should do it. The Handset/Netbook/Videoplayer Developer just has to work a little bit harder
OMAP4 uses dedicated/specialized hardware to playback 1080p video, the DSP/CPU/GPU are shutoff *completely* in those cases. And I very much doubt this dedicated hardware is programmable; it's "configurable" at best IMO. The DSP alone is not much more powerful than what you have in OMAP3, and without accelerators it barely does 720p H.264 Baseline.OMAP4 should be capable of playing any Blu-ray stream. Maybe not with it's DSPs alone, but the SGX540 (and Dual-Core ARM9) should do it. The Handset/Netbook/Videoplayer Developer just has to work a little bit harder
Hm ok, thanks for explaining. I just thought if a SGX535 with Single Core Atom can play a Blu-ray stream plus Quake 3 at the same time then a SGX540 with two Cortex-A9 must be able to do it too (for example on a Netbook). But I guess that was just very wishful thinking and I have to wait for OMAP5 in 2013OMAP4 uses dedicated/specialized hardware to playback 1080p video, the DSP/CPU/GPU are shutoff *completely* in those cases. And I very much doubt this dedicated hardware is programmable; it's "configurable" at best IMO. The DSP alone is not much more powerful than what you have in OMAP3, and without accelerators it barely does 720p H.264 Baseline.
Now, if you mean you could bypass that hardware and use the DSP, the two A9s (with NEON), and the GPU all at the same time in an incredibly complex and interwoven program, and that with this you could do 1080p H.264 High Profile @ 40Mbps... Then yeah, there might be just enough processing power to do that, but it'd be incredibly power inefficient and, more importantly, would be mind-blowingly hard to program. The chances of anyone ever bothering to do this is, IMO, literally zero.
Consider that at MWC08, about 18 months post-tapeout, the best they could demo was a partner using the DSP for 720p H.264 decode at not incredibly high bitrates - and they didn't even yet manage to use the hardware accelerators at all, AFAIK! Which is why OMAP3's demos have become increasingly more appealing over the years: the software curve. And that's with something that is at least an order of magnitude easier than what you're describing, so really please don't get your hopes up
That's interesting. So it's very different from OMAP3, since IIUC the video IP in IVA2+ are controlled by the DSP. According to high-level information provided by TI I would assume IVA3 uses a very similar architecture: the IVA3 block is made of IP + DSP.
And what if they couldn't use the hw accelerators because these were too specific to some codec algorithm that they are of no use for H.264?Consider that at MWC08, about 18 months post-tapeout, the best they could demo was a partner using the DSP for 720p H.264 decode at not incredibly high bitrates - and they didn't even yet manage to use the hardware accelerators at all, AFAIK! Which is why OMAP3's demos have become increasingly more appealing over the years: the software curve. And that's with something that is at least an order of magnitude easier than what you're describing, so really please don't get your hopes up
http://focus.ti.com/en/pdfs/wtbu/omap_4_pb_swpt034.pdf?DCMP=wtbu_omap&HQS=ProductBulletin+PR+omap4pbI guess that's theoretically possible, I'm just not sure anyone has ever done that in the history of the industry? Of course, that doesn't prove anything...
Unless you're thinking of a software company, the only SoC I can think of that was upped from 720p to 1080p is Tegra. It is certainly not my understanding that they're doing something as fancy as what I described (although I'm not far from sure of the details), but I see your point.
