OMAP4 & SGX540

[silent_guy]

That looks to me like a lack of precaching combined with blocking I/O; Google apparently has/had a tendency to overuse it and it's only slowly improving. I agree it's a clear weakness, however silent_guy explicitly said it wasn't 'slow', just awkward. So I still suspect that's not what he meant.

Actually, I think that's exactly what I meant. Not slow in terms of pure speed rendering a page. But it's indeed probably the reaction time to things. It's that every gesture subconsciously adds a bit to the aggravation of the thing not reaction the way you expect it to.

 

[tangey]

Anyone know when the 4440 suddenly turned into the 4460 ?

http://focus.ti.com/general/docs/wt...ateId=6123&navigationId=12843&contentId=53243

 

[Deleted member 13524]

Wild guess:
The 4440's 3D performance didn't end up competitive enough for other tablet solutions appearing in H2 2011, so 4440 was scratched and 4460 has a higher-clocked GPU (~450MHz?)

 

[tangey]

Wild guess:
The 4440's 3D performance didn't end up competitive enough for other tablet solutions appearing in H2 2011, so 4440 was scratched and 4460 has a higher-clocked GPU (~450MHz?)

dunno, it seems to have been just a name change to me. However there is some rumours of a 1.6Ghz 4470 being used by Archos.

 

[Deleted member 13524]

dunno, it seems to have been just a name change to me. However there is some rumours of a 1.6Ghz 4470 being used by Archos.

Wut? In Gen9?

Just a couple of posts above I mentioned that and people thought it was a Nufront chip..

 

[tangey]

Wut? In Gen9?

Just a couple of posts above I mentioned that and people thought it was a Nufront chip..

"TI has shifted from OMAP4440 to OMAP4460 naming as OMAP4 adoption expands in new markets worldwide. (March 2011)
· OMAP4440 renamed OMAP4460
· OMAP4450 renamed OMAP4470
"
http://en.wikipedia.org/wiki/Texas_Instruments_OMAP

I've seen the archos chat above, and also note the above reference to a 4470 in the wiki, which leads me to speculate that the 1.6Ghz archos device and the Omap4470 is one and the same ?

 

[Deleted member 13524]

"TI has shifted from OMAP4440 to OMAP4460 naming as OMAP4 adoption expands in new markets worldwide. (March 2011)
· OMAP4440 renamed OMAP4460
· OMAP4450 renamed OMAP4470
"
http://en.wikipedia.org/wiki/Texas_Instruments_OMAP

I've seen the archos chat above, and also note the above reference to a 4470 in the wiki, which leads me to speculate that the 1.6Ghz archos device and the Omap4470 is one and the same ?

Well, just below that there's a chart claiming the OMAP4440 has a 300MHz GPU..

 

[Exophase]

This is the only thing I can find linking OMAP4 and Archos Gen 9:

http://armdevices.net/2011/03/18/archos-gen9-tablets-1-6ghz-dual-core-3g-to-be-announced-in-june/

- ARM Cortex-A9 Dual-core 1.6Ghz, it’s the fastest processor in a tablet ever unveiled thus far. I expect this to be a customized Texas Instruments OMAP4440, but that is only because I guess Archos will continue their long time partnership with TI that they’ve had for over 10 years now.

That isn't a rumor, it's merely speculation. Archos sells Rockchip tablets, so incorporating an SoC made by a Chinese company and not TI is not at all unprecedented for them.

June just seems too early for OMAP4440 (sorry, OMAP4460?)

 

[Deleted member 13524]

This is the only thing I can find linking OMAP4 and Archos Gen 9:

http://armdevices.net/2011/03/18/archos-gen9-tablets-1-6ghz-dual-core-3g-to-be-announced-in-june/



That isn't a rumor, it's merely speculation. Archos sells Rockchip tablets, so incorporating an SoC made by a Chinese company and not TI is not at all unprecedented for them.

June just seems too early for OMAP4440 (sorry, OMAP4460?)

AFAIK, chinese chipsets will be used in ARNOVA tablets, which is a low-cost daughter-brand of Archos (kinda like ASRock when it belonged to Asus).
Gen9 is a family of products that are supposed to go all out against the iPad, Playbook, Xoom, Galaxy Tab and others.

 

[Ailuros]

Well, just below that there's a chart claiming the OMAP4440 has a 300MHz GPU..

307MHz for SGX540 in 4430 and the original claim for 4440 was +25% in GPU performance over 4430 which gives 384MHz. Lord knows what the name change for 4440 stands for exactly, but I doubt it's GPU frequency related as 384MHz is already quite high.

Wikipedia is an open online encyclopedia. If I log in and change the 4440 to 384MHz does it change anything in terms of credibility?

 

[Deleted member 13524]

Wikipedia is an open online encyclopedia. If I log in and change the 4440 to 384MHz does it change anything in terms of credibility?

?

Yes, that was exactly my point in my reply to tangey, since he quoted the Wikipedia entry for the name changing.and the existance of a 4470 (as a renaming of an unnanounced 4450).

 

[tangey]

?

Yes, that was exactly my point in my reply to tangey, since he quoted the Wikipedia entry for the name changing.and the existance of a 4470 (as a renaming of an unnanounced 4450).

I original quoted TI for the name change of 4440 to 4460, I'm sure we can trust that

I have not actually heard of a 4450, which according to the wiki was renamed to 4470 at the same time as the renaming of the 4440.

You had orginally suggested in an earlier post that the archos might be an TI Omap4, based on a quote from an article, indicating a 1.6Ghz dual core, and also on archos's past of using TI.

but we don't know of a 1.6Ghz Omap4, the wiki entry refers to an unknown omap4 as 4470 (which would likely be higher performance than the known1.5Ghz version,given its name), the wiki seems to add some credence to your Archos posting. But of course it is all chicken and egg, which came first.

 

[Deleted member 13524]

It seems the Motorola DROID Bionic, initially thought to have a Tegra 2, will have an OMAP4 instead.

 

[tangey]

Some great confirmed news on Omap4 line from computex.

4460 will run at 1.5Ghz, with SGX540 running at 384Mhz, which is ties in with my thoughts from the Omap4 datasheet here:-
http://forum.beyond3d.com/showpost.php?p=1501089&postcount=22.
In production Q3 '11. This is TI's main platform for honeycomb/ ice cream sandwich.

But the real news is 4470 which I first mentioned in post #46 in this thread, is not a just a rumour but in fact gives "a different level of graphics performance", with a single core SGX544, sampling OCT 2011.
Thats not a typo, its SGX544.

This is the first known use of a single core variant of any multi-core graphics I/P from IMG.

The video also confirms that Omap5 will run dual core 544. and sampling OCT as well. Which resolves how many graphics cores that Omap5 will contain, at least initially.

Both Omap4470 and 5530 will be demoed at CES 2012.

http://www.youtube.com/watch?v=w5Cz8AENzpU&feature=player_embedded

confirmation of 384Mhz SGX540 in Omap4460 is in this video... http://www.youtube.com/watch?feature=player_embedded&v=5YfgteAvoPY

 

[Deleted member 13524]

Any word on clocks for the SGX544 implementation in OMAP4470?

Are there performance-specific differences between SGX543 and 544, or is it just DX9L3?

If they're using vec4 shaders and the same 384MHz, the OMAP4470 should be able to match the A5 in 3D performance, right?

Is the 4470 exclusive to tablets?

 

[Mariner]

Interestingly, GSMArena seems to think that the i9101 follow-up to the SGSII will contain an OMAP 4430:

http://www.gsmarena.com/specs_of_a_..._be_running_powervr_sgx_540_gpu-news-2717.php

Probably a slightly cheaper version of the SGSII, perhaps without the new Super AMOLED screen, but it would be another design win, nonetheless.

 

[Deleted member 13524]

Funny thing, the OMAP4 version of Galaxy SII may turn out cheaper but actually a bit faster for some operations because of the dual Cortex M3 for imaging and media assist.

 

[Exophase]

If they're using vec4 shaders and the same 384MHz, the OMAP4470 should be able to match the A5 in 3D performance, right?

I've heard 250MHz for SGX543MP2 in A5, so that'd be a no (unless it's something that scales poorly, which is unusual, especially on MP2)

 

[Deleted member 13524]

I've heard 250MHz for SGX543MP2 in A5, so that'd be a no (unless it's something that scales poorly, which is unusual, especially on MP2)

Really? I thought the A5 had 8-9x the graphics power of an A4, and given the A4 had a 200MHz SGX535, a 200MHz SGX543MP2 would fit within that comparison.

 

[Lazy8s]

The A4 also likely clocked the pad's GPU higher than the phone's. My guess of ~250 MHz for the iPad 2's A5 GPU came from its fill rate score in GLBenchmark 2.0 and from Apple's tendency to use the headroom of the pad for a higher clock, realizing that the addition of the second cores kept the A5 from bumping clock speeds beyond the A4.

Considering that TI is using a single core and considering the clock speeds they achieved with the last one, the OMAP4470 should be clocking its SGX544 very high, so it'll definitely be competitive with Apple's A5 in graphics.

Now that TI has shown that they're pushing the pace and using two PowerVR variants within the same generation of OMAP, their additional Series5 XT license is clearly no indicator of a longer wait for their Series6 part.