Renesas combined Modem/app Soc for 2012.

[tangey]

Over on eetimes, there is an interview with the Renesas CEO, in which he is very upbeat about their chances of taking significant market share over time in the mobile sector.
http://www.eetimes.com/electronics-...d-growth-for-its-global-handset-market-share?

In a question regarding their 543MP product, he indicated that their app processor would be designed into various platforms, and specifically includes game consoles.

now waaaay towards the start of the NGP thread,(http://forum.beyond3d.com/showpost.php?p=1517219&postcount=113), I speculated whether Renesas would be the manufacturer of NGP's Soc, given the timeframe of the announcement of renesas's own chip. Has the NGP Soc designer been determined yet ? If not I'd suggest that the above would add credence to it being a Renesas designed/manufactured item.

Talking about why his outlook is so upbeat......

"...He says the catalyst to this boom will be LTE. Armed with a proven and trusted LTE protocol stack Renesas gained from Nokia, Kawasaki likes his odds."

The eetimes article go on to say that they'll have a 28nm combined modem/App Soc mass produced in 2012. One wonders if the recent Renesas licence announcment of further use of the PowerVR 5XT family of graphics cores is related to this chip ?

http://www.imgtec.com/corporate/newsdetail.asp?NewsID=629

 

[Arun]

In a question regarding their 543MP product, he indicated that their app processor would be designed into various platforms, and specifically includes game consoles.

now waaaay towards the start of the NGP thread,(http://forum.beyond3d.com/showpost.php?p=1517219&postcount=113), I speculated whether Renesas would be the manufacturer of NGP's Soc, given the timeframe of the announcement of renesas's own chip. Has the NGP Soc designer been determined yet ? If not I'd suggest that the above would add credence to it being a Renesas designed/manufactured item.

It's definitely a custom SoC so it could just be a coincidence it's the same GPU IP, but yeah now that I think about it, I'm not sure who else it could be anyway. The Renesas mobile group is the result of the merger between Renesas, NEC, Casio and Hitachi. The only other japanese options are Fujitsu and Panasonic but I don't think either is as likely.

Talking about why his outlook is so upbeat......

"...He says the catalyst to this boom will be LTE. Armed with a proven and trusted LTE protocol stack Renesas gained from Nokia, Kawasaki likes his odds."

Heh that's not much of a selling point, but at least it's better than Motorola:

He did sort of take blame for the delay though and dropped this little nugget today, “We were delayed on two 4G products because of network interaction issues with our protocol stack.”

The odd thing is, that just 30 seconds later when asked how they can prevent such a delay from happening again, Jha made reference to having burned bridges and “betting everything on one place.” We can only assume he’s referring to their decision to go with NVIDIA’s Tegra 2 processor

Actually, he's referring to their own internal LTE baseband. They made their own chip with probably their own protocol stack and PHY, just like Samsung/LG/Nokia did. The most successful of those is Samsung's (used in quite a few devices), LG is also shipping today (including at Verizon) via their joint chip with GCT with integrated RF (probably a SiP), and Nokia's baseband is what Renesas is talking about here.

Meanwhile, Motorola's baseband/stack is a mess and it delayed two of their flagship devices. So I don't think the problem is with Tegra 2, and one possible explanation for the change to OMAP4 in the Droid Bionic is that the original Bionic was completely cancelled and replaced by another product which was already in development and used a non-Motorola LTE baseband (probably Qualcomm). Keep in mind the goal for all these companies is building a patent portfolio for LTE and none of them is going to use proprietary basebands in the long term.

But I'd say protocol stack problems are more the exception than the rule. Many merchant solutions already have market proven LTE stacks already including Qualcomm, ST-Ericsson, and Icera. Also seemingly solid but not as proven yet are Intel (Infineon), Broadcom (Beceem), Sequans, and Altair.

You can also buy solid protocol stacks and PHY sotware from mimoOn or 4M Wireless (a few of the above use those, including 4M for Wireless IIRC) along with PHY hardware from Tensilica or (more area efficient) CEVA. The barrier to entry is much lower than it ever was in the 3G market, and interestingly the main problem by far for a new company would be the 2G/3G protocol stacks for multimode. Renesas does have a very strong position there at least.

But more importantly, Qualcomm/ST-E/Icera do extensive reuse between algorithms and the 2G/3G/4G PHYs to save a lot of die area in their *current* generation chips sampling today. Icera easily does the most, Qualcomm the least, and ST-Ericsson between the two, but these are not naive fixed-function solutions. Unlike the Nokia-designed Renesas solution which indeed is a naive fixed-function solution.

It'll be interesting to see what they come up next, but architecturally, there's not much to be confident about just yet. There's no clear gap in the worldwide LTE market for them to fill and their solution is far from extraordinary. They do have a lot of resources though, so they could definitely surprise positively in the future.

One wonders if the recent Renesas licence announcment of further use of the PowerVR 5XT family of graphics cores is related to this chip ?

Yeah, I'd think so. I'd rather bet on 544 than 554 but I could be wrong.

 

[Ailuros]

554 sounds increasingly to be a TI affair. After seeing 540 and 544 within OMAP4 (where one of the biggest increases is floating point power), I wouldn't be surprised if OMAP5 would use something like 544 in the beginning and down the line a 554.