Low-cost emerging market SoC/phone discussion

  • Thread starter Deleted member 13524
  • Start date
To follow-up the above, the IMG twitter account confirmed that the website is officially associated with IMG, but it accepts 3rd party contributions. They asked the editor to remove the article.

Make of that what you will.
 
I have experience on browsing with 512 MB memory phones now, using a particular dead/zombie browser-oriented OS.
It's likable enough but not always.

What I'm interested in is memory density: do we need to consider lpddr4 only? What I would like best is 2GB on a single DRAM chip. Pair with an advanced low end SoC on 28nm with one or two A53 clusters.

I believe USB C can be very useful on the low end : that's for a guaranteed 5 volts at three amps whereas there is much 500 mA hardware out there, whether it's the power supply or device or both.
I don't know if you need a new SoC for that, if last year's Mediatek support it. USB 2.0 type C would be great - low end phone and SD are nowhere near the 30MB/s ceiling on transfers. Not sure if I'd want hdmi or audio and whether you can get both or not.

I'd chime on display and OS but will leave it as is.
 

Of more interest is the fact that Mediatek is skipping 16nm and going straight for 10nm. And they say available in devices mid 2017

Also their entire mid-range and low-end line is still on 28nm unlike Qcom, Samsung, and Huawei who have all announced Octa Core A53 SoCs on 14/16nm.
 
The following is a transcript from the conference call after Mediateks earnings report on 28th Oct.
http://finance.yahoo.com/news/edited-transcript-2454-tw-earnings-150551848.html

Many of the analyst questions relate to the X30.

There are some hints regarding Samsung being a user.
They are hopeful of being the first 10nm app pro in the market (though Samsung might beat them). Expect it to outsell the X20. Cite improvements to CPU and GPU as key areas.

They also talk about overall improved power consumption of the soc. They see camera functionality and power consumption more important than VR compatibilty.
 
Pardon me for being too bored to actually read it....the notion that X30 will do better than X20 doesn't surprise me to be honest. X30 is the successor of two former quite successful SoC Generations for MTK. Since I'm using almost in parallel a X10 and a X20 smartphone from the same manufacturer the first thing that sticks out to any layman comparing them is how much less power the latter consumes. You don't even have to measure anything, you can feel the difference in temperatures right in our hand.
 
Pardon me for being too bored to actually read it....the notion that X30 will do better than X20 doesn't surprise me to be honest. X30 is the successor of two former quite successful SoC Generations for MTK. Since I'm using almost in parallel a X10 and a X20 smartphone from the same manufacturer the first thing that sticks out to any layman comparing them is how much less power the latter consumes. You don't even have to measure anything, you can feel the difference in temperatures right in our hand.

The impression I'm left with is that they see this soc progressing them further up the performance scale, and so getting more seats in higher-end smartphones, not least because they will be the first (or possibly the 2nd) 10nm soc to market.
 
The impression I'm left with is that they see this soc progressing them further up the performance scale, and so getting more seats in higher-end smartphones, not least because they will be the first (or possibly the 2nd) 10nm soc to market.

It's almost self explanatory that if a product brand is to scale in sale figures it'll also have to improve somewhere. The chapter now with the underlying process is more marketing than essence, since being an early adopter with an as new process can have both advantages as well as pitfalls.
 
The following is a transcript from the conference call after Mediateks earnings report on 28th Oct.
http://finance.yahoo.com/news/edited-transcript-2454-tw-earnings-150551848.html

Many of the analyst questions relate to the X30.

There are some hints regarding Samsung being a user.
They are hopeful of being the first 10nm app pro in the market (though Samsung might beat them). Expect it to outsell the X20. Cite improvements to CPU and GPU as key areas.

They also talk about overall improved power consumption of the soc. They see camera functionality and power consumption more important than VR compatibilty.

Samsung will definitely beat them..they will be shipping in Q1 and by all indications MT will be shipping only in Q2. Apple is also rumoured to ship by late Q1. Outselling the X20 isnt going to be too difficult IMHO.
Pardon me for being too bored to actually read it....the notion that X30 will do better than X20 doesn't surprise me to be honest. X30 is the successor of two former quite successful SoC Generations for MTK. Since I'm using almost in parallel a X10 and a X20 smartphone from the same manufacturer the first thing that sticks out to any layman comparing them is how much less power the latter consumes. You don't even have to measure anything, you can feel the difference in temperatures right in our hand.

I dont quite agree that the X20 has been successful..its been used by only a handful of customers and the shipments have been low too. Snapdragaon 650/652 have been far more successful..despite being on 28nm. Mediatek as a whole has been underperforming and I'll quote from the transcript here "Gross margin for the quarter was 35.2%, which was unchanged from previous quarter and down 7.5 percentage points year over year.". The 10nm process is certainly an advantage though and there is a chance that X30 may do well..but a 14nm refresh of the S650/652 could again disrupt the market.
 
Samsung will definitely beat them..they will be shipping in Q1 and by all indications MT will be shipping only in Q2. Apple is also rumoured to ship by late Q1. Outselling the X20 isnt going to be too difficult IMHO.
As I said the transcript varies between hoping to be the first, and being one of the first.

Apple socs aren't a competitor to Mediatek. Apple only supplies Apple. So in relation to their stated hopes of getting sales based on having a leading process, Apple shipments dates are irrelevant. And anyhow, apple won't be shipping 10nm in a smartphone until next Sept/Oct.
 
Samsung will definitely beat them..they will be shipping in Q1 and by all indications MT will be shipping only in Q2. Apple is also rumoured to ship by late Q1. Outselling the X20 isnt going to be too difficult IMHO.

*ahem* http://www.fudzilla.com/news/processors/42014-qualcomm-confirms-a-10nm-snapdragon


I dont quite agree that the X20 has been successful..its been used by only a handful of customers and the shipments have been low too. Snapdragaon 650/652 have been far more successful..despite being on 28nm. Mediatek as a whole has been underperforming and I'll quote from the transcript here "Gross margin for the quarter was 35.2%, which was unchanged from previous quarter and down 7.5 percentage points year over year.". The 10nm process is certainly an advantage though and there is a chance that X30 may do well..but a 14nm refresh of the S650/652 could again disrupt the market.

Problem here being that X20 and any Helio Xxx SoC is representing MTK's "high end" smartphone SoCs which is typically high margin low volume compared to anything else. You'll get a clue only of their sales if MTK would bother to release revenue figures per market and market segment which is highly unlikely. What I've heard in the background is that Helio SoCs are step by step in an upswing and considering that X10 didn't exactly take the market by storm I can believe it unless someone has some detailed specific data about it. Else: even if X20 did fine, if they didn't manage to sell the volume they had projected for, then flat or decline isn't much of a surprise.

Other than that, yes QCOM did fine http://www.fudzilla.com/news/mobile/42013-qualcomm-bets-street-with-6-2-b-revenue-in-q4-16 and I'm VERY surprised no one mentioned here the HUGE acquisition of NXP for wooping $47 billion.
 
As I said the transcript varies between hoping to be the first, and being one of the first.

Apple socs aren't a competitor to Mediatek. Apple only supplies Apple. So in relation to their stated hopes of getting sales based on having a leading process, Apple shipments dates are irrelevant. And anyhow, apple won't be shipping 10nm in a smartphone until next Sept/Oct.

I meant 10nm in general..and I forgot to mention Qualcomm actually as rumours point to them shipping in Q1 as well.

I dont think either of those will actually have any affect on MTK or X30 though. X30 will not compete with the 10nm Exynos or 10nm Snapdragon just like X20 does not compete with E8890 or S820 today. Its more of a upper mid range product if anything and will probably compete with the S650/652 refresh (I'm assuming there is one)

Oops..yes I forgot to mention Qualcomm..they should be shipping in Q1 as well.
Problem here being that X20 and any Helio Xxx SoC is representing MTK's "high end" smartphone SoCs which is typically high margin low volume compared to anything else. You'll get a clue only of their sales if MTK would bother to release revenue figures per market and market segment which is highly unlikely. What I've heard in the background is that Helio SoCs are step by step in an upswing and considering that X10 didn't exactly take the market by storm I can believe it unless someone has some detailed specific data about it. Else: even if X20 did fine, if they didn't manage to sell the volume they had projected for, then flat or decline isn't much of a surprise.

Other than that, yes QCOM did fine http://www.fudzilla.com/news/mobile/42013-qualcomm-bets-street-with-6-2-b-revenue-in-q4-16 and I'm VERY surprised no one mentioned here the HUGE acquisition of NXP for wooping $47 billion.

Actually..according to the transcript..they kind of say that X20 margins are on the low side. They expect X30 to improve by several percentage points. Yes true they dont release revenue figures by market segment so its difficult but usually the sales figures for specific phones give us somewhat of a clue. For example sales of the Indian Redmi Note 3 and the Le 2 are good indicators of S650/652 sales (record sales in the millions). I haven't heard anything about the sales of the phones with X20. The number of customers for the X20 is also very limited compared to S650/652 so again quite telling. Overall Mediatek has been facing intense competition from Qualcomm at the top and Spreadtrum at the bottom and hence the margins have been falling.

We missed the smaller acquisition of Brocade by Broadcom too - http://www.wsj.com/articles/chip-maker-broadcom-to-buy-brocade-for-5-5-billion-1478089036 (And remember that Broadcom itself was purchased by Avago relatively recently!)
 
X30 will not compete with the 10nm Exynos or 10nm Snapdragon just like X20 does not compete with E8890 or S820 today. Its more of a upper mid range product if anything and will probably compete with the S650/652 refresh

GSMArena offers an apples-to-apples comparison between the X20 and Snapdragon 650, with their Xiaomi Redmi Note 4 review (comparing to Redmi Note 3 S650):

http://www.gsmarena.com/xiaomi_redmi_note_4-review-1504p5.php

Yeah, except for the multi-core test in Geekbench (it is after all a 6-core versus a 10-core solution) they're very close. Single-core and GPU results are so similar it's embarrassing.
One could argue the X20 could have lower power consumption because it's actually a 3-module "big.LITTLE.LITTLER" arrangement and it's done using 20nm (unlike S650's old 28nm), but turns out the Note 3 S650 has longer battery life.
 
Back
Top