Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/snapdragon-810-overheating-issue-or-not-an-issue.5393/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Snapdragon 810 overheating issue (or not an issue!)

hist78

Well-known member
I just read an UDN.COM report saying HTC will introduce their new HTC One M9 (Hima) at Mobile World Congress (MWC), March 2~5 in Barcelona. An interesting thing in that report is that this new phone will come with Qualcomm's Snapdragon 810 and the alleged overheat problem has been resolved. It went on saying Samsung still insists there is an overheat problem and pressures Qualcomm to move more orders from TSMC to Samsung otherwise Samsung will use their own processor, Exynos.

Another report from Bloomberg mentioned Samsung to drop Snapdragon 810:
Samsung Drops Qualcomm Chip From Next Galaxy Smartphone - Bloomberg

And from Reuters: LG doesn't think Sanpdragon 810 has the overheat issue:
"I don't understand why there is a issue over heat" ~ Woo Ram-chan, LG vice president for mobile product planning
LG Electronics says no overheating issue with new Qualcomm processor | Reuters


So does Snapdragon 810 really have the overheat problem at this moment?
 
You need to realize that we're talking about different revisions of the chip. All reports are true but you need to learn the context.

LG stating that the Flex2 doesn't overheat might just as well be technically true but then what throttled performance did they have to go down to?
 
You need to realize that we're talking about different revisions of the chip. All reports are true but you need to learn the context.

LG stating that the Flex2 doesn't overheat might just as well be technically true but then what throttled performance did they have to go down to?

Agree. But I assume a workable revision should be available to all Qualcomm's major customers, such as LG, HTC, Xiaomi, and Samsung. I don't think Qualcomm will withhold a solution from Samsung, one of their biggest customer.

In terms of imposing processor throttling to control the heat, that might be true or might not be true. The industry will know the truth once LG releases their G Flex 2 phone using Snapdragon 810 on January 30 in Korea.
 
Last edited:
Can someone post a link to either Qualcomm or Samsung officially commentating on this? Or is this just rumors from analysts and media? I spoke to multiple friends at Qualcomm after this story broke and was told it was absolutely false.
 
Agree. But I assume a workable revision should be available to all Qualcomm's major customers, such as LG, HTC, Xiaomi, and Samsung. I don't think Qualcomm will withhold a solution from Samsung, one of their biggest customer.

In terms of imposing processor throttling to control the heat, that might be true or might not be true. The industry will know the truth once LG releases their G Flex 2 phone using Snapdragon 810 on January 30 in Korea.
The problem with their fixed revision is that it seems to have missed the Q1 flagship launch window (S6 especially). The early adopters will probably use the broken edition.
 
I just heard that Samsung is looking to dump their Exynos group to a major fabless company. Care to guess who?
 
I just heard that Samsung is looking to dump their Exynos group to a major fabless company. Care to guess who?
If true, I sure hope it's not Qualcomm. The industry doesn't need even more consolidation. Sadly I can't name any good fit that would benefit the consumer.

I also find it very weird to do this just when things are looking extremely positive for them.
 
Last edited:
The entire article is based on a false premise. They claim that Exynos 7420 is 14nm, while in fact Samsung's own website says Exynos 7 series is 20nm:

Samsung Exynos

Sure, there is at least one Exynos 7 chip, the 7410, built at 20nm, so you are right. Anyway, I have been told that Samsung has really a 14nm pilot line based on one Exynos 7 device (7420?), so there is nothing false in that premise.
 
Sure, there is at least one Exynos 7 chip, the 7410, built at 20nm, so you are right. Anyway, I have been told that Samsung has really a 14nm pilot line based on one Exynos 7 device (7420?), so there is nothing false in that premise.

Ok, you might be right.
But still, given that Exynos 5433/7410 only really started shipping last November, I wonder when the 7420 will start shipping, maybe summer?
 
Recent history tells us that every time Samsung announces a phone with an Exynos chip inside, they introduce a different SKU in the US with a Qualcomm part to get through carrier qualifications in a reasonable amount of time. Until Verizon or AT&T buy in, Samsung doesn't have a lot of leverage.
 
Recent history tells us that every time Samsung announces a phone with an Exynos chip inside, they introduce a different SKU in the US with a Qualcomm part to get through carrier qualifications in a reasonable amount of time. Until Verizon or AT&T buy in, Samsung doesn't have a lot of leverage.
GSM carriers such as AT&T were planned with a Exynos and some other modem anyway, recently CDMA carriers will now also receive a 7420 version with a dedicated Qualcomm modem as the S810 was totally dropped from the initial release.
 
GSM carriers such as AT&T were planned with a Exynos and some other modem anyway, recently CDMA carriers will now also receive a 7420 version with a dedicated Qualcomm modem as the S810 was totally dropped from the initial release.


I believe Qualcomm charge a premium for that. I mean if you pair a Qualcomm SoC with Qualcomm modem the price is much cheaper than if you pair the same modem with a 3rd party SoC. The EU is currently investigating Qualcomm for that practice.
 
Last edited:
I believe Qualcomm charge a premium for that. I mean if you pair a Qualcomm SoC with Qualcomm modem the price is much cheaper than if you pair the same modem with a 3rd party SoC. The EU is currently investigating Qualcomm for that practice.
And that's why Samsung will only do this with Verizon and Sprint and avoid Qualcomm for everything else.
 
And that's why Samsung will only do this with Verizon and Sprint and avoid Qualcomm for everything else.
Make sense. But why Samsung doesn't go further to use non-Qualcomm solutions on phones made by Samsung for Verizon and Sprint? Unless they don't have a viabel solution yet in terms of technology, quality, and cost.
 
Make sense. But why Samsung doesn't go further to use non-Qualcomm solutions on phones made by Samsung for Verizon and Sprint? Unless they don't have a viabel solution yet in terms of technology, quality, and cost.
Because there's no other option with CDMA compatibility.

In any case, if someone was still doubting the overheating reports, cat's out of the bag:
AnandTech | The HTC One M9 Review: Part 1

Looks like the same broken revision as initially on the Flex2. Qualcomm's in a world of hurt right now.
 
Because there's no other option with CDMA compatibility.

In any case, if someone was still doubting the overheating reports, cat's out of the bag:
AnandTech | The HTC One M9 Review: Part 1

Looks like the same broken revision as initially on the Flex2. Qualcomm's in a world of hurt right now.

Either the S810 overheat is an issue or not, the article you mentioned might not provide the clear root cause that link to the S810.

Quote from the article you mentioned:

"Case in point, our GFXBench 3.0 battery life results were significantly altered by the update. With the initial version of the phone’s software we hit 1.73 hours – the phone ran fast but almost unbearably hot – and after the software update the One M9 is over 3 hours on the same test with a maximum temperature of 45C, a still-warm but certainly much cooler temperature, as seen in the photo above."

Thank you for sharing the link. I think for them to run battery test for 1.73 hour to drain the whole battery, they have to run heavy duty load constantly in that duration. But what exactly attributed to the root cause about the excess heat? They didn't attempt to find it out (not their goal) and I agreed it's hard to pinpoint out the heat sources. And, a 1.73 hour battery usage is not a real life situation. It's for testing battery life.
 
Back
Top