The Samsung Galaxy S9 and S9+ Review: Exynos and Snapdragon at 960fps
by Andrei Frumusanu on March 26, 2018 10:00 AM ESTBattery Life - A Stark Contrast
We extensively covered the performance of the Snapdragon and Exynos Galaxy S9’s – what remains to be seen is how that performance affects battery life in our standard tests. Performance between the regular and larger Galaxy S9 doesn’t change, however battery life may differ based on the variant as the regular S9 sports a 3000mAh (11.55Wh) battery versus the 16% higher 3500mAh (13.47Wh) of the Galaxy S9+. Naturally the 14.2% bigger screen area of the S9+ offsets some of that advantage.
Unfortunately for this review we couldn’t get identical variants of the different SoC Galaxy S9s – our S9+ is a Snapdragon 845 unit while our S9 is an Exynos 9810 unit, so we weren’t able to execute a true apples-to-apples comparisons between the SoC variant, however as we’ll see the delta between the units is large enough that it won’t change the conclusion.
Our web browsing test tries to mimic real-world usage patterns of browsing websites. This includes iterating through a list of websites and scrolling through them. In this test, screen efficiency and battery size play a role, but also we balanced it so that it also sufficiently stresses the SoC (CPU, GPU and display pipeline) as well.
The Snapdragon 845 Galaxy S9+ posted excellent battery life in our test and lands only third to the iPhone 8 Plus and the Mate 9. Unfortunately we never tested the S8+ to see the generational difference, but it shouldn’t be too different from the S835 regular S8 at around the 10 hour mark.
The Exynos 9810 Galaxy S9 absolutely fell flat on its face in this test and posted the worst results among our tracking of the latest generation devices, lasting 3 hours less than the Exynos 8895 Galaxy S8. This was such a terrible run that I redid the test and still resulted in the same runtime.
I investigated the matter further to try to see if this was caused by the high energy usage of the M3 cores – and it seems it is. Enabling the “CPU limiter” (S9 PS result in the graphs) which is found in the battery optimisation options of Samsung’s firmware greatly throttles the M3 cores down to 1469 MHz, memory controller to half speed and also seemingly changes some scheduler settings to make them more conservative. This results in peak performance equal to the Exynos 8895- however the scheduler alterations also noticeably slow down UI responsiveness so it’s actually a worse experience. Nevertheless, backing off on performance results in regaining almost 3 hours.
This is such a terrible battery performance of the Exynos 9810 variant that it again puts even more clout into the new SoC. My theory as to why this happens is that not only do the higher frequency state require more energy per work done than competing SoCs – because this is a big CPU complex there’s also lots of leakage at play. The DVFS system being so slow might actually be bad for energy here as we might be seeing the opposite of race-to-sleep, walk-to-waste. The fact that Apple’s SoCs don’t have any issues with battery life in this test showcases that it’s not an inherent problem of having a high-power micro-architecture, but rather something specific to the Exynos 9810.
In PCMark the disadvantage of the Exynos 9810 S9 isn’t as pronounced as in the web test, however it’s again a regression to the Exynos 8895 S8 – all while not posting a meaningful performance advantage over its predecessor that might explain the lower battery life.
The Snapdragon 845 Galaxy S9+ fared relatively well, even though it’s not quite as good as other devices.
In my personal every-day usage I can’t saw that I noticed a massive disadvantage in battery life on the Galaxy S9, however my everyday usage is relatively light and I haven’t had enough time with the phone yet as a daily driver to make a final judgment. I did notice that the Exynos 9810 does shows signs of suffering in heavy tasks. Instances of Gmail syncing my inbox with a new account did once result in a warm phone while the Snapdragon 845 Galaxy S9 did not showcase this characteristic.
I can’t fault the Snapdragon S9+ in the time I had it, but again I haven’t had enough real time with it to really judge it subjectively. As far as AnandTech testing goes, the data speaks for itself and based on what I’ve seen I strongly do not recommend the Exynos variant of the Galaxy S9 in its current state, especially if you’re a regular user of heavy apps.
Over the coming weeks I’m planning to try to dive into the workings of the Exynos 9810 and post a follow-up article on whether it’s possible to improve both in terms of performance as well as battery life if one changes the way the SoC’s scheduler and DVFS works. In the eventuality that Samsung updates its firmware to resolve these large issues with the Exynos Galaxies, then we’ll revisit the matter as soon as possible.
Facebook
Twitter
RSS
190 Comments
View All Comments
id4andrei - Tuesday, March 27, 2018 - link
All reviewers go gaga for geekbench scores with iphones/ipads as well. In this case the GB scores prove that at least in chip design Samsung has made a huge leap. As the review has outlined, the problem lies with the scheduler and DVFS which Samsung can and should address.If "Samdung" is so bad at hardware design, how do you call Apple's high priced iphones of the last 3 years that could not sustain chip performance and had to be throttled so as to not crap out. All initial reviews were glowing but they were all impervious to the impeding throttling.
name99 - Tuesday, March 27, 2018 - link
Dude, you really do yourself no favors by struggling so hard to criticize Apple.Apple's throttling has NOTHING to do with the CPU per se (ie the CPU is not generating excessive heat beyond spec, or because it has been running too fast for too long), it has to do with the BATTERY and with a concern that, if CPU performance were to spike the battery could not supply enough current.
Very different problem, nothing to do with the CPU design. A real problem yes but totally irrelevant to the issues being discussed here.
Matt Humrick - Wednesday, March 28, 2018 - link
Apple's big CPU and GPU are susceptible to thermal throttling when running sustained workloads too.Also, having to throttle a processor within a year of sale because its transient current requirements overwhelm the power delivery system is most definitely a design flaw.
Icehawk - Friday, April 6, 2018 - link
My wife’s 6S is still working at 100% after several years, I get the feeling the amount of people affected is overblown as pretty much anything anti-Apple is. I do think Apple needs to look at a better way of dealing with this but it’s also not the armeggedon somemake it out to be. I am far from a Apple fanboy but I do like their iOS products but I am sure someone will make a retort of that nature. I’d say the same thing about the Samsung chip - not great but it is performant, perhaps if we stop thinking each year a new phone should blow us away it would help us be more realistic.Lavkesh - Tuesday, March 27, 2018 - link
"In this case the GB scores prove that at least in chip design Samsung has made a huge leap" - Please explain huge leap here? The new chip barely outperforms the older SOC.ZolaIII - Monday, March 26, 2018 - link
I am very disappointed with both SoC's. Qualcomm wasted so much space on bad L4 cache which only added to latency & generally wasted more. The 30% is enormous even if new A75 cores are 35% bigger (would be 50% with ARM's L2 reference cache size) I don't know about A630 vs A540 size but if it grown-up let's say 10% the cores & GPU would together accommodate for around 15~20% leaving L3 & L4 responsible for the rest. Would be much better they used it for GPU as it could had been 2x the size then. I am also very disappointed with new cache hierarchy as it turns out to be stupid and a waist of silicone. Seams to me neither SoC used good scheduler nor scheduling by the looks of things it seems Samsung used the CAF HPM sched settings for Snapdragon SoC very aggressive patched interactive without any restraints whatsoever & no hotplug whatsoever which is very south from optimal, reference QC platform seams to had at least used hotplug (as their is no other way to explain the difference of almost 1W in GPU testing as two vs four A75's active). On the other hand seems Samsung used Power aware schaduler instead HPM & very granulated hotplug producing very bad results as those are directly confronted two things & when splashed together can only result in catastrophic result. I prefer HPM configured to be used with limited task packing and a high priority tasks enabled with significant increase of time interval for it (so that it can skip CPU sched limit), for CPU sched interactive traditional not patched with tree step load limitations (idle so that it doesn't jump erratic on any back shade task, ideal that is considered as best sustainable leakage for given lithography & max sustainable for two core's [only on big cores] i also use boost enabled & set to ideal frequency one [same as in interactive]). Preferred to use core_ctl hotplug disabled for the two little & two big cores so that they never get switched off from it. I won't go further in details about it hire as its pointless. I find this idea balanced between always available/needed/total performance as most of the times two of each course are enough for most of tasks & if not it's not a biggie to wait for other two to kick in. There is a minor drow back in responsiveness on lite task's but actually it works as fast as possible on hard one's flagged as heavy tasks like for instance Chrome rendering. It's also very beneficial to GPU workloads where even switching of two little core's and giving even 100~150mv headroom to GPU means much.Sorry for getting a bit deep regarding how complete scheduling mechanism should be done but I had an urge to explain how it should be done as it's so terrible done in the both cases examined hire.
tuxRoller - Wednesday, March 28, 2018 - link
It's not at all clear that the hpm is meaningfully better (much faster or much more power efficient) than a proper schedtune + energy model implementation.Scheduling is just ridiculously hard. Adding the constraints of: soft-realtime requirements, minimal battery usage, AND an asmp and you've got the current situation where there's not yet a consensus design. We are, however, starting to see signs of convergence, imho.
zeeBomb - Monday, March 26, 2018 - link
I came...and I finally sawphoenix_rizzen - Monday, March 26, 2018 - link
Ouch. The Exynos S9 is just barely better than the Exynos S7. :( And that's what Canada's going to get.Here's hoping they can improve things via software updates. Was considering the S9 to replace the wife's now dead S6. She's been using my S7 for the past two months while I limp along with a cracked-screen Note4. Other than the camera and screen, this isn't looking like much or an upgrade for being two generations newer.
Maybe we'll give the ZTE, Huawei, and Xiaomi phones another look ...
mlauzon76 - Monday, March 26, 2018 - link
Samsung Exynos 9810 (Europe & Rest of World)Canada is the 'rest of [the] world', but we don't get that version, we never get anything with the Exynos processor, we get the following one:
Qualcomm Snapdragon 845 (US, China, Japan)