Considering Anandtech not only identified the performance difference, but determined the cause and demonstrated a fix, you would hope Samsung would prioritize providing Anandtech samples of their phones.
While I personally detest the practice, it's easy to understand why Samsung does this. They know better than AnandTech the deficiencies of their phones, they don't need AT to tell them. The key point is they don't want AT to tell consumers, because Samsung wants to sell phones. Especially during the launch when consumer sentiment and momentum is at its peak, Samsung doesn't want negative coverage, and so that would be a deciding factor why it would withhold phones from AT.
Doesn't speak well of Samsung to do that, and one would think this wouldn't be an issue if they actually had confidence in the S10. The unimpressive performance of the Exynos vs. Snapdragon is probably one of the reasons why.
Am I missing something? It just says it's not the fastest. At current levels of performance I doubt that difference can be a dealbreaker for anyone. And if it was then everybody would flock to the fastest SoC.
These days what's around that SoC matters even more. Screen, battery, camera(s), OS updates, etc.
Once upon a time, Exynos was exciting and a dream for me as EU citizen. Now, Exynos performs under expectations. It can't even decode 60fps without stutter. I hate this chip.
This is the smart take. Samsung should just ship Exynos and support it like Apple supports their chips. Also, then there are no benchmarks one to one to worry about. And it blunts Qualcomm's power that is holding Android back. It's really the only way to catch up to Apple in security with updates.
Only because of Andrei's ingenuity in borrowing someone else's S10. Samsung's strategy has half-worked because AT can only put out this snapshot during the launch instead of a more detailed review like every other publication (although to be fair, they never go as deep as AT anyway).
Let's keep it real, I highly doubt that the highly payed engineers at samsung didn't figure out those relatively simple solutions for the 9810, I am pretty sure they got a good reasoning why they setup the SOC control the way they did.
In these kind of situations you need to be aware of the cultural differences. I've got the distinct impression samsung is being run in a very strict hierachy. If some high level manager says something, that's whats going to happen. And nobody is going to get into a discussion. Example: We have this new ssd controller that destroys all the competition. No matter the chip tyoe. Lets put it in all devices and keep the pricelevel in line with the competition. Nobody at Intels would accept that situation.
And in this case, I can imagen some manager saying. Those anandtech guys. Boycot them... Done boss.
It was the worst Soc and software implementation i've ever seen in any high-end phone, ever. Completely unacceptable, especially from such a massive company. Not fit for sale. The whole team who worked on it should have been fired. I had to get a Snapdragon 845 S9+ import. That appalling S9 Exynos SoC should have been a way bigger story on other tech news sites, but you can't trust "journalist" filth.
The Exynos 9820 is still looking bad too, but at least it's looking a lot better than last years SoC, although that really isn't saying much.
You make good points though you seem a little antagonistic. In my opinion in light of current events anyone who characterizes all journalists as "filth" casts themselves in a questionable light.
The quotes were around journalist, not filth. He was calling people who claim to be journalists, but are actually just mouthpieces for some big company's marketing department, filth.
Speaking the undeniable fact that a lot of journalists these days are no better than paid trolls, or filth, puts the person in GOOD light with SMART people.
That "fact" is extremely deniable because you didn't define your claim. Is "a lot" a specific number or a proportion? Which area of journalism are you referring to? As it is you have said nothing whilst managing to *sound* as though you made a point, which is not the same as making one.
I'd agree that a majority of tech "journalism" consists of creatively rewriting press releases and spewing platitudes about the latest product. You're saying "a lot of journalists" are no better than filth, though, and that's a much wider claim. It sounds akin to the "liberal media" propaganda foghorn being blown wherever ignorance reigns. If that's not the claim you're making then it would help to be more specific to avoid sounding like it is.
It's really sad that these phones are still made with different SoCs depending on where you buy it, and without a choice by the buyer, really.
Just make them all equal and figure out that nonsense with US Cellular networks. Just make all of them Snapdragon or make your own SoC actually work and use that everywhere. It's really a bad sign that they can't even make a phone that works globally.
I just ran this same benchmark on an s9+ with Exynos processor and the numbers I get match the snapdragon S9 version posted here. This is the score I get: Total score: 6902 Web browsing 2.0: 7316 Video editing: 5856 Writing 2.0: 6313 Photo editing 2.0: 11557 Data Man: 5012
Wow. So, you are saying that Sammy managed to fix that mess of a scheduler and the chip itself to a degree that it can compete with it's SD counterpart, only with software updates?? Nice... Don't take me wrong, but i'll have to test that out today on my friends S9+ and report back... I guess there is still hope Sammy can do the same with the Exynos 9820...
Are you using Android 9 (pie)? Anandtech didn't retest the Exynos s9+ with the latest firmware. Can you test the difference between the different firmwares?
I have two S9s: an s9+ and an s9 and neither score anywhere close to the numbers Braincruser posted. S10+ does seem like a massive upgrade in terms of performance. App loading times also seem much improved based on the youtube speed tests I've seen. Shame it didn't ship with UFS 3.0 though.
This is not stock and with custom kernel. Please, state that when you post results. Also, while you are close (tho still lacking) to the sd 845 - the sd845 will drain A LOT less power than your exynos with this kernel/SOC control.
True - am somewhat upset that Europe will get the lesser performing model and probably end up paying more.
Another downside could be that consumers read tests / comparisons made with models using the Snapdragon 855 while they will get the Exyons model at retail, so this may be misleading.
You could get any version you want on if not on eBay than most certainly on Taobao(there are export/international shipping options though it gets a bit tricky), I never go for the warranty, I think the price cut I get compared to the official fixed pricing is more than worth it.
I'm sure if they were able, they wouldn't use the snapdragon at all. I'm sure they would prefer to use their own chips in all phones but it's a contractual obligation to use Qualcomm chips in the US, and I think they still have a few more years to go on that's front
I think a part of this is the nature of competition in the different regions as well. In the USA, to combat Apple, Samsung needs to put in the best performing components, while Apple is not as entrenched in other countries, so anti-Apple or anti-American products may allow Samsung to do well, even with the Exynos. It may also be a supply issue of how many Snapdragon chips are actually available while product manufacturing ramps up.
9820 is better, but that isn't helping much as the sd855 eats it for breakfast... it's seriously faster and smoother (from what I read around the web) + the GPU is putting a lot more work and do it with great frame timing while the exynos is all over the place and got bad drivers.
Given the number of devices sold with a 9810 inside, the consumer would seem to disagree without regarding the device's fitness for purpose.
The vast majority of users don't judge the speed of their Exynos device against the speed of the Snapdragon variant. The units are sold in different markets, so most users will never experience both variants. Users typically implicitly judge their new device relative to their previous device, which is obviously a low bar, but one which leaves almost all general consumers satisfied.
Given the success of the 9810, it's hard to imagine why Samsung might sack the engineers responsible for developing it. Some of those same people were almost certainly involved in the development of this year's 9820.
Having read the above article, I don't come away with the conclusion that the 9820 is "looking bad", either. The equivalent Snapdragon is still faster in most regards, yes, but that doesn't matter one jot. The only thing that matters in the real world is whether the chip is fast enough for the workload given to it, not whether somewhere else in the world there exists an even faster chipset.
It may not "matter" as much, but knowing that just by living in another part of the world, I would get a better phone for the same price, still leaves a bad aftertaste with me.
Thanks for the update. Some editing: Under the writing scores - "The Snapdragon 855 Galaxy S10 falls in line with the QRD’s performance, which [is] excellent."
And under battery life TBD: "While the new Exynos 9820 can’t keep up to the Snapdragon 855 in terms of performance, it’s [no] longer such [a] stark difference as we saw last year."
I love how some people call any product that isn't "the best", "trash". The 9820 is a significant step forward, and while it wouldn't sell well on its own, unless it is unstable and prone to true failure(causes the OS to crash due to hardware flaws), devices that use it ARE functional. Honestly, if you were qualified when it comes to chip design, and feel YOU could make a better product, why not apply to Samsung for a high paying design position and "show them how its done"?
It's not that good, the exynos 9810 got 6030 overall - the 9820 bump is just not that much while the sd855 is quite significantly better. Waiting for the full comparison, but I think the sd855 will prove it's a lot lot better.
First paragraph: "Specifically we covered the quite large differences between units offered with Qualcomm’s Snapdragon 845 and Samsung’s own Exynos 9820." Should be 9810
Anyway, I am very much looking forward to your review. Keep up the good work, there is no one else on the internet doing such in depth analysis publicly. Thaynk you for that.
Exynos might be catching up with Snapdragon, but it's frankly hilarious to see both SoCs (and practically everything else) get bitch-slapped into the shadow realm by Apple. It seems like the Snapdragon 855 can't even beat the A10 in my iPhone 7 Plus that I bought more than two years ago, nevermind the latest and greatest chips.
Yep in worthless unoptimized benchmark. But in realworld comparison, your iPhone 7 is not even faster than SD821 phones in comparison. At best it is as fast as $150 Android phone nowadays.
I am willing to bet money that once you actually use the device for more than a few days, you'll notice the difference. And it may be because of how Android handles RAM. Also, Apple devices use some of the fastest storage tech in phones, so load times and general responsiveness will be way up.
you're right, if you're strictly talking CPU, you may not notice the difference, but combined, you'll definitely notice the iPhone's performance over this 150$ Android phone.
This is simply not true, I have exynos note 9, exynos s7e, HTC M8 and iphone 6s - the 6s is smoother overall than the note 9 running android 9. 6s is from 2015 and with A9. Apple did massive jumps in A11 and A12 regarding performance.
That's the advantage of vertical integration. They can build a SoC that performs exactly how they want, an OS built exactly for that SoC and compilers that only need to be optimized for a few very similar SoCs instead of just broad architecture-level tweaks.
Nothing can be further from the truth.If one wants to believe benchmarks SD 855 is overall a little bit better than a12 bionic, but the difference is so small, it is most probably a rounding error. Of course if it were not for your beloved Apple aand its lame secrecy every company would publish the theoretical performance of their chips and benchmarks would only be needed as an adjunct. And the reason for SD 855 being almost identical with a12 is that its GPU is so far ahead , as the CPU of a12 is ahead from any other mobile chip. But at these performance levels , it really does not matter. There's so much you can do with a cpu and a gpu on a smartphone.
so all the hype around apple, at least concening hardware is unjustified to put it very , very kindly. But when one would like to stimulate sales they can claim anything. If for nothing else this goes to show , how much overpices are iphones and samsung's huawei's flagships. You must be not well in your mind or a social show off to buy them
Only one benchmark against iPhones and the iPhones are twice as fast. The 2-year-old iPhone 7 is tied with the fastest Android device. LOL. Surely not everyone that buys phones is locked into a platform, I think it would be beneficial to include more cross-platform benchmarks in the future.
They include cross platform tests in the full review, where possible. Apple's CPU performance has been well ahead of anything from Qualcomm/Samsung for a while now. The GPU performance has been better on the Android side though, IIRC.
Wrong, the A12 GPU makes a joke of all the other GPU's (sd845/exynos 9810/kirin 980/exynos9820). The only GPU that will get close is the sd855, but from what is available now as benches - the sd855 PEAK performance is on the level of A12 SUSTAINED performance and quite bit lower than the A12 PEAK. Let's wait for the full review tho, as this is from the QRD.
I still use a 7 plus as my work phone and it is definitely slower than my pixel 2 XL. The benchmarks are meaningless to real world performance cross platform.
Luckily although Asia gets the Exynos S10, it also likely comes with a better headphone Cirrus DAC vs the Aqstic Snapdragon DAC on the SD855 version. At least this was the case with the S8 and S9.
Really? Any measurements you can point me to? Doesn't seem to match user feedback in the forums I've read here and there. Certainly the exynos note 9 /S9s I tested were the arguably best headphone sound I've ever heard and I've personally been paying close attention to this aspect of phone for a long time. Of cos my sensing of things is subjective, would love to see some measurements
Starting from sd835, qualcomm really improved their DAC. The one in sd845 is super good (if not better, atleast on par with the exynos 9810 one), I would expect even better results for the sd855.
Actually, they started with aqstic dac in snapdragon 820. I should know becos that's what my htc 10 uses (and the cirrus dac in the iPhone 6s kicked its butt) . In the absence of comparative measurements this debate can't be settled. Cirrus's track record in audio is considerably. More illustrious than qualcomms. I'm not inclined to buy into assertions regarding the latters superiority without numbers.
I like the cirrus/wolfson DAC in the exynos better as it's neutral and of high-quality. So I am not the one to go in a discussion about that, just stated what I know from around the web and yes - some of the info got numbers to it. You can search about it.
Also, aqstic DAC started in sd820, but what does have to do with the DAC in sd845 and sd855? It's like stating sd801 is the same as sd855... snapdragon for both after all. The DAC name does nothing to it's performance.
I mentioned 820 being the start of aqstic merely as a factual correction. I certainly don't think that they havnt improved on it since. Notice I made a point of comparing my htc 10 aqstic dac to same generation cirrus dac in iphone 6s. The point I was trying to make was that cirrus has more of an audio pedigree (which we both actually agree on)
Also 'you can search about it' is not a valid piece of evidence. I have most certainly searched for it and found nothing except subjective consensus that exynos galaxies tend to sound better than snapdragon galaxies.
So since the exynos is the only chipset available in the UK, getting the mate 20 pro was a good move? Glad to hear it, thanks for removing my buyers remorse. I was not expecting the base model of the S10+ to have 8gb. And finding an 8gb ram M20 was difficult
Thanks for this Andrei. In my country, pre-ordering the S10 gets you 1/3rd of the phone purchase price in freebies (Tab A, Galaxy Buds). Tmrw is the last day of the pre-order, so this review is VERY timely. Now I know to avoid, and get a phone with Kirin 980 or SD855 instead..
From the article (emphasis mine): "It’s to be noted that the comparisons I’m making today ARE ALL ON THE NEW ANDROID 9 FIRMWARES – I don’t have updated figures for the Exynos S9 or the Snapdragon Note9, BUT HAVE THE LATEST NUMBERS ON THE SNAPDRAGON S9 AND EXYNOS NOTE9"
Saw that, but as you can see some of the numbers are off with a big margin. I have the exynos note 9 running stock android 9 csb3 currently (the latest is csb5 for some regions, few days old). Not like it will make a lot of difference if the scores are correct + the conclusion will be the same overall, but those are not, I had run pcmark 10 times in 3 different totally stock firmwares (one is of the Germany region when it was first released, second are Jan/Feb for my region after a clean flash/wipe). All of them are higher than the ones in the article. The scores posted by me are not the highest I got either, just the ones I run 5m before I post.
Just ran speedometer 2.0 and the score is 37.5. In this article it's stated 34.5, the score I got is 8.7% higher (two runs).
Yes, the conclusion is the same as I said already, but he put the exynos 9810 in far grimmer light that it really is (and believe me as I tell you that the exynos 9810 is pure garbage of SOC that is more of a midrange category vs high-end one).
It's definitely nice to see significant performance upgrades with this generation of SoC. On the other hand, are they as significant when we consider past years ? It seems to me like delivering a big upgrade after a few generations of what I consider "chilling out".
I see EUV as quite a performance boost... if it can get ramped up.
Thanks Andrei! Shame on Samsung for the attempt to freeze you and AT out of the pool of early reviewers - that act by itself tells me a lot about Samsung's own doubts about their new Exynos flagship. Maybe they should stop throwing good money after bad, retire Mongoose, and just go with plain vanilla A76 big cores; that worked well for HiSilicon/Huawei. My second takehome from this is that Qualcomm's strategy of having one of their big cores run really fast to speed up single-threaded processes seems to work, and not just in theory. I expect that the 1Fast&Big+1-3Big + 4-6 Little core design will spread both to other QC SoCs and other makers SoC designs.
I know I'm in the minority, but historically the Samsung sensors had better color reproduction, and the video encode block was superior as well (Snapdragon's was bugged and produced excessive macroblocking in 4K a few years ago). I would be very interested in a side by side camera comparison, specifically looking at color reproduction and advanced video features (HDR10+ video capture may show noteable differences).
This is what happens when you design an SOC for Geekbench single core benchmark just to match Apple but fall apart as soon as you engage the second core. Samdung again doing what it is best at.
Sitting here reading this on my S6 edge wondering if I should upgrade to a S8. These all have some nice tech no doubt, but the only way you can replace a laptop -- is with a laptop. You just can't use these devices in any meaningful capacity and not drain the battery, regardless how it's sold. We're going backwards with these ever-increasing sizes.
BS irrelevant methodology and test suite, as proven many times already. The E9820 is better than K980 in all metrics save, perhaps, the energy efficiency.
The "many times" that Andrei's selection of benchmarks has been proven as BS must've completely gone over my head, because as far as I know, no one has ever tried to come up with a benchmark suite that manages to resemble real life workloads in a more accurate manner.
"Many thanks to our lab colleagues for the quick turnaround and getting us the die information and die photos of the AP/Modem die HG11-PC761-2. The die size (seal) is 8.48 mm x 8.64 mm = 73.27 mm2, on the TSMC 7FF fabbed die representing a 17.9% die shrink when compared to the Samsung 10LPP fabbed Snapdragon 845 die HG11-P7872-2."
This is smaller than A12 (83.27 mm²) and the K980 (74.13 mm²) while achieving better integration than the A12 (with modem on SoC in the 855) and better performance than the K980 in quite a few of the tests above. I imagine the subsequent part announced to have an integrated 5G modem will be closer to ~90 mm² with the standard GPU performance bump, a modification of next year's ARM high end reference designs and hexagon improvements.
Samsung announced that before releasing S10 they’re going to use the fastest storage of any android phones around UFS 3.0 don’t know if it’s true or just marketing move from Samsung?!. Wondering why Samsung didn’t add DDR5 Ram instated of DDR4x.
I’m looking forward to the full review now that the phone is publicly available. I suspect that the apparently underwhelming performance is due to a power management policy. For example, it could slow the cores to their most energy efficient speed when it determines that there isn’t a user actively interacting with the phone, like when running a scripted series of benchmarks. Sort of the opposite of automatic core boosting when certain benchmarks are detected.
Power management is really important. If they wanted, they could had unleash the exynos 9810 too to achieve fast performance/smooth experience/fast response, but this would had lead to insane power consumption to absurd levels, a lot of heat and fast thermal throttling. They can't, the M3 cores consumption is just too high and this is deeply reviewed by Andrei. I guess exynos 9820 won't be a lot different + the control of those SOCs is really not that great at all, samsung is still learning.
Galaxy S10+ Exynos got an update today making the score better, Data manipulation is a little worse but negligable, Photo editing still not reach SD levels, But overall an improvment.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
107 Comments
Back to Article
Ironhidee - Tuesday, February 26, 2019 - link
No more review units for you, Looks like Anandtech got blacklisted by Samsung for covering the massive issue in Exynos chips last year.quorm - Tuesday, February 26, 2019 - link
Considering Anandtech not only identified the performance difference, but determined the cause and demonstrated a fix, you would hope Samsung would prioritize providing Anandtech samples of their phones.warreo - Tuesday, February 26, 2019 - link
While I personally detest the practice, it's easy to understand why Samsung does this. They know better than AnandTech the deficiencies of their phones, they don't need AT to tell them. The key point is they don't want AT to tell consumers, because Samsung wants to sell phones. Especially during the launch when consumer sentiment and momentum is at its peak, Samsung doesn't want negative coverage, and so that would be a deciding factor why it would withhold phones from AT.Doesn't speak well of Samsung to do that, and one would think this wouldn't be an issue if they actually had confidence in the S10. The unimpressive performance of the Exynos vs. Snapdragon is probably one of the reasons why.
levizx - Wednesday, February 27, 2019 - link
And they think not providing samples would stop AT from telling the truth?goatgaaru - Wednesday, February 27, 2019 - link
he didn't say it would... but it would prevent them from saying the truth for the first few days post launch... It's like you didn't read what he saidlevizx - Wednesday, February 27, 2019 - link
Except it's still "first few days", and here we are.close - Wednesday, February 27, 2019 - link
Am I missing something? It just says it's not the fastest. At current levels of performance I doubt that difference can be a dealbreaker for anyone. And if it was then everybody would flock to the fastest SoC.These days what's around that SoC matters even more. Screen, battery, camera(s), OS updates, etc.
GlossGhost - Saturday, March 2, 2019 - link
Once upon a time, Exynos was exciting and a dream for me as EU citizen. Now, Exynos performs under expectations. It can't even decode 60fps without stutter. I hate this chip.nico_mach - Sunday, March 3, 2019 - link
This is the smart take. Samsung should just ship Exynos and support it like Apple supports their chips. Also, then there are no benchmarks one to one to worry about. And it blunts Qualcomm's power that is holding Android back. It's really the only way to catch up to Apple in security with updates.warreo - Wednesday, February 27, 2019 - link
Only because of Andrei's ingenuity in borrowing someone else's S10. Samsung's strategy has half-worked because AT can only put out this snapshot during the launch instead of a more detailed review like every other publication (although to be fair, they never go as deep as AT anyway).cha0z_ - Friday, March 1, 2019 - link
Let's keep it real, I highly doubt that the highly payed engineers at samsung didn't figure out those relatively simple solutions for the 9810, I am pretty sure they got a good reasoning why they setup the SOC control the way they did.Foeketijn - Tuesday, March 5, 2019 - link
In these kind of situations you need to be aware of the cultural differences.I've got the distinct impression samsung is being run in a very strict hierachy.
If some high level manager says something, that's whats going to happen. And nobody is going to get into a discussion.
Example: We have this new ssd controller that destroys all the competition. No matter the chip tyoe. Lets put it in all devices and keep the pricelevel in line with the competition. Nobody at Intels would accept that situation.
And in this case, I can imagen some manager saying. Those anandtech guys. Boycot them... Done boss.
B3an - Tuesday, February 26, 2019 - link
It was the worst Soc and software implementation i've ever seen in any high-end phone, ever. Completely unacceptable, especially from such a massive company. Not fit for sale. The whole team who worked on it should have been fired. I had to get a Snapdragon 845 S9+ import. That appalling S9 Exynos SoC should have been a way bigger story on other tech news sites, but you can't trust "journalist" filth.The Exynos 9820 is still looking bad too, but at least it's looking a lot better than last years SoC, although that really isn't saying much.
Thud2 - Tuesday, February 26, 2019 - link
You make good points though you seem a little antagonistic. In my opinion in light of current events anyone who characterizes all journalists as "filth" casts themselves in a questionable light.Lord of the Bored - Wednesday, February 27, 2019 - link
The quotes were around journalist, not filth. He was calling people who claim to be journalists, but are actually just mouthpieces for some big company's marketing department, filth.peevee - Monday, March 4, 2019 - link
Speaking the undeniable fact that a lot of journalists these days are no better than paid trolls, or filth, puts the person in GOOD light with SMART people.Spunjji - Friday, March 22, 2019 - link
That "fact" is extremely deniable because you didn't define your claim. Is "a lot" a specific number or a proportion? Which area of journalism are you referring to? As it is you have said nothing whilst managing to *sound* as though you made a point, which is not the same as making one.I'd agree that a majority of tech "journalism" consists of creatively rewriting press releases and spewing platitudes about the latest product. You're saying "a lot of journalists" are no better than filth, though, and that's a much wider claim. It sounds akin to the "liberal media" propaganda foghorn being blown wherever ignorance reigns. If that's not the claim you're making then it would help to be more specific to avoid sounding like it is.
nevcairiel - Tuesday, February 26, 2019 - link
It's really sad that these phones are still made with different SoCs depending on where you buy it, and without a choice by the buyer, really.Just make them all equal and figure out that nonsense with US Cellular networks. Just make all of them Snapdragon or make your own SoC actually work and use that everywhere. It's really a bad sign that they can't even make a phone that works globally.
Braincruser - Tuesday, February 26, 2019 - link
I just ran this same benchmark on an s9+ with Exynos processor and the numbers I get match the snapdragon S9 version posted here.This is the score I get:
Total score: 6902
Web browsing 2.0: 7316
Video editing: 5856
Writing 2.0: 6313
Photo editing 2.0: 11557
Data Man: 5012
1nterceptor_013 - Wednesday, February 27, 2019 - link
Wow. So, you are saying that Sammy managed to fix that mess of a scheduler and the chip itself to a degree that it can compete with it's SD counterpart, only with software updates?? Nice... Don't take me wrong, but i'll have to test that out today on my friends S9+ and report back... I guess there is still hope Sammy can do the same with the Exynos 9820...cha0z_ - Thursday, February 28, 2019 - link
No, this is with custom kernel and power consumption way above the sd845 for the same workload.Rudde - Wednesday, February 27, 2019 - link
Are you using Android 9 (pie)? Anandtech didn't retest the Exynos s9+ with the latest firmware. Can you test the difference between the different firmwares?Andrei Frumusanu - Wednesday, February 27, 2019 - link
The Note9 Exynos scores as detailed were running Android 9.Monty1401 - Thursday, February 28, 2019 - link
Was the benchmark on the mate 20 being run in performance mode?Monty1401 - Thursday, February 28, 2019 - link
That's interesting, have you installed custom roms or a new kernel? I've just run it myself and got much worse:Work: 5127
Web browsing: 5049
Video: 5005
Writing: 4982
Photo editing: 6568
Data: 4285
Toss3 - Saturday, March 2, 2019 - link
I have two S9s: an s9+ and an s9 and neither score anywhere close to the numbers Braincruser posted. S10+ does seem like a massive upgrade in terms of performance. App loading times also seem much improved based on the youtube speed tests I've seen. Shame it didn't ship with UFS 3.0 though.cha0z_ - Thursday, February 28, 2019 - link
This is not stock and with custom kernel. Please, state that when you post results. Also, while you are close (tho still lacking) to the sd 845 - the sd845 will drain A LOT less power than your exynos with this kernel/SOC control.Irata - Wednesday, February 27, 2019 - link
True - am somewhat upset that Europe will get the lesser performing model and probably end up paying more.Another downside could be that consumers read tests / comparisons made with models using the Snapdragon 855 while they will get the Exyons model at retail, so this may be misleading.
haukionkannel - Wednesday, February 27, 2019 - link
It is better to get snapdragon 845 that exynos 9820... it tells something,s.yu - Thursday, February 28, 2019 - link
You could get any version you want on if not on eBay than most certainly on Taobao(there are export/international shipping options though it gets a bit tricky), I never go for the warranty, I think the price cut I get compared to the official fixed pricing is more than worth it.[email protected] - Thursday, February 28, 2019 - link
I'm sure if they were able, they wouldn't use the snapdragon at all. I'm sure they would prefer to use their own chips in all phones but it's a contractual obligation to use Qualcomm chips in the US, and I think they still have a few more years to go on that's frontTargon - Friday, March 1, 2019 - link
I think a part of this is the nature of competition in the different regions as well. In the USA, to combat Apple, Samsung needs to put in the best performing components, while Apple is not as entrenched in other countries, so anti-Apple or anti-American products may allow Samsung to do well, even with the Exynos. It may also be a supply issue of how many Snapdragon chips are actually available while product manufacturing ramps up.cha0z_ - Friday, March 1, 2019 - link
9820 is better, but that isn't helping much as the sd855 eats it for breakfast... it's seriously faster and smoother (from what I read around the web) + the GPU is putting a lot more work and do it with great frame timing while the exynos is all over the place and got bad drivers.ianmacd - Friday, March 1, 2019 - link
Given the number of devices sold with a 9810 inside, the consumer would seem to disagree without regarding the device's fitness for purpose.The vast majority of users don't judge the speed of their Exynos device against the speed of the Snapdragon variant. The units are sold in different markets, so most users will never experience both variants. Users typically implicitly judge their new device relative to their previous device, which is obviously a low bar, but one which leaves almost all general consumers satisfied.
Given the success of the 9810, it's hard to imagine why Samsung might sack the engineers responsible for developing it. Some of those same people were almost certainly involved in the development of this year's 9820.
Having read the above article, I don't come away with the conclusion that the 9820 is "looking bad", either. The equivalent Snapdragon is still faster in most regards, yes, but that doesn't matter one jot. The only thing that matters in the real world is whether the chip is fast enough for the workload given to it, not whether somewhere else in the world there exists an even faster chipset.
nevcairiel - Friday, March 1, 2019 - link
It may not "matter" as much, but knowing that just by living in another part of the world, I would get a better phone for the same price, still leaves a bad aftertaste with me.GlossGhost - Saturday, March 2, 2019 - link
I couldn't agree with you more, as I said above, this chip performs very poor.zeeBomb - Saturday, March 2, 2019 - link
What was the massive issue? Something to do with thermals?Bloorf - Tuesday, February 26, 2019 - link
Thanks for the update. Some editing:Under the writing scores - "The Snapdragon 855 Galaxy S10 falls in line with the QRD’s performance, which [is] excellent."
And under battery life TBD: "While the new Exynos 9820 can’t keep up to the Snapdragon 855 in terms of performance, it’s [no] longer such [a] stark difference as we saw last year."
p3ngwin1 - Tuesday, February 26, 2019 - link
Aso this sentence :"The performance of the Snapdragon 855 is a *big* better than the QRD we tested in January..."
I suspect that should be "...855 is a bit better than the ..." :)
Nice preview article on the performance of the 855 though, looking further to a possible deeper dive in future.
SirPerro - Tuesday, February 26, 2019 - link
So essentially the SD855 is better in every single test?The_Assimilator - Wednesday, February 27, 2019 - link
Exynos is trash, what did you expect? Hell, the older 845 manages to beat the 9820 half the time.s.yu - Thursday, February 28, 2019 - link
Yeah...I still had good impressions from back in the 7420 days. Looks like these two generations are trash.Targon - Friday, March 1, 2019 - link
I love how some people call any product that isn't "the best", "trash". The 9820 is a significant step forward, and while it wouldn't sell well on its own, unless it is unstable and prone to true failure(causes the OS to crash due to hardware flaws), devices that use it ARE functional. Honestly, if you were qualified when it comes to chip design, and feel YOU could make a better product, why not apply to Samsung for a high paying design position and "show them how its done"?Lodix - Tuesday, February 26, 2019 - link
What are the overall scores ??Andrei Frumusanu - Tuesday, February 26, 2019 - link
7767 vs 9580Lodix - Tuesday, February 26, 2019 - link
Thanks :)Good improvement from Samsung, but still lagging a generation behind the Snapdragon counterpart. Hopping for future software improvements.
cha0z_ - Friday, March 1, 2019 - link
It's not that good, the exynos 9810 got 6030 overall - the 9820 bump is just not that much while the sd855 is quite significantly better. Waiting for the full comparison, but I think the sd855 will prove it's a lot lot better.Dodozoid - Tuesday, February 26, 2019 - link
First paragraph:"Specifically we covered the quite large differences between units offered with Qualcomm’s Snapdragon 845 and Samsung’s own Exynos 9820."
Should be 9810
Anyway, I am very much looking forward to your review. Keep up the good work, there is no one else on the internet doing such in depth analysis publicly. Thaynk you for that.
Edwardmcardle - Tuesday, February 26, 2019 - link
Thanks for that. Looking forward to the main event!jjj - Tuesday, February 26, 2019 - link
The benchmarks are misleading as you only list high end SoCs.In PCMark you get up to 7.5k with a P60 and maybe 6.8k overall with a SD710.
r3loaded - Tuesday, February 26, 2019 - link
Exynos might be catching up with Snapdragon, but it's frankly hilarious to see both SoCs (and practically everything else) get bitch-slapped into the shadow realm by Apple. It seems like the Snapdragon 855 can't even beat the A10 in my iPhone 7 Plus that I bought more than two years ago, nevermind the latest and greatest chips.melgross - Tuesday, February 26, 2019 - link
And there’s more software that will compare more functions between Android and iOS based products, but it’s not used as much as it should be.joms_us - Tuesday, February 26, 2019 - link
Yep in worthless unoptimized benchmark. But in realworld comparison, your iPhone 7 is not even faster than SD821 phones in comparison. At best it is as fast as $150 Android phone nowadays.jordanclock - Wednesday, February 27, 2019 - link
Got any sources to cite for those claims?Xyler94 - Wednesday, February 27, 2019 - link
I am willing to bet money that once you actually use the device for more than a few days, you'll notice the difference. And it may be because of how Android handles RAM. Also, Apple devices use some of the fastest storage tech in phones, so load times and general responsiveness will be way up.you're right, if you're strictly talking CPU, you may not notice the difference, but combined, you'll definitely notice the iPhone's performance over this 150$ Android phone.
cha0z_ - Friday, March 1, 2019 - link
This is simply not true, I have exynos note 9, exynos s7e, HTC M8 and iphone 6s - the 6s is smoother overall than the note 9 running android 9. 6s is from 2015 and with A9. Apple did massive jumps in A11 and A12 regarding performance.Irish910 - Saturday, March 9, 2019 - link
Seriously??? You wanna talk benchmarks?!Apple A10 scores 3500+ single geekbench and 6000 multi. What does the new SD855 score single??
Oh ya, around 3500. Comparing the A10 to a junk 821 is like comparing your brain to a birds brain.
jordanclock - Wednesday, February 27, 2019 - link
That's the advantage of vertical integration. They can build a SoC that performs exactly how they want, an OS built exactly for that SoC and compilers that only need to be optimized for a few very similar SoCs instead of just broad architecture-level tweaks.IUU - Thursday, February 28, 2019 - link
Nothing can be further from the truth.If one wants to believe benchmarks SD 855 is overall a little bit better than a12 bionic, but the difference is so small, it is most probably a rounding error. Of course if it were not for your beloved Apple aand its lame secrecy every company would publish the theoretical performance of their chips and benchmarks would only be needed as an adjunct. And the reason for SD 855 being almost identical with a12 is that its GPU is so far ahead , as the CPU of a12 is ahead from any other mobile chip. But at these performance levels , it really does not matter. There's so much you can do with a cpu and a gpu on a smartphone.IUU - Thursday, February 28, 2019 - link
so all the hype around apple, at least concening hardware is unjustified to put it very , very kindly. But when one would like to stimulate sales they can claim anything. If for nothing else this goes to show , how much overpices are iphones and samsung's huawei's flagships. You must be not well in your mind or a social show off to buy themtechguymaxc - Tuesday, February 26, 2019 - link
Only one benchmark against iPhones and the iPhones are twice as fast. The 2-year-old iPhone 7 is tied with the fastest Android device. LOL. Surely not everyone that buys phones is locked into a platform, I think it would be beneficial to include more cross-platform benchmarks in the future.Thermogenic - Tuesday, February 26, 2019 - link
They include cross platform tests in the full review, where possible. Apple's CPU performance has been well ahead of anything from Qualcomm/Samsung for a while now. The GPU performance has been better on the Android side though, IIRC.cha0z_ - Friday, March 1, 2019 - link
Wrong, the A12 GPU makes a joke of all the other GPU's (sd845/exynos 9810/kirin 980/exynos9820). The only GPU that will get close is the sd855, but from what is available now as benches - the sd855 PEAK performance is on the level of A12 SUSTAINED performance and quite bit lower than the A12 PEAK. Let's wait for the full review tho, as this is from the QRD.Featherinmycap - Tuesday, February 26, 2019 - link
2 year old iPhone 7? You mean 3 year ago iPhone 7. This year it's XR/XS; Last year is X and 8; year before that iPhone 7.Speedfriend - Wednesday, February 27, 2019 - link
I still use a 7 plus as my work phone and it is definitely slower than my pixel 2 XL. The benchmarks are meaningless to real world performance cross platform.Ahadjisavvas - Tuesday, February 26, 2019 - link
Is there any way to benchmark the npu inside the exynos 9820 to see if its performance is actually good?SSTANIC - Tuesday, February 26, 2019 - link
Go Andrei!! Massive respect.NICOXIS - Tuesday, February 26, 2019 - link
The sd845 S9 is faster than the 9820 S10 ???Vitor - Tuesday, February 26, 2019 - link
I'm actually impressed by the Kirin 980.Lau_Tech - Wednesday, February 27, 2019 - link
Luckily although Asia gets the Exynos S10, it also likely comes with a better headphone Cirrus DAC vs the Aqstic Snapdragon DAC on the SD855 version. At least this was the case with the S8 and S9.Andrei Frumusanu - Wednesday, February 27, 2019 - link
The Snapdragon DAC was better in the S9.Lau_Tech - Wednesday, February 27, 2019 - link
Really? Any measurements you can point me to? Doesn't seem to match user feedback in the forums I've read here and there. Certainly the exynos note 9 /S9s I tested were the arguably best headphone sound I've ever heard and I've personally been paying close attention to this aspect of phone for a long time. Of cos my sensing of things is subjective, would love to see some measurementscha0z_ - Wednesday, February 27, 2019 - link
Starting from sd835, qualcomm really improved their DAC. The one in sd845 is super good (if not better, atleast on par with the exynos 9810 one), I would expect even better results for the sd855.Lau_Tech - Wednesday, February 27, 2019 - link
Actually, they started with aqstic dac in snapdragon 820. I should know becos that's what my htc 10 uses (and the cirrus dac in the iPhone 6s kicked its butt) . In the absence of comparative measurements this debate can't be settled. Cirrus's track record in audio is considerably. More illustrious than qualcomms. I'm not inclined to buy into assertions regarding the latters superiority without numbers.cha0z_ - Friday, March 1, 2019 - link
I like the cirrus/wolfson DAC in the exynos better as it's neutral and of high-quality. So I am not the one to go in a discussion about that, just stated what I know from around the web and yes - some of the info got numbers to it. You can search about it.Also, aqstic DAC started in sd820, but what does have to do with the DAC in sd845 and sd855? It's like stating sd801 is the same as sd855... snapdragon for both after all. The DAC name does nothing to it's performance.
Lau_Tech - Friday, March 1, 2019 - link
I mentioned 820 being the start of aqstic merely as a factual correction. I certainly don't think that they havnt improved on it since. Notice I made a point of comparing my htc 10 aqstic dac to same generation cirrus dac in iphone 6s. The point I was trying to make was that cirrus has more of an audio pedigree (which we both actually agree on)Also 'you can search about it' is not a valid piece of evidence. I have most certainly searched for it and found nothing except subjective consensus that exynos galaxies tend to sound better than snapdragon galaxies.
MyFluxi - Wednesday, February 27, 2019 - link
So since the exynos is the only chipset available in the UK, getting the mate 20 pro was a good move? Glad to hear it, thanks for removing my buyers remorse. I was not expecting the base model of the S10+ to have 8gb. And finding an 8gb ram M20 was difficultzer0hour - Wednesday, February 27, 2019 - link
Thanks for this Andrei. In my country, pre-ordering the S10 gets you 1/3rd of the phone purchase price in freebies (Tab A, Galaxy Buds). Tmrw is the last day of the pre-order, so this review is VERY timely. Now I know to avoid, and get a phone with Kirin 980 or SD855 instead..cha0z_ - Wednesday, February 27, 2019 - link
The exynos note 9 scores are not current. With android 9 the SOC control is changed and scores as follow:Work 2.0 performance score (i.e. overall): 6015
Web browsing 2.0 score: 5523
Video editing score: 5222
Writing 2.0 score: 5912
Photo editing 2.0 score: 8766
Data manipulation score: 5265
Multiple runs keeps it around those values. GPU is around 10-12% faster in all scenarios too (probably new GPU driver).
cha0z_ - Wednesday, February 27, 2019 - link
P.S. This is not the best score, but the last one. For example in some of the benches I had over 6000 for the Writing 2.0 score.tuxRoller - Wednesday, February 27, 2019 - link
From the article (emphasis mine):"It’s to be noted that the comparisons I’m making today ARE ALL ON THE NEW ANDROID 9 FIRMWARES – I don’t have updated figures for the Exynos S9 or the Snapdragon Note9, BUT HAVE THE LATEST NUMBERS ON THE SNAPDRAGON S9 AND EXYNOS NOTE9"
cha0z_ - Thursday, February 28, 2019 - link
Saw that, but as you can see some of the numbers are off with a big margin. I have the exynos note 9 running stock android 9 csb3 currently (the latest is csb5 for some regions, few days old). Not like it will make a lot of difference if the scores are correct + the conclusion will be the same overall, but those are not, I had run pcmark 10 times in 3 different totally stock firmwares (one is of the Germany region when it was first released, second are Jan/Feb for my region after a clean flash/wipe). All of them are higher than the ones in the article. The scores posted by me are not the highest I got either, just the ones I run 5m before I post.tuxRoller - Wednesday, March 6, 2019 - link
So you've got one of the good phones. What's that to do with the scores Andrei achieved with the tested device?cha0z_ - Thursday, February 28, 2019 - link
Just ran speedometer 2.0 and the score is 37.5. In this article it's stated 34.5, the score I got is 8.7% higher (two runs).Yes, the conclusion is the same as I said already, but he put the exynos 9810 in far grimmer light that it really is (and believe me as I tell you that the exynos 9810 is pure garbage of SOC that is more of a midrange category vs high-end one).
GC2:CS - Wednesday, February 27, 2019 - link
It's definitely nice to see significant performance upgrades with this generation of SoC.On the other hand, are they as significant when we consider past years ? It seems to me like delivering a big upgrade after a few generations of what I consider "chilling out".
I see EUV as quite a performance boost... if it can get ramped up.
cknobman - Wednesday, February 27, 2019 - link
Gary Explains did a full test of the S10+ SD vs Exynos here: https://www.youtube.com/watch?v=MgRnbm9wrzgThe Exynos was down right embarrassing!!
eastcoast_pete - Wednesday, February 27, 2019 - link
Thanks Andrei! Shame on Samsung for the attempt to freeze you and AT out of the pool of early reviewers - that act by itself tells me a lot about Samsung's own doubts about their new Exynos flagship. Maybe they should stop throwing good money after bad, retire Mongoose, and just go with plain vanilla A76 big cores; that worked well for HiSilicon/Huawei.My second takehome from this is that Qualcomm's strategy of having one of their big cores run really fast to speed up single-threaded processes seems to work, and not just in theory. I expect that the 1Fast&Big+1-3Big + 4-6 Little core design will spread both to other QC SoCs and other makers SoC designs.
beginner99 - Wednesday, February 27, 2019 - link
At this point I really wonder whats the point of the exynos SOC?wrkingclass_hero - Wednesday, February 27, 2019 - link
I know I'm in the minority, but historically the Samsung sensors had better color reproduction, and the video encode block was superior as well (Snapdragon's was bugged and produced excessive macroblocking in 4K a few years ago).I would be very interested in a side by side camera comparison, specifically looking at color reproduction and advanced video features (HDR10+ video capture may show noteable differences).
Lavkesh - Wednesday, February 27, 2019 - link
This is what happens when you design an SOC for Geekbench single core benchmark just to match Apple but fall apart as soon as you engage the second core. Samdung again doing what it is best at.NoSoMo - Wednesday, February 27, 2019 - link
Sitting here reading this on my S6 edge wondering if I should upgrade to a S8. These all have some nice tech no doubt, but the only way you can replace a laptop -- is with a laptop. You just can't use these devices in any meaningful capacity and not drain the battery, regardless how it's sold. We're going backwards with these ever-increasing sizes.AceMcLoud - Thursday, February 28, 2019 - link
Slowly catching up to the iPhone 7. Great progress 👍darkich - Thursday, February 28, 2019 - link
BS irrelevant methodology and test suite, as proven many times already.The E9820 is better than K980 in all metrics save, perhaps, the energy efficiency.
N Zaljov - Thursday, February 28, 2019 - link
The "many times" that Andrei's selection of benchmarks has been proven as BS must've completely gone over my head, because as far as I know, no one has ever tried to come up with a benchmark suite that manages to resemble real life workloads in a more accurate manner.Raqia - Thursday, February 28, 2019 - link
Techinsights posted some initial info about the 855:https://techinsights.com/about-techinsights/overvi...
"Many thanks to our lab colleagues for the quick turnaround and getting us the die information and die photos of the AP/Modem die HG11-PC761-2. The die size (seal) is 8.48 mm x 8.64 mm = 73.27 mm2, on the TSMC 7FF fabbed die representing a 17.9% die shrink when compared to the Samsung 10LPP fabbed Snapdragon 845 die HG11-P7872-2."
This is smaller than A12 (83.27 mm²) and the K980 (74.13 mm²) while achieving better integration than the A12 (with modem on SoC in the 855) and better performance than the K980 in quite a few of the tests above. I imagine the subsequent part announced to have an integrated 5G modem will be closer to ~90 mm² with the standard GPU performance bump, a modification of next year's ARM high end reference designs and hexagon improvements.
buggz - Sunday, March 3, 2019 - link
Is Samsung still limiting access to the complete camera API?128bit - Sunday, March 3, 2019 - link
Samsung announced that before releasing S10 they’re going to use the fastest storage of any android phones around UFS 3.0 don’t know if it’s true or just marketing move from Samsung?!. Wondering why Samsung didn’t add DDR5 Ram instated of DDR4x.cha0z_ - Tuesday, March 5, 2019 - link
s10e,s10,s10+ are all with UFS 2.1 (same as the note 9, for example). The galaxy fold got UFS 3.0 and for now only that phone.bartoni - Friday, March 8, 2019 - link
I’m looking forward to the full review now that the phone is publicly available. I suspect that the apparently underwhelming performance is due to a power management policy. For example, it could slow the cores to their most energy efficient speed when it determines that there isn’t a user actively interacting with the phone, like when running a scripted series of benchmarks. Sort of the opposite of automatic core boosting when certain benchmarks are detected.cha0z_ - Monday, March 11, 2019 - link
Power management is really important. If they wanted, they could had unleash the exynos 9810 too to achieve fast performance/smooth experience/fast response, but this would had lead to insane power consumption to absurd levels, a lot of heat and fast thermal throttling. They can't, the M3 cores consumption is just too high and this is deeply reviewed by Andrei. I guess exynos 9820 won't be a lot different + the control of those SOCs is really not that great at all, samsung is still learning.Roy2002 - Saturday, March 9, 2019 - link
Kirin 980 is very competitive, unexpected.Keks - Thursday, March 14, 2019 - link
Terrible and not real review. PCMark is proved trash. Check Geekbench. Both Exynos 9820 and 9810 are leading over Snapdragon and Kirin.yeeeeman - Thursday, March 14, 2019 - link
How much until we get the actual review?yeeeeman - Tuesday, March 19, 2019 - link
Any update?yeeeeman - Wednesday, March 20, 2019 - link
Andrei, suntem nerabdatori :) Cam cat mai ai de lucrat la articolul final?GhostArashi - Saturday, March 30, 2019 - link
Galaxy S10+ Exynos got an update today making the score better, Data manipulation is a little worse but negligable, Photo editing still not reach SD levels, But overall an improvment.Overall: 8712
Web: 9423
Video: 5753
Writing: 9771
Photo: 15705
Data: 6002
bandinelli - Tuesday, September 17, 2019 - link
The benchmarks are misleading as you only list high end SoCs.In PCMark you get up to 7.5k with a P60 and maybe 6.8k overall with a SD710 https://free1080pporn.com/category/fake-celebrity-...