The Samsung Galaxy S8’s headline features are its edge-to-edge Infinity Display and striking new design. Of course it still comes packed with the latest hardware and technology like previous Galaxy phones, including iris recognition, wireless charging, and a flagship SoC. Actually, there are two different SoCs for the S8 and S8+. Most regions around the world will get Samsung's Exynos 8895, while regions that require a CDMA modem, such as the US and China, will get Qualcomm's Snapdragon 835. Both SoCs are built on Samsung's 10nm LPE process and are paired with 4GB of LPDDR4 RAM and 64GB of UFS NAND.

While no market receives both types of phones through official channels, with the wonders of modern shipping, anyone with a bit of time and patience would have little trouble tracking down the out-of-region version of the phone. Consequently, for the nerdy among us, we simply have to ask: how do these dueling SoCs compare? Which SoC – and consequently which phone – is better?

Today we’ll delve into the performance differences between the Snapdragon 835 and Exynos 8895 to help answer those questions. We'll also see how well they work with the Galaxy S8’s other hardware and software when we evaluate its system performance, gaming performance, and battery life.

Samsung Galaxy S8 Series
	Samsung Galaxy S8	Samsung Galaxy S8+
SoC	Qualcomm Snapdragon 835 (US, China, Japan) 4x Kryo 280 Performance @ 2.36GHz 4x Kryo 280 Efficiency @ 1.90GHz Adreno 540 @ 670MHz Samsung Exynos 8895 (rest of world) 4x Exynos M2 @ 2.31GHz 4x Cortex-A53 @ 1.69GHz ARM Mali-G71 MP20 @ 546MHz
Display	5.8-inch 2960x1440 (18.5:9) SAMOLED (curved edges)	6.2-inch 2960x1440 (18.5:9) SAMOLED (curved edges)
Dimensions	148.9 x 68.1 x 8.0 mm 155 grams	159.5 x 73.4 x 8.1 mm 173 grams
RAM	4GB LPDDR4 (US)
NAND	64GB (UFS) + microSD
Battery	3000 mAh (11.55 Wh) non-replaceable	3500 mAh (13.48 Wh) non-replaceable
Front Camera	8MP, f/1.7, Contrast AF
Rear Camera	12MP, 1.4µm pixels, f/1.7, dual-pixel PDAF, OIS, auto HDR, LED flash
Modem	Snapdragon X16 LTE (Integrated) 2G / 3G / 4G LTE (Category 16/13) Samsung LTE (Integrated) 2G / 3G / 4G LTE (Category 16/13)
SIM Size	NanoSIM
Wireless	802.11a/b/g/n/ac 2x2 MU-MIMO, BT 5.0 LE, NFC, GPS/Glonass/Galileo/BDS
Connectivity	USB Type-C, 3.5mm headset
Features	fingerprint sensor, heart-rate sensor, iris scanner, face unlock, fast charging (Qualcomm QC 2.0 or Adaptive Fast Charging), wireless charging (WPC & PMA), IP68, Mobile HDR Premium
Launch OS	Android 7.0 with TouchWiz

Our initial look at Snapdragon 835 revealed that its Kryo 280 performance cores are loosely based on ARM’s Cortex-A73 while the efficiency cores are loosely based on the Cortex-A53. Samsung's Exynos 8895 also has an octa-core big.LITTLE CPU configuration, but uses four of its own custom M2 cores paired with four A53 cores. Samsung introduced its first custom CPU core, the M1, last year. Compared to ARM’s A72, integer IPC was similar but the M1 trailed the A72 in efficiency. The M2 does not appear to be a radical redesign, but rather a tweaked M1 that offers the usual promises of improved performance and efficiency. Are the changes enough to top Qualcomm’s flagship SoC?

Battery life is one of the most important metrics for a smartphone. A bunch of cool features and lightning quick performance will do little to temper your frustration if the phone is dead by lunchtime. This was an issue for the Galaxy S6, which came with a small-capacity battery that contributed to its at-times disappointing battery life. Samsung increased their battery capacity for the S7 models, but there’s no further increase for the S8s. The smaller S8 retains the same 3000 mAh capacity as the S7, while the the S8+ drops 100 mAh compared to the S7 edge. Any improvement to battery life for this generation will need to come from more efficient hardware, and indeed at least for Qualcomm, this is precisely the angle they've been promoting to hardware developers and the public alike.

Previous Galaxy phones delivered good performance, but shortfalls in one or more performance metrics have kept them from being a class leader. Will the updates to the S8’s hardware and software finally smooth away these performance wrinkles? Will efficiency improve with the new 10nm SoCs? Did Samsung reduce power consumption in other areas? It’s time to take a closer look at the Galaxy S8.

CPU Performance

Before we evaluate the Galaxy S8’s system-level performance and battery life, we’ll run some lower-level tests to examine CPU integer and floating-point IPC and memory system throughput and latency. The first test will be SPECint2000, the integer component of the SPEC CPU2000 benchmark developed by the Standard Performance Evaluation Corporation. This collection of single-threaded workloads allows us to compare IPC for competing CPU microarchitectures. The scores below are not officially validated numbers, which requires the test to be supervised by SPEC, but we’ve done our best to choose appropriate compiler flags and to get the tests to pass internal validation.

SPECint2000 - Estimated Scores ARMv8 / AArch64
	Exynos 8895 (Galaxy S8)	Snapdragon 835 (Galaxy S8)	Snapdragon 821 (LeEco Le Pro3)	Kirin 960 (Mate 9)
164.zip	1120	1207	1273	1217
175.vpr	3889	3889	1687	4118
176.gcc	2000	1930	1746	2157
181.mcf	1268	1146	1200	1118
186.crafty	2083	2222	1613	2222
197.parser	1125	1364	1059	1395
252.eon	3333	3333	3714	3421
253.perlmk	1698	1714	1513	1748
254.gap	-	1864	1594	1930
255.vortex	2235	1900	1712	2111
256.bzip2	1351	1376	1172	1402
300.twolf	2113	2419	847	2479

The peak CPU operating frequencies for all the SoCs in the table above fall within 2% of each other, making it easier to compare IPC. It’s interesting to see how close the Exynos 8895’s M2 cores are to the Snapdragon 835’s Kryo 280 cores in integer performance. Each core separates itself in a couple tests—the S835 is faster in parser (21%) and twolf (14%) while the E8895 is faster in vortex (18%)—but generally the performance differences are less than 10%.

Both the M2 core inside E8895 and the Cortex-A73 core inside S835/K960 can dispatch 4 µops/cycle; however, there are some differences between their execution pipelines. The M1 (and from what I can tell the M2) has 2 simple ALU/INT pipes for basic operations, such as additions and shifts, and 1 complex pipe for muliplication/division. The A73 has 2 complex ALU/INT pipes. While both can handle basic operations, only one ALU handles integer multiplication and multiply-accumulate operations, while the other focuses on integer division. This means the M2 and A73 can both perform 2 basic operations in parallel but cannot perform 2 of the same complex operations in parallel. The A73 can dual issue a MUL/MAC alongside a divide/add/shift, which the M2 cannot do, but the M2 can issue 2 basic operations alongside 1 complex instead of a 1/1 split for A73. Obviously, any increase in throughput resulting from these differences will be highly workload dependent.

The chart above accounts for differences in CPU frequency by dividing the estimated SPECint2000 ratio score by CPU frequency, which makes it clear to see that, in this group of tests at least, there’s no IPC difference between Galaxy S8 models running Snapdragon 835 or Exynos 8895. We also see that there’s essentially no significant difference between Kirin 960’s A73 CPU core and Snapdragon 835’s semi-custom A73 core. Otherwise, the S835/E8895 hold a 26% IPC advantage over the previous generation Snapdragon 820/821, whose fully-custom Kryo CPU performs on par with the older ARM A57 core when running integer workloads.

While the in-order A53 CPU works very well as a lower-power companion core, building octa-core A53 SoCs does not make much sense. The bigger cores provide a significant performance advantage over the A53 – the older A57 is almost twice as fast here while the A73/M2 offer 2.2x more IPC – yielding a better overall user experience when dealing with the short, bursty workloads common to smartphone use cases. And for those of you still using a phone with a Snapdragon 801 SoC and wondering if buying a new flagship phone would deliver a noticeable performance gain, the answer is yes. The Krait 400 CPU in S801 performs about the same as the A53 in these integer workloads.

Geekbench 4 - Integer Performance Single Threaded
	Exynos 8895 (Galaxy S8)	Snapdragon 835 (Galaxy S8)	Snapdragon 821 (LeEco Le Pro3)	Kirin 960 (Mate 9)
AES	971.2 MB/s	905.4 MB/s	535.8 MB/s	911.6 MB/s
LZMA	3.20 MB/s	2.84 MB/s	2.20 MB/s	3.03 MB/s
JPEG	17.2 Mpixels/s	16.0 Mpixels/s	21.7 Mpixels/s	16.1 Mpixels/s
Canny	29.1 Mpixels/s	22.5 Mpixels/s	31.2 Mpixels/s	22.5 Mpixels/s
Lua	1.63 MB/s	1.70 MB/s	1.43 MB/s	1.72 MB/s
Dijkstra	1.25 MTE/s	1.58 MTE/s	1.41 MTE/s	1.53 MTE/s
SQLite	42.2 Krows/s	51.2 Krows/s	36.5 Krows/s	51.6 Krows/s
HTML5 Parse	7.63 MB/s	8.48 MB/s	7.48 MB/s	7.99 MB/s
HTML5 DOM	2.50 Melems/s	2.18 Melems/s	0.84 Melems/s	2.15 Melems/s
Histogram Equalization	50.3 Mpixels/s	49.9 Mpixels/s	52.3 Mpixels/s	48.6 Mpixels/s
PDF Rendering	57.4 Mpixels/s	47.2 Mpixels/s	53.6 Mpixels/s	44.6 Mpixels/s
LLVM	231.8 functions/s	250.1 functions/s	164.8 functions/s	260.4 functions/s
Camera	6.58 images/s	5.47 images/s	7.17 images/s	5.45 images/s

The updated Geekbench 4 workloads give us a second look at integer IPC. Once again the performance difference between the S835 and E8895 is generally between 5% to 10%, although there is a bit more variation in these tests. The E8895 pulls ahead in Canny (29%), PDF Rendering (22%), and Camera (20%), while the S835 holds the advantage in Dijkstra (26%) and SQLite (21%).

After accounting for differences in CPU frequency, the E8895’s M2 core shows a minimal 5% advantage over S835’s Kryo 280 core in the Geekbench integer suite. As expected, the S835 and Kirin 960 perform the same, with both showing a negligible gain relative to the previous generation SoCs with A72 CPU cores. The integer IPC gap between the S835 and S820/S821 narrows to only 11% in Geekbench 4. The S820/S821 is still no better than SoCs with the A57 core despite posting better results in many of the individual integer tests. Its poor performance in LLVM and HTML5 DOM account for its lower overall score.

We tested the IPC of the A53 core using two different SoCs; in the Snapdragon 625 the A53 performs 10% better than its counterpart in the Kirin 655, primarily because of the Snapdragon’s lower memory latency (in-order cores are particularly sensitive to latency). Both A53 examples still manage to outperform the S801 however, with a margin ranging from 13% to 24%.

Geekbench 4 - Floating Point Performance Single Threaded
	Exynos 8895 (Galaxy S8)	Snapdragon 835 (Galaxy S8)	Snapdragon 821 (LeEco Le Pro3)	Kirin 960 (Mate 9)
SGEMM	13.4 GFLOPS	11.0 GFLOPS	12.2 GFLOPS	10.5 GFLOPS
SFFT	4.02 GFLOPS	2.76 GFLOPS	3.26 GFLOPS	2.88 GFLOPS
N-Body Physics	924.5 Kpairs/s	844.5 Kpairs/s	1183.3 Kpairs/s	832.6 Kpairs/s
Rigid Body Physics	6234.9 FPS	5941.6 FPS	7169.6 FPS	5879.2 FPS
Ray Tracing	203.7 Kpixels/s	220.6 Kpixels/s	297.7 Kpixels/s	221.8 Kpixels/s
HDR	9.49 Mpixels/s	8.13 Mpixels/s	11.3 Mpixels/s	8.10 Mpixels/s
Gaussian Blur	27.2 Mpixels/s	22.2 Mpixels/s	48.0 Mpixels/s	23.7 Mpixels/s
Speech Recognition	15.3 Words/s	13.2 Words/s	11.5 Words/s	12.8 Words/s
Face Detection	583.4 Ksubs/s	512.4 Ksubs/s	681.4 Ksubs/s	497.1 Ksubs/s

While integer IPC is essentially the same between E8895’s M2 CPU and S835’s Kryo 280, the E8895’s floating-point IPC is notably higher, surpassing the S835 in every test except Ray Tracing. Its advantage over the S835 is particularly pronounced in the SFFT (46%), SGEMM (22%), and Gaussian Blur (23%) tests. The E8895 even manages to outperform the S820/S821 at times, which still has the best overall floating-point IPC of current SoCs, taking the lead in the SGEMM, SFFT, and Speech Recognition workloads.

After taking the geometric mean of the Geekbench 4 floating-point subtest scores and dividing by CPU frequency, the E8895 holds a 17% IPC advantage over the S835. The S820/S821’s IPC is still higher than both the E8895 and S835 by 11% and 31%, respectively. The S835’s Kryo 280 CPU delivers the same floating-point performance in these workloads as the A72 and A73, with all 3 CPUs showing a minimal 3-5% advantage over the older A57. Altogether, the past 3 generations of big CPU cores only show a 35% difference in floating-point IPC in Geekbench 4.

Memory Performance

The Geekbench memory tests show some performance differences between the Galaxy S8’s two SoCs. The E8895 performs significantly better than the S835 in the Memory Copy test that uses the memcpy() routine with randomized offsets; however, the S835 delivers higher bandwidth when streaming to system memory along with a 9% lower latency figure.

Geekbench 4 - Memory Performance Single Threaded
	Exynos 8895 (Galaxy S8)	Snapdragon 835 (Galaxy S8)	Snapdragon 821 (LeEco Le Pro3)	Kirin 960 (Mate 9)
Memory Copy	6.74 GB/s	4.32 GB/s	8.05 GB/s	4.60 GB/s
Memory Latency	147.5 ns	134.4 ns	150.2 ns	140.4 ns
Memory Bandwidth	14.95 GB/s	16.87 GB/s	13.93 GB/s	17.23 GB/s

It’s always interesting to look further back in time to see how performance has improved over several generations. In this case, the memory bandwidth test shows the biggest gains when comparing the Galaxy S8 to the Galaxy S5 (S801) and Galaxy S6 (E7420), where the older SoCs can only muster 6.95GB/s and 7.51GB/s, respectively, compared to at least 14.95GB/s for the S8 (E8895). The gains in the Memory Copy test are not as dramatic, with the Snapdragon version of the S8 (4.32GB/s) showing a small improvement over the S5 (3.98GB/s) and S6 (3.35GB/s). The transition from LPDDR3 to higher-frequency LPDDR4 DRAM along the way certainly helped boost performance, as did improvements in CPU microarchitecture (the AGUs in particular).

The S835’s Kryo 280 CPU comes with twice the L1 cache as the E8895’s M2 CPU, 64KB versus 32KB, respectively. The Kryo 280’s L1 latency remains steady at 1.28ns, the same as the A73 core in the Kirin 960, which is about 26% better than the M2’s 1.74ns latency figure. The S835 extends this same latency advantage to the L2 cache as well. The S835’s latency advantage over the E8895 shrinks to 9%, nearly matching the Geekbench 4 result, when accessing main memory at the upper limit of our own internal test.

Overall, the E8895’s M2 CPU core holds a small IPC advantage over the S835’s Kryo 280 core in floating-point workloads, but the two cores are pretty evenly matched when working with integers. There are a few specific workloads where each microarchitecture shines, but the theoretical performance difference between the Galaxy S8’s two SoC choices is not as large as expected.

System Performance

Now that we have a better understanding of how the Galaxy S8’s two SoC options perform, it’s time to see how well the S8’s hardware and software work together. To evaluate overall system performance, we turn to the PCMark Work 2.0 suite, which tests the combined effects of the CPU, GPU, RAM, and NAND storage. Because it uses standard Android API calls and runs several different real-world workloads that elicit realistic behavior from the CPU governor (unlike synthetic tests that simply run one or more CPUs at max frequency), it’s a good indicator of everyday performance.

Right away we see a very noticeable difference in overall performance between the two different S8 versions. The S8 with Snapdragon 835 tops the chart, scoring better than the previous champion, the Huawei Mate 9, and outpacing the Exynos 8895 version by 29%. Not only is the E8895 S8 slower than its brother, it also fares no better overall than the older Galaxy S7 (S820). Both S8’s are faster than the Galaxy S6, however, with the S835 version scoring 48% better overall and the E8895 version squeaking past with a 15% margin. Compared to the Galaxy S5 and its S801 SoC, the S8 is 100% and 55% faster, respectively.

In the Web Browsing test, the hierarchy at the top half of the chart is largely the same, although the performance deltas are less than what we saw for the overall PCMark scores. The S835 S8 claims the top spot once again and outperforms its E8895 counterpart by 19%. The Galaxy S8 (S835) and Galaxy S5 bracket the results, with the former showing a 66% advantage. The Exynos S8 is a little slower than the Snapdragon version (19%), but is at least faster than all the previous Galaxy phones by as much as 40% relative to the S5. LG phones typically do not perform as well in this test because threads spend more time running on the lower-performing A53 or efficiency CPU cores compared to other phones, presumably a deliberate decision by LG to reduce power consumption.

The Galaxy S8 (S835) delivers the best performance in the Writing test, whose workload elicits frequent, short bursts of activity from the big CPU cores. Our previous testing showed little difference in integer IPC between the S8’s two SoCs, but the S835 model performs 49% better than the S8 with E8895. The primary difference between the two models is how the scheduler migrates threads between the small and big CPU clusters. With the E8895, Samsung is a bit more conservative with moving threads onto the M2 cores. Both S8s are faster than the LG G6 and all the phones using a Snapdragon 820 SoC, though. The S8 is also between 2x and 3x faster than the Galaxy S5.

The Video Editing test’s workload is not very strenuous, so we do not see much performance variation; however, the Photo Editing test, which applies a number of different photo effects and filters using both the CPU and GPU, shows more interesting results. Qualcomm’s Adreno GPUs stand out here by combining strong ALU performance with driver optimizations. The Galaxy S8 (S835) sits among the phones using S820/S821 at the top of our chart, all sitting above the Mate 9, which is the first phone in this group to use an ARM Mali GPU. Compared to the E8895 model, the S8 (S835) performs 56% better. Curiously, the S8 (E8895) and its 20-core Mali-G71 GPU is considerably slower than the Mate 9’s 8-core Mali-G71 when running this workload.

* Note: The OnePlus 3T and Huawei P10 may be using an unpatched version of the F2FS filesystem that artificially inflates write performance.

It’s common to source NAND from multiple vendors, a practice forced into the spotlight this year with NAND chips in short supply. Earlier this year, Huawei took some heat after customers began complaining of poor storage performance in its top-tier Mate 9 and P10 phones. Huawei admitted that it was using lower-performing components due to supply shortfalls, with some customers claiming they received eMMC NAND instead of UFS.

Unfortunately, even Samsung, which is one of the major NAND suppliers, is not immune to this dilemma, sourcing NAND from at least two suppliers for the Galaxy S8. Our E8895 S8 uses Samsung UFS 2.1 NAND and performs about 2x faster in the sequential read and write tests than our S835 S8 that uses UFS 2.0 NAND from Toshiba. The S8’s (S835) write performance, both sequential and random, is particularly poor. This is not necessarily a deal breaker, however, because read performance has a greater impact on user experience.

Storage Performance (AndroBench 5.0)
	Seq. Read (KB)		Seq. Write (KB)		Random Read (KB)		Random Write (KB)
	1024	2048	1024	2048	8	16	8	16
Galaxy S8 (S835)	320.91	388.53	66.55	90.41	21.26	37.29	5.53	6.90
Galaxy S8 (E8895)	583.66	674.02	101.31	136.19	32.16	64.59	7.00	12.62
Galaxy S7 (S820)	275.80	324.24	90.52	115.43	19.76	39.65	5.76	10.56
Huawei P10	499.35	541.97	185.50	185.18	55.73	96.00	84.86	139.32
LG G6	314.68	346.02	102.62	112.12	19.24	37.37	5.87	10.84

* Note: The Huawei P10 write values are crossed out because they likely are not accurate.

The table above shows how storage performance scales with larger IO sizes. While the values we use in our general tests above represent the most common IO sizes, other workloads, notably accessing media files like photos and video, will utilize larger IO transfers, so it’s important to see how a phone performs over a wider range of scenarios. 

Throughput generally increases with IO size, and this is certainly true for the S8 (E8895), whose sequential read speed reaches 674 MB/s for 2 MB transfers, surpassing even the speedy P10. Performance for the S8 (S835) scales too, but its write performance remains consistently lower than other flagships; however, at larger IO sizes its throughput surpasses the Galaxy S7 (S820) when reading from storage.

In the JavaScript tests, both the E8895 and S835 versions of the Galaxy S8 perform similarly, with their scores differing by no more than 8%. This is not surprising considering how close the M2 and Kryo 280 CPUs are in peak frequency and integer IPC. Both S8’s perform well relative to other flagship phones (all of which had the latest updates applied and were running Chrome 57), but do not distinguish themselves as the fastest currently available.

At the other end of the scale, the Galaxy S5 manages to outperform the Moto Z Play Droid (S625) and its octa-core A53 CPU configuration. The S8 offers a much faster browsing experience, of course, not just for JavaScript but also page rendering and scrolling, outperforming the S5 by an average of 2.1x in these three tests.

The Galaxy S8 performs well overall, but isn't consistently leading the pack and thus unable to cleanly establish itself as the premiere flagship in terms of everyday performance. The S835 model, despite its poor write performance, is clearly the faster of the two models when using real-world apps. I did not have much of an opportunity to use the E8895 S8 since it was a time-limited loaner, but the user interface of the S835 model remains quite fluid while swiping and scrolling through the home screen, app drawer, recent apps menu, Edge panel, and browser. The S8 (S835) is the fastest and most fluid feeling Galaxy phone I’ve ever used.

GPU & Gaming Performance

Sitting alongside the different CPU cores in the Galaxy S8’s two SoCs are two different GPU configurations. The Snapdragon 835 includes an Adreno 540 GPU that uses the same basic architecture as the Adreno 530 found in Snapdragon 820/821. While the new Adreno 540 remains a black box, Qualcomm says it improved performance and efficiency by eliminating some bottlenecks, tweaking the register file and ALUs, and improving depth rejection. Qualcomm also used the move to 10nm to raise the max GPU frequency to 710MHz, a roughly 14% increase over S820’s peak operating point; however, Samsung caps the Adreno 540 in the Galaxy S8 to 670MHz for both the FHD+ and WQHD+ resolution settings.

The S8’s Exynos 8895 SoC comes with an ARM Mali-G71 GPU that uses ARM’s latest Bifrost architecture. We first saw the Mali-G71 in action when we looked at Huawei’s Mate 9 and P10, which both use the Kirin 960 SoC from HiSilicon. Unlike Huawei’s offerings that use the G71 in an 8-core configuration with a peak operating point of 1037MHz, Samsung went wide and slow for its E8895, with 20 cores running at up to 546MHz.

A Bifrost GPU core can process 1 pixel per clock and up to 12 FP32 FMAs. After accounting for differences in core count and frequency, the S8’s E8895 holds a 32% theoretical throughput advantage over the Kirin 960 in Huawei’s flagships. In the GFXBench ALU 2 test (run offscreen at a fixed 1080p resolution), however, the S8 (E8895) does even better, managing to outperform the Mate 9 and P10 by 59%. The S8 is using a newer GPU driver than Huawei’s phones, which likely accounts for some of this additional performance. The S8 with E8895 is also 27% faster here than the S835 version, which is a bit of an upset considering Adreno’s historically strong ALU performance. The S8 (E8895) even bests the iPhone 7 in this test.

The Adreno 540 in the S835 version of the S8 is not much faster than the Adreno 530 in the Snapdragon 820 phones in this synthetic ALU test, giving up about 5fps relative to the S835 mobile development platform we previously tested due to its lower GPU frequency. Our previous testing also showed that the Adreno 540’s microarchitecture tweaks provide no advantage here, because Adreno 530 and 540 give the same performance at the same frequency. The S8 (S835) is still more than 3x faster than the Galaxy S6 and S5, though.

The two different S8 versions swap positions in the offscreen texturing test, with the S8 (S835) pulling ahead of the E8895 version by 22%. The S8 (E8895) offers about the same level of performance as the iPhone 7 and phones using the S820/S821. It also holds a 31% advantage over the Mate 9’s G71 GPU, which happens to be very close to the 32% theoretical value based on core count and frequency.

In the synthetic tests above, the E8895 S8 had an advantage over the S835 S8 in ALU performance, but the S835 version had the edge in texturing. Now it’s time to see if this holds true while running some strenuous 3D workloads. First up is the older OpenGL ES 2.0-based GFXBench T-Rex game simulation, where the last couple generations of flagship phones have not only been hitting the 60fps V-Sync limit when running at their native onscreen resolution but averaging 60fps over the duration of the test. The S8 is no exception, averaging 60fps while running at their highest WQHD+ resolution. Even though the Galaxy S7, S6, and S5 have lower-resolution displays, they still fall short of the 60fps barrier.

Moving to the 1080p offscreen test, the Galaxy S8 (E8895) tops the chart, pulling ahead of the S835 version by 11% and the Kirin 960 in the Mate 9 and P10 by 20%. Compared to the previous generation S820 phones, including the Galaxy S7, the S8 is either 14% (S835) or 27% (E8895) faster. It’s also interesting to see that the S8’s peak performance is more than 4x higher than the S5 in T-Rex, which is the least strenuous of our game simulation tests (and the only one the S5 can even run).

The GFXBench Car Chase game simulation uses a more modern rendering pipeline and the latest features, including tessellation, found in OpenGL ES 3.1 plus Android Extension Pack (AEP). Like many current games, it stresses ALU performance to deliver advanced effects.

In the onscreen test, with the S8 set to its highest WQHD+ resolution, the two versions perform roughly the same. The E8895 S8’s small advantage is due to a slight difference in resolution: on the E8895 S8 the game ran at 2560x1440, keeping the nav buttons visible, while the S835 version defaulted to running the game full screen at 2678x1440. As expected, the S8 is faster than the S7, which uses a WQHD 2560x1440 resolution, by at least 15% compared to the S820 version and a more noticeable 57% compared to the E8890 version and its Mali-T880MP12 GPU. Stepping back one more generation to the Galaxy S6 shows that peak performance has more than doubled. It should be noted that all of the phones above the S8 in this chart benefit from using lower-resolution 1080p displays. When set to its FHD+ display mode, the S8 does outperform the OnePlus 3T by 1-2fps.

The S8 jumps to the top of the offscreen chart where resolution is no longer a factor. The E8895 version outperformed the S835 version in the GFXBench ALU 2 test, so it’s not too surprising to see the same hierarchy in this workload, although the E8895 S8’s margin of victory is narrower at only 8%. The S8 (E8895) is also faster than the Mate 9 and P10 by 55%, almost the same difference we saw in the GFXBench ALU 2 test. Again, at least some of this advantage comes from the S8’s newer GPU driver. With the S835 inside, the S8 is at least 16% faster than the S7 (S820) and the other S820 phones.

3DMark Sling Shot Extreme uses either OpenGL ES 3.1 on Android or Metal on iOS and stresses the GPU and memory subsystems by rendering offscreen at 1440p (instead of 1080p like our other tests).

The Galaxy S8 delivers the highest peak graphics performance in this test. The E8895 version performs as well as the iPhone 7, while the S835 version does even better, topping the Exynos SoC by 17%. In GFXBench Car Chase, which also stresses ALU performance, the E8895’s 20-core Mali-G71 GPU outperformed the S835’s Adreno 540, but the order flips in this workload. In the first graphics subtest, which emphasizes geometry processing and uses simpler shaders, the S835 is 9% faster than E8895, while in the second graphics subtest, which uses more mathematically complex shaders and adds volumetric illumination, the S835 is 21% faster than E8895.

Compared to the previous generation, the S8 (E8895) is only 9% faster than the S7 (E8890) in the combined graphics test, which is a little disappointing considering the E8890 uses the older Mali-T880 GPU with only 12 cores. The gap between the S8 (E8895) and the S7 (S820) is not much different, but it is 2.6x faster than the Galaxy S6.

The demanding Basemark ES 3.1 game simulation uses either OpenGL ES 3.1 on Android or Metal on iOS. It includes a number of post-processing, particle, and lighting effects, but does not include tessellation like GFXBench 4.0 Car Chase.

The iPhones take the lead in the onscreen test, partially because of their lower-resolution displays and partially because they are using Apple’s Metal graphics API, which dramatically reduces driver overhead when issuing draw calls. The Mate 9 and P10 also pull ahead of the S8 when running this test onscreen purely because their displays top out at 1080p. The S8 does deliver the best onscreen performance among the QHD resolution phones, with the S835 version outpacing the Galaxy S7 (S820), LG G6, Pixel XL, and other S820 phones by at least 18%. The E8895 S8 performs particularly well, posting a result 56% higher than the S835 version.

Hardware comparisons are a little easier when rendering at a fixed resolution offscreen. The S8 (E8895) is the fastest Android phone in this test, and its wider Mali-G71 GPU configuration bests the Kirin 960’s high-frequency approach by 16%. This is about half the E8895’s theoretical advantage when looking solely at compute/texturing resources, suggesting the E8895 is bottlenecked elsewhere. The S8 (E8895) is also considerably faster than the S835 version in this test, with the gap growing to 50%. We’ve already seen the E8895 outperform the S835 in other shader-intensive workloads, and ARM’s Mali GPUs historically handle this test’s workloads well, so this is not a huge surprise.

In addition to the results shown above, Basemark ES 3.1 also measures the performance of specific graphical features. The E8895 outperforms the S835 in all of these subtests, but it does particularly well when performing SSAO (screen space ambient occlusion), a technique used for calculating soft shadows, where it’s 58% faster than the S8 (S835). The delta between the E8895 and S835 versions shrinks to only 16% in the post-processing test (depth of field, antialiasing, etc) and 8% in the particle instancing test.

Overall the Galaxy S8 delivers excellent peak graphical performance. It offers a significant performance uplift over the Galaxy S5 and S6, although its gains over the S7 and last year’s crop of S820/S821 flagships are not as impressive. The performance delta between the E8895 and S835 versions of the S8 varies depending on workload, but the E8895 S8 is faster in most of our tests.

Battery Life

When Samsung introduced the Galaxy S7 last year, we were happy to see a larger 3000 mAh (11.55 Wh) battery inside, a significant increase over the 2550 mAh (9.81 Wh) unit inside the Galaxy S6. There is no further capacity increase this year, however. The S8 retains the same 3000 mAh (11.55 Wh) battery, which is the same capacity found in the HTC U11. The LG G6, perhaps the S8’s closest competitor, comes with a larger 3300 mAh (12.54 Wh) unit, which could help keep it running a little longer.

To see how Samsung’s two SoC choices affect the S8’s battery life, and see how it stacks up to its competitors, we’ll run it through our standard suite of battery tests. To make the tests accurate and repeatable, we control as many variables as possible, including minimizing background tasks and calibrating each display to 200 nits at 100% APL. All of the Android phones in the charts below except for the S7 (E8890) have all available software updates applied and are running the same version of Chrome.

The Galaxy S8 does quite well in our Wi-Fi browsing test that loads, pauses, and then scrolls through a set of popular websites while connected to Wi-Fi with the cellular radio turned off. We do not see much difference in runtime (only 25 minutes or 4%) between the E8895 version and the S835 version here, mainly because this workload does not utilize the big CPU cores that much, and the display tends to be the greatest power consumer in this scenario. Shutting down just shy of the 10 hour mark, the S8 outlasts the iPhone 7 by either 17 or 41 minutes, a noteworthy accomplishment considering the S8 uses a larger AMOLED display that needs to cope with relatively high APL web content (lots of white backgrounds). The S8 (S835) also lasts nearly 1.25 hours longer than the LG G6 (S821), despite the latter phone having a bigger battery.

Of the two Exynos powered Galaxies, the E8895 S8 lasts 1.5 hours longer than the E8890 S7. The E8890 is hardly the most power efficient SoC, so this likely accounts for some but not all of the difference. The gap between the Snapdragon Galaxies is even bigger, with the S835 S8 browsing for 1.9 hours longer than the S820 S7.

PCMark’s mixed workloads exercise the CPU, GPU, storage, and memory subsystems, giving us a more complete evaluation of battery life. Huawei’s Mate 9 runs longer than the S8 because of its larger 4000 mAh battery, but the S8 actually does quite well, with no appreciable difference in runtime between the two S8 models. Unfortunately, I was not able to take accurate power measurements, so we can only use average platform power consumption over the duration of the test for comparison, but the S8, regardless of SoC choice, uses considerably less power than the S5, S6, or S7 (S820). After factoring in overall PCMark performance, the S8 (S835) has the highest average efficiency of any device in this chart—about 24% better than the S8 (E8895). In the Wi-Fi Web browsing test, we saw very little difference between the two SoCs under a relatively light CPU load. The S8 (S835) manages to distinguish itself in PCMark primarily because of its more efficient GPU (Photo/Video Editing tests), which did not factor into the Wi-Fi browsing results. While not as efficient as the S835 version or the Mate 9 and its Kirin 960 SoC, the S8 (E8895) is still better than the LG G6 and the older Galaxy phones.

The S8’s settings menu offers a choice between several preset performance modes that affect sound quality, screen brightness, and screen resolution, among other things. For the results shown in the charts above, the S8s used the default “Optimized” setting that reduces the screen resolution to FHD+ (2220x1080). Running the PCMark battery test at its highest WQHD+ (2960x1440) setting has no appreciable affect on battery life, however, regardless of SoC.

Note: Both Galaxy S8 models use the default FHD+ (2220x1080) resolution setting in this test

In the GFXBench Manhattan 3.1 battery life test, which predicts runtime while playing games, there’s a significant disparity between the S8’s two SoCs. The S835 S8 lasts about an hour longer than the E8895 S8, even though both deliver nearly identical steady-state performance during most of the test. Using the average platform power to calculate efficiency (performance per watt) after the two phones reach steady-state shows the S835 S8 with a 39% advantage over the E8895 S8. Our E8895 example does seem to have a poor bin for both the CPU and GPU, an issue that seems fairly common with the early batch of 10nm SoCs, which likely skews the results somewhat, but it cannot account for this large of a gap.

This is the second SoC example we’ve seen using ARM’s Mali-G71 GPU—the first being HiSilicon’s Kirin 960—and performance per watt is disappointing for both in workloads that really stress the GPU. The GPU configurations are significantly different too: The Kirin 960 uses 8 cores running at up to 1037MHz on TSMC’s 16nm FFC process, while the Exynos 8895 uses 20 cores running at up to 546MHz on Samsung’s 10nm LPE process. Without more detailed data, this is just an interesting observation rather than a definitive statement about the G71’s or E8895’s power consumption.

While the E8895 S8’s battery life in this test falls within the 3-3.5 hour average for flagship phones, the S835 S8’s 4.1 hour mark is pretty impressive. It’s certainly not the longest lasting, but the phones that run longer deliver lower performance. For example, the S8 (S835) offers double the sustained performance of either S7 model while lasting 20 minutes longer. The Galaxy S6 outlasts the S8 (S835) by a slim margin but only because it throttles back performance so severely.

GPU Thermal Stability

The E8895 S8 achieves better peak performance than the S835 model initially, but begins to throttle back the GPU frequency just shy of 5 minutes to stay within its thermal limits. The S835 S8 maintains peak performance for about 13 minutes before throttling. Interestingly, both S8 models deliver the same performance after reaching steady-state 21 minutes into the test. The E8895 model consumes more power (and generates more heat) at the same performance level as the S835 model in this test, so it appears the E8895 S8 is using a higher thermal limit, although I was unable to confirm this in the short time I had with it.

Both Galaxy S8 models have higher sustained performance than other flagships. The Mate 9, for example, loses 38% of its peak performance after just 8 minutes, dropping to 21fps from 34fps, and remains at 19fps after about 30 minutes. The LG G6 starts at 16fps and begins throttling after 4 minutes before settling at 10fps after 13 minutes. The previous generation Galaxy S7 (S820) yields similar results to the G6: It starts at 16fps and starts throttling at the 13 minute mark, eventually settling at 8fps after 21 minutes.

Battery Charging

The Galaxy S8 comes with Samsung’s Adaptive Fast Charging technology along with both WPC and PMA wireless charging. Unlike some companies, such as Huawei and Motorola, that are pulling more than 20W of power at the battery, Samsung remains a bit more conservative. The included wall charger is rated for 5V/2A (10W) and 9V/1.67A (15W) operation—same as the Galaxy S7—which translates to a peak of 9.3W at the battery when charging with the screen off. Turning the screen on reduces the peak charging power to just 4.3W, so the S8 will not charge nearly as fast while you are using it.

The peak charging period lasts for 1 hour and 9 minutes before ramping down exponentially. This gets the S8 to 25% in 21 minutes and 50% in 42 minutes, very similar to the LG G6 and its almost 10% larger battery, which gets to 25% in 25 minutes and 50% in 46 minutes; however, the S8 is fully charged in just under 2 hours where the G6 takes 2 hours and 52 minutes. As another point of comparison, Huawei’s Mate 9, which has a much larger 4000 mAh battery, pulls up to 20.2W at the battery and reaches 25% capacity in about 12.5 minutes, 50% in about 25.5 minutes, and 100% in 2 hours and 5 minutes.

Final Words

Any discussion about the Galaxy S8’s performance begins with its two SoCs, which have some things in common, but a lot more differentiating the two. At a low level, both are built on the same Samsung 10nm LPE process. But past that, what the chip designers at Qualcomm and Samsung LSI built with that process are at times very different.

In terms of processing elements, the Snapdragon 835 uses four semi-custom Cortex-A73 CPU cores for its big cluster, while the Exynos 8895 employs four of Samsung’s custom M2 cores. Our lower-level tests show almost no overall difference in integer IPC between the two CPU cores, with each microarchitecture showing a small advantage in a few, very specific workloads. The M2 in the E8895 delivers better overall floating-point IPC, but on the whole there is not a big difference in CPU performance between the S835 and E8895, thanks to their similar IPCs and clockspeeds.

Focusing solely on the hardware’s capabilities ignores a vital piece of the puzzle, however. Software plays an important role too, particularly the parameters that control a phone’s CPU scheduling and DVFS systems. OEMs fine tune these parameters to find the right balance between performance, power consumption, and thermal limits. It’s only when running system-level tests such as PCMark, which runs more realistic workloads that use standard Android API calls, where these effects become evident and where we see a noticeable difference in performance between the two S8 models. The S835 S8 performs almost 30% better than the E8895 model overall in PCMark, with a 49% advantage in the Writing test where thread migration between the little and big clusters plays a prominent role. The storage performance of our E8895 S8 sample, which came with Samsung UFS 2.1 NAND, was significantly better than our S835 S8’s Toshiba UFS 2.0 NAND, however.

When it comes to running apps, the E8895 S8’s performance is comparable to last year’s flagships, while the S835 S8 is among the fastest currently available. There’s another aspect of performance, though, that’s more difficult to measure: user interface responsiveness and fluidity. This is an area where Galaxy phones have struggled in the past. I only had access to the E8895 S8 for a brief period (all of which was used for testing and collecting data), and I did not have the S835 model at the same time for a side-by-side comparison, so I’ll reserve my subjective opinion about UI performance to the S835 model. Overall I found it to be very fluid. Not quite as smooth as Google’s Pixel, but noticeably better than the Galaxy S7 (S820), which never felt as fast as some of its peers. The S835 S8’s performance perfectly mirrors the smooth and fluid design of its chassis.

Both models deliver excellent graphics performance, although the E8895 model and its 20-core Mali-G71 GPU is a little faster in most workloads. The flipside is that the S835 model’s Adreno 540 offers much better efficiency, prolonging battery life by an extra hour in our GFXBench Manhattan ES 3.1 battery test.

Peak performance is good for bragging rights, but what really matters when playing the most demanding games is sustained performance. Interestingly, both S8 models deliver the same steady-state performance after throttling GPU frequency to stay within their thermal limits. While neither SoC can maintain peak frequency for very long, sustained performance is still excellent, which is important if you want to use the S8 with Samsung's Gear VR system.

Battery life has also improved significantly from the S7 to the S8, even though there’s been no change in battery capacity. This comes thanks in large part to Samsung's 10nm LPE process, which has allowed chip designers to rebalance their designs to curtail power consumption while still offering a modest performance increase. Overall Samsung has definitely improved overall efficiency for this generation, however the S835 model has a clear advantage over the E8895 S8. This is particularly obvious when looking at GPU power consumption.

If you’re upgrading from a previous Android or Galaxy phone, especially one that predates the S7, the Galaxy S8’s performance and battery life will not disappoint, no matter which SoC is used. Between these two, however, across all of the tests I've run, the S835 model is certainly the better of the two in terms of those metrics.