A few days back, one of our readers shared a link of a comparison between the Helio P70 and Snapdragon 660. It felt really bad seeing a comparison done solely on the basis of synthetic benchmarks.
So, I decided to do a comparison that would do justice to these 2 amazing processors.
Today, I’ll be comparing different features of these two SOCs not just on the basis of theoretical knowledge but backed up with real-world testing over the last few months.
This should give you a better idea of how well does the specification directly translates into real-world performance. My aim here is to help consumers make the right decision based on how they use their smartphone.
But before we get started, here’s a reminder that there also exists an underclocked Snapdragon 660 with a clock speed of 1.95 GHz that can be found in devices like Realme 2 Pro and Asus Zenfone Max Pro M2.
But in this review, we will be comparing the actual Snapdragon 660 (2.2 GHz) that is present in smartphones like Xiaomi Redmi Note 7, Mi A2, and Nokia 7 Plus.
Or you can say that we are directly comparing the performance of Redmi Note 7 against the Realme 3.
First, let’s have a look at the specification.
||Samsung’s 14nm LPP FinFET
||TSMC’s 12nm FinFET
||4x Kryo 260 (Cortex-A73) + 4x Kryo 260 (Cortex-A53)
||4x Cortex-A73 + 4x Cortex-A53
|Max Clock Speed
||Yes, via Qualcomm AI Engine
||Multi-Core AI Engine + NeuroPilot
|Max. Download Speed
||Yes, Dual VoLTE
||Yes, Dual VoLTE
||Quick Charge 4
||Pump Express 4.0
When going through the specification, both the SOCs seem quite similar. One of the major differences is that Snapdragon 660 is fabricated by Samsung’s 14nm Process while MediaTek’s Helio P70 is fabricated using TSMC’s 12nm Process.
We will get back to the difference between these fabrication technologies in the next section.
The CPU Core Configuration on both the chips is pretty similar except for the fact that the CPU Cores of Snapdragon 660 are enhanced variants of the existing Cortex-A73 & Cortex-A53 architectures. You can also call them semi-custom cores.
But in case of Helio P70, they have directly implemented the ARM’s designs without any customizations.
The maximum frequency of the 660 can reach up to 2.2 GHz while it is limited to 2.1 GHz in case of P70. The MediaTek Helio P70 also has a better AI Engine.
In the GPU department, we have the Adreno 512 and Mali-G72 MP3 GPUs. Both of them are quite good when it comes to 1080p Gaming and can run most Android games without any hiccups.
Both have a similar Memory Configuration, support Fast Charging and support Dual 4G VoLTE.
Snapdragon 660 has a faster Modem and it also supports the latest Bluetooth v5.
I’ve divided the comparison into 7 different sections and I’ll be giving a score to each of them in every section. In the end, we will combine the scores and see which SOC comes out to be the winner.
Most of you that know the basics about Processors will probably guess that 12nm is more Power efficient than 14nm because lower nanometer number is better, right?
Wrong! Well, it’s a lot more complicated than that.
Once there was a time when these nanometer values actually used to mean something and had a physical significance. But due to the slowdown of Moore’s Law and to mislead their consumers, most companies have started naming these fabrication nodes as per their marketing strategy.
In fact, I have written a whole article where I’ve explained all this. You can check it out here.
Yes, 12nm sounds so much better than 14nm, but it is one of those strategies that brands do to mislead their consumers.
TSMC’s 12nm Process is just a marketing gimmick and an enhanced version of their 16nm Node. After they enhanced their 16nm Node, it became somewhat equivalent to Samsung’s 14nm Node. However, they decided to name it 12nm to gain an edge during marketing.
Even some of the biggest Tech Youtubers & Bloggers become a victim of this false marketing and end up telling their audience than 12nm is much better than 14nm.
In fact, Samsung’s 14nm isn’t a true 14nm either. The only brand that has the highest transistor density and power efficiency at 14nm is none other than Intel.
Here’s a graph showing different fabrication nodes of brands like Intel, Samsung, and TSMC alongside their Transistor Densities.
You can clearly see that at each Fabrication Node, every brand has a different transistor density. What Intel calls as 10nm is similar to what Samsung or TSMC call as 7nm.
The transistor density of TSMC’s 12nm is still slightly higher than Samsung’s 14nm Node, but the Power efficiency of Samsung’s 14nm still has a slight edge. This can be confirmed using a long-term Battery Test.
Let’s look at GSMArena’s Battery Endurance Tests, of Realme U1 (Helio P70), Realme 2 Pro (Underclocked SD660), and Redmi Note 7 (SD660). Here are the results.
|Realme 2 Pro
|Redmi Note 7
In the last column, I have divided the Endurance Rating with Battery Capacity and multiplied with 1000 to obtain how many hours does 1000 mAh battery in each phone lasts. You can clearly see that the results of all three phones in Battery tests are pretty similar.
The battery of underclocked SD660 lasts the longest due to a slower clock speed while Helio P70 has the lowest power efficiency due to the reasons I have explained above.
Hence, we can conclude that the power efficiency of Samsung’s 14nm Process is on par or slightly better than TSMC’s 12nm Node.
But I won’t go too harsh on Helio P70 here and will give it the same score as Snapdragon 660 because the difference in results isn’t that significant.
Here are the scores in this section.
Both the SOCs have a pretty good CPU Performance, thanks to the Cortex-A73 Cores in their high-performance cluster.
While Helio P70 directly implements the ARM’s Cortex-A73 & Cortex-A53 CPU designs, Snapdragon 660 uses the same architecture with some optimizations of its own. Qualcomm calls this new architecture as a Semi-Custom Kryo 260 Core.
Well, the performance gains due to the optimizations in Kryo 260 aren’t that significant, but that does improve the overall stability and user experience of the device.
What makes a difference here is the clock speeds. The maximum CPU frequency of Snapdragon 660 & Helio P70 is 2.2 GHz & 2.1 GHz respectively. This is why SD660 devices score higher in Geekbench Tests.
Hence, the Snapdragon 660 gives 5-10% higher performance during heavy usage. But remember, this only applies during heavy usages such as gaming and other intensive workloads.
The difference in real-world performance or while performing daily-life tasks is negligible.
The gaming performance of both the processors is incredible as well. But once again, the Adreno 512 GPU has a slight edge over Mali-G72 MP3.
I know there are many people that usually assume every Mali GPU to be bad. But this is not always the case. You can refer to our Mobile GPU Rankings to understand which GPU is better than the other.
Both the Adreno 512 and Mali-G72 MP3 can run games like PUBG Mobile at high settings without any lags. I can highly recommend you both these SOCs if gaming performance is your priority.
Due to slightly better FPS and consistent performance of Adreno 512, Snapdragon 660 is once again is the winner by a narrow margin.
The latest LPDDR4x RAM is supported by both these SOCs. Also, Dual-Channel Memory is supported in both the products.
Additionally, Snapdragon 660 also supports the older LPDDR4 RAM while P70 supports both LPDDR4 & LPDDR3 RAM.
It’s a choice of your smartphone manufacturer to decide which Memory module they want to pair up with that smartphone. Hence, make sure to go through the phone’s specification to find out which RAM module is used.
In Snapdragon 660, the Memory frequency is 1,866 MHz while it is 1,800 MHz in case of Helio P70. I guess it is fair to give them equal points in this case.
Helio P70 has a dedicated AI Processing Engine coupled with MediaTek NeuroPilot that delivers superior AI Performance. Though, Snapdragon 660 doesn’t have a dedicated AI Processor, it uses its existing CPU and GPU paired up with Hexagon Vector Processor to drive AI related tasks.
Nevertheless, its an easy win for Helio P70 in this section. Even the AI related benchmarks confirm the same theory.
But how much does AI Performance actually impact our smartphone usage?
Well, some of the most common tasks that benefit from the AI Performance are: Face Unlocking, Capturing Portrait shots, and AI Scene Detection.
It totally depends on the users whether they use these features on a regular basis or not.
Snapdragon 660 comes with a superior X12 Modem that delivers download speeds up to 600 Mbps and Upload speeds up to 150 Mbps. The maximum downlink speed in Helio P70 is limited to 300 Mbps.
Dual 4G VoLTE technology is supported by both the SOCs.
There are some other factors as well that might not affect every user but I can’t skip them either.
Let’s start with the Bluetooth version. Many of you probably use Bluetooth Headphones and Fitness Bands that are connected with phones for hours.
The Snapdragon 660 uses the latest Bluetooth v5.0 which consumes less power and delivers higher bandwidth than the Bluetooth v4.2 in Helio P70. Furthermore, the range of Bluetooth v5.0 is higher.
Though this should not affect every user, if you use a Bluetooth device regularly, this point may apply to you.
Another factor is Custom ROMs and development on XDA Developers.
If you’re a power user who likes to root his/her phone and flash Custom ROMs, then a device with Snapdragon 660 will be a better choice due to the availability of Kernel Sources. Some MediaTek devices also get good support at XDA Developers but it is generally not as great as compared to popular Snapdragon devices.
So, which SOC is the winner?
After summing up the scores, here’s what we obtain.
That’s pretty close, right?
Yes, some of you can consider it a draw but that fact that Snapdragon 660 has better CPU & GPU performance is already good enough to declare it the winner. If you already checked our Smartphone Processor Rankings, you probably knew that this was coming.
Nevertheless, both these SOCs are great for midrange devices and I’ll highly recommend them if performance is what you need the most.
This content was originally published here.