ARM A53/A57/T760 investigated - Samsung Galaxy Note 4 Exynos Review
by Andrei Frumusanu & Ryan Smith on February 10, 2015 7:30 AM EST20nm Manufacturing Process
Both Samsung Semiconductor and TSMC delivered their first 20nm products in Q3 2014, but they don't represent the same jump in efficiency. Samsung's 28nm HKMG process varied a lot from TSMC's 28nm HPM process. While Samsung initially had a process lead with their gate-first approach when introducing 32nm HKMG and subsequently the 28nm shrink, TSMC went the route of gate-last approach. The advantage of the gate-last approach is that it allows for lower variance in the manufacturing process and being able to allow for better power characteristics. We've seen this as TSMC introduced the highly optimized HPM process in mobile. Qualcomm has been the biggest beneficiary as they've taken full advantage of this process jump with the Snapdragon 800 series as they moved from 28nm LP in previous SoCs.
In practical terms, Samsung is brought back on even terms with TSMC in terms of theoretical power consumption. In fact, 28nm HPM still has the same nominal transistor voltage as Samsung's new 20nm process.
Luckily Samsung provides useful power modeling values as part of the new Intelligent Power Allocation driver for the 5422 and 5430 so we can get a rough theoretical apples-to-apples comparison as to what their 20nm process brings over the 28nm one used in their previous SoCs.
I took the median chip bin for both SoCs to extract the voltage tables in the comparison and used the P=C*f*V² formula to compute the theoretical power figure, just as Samsung does in their IPA driver for the power allocation figures. The C coefficient values are also provided by the platform tables.
We can see that for the A15 cores, there's an average 24% power reduction over all frequencies, with the top frequencies achieving a good 29% reduction. The A7 cores see the biggest overall voltage drop, averaging around -125mV, resulting in an overall 40% power reduction and even 56% at the top frequency. It's also very likely that Samsung has been tweaking the layout of the cores for either power or die size; we've seen this as the block sizes of the CPUs have varied a lot between the 5410, 5420 and 5422, even though they were on the same process node.
While these figures provide quite a significant power reduction by themselves, they must be put into perspective with what Qualcomm is publishing for their Krait cores. The Snapdragon 805 on a median speed bin at 2.65GHz declares itself with a 965mW power consumption, going down to 57mW at 300MHz. While keeping in mind that these figures ignore L2 cache power consumption as Qualcomm feeds this on a dedicated voltage rail, it still gives us a good representation of how efficient the HPM process is. The highest voltages on the S805 are still lower than the top few frequencies found on both the 5430 and the 5433.
20nm does bring with itself a big improvement in die size. If we take the 5420 as the 28nm comparison part and match it against the 5430, we see a big 45% decrease on the A7 core size, and an even bigger 64% reduction on the A15 core size. The total cluster sizes remain relatively conservative in their scaling while shrinking about 15%; this is due to SRAM in the caches having a lower shrinking factor than pure logic blocks. One must keep in mind that auxiliary logic such as PLLs, bus interfaces, and various other small blocks are part of a CPU cluster and may also impact the effective scalability. Samsung also takes advantage of artificially scaling CPU core sizes to control power consumption, so we might not be looking at an apples-to-apples comparison, especially when considering that the 5430 is employing a newer major IP revision of the CPU cores.
| Exynos 5420 vs Exynos 5430 block sizes | ||||
| Exynos 5420 | Exynos 5430 | Scaling Factor | ||
| A7 core | 0.58mm² | 0.4mm² | 0.690 | |
| A7 cluster | 3.8mm² | 3.3mm² | 0.868 | |
| A15 core | 2.74mm² | 1.67mm² | 0.609 | |
| A15 cluster | 16.49mm² | 14.5mm² | 0.879 | |
The Mali T628 between the 5420 and the 5430 actually had an increase in die size despite the process shrink, but this is due to a big increase in the cache sizes.
Samsung regards their 20nm node as very short-lived and the 5430 and 5433 look to be the only high volume chips that will be coming out on the process as their attention is focused on shipping 14nm FinFET devices in the next few months. In fact at the Samsung Investor Forum 2014 they announced mass production of a new high-end SoC has already begun mid-November and will be ramping up to full volume in early 2015. I suspect this to be the Exynos 7420 as that is the successor SoC to the 5433.
All in all, the argument that this 20nm chip should be more power efficient than the competitors' 28nm is not completely factual and doesn't seem to hold up in practice. The process still seems young and unoptimized compared to what TSMC offers on 28nm.
Before we get to the performance and power figures, I'm handing things over to Ryan as we take a look at the architectural changes, starting with an analysis of the Cortex A53.
Facebook
Twitter
RSS
135 Comments
View All Comments
MrSpadge - Tuesday, February 10, 2015 - link
Excellent work, guys! But let me point out something:1. Your Geekbench numbers show spectacular gains for the A53. They're actually so good that the multi-threading speed-up almost always exceeds the number of cores (judging by eye, have not run the numbers). For A7 the speedup factor is a bit less than 4, right where it should be. It certainly looks like the big cores were kicking in, at least partly.
2. You wonder that power scales sub-linearly with more cores being used. Actually this is just what one would expect in the case of non-ideal multi-thread scaling (as the A7 shows in Geekbench). The load of all ccores may be the same, percentage-wise, but when running many threads these are competing for L2 cache and memory bandwidth. There will be additional wait times under full load (otherwise the memory subsystem is vastly oversized), so each core is busy but working a little less hard.
3. The lower performance per W of the A53: this may well be true. But if the big cores did interfere, it could certainly explain the significantly increased power draw.
4. You hope for A53 at 1.5 - 1.7 GHz. I agree for CPUs which use only A53's. But as little cores this is not necessary. If you push a design to higher frequencies at the same technology node you always pay in terms of power efficiency (unless you start bad / unoptimized). Better let the little cores do what they're best at: be efficient.
... I'm off to the next page :)
MrS
Andrei Frumusanu - Tuesday, February 10, 2015 - link
Regarding your concerns on the big cores turning on: They did not. I made sure that they're always offline in the little testing. As to why GeekBench scaled like that, I can't answer.jjj - Tuesday, February 10, 2015 - link
Geekbanch scales like that sometimes, TR looked at this phone last week and they notice the Shield tab scaling better than 4x in some tests.For the A53 better scaling over A7 maybe it's the memory subsystem? No clue how heavy GB is on it but the A53 does have more memory BW and maybe faster NAND (you don't have storage tests for the Alpha ).
MrSpadge - Tuesday, February 10, 2015 - link
Sometimes one gets such results if the MT code path uses newer libraries to get multi threading, but also introduce other optimizations the tester is not aware of. Or they have to rewrite code/algorithm to make it multi-threaded and thereby create a more efficient version "by accident". I don't know if any of this applies to GB, though.tipoo - Tuesday, February 10, 2015 - link
It's done that before, I think the cache is almost certainly involved. Maybe it has good cache reusability for multicore in its testing.SydneyBlue120d - Tuesday, February 10, 2015 - link
Excellent article, thanks a lot for that.I'd like to dig deeper about why Samsung is ruining the wonderful Wolfson DAC, I remember that even the legendary François Simond shared the issue with Samsung Developers without getting a response, maybe You will be able to have a response :) Also, I remember an audio quality test suite, why didn't You use it in this review? Thanks a lot.
Andrei Frumusanu - Tuesday, February 10, 2015 - link
We realized testing audio without having the proper equipment is a futile exercise and does not really portray audio quality of a device. Only Chris has the professional equipment to objectively test audio, such as done in the iPhone6 review: http://www.anandtech.com/show/8554/the-iphone-6-re...PC Perv - Tuesday, February 10, 2015 - link
Thank you so much for ditching SunSpider, and thank you for explaining why. It has gone way too long.PC Perv - Tuesday, February 10, 2015 - link
I know you are trying to be nice to your colleague and give them a benefit of doubt, but to mercilessly critical eyes of mine your colleagues (Joshua Ho & Brandon Chester) have shown and time and time again they are not qualified to review Android products. I hope Mr. Frumusanu and Mr. Smith will take charge of this area in the future and limit those two to Apple product reviews.I would love to read quality reviews and analysis like this instead of their tamper tantrum.
PC Perv - Tuesday, February 10, 2015 - link
For example, in the comment section here,http://www.anandtech.com/comments/8795/understandi...
[Q]The increase in brightness for AMOLEDs at 80% APL rather than 100% APL is not very significant, and changing testing to accommodate AMOLED's idiosyncrasies doesn't seem like a good idea either. To put it in perspective, even if I had tested the Nexus 6 at 80% APL in the review my conclusion about the brightness being sub-par would have been exactly the same.[/Q]
That is what Brandon Chest had to say about brightness of the Nexus 6's screen. So arrogant, so biased. But according to the scientific data provided by this article (pg. 10), APL can indeed make a huge difference in AMOLED screen's brightness, given power target. For example, at display power target is 1W, 70% APL will raise the brightness by approximately 40%. At 50% APL, max brightness practically doubles.
I was appalled by the disparaging and arrogant attitude by Brandon Chester when it comes to correctly evaluating Android devices. He also made similarly nonsensical "arguments" in his tablet recommendation article for the holidays.