Who Controls the User Experience? AMD’s Carrizo Thoroughly Tested
by Ian Cutress on February 4, 2016 8:00 AM ESTPower Consumption: Big Improvements to Video Playback
It was teased earlier in the review, but it makes sense at this stage to talk about power consumption.
With a system as complex as a modern APU or SoC, the initial plans for this review involved getting a development system with the right shunts and hooks to measure the core and graphics power separately in a thermally unconstrained environment for both Kaveri and Carrizo, but unfortunately the parts didn’t come together at the time they were needed. Instead we had access to a Watts Up PRO, a power outlet based monitor with some recording capabilities. While the hardware was not ideal for what we wanted to test, it provided a large chunk of interesting data.
We did a number of tests with data monitoring enabled on both the HP Elitebooks. When AMD released Carrizo, a lot of fuss was made about video playback for several reasons. Firstly, Carrizo implements an adjusted playback pathway for data so instead of moving data from the decoder to the GPU to the display controller, it moves data directly from decoder to display, saving power in the process.
AMD also listed the video playback power of Carrizo (as compared to Kaveri) as significantly reduced. In the example above in the top right, the Kaveri APU is consuming nearly 5W, whereas Carrizo will consume only 1.9W for 1080p content.
The other video playback optimization in Carrizo is the Unified Video Decoder. The bandwidth and capability of the UVD is increased four-fold, allowing the system to ‘sleep’ between completed frames, saving power.
Video Playback, 1080p30 h264
For the first test, we took a 1920x1080 resolution h264 video at 30 FPS (specifically Big Buck Bunny) and recorded the power consumption for playback.
The difference here is striking. The Carrizo system in this instance has sustained power consumption lower than that of the Kaveri system. Overall the Kaveri system draws 11W over idle to play back our test video while the Carrizo system only draws 6.8W over idle for the same task. Put another way, the load power cost at the wall for watching 1080p video is about 4W lower on Carrizo as compared to Kaveri, which is close to what AMD claimed in the first slide above (and note we’re measuring at the wall, so chances are there are other chipset optimizations being done under the hood).
Video Playback, 2160p30 h264
The same video but in 4K format was also tested on both systems. It is at this point I should say that the Kaveri system was unable to play the 4K video properly (Kaveri doesn't officially support 4K decoding to begin with), and would only show about 20% of the frames. Audio was also affected.
In this case power consumption is above that of the 1080p video, and both systems require around 11 watts from idle to sustained performance. The added benefit with Carrizo though is that you can actually watch the video.
Other Power Benchmarks
We also ran power tests on a set of our regular benchmarks to see the results.
Three-Dimensional Particle Movement
In our 3DPM test, we typically script up a batch of six runs and take the average score. For this we did it to the single thread and multithreaded environments.
In single threaded mode, two interesting things occurred. First, as we expected, the Carrizo system can idle lower than the Kaveri. Second is that the Carrizo system actually goes into a higher power state at load by almost 4W. This means that the delta (Load to Idle) is 8W higher for Carrizo than Kaveri.
It is easy to take away from this that Carrizo, as an APU, uses more power. But that is not what is happening. Carrizo, unlike Kaveri, integrates the chipset onto the same die as the APU (better integration, saves power), but it also means that it is essentially shut off at idle. Part of Carrizo’s optimizations is power management, so the ability to shut something down and fire it back up again gives a larger low-to-high delta automatically. Essentially, more things are turning on. The fact that the Carrizo power numbers are higher than Kaveri during the benchmark is correlated by the performance, despite Kaveri having the higher TDP.
For the multithreaded test, both systems settle to similar power consumptions as the single threaded test, although the Carrizo system has a much more varied power profile, which also finishes the benchmark earlier than the Kaveri.
Octane and Kraken
For the web tests, we expect them to be partially threaded but because they probe a number of real-world and synthetic tests, there should be some power variation.
Octane is actually relatively flat, instigating similar power profiles to both. Again, it looks like that Carrizo expends more energy to do the same amount of work, however it is easy to forget that the Carrizo idle power state is lower due to optimizations.
With Kraken we also get a flat profile, although one could argue that we’re seeing a classic case of running quick and finishing the benchmark sooner vs. a more sedate path.
WebXPRT
This graph was shown earlier in the review, but let’s look at it again, as it is a good example of a bursty workload:
With average power numbers only a few watts above the idle numbers, both systems do a good job on overall power though again it is easy to think that the larger delta of the Carrizo numbers means that the APU is consuming more power. This is where if you try and calculate the actual energy consumed for each system, you get stupid numbers: 1208.7 joules for the Kaveri and 1932.8 joules for the Carrizo. Without starting from the same platform (or without taking numbers direct from the cores), there are obviously other things at play (such as Carrizo’s capability to control more power planes).
WinRAR
Our final power test is WinRAR, which is characterized as a variable threaded load involving lots of little compressible web files and a handful of uncompressible videos.
In this instance I was surprised to see both systems perform similarly. The HP Elitebook G2 actually has the upper hand here, as it is equipped with dual channel memory. WinRAR is a very memory bandwidth affected benchmark, so the G2 has an upper hand in performance but will also balance between drawing more power for two modules or running in a more efficient mode if there is sufficient data at the CPU.
Facebook
Twitter
RSS
175 Comments
View All Comments
ncsaephanh - Monday, February 8, 2016 - link
Can you guys do a podcast on this article? Would love to hear you guys discuss it and also answer questions/comments on the article.ET - Monday, February 8, 2016 - link
Nice to see a Carrizo article finally, although it's rather disappointing, for example because only single channel was tested.You talked about solutions, here's how I see what AMD and publications like Anandtech need to do (I'm using Carrizo as an example, but it's a lesson for the future):
AMD: When Carrizo is available in a laptop, send one to Anandtech. Immediately. If you have a prototype before that, send that. We want to learn about the chip as quickly as possible, not have to wait months looking for nuggets of information on the web.
Anandtech: Benchmark the hell out of the laptop. If there's single channel with a dual channel option, show a comparative benchmark, but concentrate on dual. We're enthusiasts, we'll install a second DIMM to get better performance. For benchmarks, basic system performance and a plethora of games, and comparison to Intel, plus battery life. Deep dives are nice, but I'd rather have a quick overview of what the system is suitable for, and what kind of gaming it can achieve.
AMD: Desktop first! I know that laptops are where the money is, but desktop is where the enthusiasts are, and if your chip is worth anything, fans and publications like Anandtech will pair it with the fastest memory, configure it with the best TDP, and see what it's really capable of. OEM limitations will not get in the way.
AMD: Fans first! That's pretty much a repeat of the previous point, but AMD, you still have fans, and they are your best customers, not the OEM's or the clueless general public. If you make something that you think is good and you let your fans learn of it and get hold of it, they will tell you what they think and they will tell others. If you leave them in the dark, they will end up losing their enthusiasm.
Anandtech: Follow up on AMD stuff. It may be hard to get the latest AMD chips if AMD isn't helping, but at least let us know you're on it. An occasional news item telling us that you've tried to get some laptops for testing or whatnot will tell us that you're on it, and hopefully shame AMD and the OEM's enough to get a move on.
Personally, I would likely have bought a Carrizo system if there was one of similar size to my old Thinkpad X120e (which I still use, even if I'm not that happy with its speed). I might have bought a Carrizo for my HTPC if I could and I knew it provided decent enough performance.
sofocle10000 - Monday, February 8, 2016 - link
I just signed in to state that Asus had nice business/multimedia notebooks (I used N60DP/N56DP and I actually use an N551ZU - all based on AMD), and although my actual N551ZU is only based on the top of the line Kaveri, it is an exceptional machine for normal use/light gaming...Customers play a big part in the AMD problem, but if there were more incentives (take my current N551ZU, which is a great notebook for ~750-850 $, and if configured with an SSD, you could hardly tell it apart most of the time from the Intel i5H/i7QH + GTX 950M variants), not only a great price, but a better build quality, display, sound system the the market average, some of them would actually pay more attention to the AMD.
The OEM's should have a more defined bottom line for the AMD notebooks - were dual channel memory and a better display, a hybrid SSHD or a SSD are a must, especially for the models in the upper part of the price range 400-700 $...
dragosmp - Monday, February 8, 2016 - link
@Ian - great article, really a good example of investigative journalism. I'm happy this kind of articles are being revived, but being a reader of Tom's I see where this may be coming from.As the "guy that says what laptop/phone to buy" to my family and friends I have to say your findings and conclusions speak to me very clearly - AMD has a system-problem, not so much a CPU-problem (though some may argue differently). AMD chips are fed into cheap looking/feeling PCs with far too many corners cut, but this is how under 700$ market looks like. Could AMD's OEMs sell a 600$ 13" PC to compete with the CoreM UX305? I think not, simply because AMD's CPUs (who consume more) need thicker chassis with stronger cooling and a beefier battery and that costs money - so there's less available for the UX; even if the OEM accepted lower margins on the AMD PC, or AMD to sell the CPU at bargain prices, that design compared to the UX305 would be thicker and likely noisier.
If Zen is good, I could see it in a Mac as Apple has a history of doing good software. Or AMD should build their own surface line and set an example of what can be done.
Gunbuster - Monday, February 8, 2016 - link
People buy the cheapest $300 laptop they can get or something premium. Who are they targeting with these mid-rangers?farmergann - Tuesday, February 9, 2016 - link
Wife uses her Y700 for school and a few hours of photo editing every week. Exactly what she wanted. This article did a worthless job of representing the actual Y700 w/fx8800p you can pick up at Best Buy for $665-830. Everything is fantastic about it save for the TB HDD which I immediately replaced with a Samsung 850 Pro I had laying around.Somehow, this "investigative" nonsense missed the fact the U.S. Y700 has a superb little IPS screen with Freesync to go along with a surprisingly (truly) good sound system and -despite the author's claim- dual channel ram. Just for grins I've played BF3 and a few other games - none of which had issues. Great low/mid-range laptop with plenty of chops.
every1hasaids - Tuesday, February 9, 2016 - link
Nope, the US model is absolute garbage. They skimped on the VRMs and the laptop subsequently throttles in moderately intensive CPU tasks. Example, try running Cities: Skylines with a decent sized city and tell me that it doesn't stutter after about 20 seconds of play and every 5 seconds or so after that. The stutters which coincide with the CPU being utilized near 100% and the frequency dropping per resource monitor and Afterburner all the way down to 1.6ghz... Also I don't know what you're talking about with the Freesync capability, I could not get it to work after reading elsewhere that it may be possible.The main issue with a product like the Y700 is that the intel variant is only a couple hundred bucks more and you get a genuine quad core with HT, dual channel DDR4-2133 and comparable discreet graphics. Oh, and it has no trouble with voltage supply. Not to mention that the m.2 interface is PCI-E as opposed to SATA on the AMD model. It just doesn't make sense to purchase a far inferior product for only $200 less at the price point these models occupy.
farmergann - Thursday, February 11, 2016 - link
Cities: Skylines? LOL, that's about as rich as whining about Starcraft 2 performance on an FX Octacore - what were you expecting exactly? For people not looking to shove a laughably CPU bound title down a 35W laptop's throat, the FX8800p with user installed SSD is a far better choice, sorry guy.Peichen - Monday, February 8, 2016 - link
Wow, that's wasting a lot of time and words reviewing a product no one will buy. AMD needs to exist to keep the cheap Intel stuff dirt cheap but I don't feel anyone should waste time reviewing AMD CPU products. 10 years of marketing hype and under-delivery means AMD is actually slower than ever compares with Intel.I bought 2 AMD CPU over the last 6/7 years and frankly I wish I spend more buying Intel because I wouldn't have to spend time and money as often upgrading the CPU.
Danvelopment - Monday, February 8, 2016 - link
The way I see it, AMD needs to stop comparing themselves with themselves and needs to compare themselves with the competition. People don't understand the improvements if they aren't involved with the predecessor.They produce a reasonable product that performs at 60-80% of the competition at 50% of the price.
Good designs are produced for the competition, that could fundamentally have their parts, and they're losing on the design front.
And strangely, for similar products the AMD machines are the same cost, even though the difference is the chip (at halfish the price).
Can they not work to develop an easier transition method for OEM's to produce this-or-that designs that allow end users to pick AMD or Intel during the selection process. Tier them like Dell does for the various Intel processors but have them consistently show up as the cheapest option $100 off a $500 laptop is a decent drop and if the chip and PCB is $150 cheaper to produce the OEM still wins).
Differentiating the product creates too many variables people don't understand, and creates the issue above, CPU brand aversion on entire product stacks with no common ground.
I'd say take a long, hard look at current machines, and develop a method of getting their chips into them as an option, without OEMs designing a product from the ground up.
I'd certainly consider AMD if I could just select it as an option that knocks $100 off on the low cost tier laptop in my workplace.