Investigating Cavium's ThunderX: The First ARM Server SoC With Ambition
by Johan De Gelas on June 15, 2016 8:00 AM EST- Posted in
- SoCs
- IT Computing
- Enterprise
- Enterprise CPUs
- Microserver
- Cavium
Energy Consumption
A large part of the server market is very sensitive to performance-per-watt. That includes the cloud vendors. For a smaller part of the market, top performance is more important than the performance/watt ratio. Indeed for financial trading, big data analyses, database, and some simulation servers, performance is the top priority. Energy consumption should not be outrageous, but it is not the most important concern.
We tested the energy consumption for a one-minute period in several situations. The first one is the point where the tested server performs best in MySQL: the highest throughput just before the response time goes up significantly. Then we look at the point where throughput is the highest (no matter what response time). This is is the situation where the CPU is fully loaded. And lastly we compare with a situation where the floating point units are working hard (C-ray).
SKU | TDP (on paper) spec |
Idle W |
MySQL Best throughput at lowest resp time (W) |
MySQL Max Throughput (W) |
Peak vs idle (W) |
Transactions per watt |
C-ray W |
Xeon D-1557 | 45 W | 54 | 99 | 100 | 46 | 73 | 99 |
Xeon D-1581 | 65 W | 59 | 123 | 125 | 66 | 97 | 124 |
Xeon E5-2640 v4 | 90 W | 76 | 135 | 143 | 67 | 71 | 138 |
ThunderX | 120 W | 141 | 204 | 223 | 82 | 46 | 190 |
Xeon E5-2690 v3 | 135 W | 84 | 249 | 254 | 170 | 47 | 241 |
Intel allowed the Xeon "Haswell" E5 v3 to consume quite a bit of power when turbo boost was on. There is a 170W difference between idle and max throughput, and if you assume that 15 W is consumed by the CPU in idle, you get a total under load of 185W. Some of that power has to be attributed to the PSU losses, memory activity (not much) or fan speed. Still we think Intel allowed the Xeon E5 "Haswell" to consume more than the specified TDP. We have noticed the same behavior on the Xeon E5-2699 v3 and 2667 v3: Haswell EP consumes little at low load, but is relatively power hungry at peak load.
The 90W TDP Xeon E5-2640v4 consumes 67W more at peak than in idle. Even if you add 15W to that number, you get only 82W. Considering that the 67W is measured at the wall, it is clear that Intel has been quite conservative with the "Broadwell" parts. We get the same impression when we tried out the Xeon E5-2699 v4. This confirms our suspicion that with Broadwell EP, Intel prioritized performance per watt over throughput and single threaded performance. The Xeon D, as a result, is simply the performance per watt champion.
The Cavium ThunderX does pretty badly here, and one of the reason is that power management either did not work, or at least did not work very well. Changing the power governor was not possible: the cpufreq driver was not recognized. The difference between peak and idle (+/- 80W) makes us suspect that the chip is consuming between 40 and 50W at idle, as measured at the wall. Whether is just a matter of software support or a real lack of good hardware power management is not clear. It is quite possibly both.
We would also advise Gigabyte to use a better performing heatsink for the fastest ThunderX SKUs. At full load, the reported CPU temperature is 83 °C, which leaves little thermal headroom (90°C is critical). When we stopped our CRAC cooling, the gigabyte R120-T30 server forced a full shutdown after only a few minutes while the Xeon D systems were still humming along.
82 Comments
View All Comments
Daniel Egger - Wednesday, June 15, 2016 - link
I could hardly disagree more about the remote management of SuperMicro vs. HP. Remote management of HP is *the horror*, I've never seen worse and I've seen a lot. It's clunky, it requires a license to be useful (others do to but SuperMicro does not have such nonsense), the BCM tends to crash a lot (which is very annoying for a remote management solution), boot is even slower than all other systems I know due to the way they integrate the BIOS and remote management on the system and it also uses Java unless you have Windows machines around to use the .NET version.For the remote management alone I would chose SuperMicro over most other vendors any day.
JohanAnandtech - Thursday, June 16, 2016 - link
I found the .Net client of HP much less sluggish, and I have seen no crashing at all. I guess there is no optimal remote management client, but I really like the "boot into firmware" option that Intel implemented.rahvin - Thursday, June 16, 2016 - link
Not only that but Supermicro actually releases updates for their BCM's. I had the same shocked reaction to the HP claim. Started to wonder if I was the only one that thought supermicro was light years ahead in usability.I should note that Supermicro's awful Java tool works on Linux as well as windows. Though it refuses to run if your Java isn't the newest version available.
pencea - Wednesday, June 15, 2016 - link
All these articles and yet still no review for the GTX 1080, while other major sites have already posted their reviews of both 1070 & 1080. Guru3D already has 2 custom 1080 and a custom 1070 review up.Ryan Smith - Wednesday, June 15, 2016 - link
It'll be done when it's done.pencea - Wednesday, June 15, 2016 - link
Unacceptably late for something that should've been posted weeks ago.Meteor2 - Thursday, June 16, 2016 - link
Will anyone read it though? Your ad impressions are going to suffer.Ryan Smith - Thursday, June 16, 2016 - link
Maybe. Maybe not. But it's my own fault regardless. All I can do is get it done as soon as I reasonably can, and hope it's something you guys find useful.name99 - Thursday, June 16, 2016 - link
Give it a freaking rest. No-one is impressed by your constant whining about this.pencea - Thursday, June 16, 2016 - link
Not looking to impress anyone. As a long time viewer of this site, I'm simply disappointed that a reputational site like this is constantly late for GPU reviews.