One of the interesting demos of the recently announced Snapdragon 855 SoC here at Qualcomms Tech Summit was to do…
One of the interesting demos of the recently announced Snapdragon 855 SoC here at Qualcomms Tech Summit was to do with power. In a dedicated area, Qualcomm had set up two pairs of devices, with Snapdragon 855 and 845 chips inside, and had images of power on two large screens designed to show the generation gain of the new chip that has been promised more performance on a smaller process node .
For power consumption values, Qualcomm had connected a few meters until the battery was connected to each device. This gives the total system strength, not just the power of the chipset, so it includes the monitor, the connection, and anything else that can pull a microjoule or two.
On the first pair of devices, Qualcomm ran through a script of random web-based search. The script is allowed for automation, and it was continuously uploaded a set of known web pages without cache search. This is a pretty bright test, but it spans the radio a fair bit and a shifting display that has a very wide range.
Overall, the S845 system used an average power of 2219mW, while the new S855 system was 1778mW. In the big planning of things, this is 20% less power.
For the second demonstration, Qualcomm went to play. Stopping PUBG, we were told that the two devices were fixed to a 40 FPS mode with identical quality settings. This was also a manuscript test with the player jumping out of the plane on the ground and ending out. This of course requires some radio to connect to the online server, and there would be a slight unevenity between the two games based on the server latency.
The values for these tests put the S845 in an average power of 3276 mW and the new S855 at 2543 mW. The difference in this test would be 22%.
All this is good and good, although a big approach came: none of the screens were brightness calibrated. The Qualcomm representative said the brightness of the devices was “eyeballed” to be in the same region, but based on personal experience, it’s never a guarantee and they can be as much as 20-30%. The panels were also not from the same manufacturer, so there is also a question of panel efficiency.
All said, it may not be unexpected (or rather, displaying power at this time is unexpected). There must be some confusion that Qualcomm has to meet expectations, but the final setting must also take place. In addition, it will be up to Qualcomm’s OEM partner to put SoC into an environment that may be effective for radio, monitors, thermals, etc.
We wait until we take some hands with the hardware until we make any assessments. Funnily, I have my display calibration tool with me. To be fair, Qualcomm said that they would allow me to equalize the screen brightness of both, which was a good gesture, but we did not reflect the workloads they used and there was no chance that I could run our power at this point. Stay tuned, as efficiency will be an interesting story here.