- This topic has 0 replies, 1 voice, and was last updated 8 years, 7 months ago by
.
-
Posts
-
If you’ve created large (e.g. 200 APs or larger) WiFi modeling projects, you know how slow the software can be…regardless of vendor. There are so many simultaneous calculations required that it would take dozens of cores to speed it up in a meaningful way. Even with my Quad-Core Xeon 3.4GHz CPU, with 32GB of RAM, it takes 3-5 minutes to refresh after a single change when I’m modeling multi-thousand AP projects.
The answer: GPUs, such as Nvidia’s Tesla P100.
http://www.nvidia.com/object/tesla-p100.html
Cheap? No. Worth it if you’re modeling very large projects? Yes. Supported by today’s modeling software vendors? No. 🙁 It’s not enough to increase the CPU cores. One of my customer has a 16-core computer, and he describes it as “adequate” to process the ~1600 AP project.
With thousands of cores, GPUs can add lots of speed and the ability to model more accurately (via more granular resolution).
Feel free to beg, scream, yell, and whine (alot) to your favorite modeling software vendor until they finally agree that this is a good idea. 🙂
Full disclosure: one of my customers (we’ll call him Josh) gave me this idea. It’s simply a great idea, so I started reading and researching further.
What do you think?
- You must be logged in to reply to this topic.