« The State of LTE Carrier Aggregation in Practice | Main | Peering: Who Pays Whom and Why? Level-3's Point of View »


William Hern

I too have been tracking the improving JavaScript performance of mobile devices using SunSpider.

Another factor in benchmark performance is the amount of RAM in the device. The JavaScript garbage collector imposes a significant performance hit so the less frequently it has to run the better. I've seen a near doubling of performance on the benchmark in the case where two devices had similar processors but one had double the memory of the other.

I've not looked closely at the JavaScript code for SunSpider to figure out how much memory the benchmarks could use, but I suspect that memory will become less of a factor now that devices routinely come with a couple of gigs of RAM.

The comments to this entry are closed.

My Photo

The Books to this Blog

Secure Hotel Wi-Fi Sharing

My Pictures on Flickr

  • www.flickr.com
    martin.sauter's photos More of martin.sauter's photos


  • Clicky
    Clicky Web Analytics