W3C home > Mailing lists > Public > public-device-apis-log@w3.org > May 2017

Re: [sensors] Define processing model

From: Alexander Shalamov via GitHub <sysbot+gh@w3.org>
Date: Wed, 24 May 2017 13:25:47 +0000
To: public-device-apis-log@w3.org
Message-ID: <issue_comment.created-303722767-1495632346-sysbot+gh@w3.org>
> Without, you need all 1000 to be copied to the application layer. I can't imagine this has no perf consequences.

You need sensor running at high sampling rate and grab data from sensor's registers when needed, if output data rate is non-deterministic (rAF), access on every frame can be used. If cannot be done synchronously, like in browser case, privileged process can do many different tricks, memory mapped I/O, or simply implement polling at faster rate than refresh rate, then copy latest reading to shared memory. Currently, sensor reading size is 48 bytes, memory I/O bandwidth is gigabytes per second. Updating shared buffer, even at 240Hz, would have no performance impact.

-- 
GitHub Notification of comment by alexshalamov
Please view or discuss this issue at https://github.com/w3c/sensors/issues/198#issuecomment-303722767 using your GitHub account
Received on Wednesday, 24 May 2017 13:25:54 UTC

This archive was generated by hypermail 2.3.1 : Monday, 23 October 2017 12:18:53 UTC