Personal computing discussed

Moderators: morphine, SecretSquirrel

Topic Author
Posts: 72
Joined: Thu Aug 14, 2014 8:46 pm

Frametime Results Interpretation

Sat Sep 30, 2017 10:42 pm

I looked on Jeff's Twitter, and found this tweet. This reminded me of something that I thought of earlier.

In the tweet that Jeff referenced, the 1080 seems to render less frames than the RX Vega 64, despite them running the same exact benchmark run (presumably). I've also noticed this in TR's own reviews to an even greater extent-two cards running the same benchmark may render significantly different amounts of frames. Can someone please explain to me why the amount of frames rendered by each card seems to vary so much?

Please bear with me if this is a dumb question. I in no way was using this to cast doubt on the reliability of the data, I was just curious to understand why this was the case.
Posts: 28
Joined: Sat Nov 24, 2007 3:19 pm

Re: Frametime Results Interpretation

Sun Oct 01, 2017 1:42 am

The total number of frames rendered is a product of the frames per second that each card is able to render. Higher fps = more total frames rendered. Thus, the card with the most frames rendered can be considered the "best" performing card on that benchmark, although you also have to take into consideration any spikes in the frametime chart, which represent momentary hiccups in the card's performance.
Emperor Gerbilius I
Posts: 6180
Joined: Fri Oct 15, 2004 10:41 pm

Re: Frametime Results Interpretation

Sun Oct 01, 2017 2:46 am

Well, more frames rendered -> higher frames per second, but the real kicker is analyzing frametimes themselves:

One of the primary reasons for measuring frametimes was AMD's pretty horrendous Crossfire implementation up to the HD6000-series when compared to Nvidia's SLI. Here, you were getting more frames by using two or more AMD cards, however frametimes were so erratic, ranging from high to low, that a single card could many times provide an experience that was as good or even better than Crossfire.

To measure this, frametime graphs as well as percentiles, which are represented as the number of frames that take longer than a particular monitor refresh interval (60Hz -> 16.6ms, 120Hz -> 8.3ms, etc.). TechReport and other reviewers use frametime analysis to show how 'smooth' performance is, and this perspective is more useful than talking about framerates.
Canon 6D|24-105/4L |70-300/4-5.6 IS USM|50/1.4 USM|50/2.5 CM|50/1.8 STM|85/1.8 USM|Samyang/Bower 14/2.8|24/2.8 IS|Sigma 150-600 | C
Canon EOS-M|11-22 IS|22/2|EF-M 18-55
Maximum Gerbil
Posts: 4157
Joined: Sat Nov 25, 2006 3:21 pm
Location: Kansas

Re: Frametime Results Interpretation

Sun Oct 01, 2017 2:54 am

i7 6700K - Z170 - 16GiB DDR4 - GTX 1080 - 512GB SSD - 256GB SSD - 500GB SSD - 3TB HDD- 27" IPS G-sync - Win10 Pro x64 - Ubuntu/Mint x64 :: 2015 13" rMBP Sierra :: Canon EOS 80D/Sony RX100

Who is online

Users browsing this forum: No registered users and 3 guests