> Does your benchmark measure the time to actually display everything on every page (what the human looking at all the pages must do), or just the time until th

> Does your benchmark measure the time to actually display everything on every page (what the human looking at all the pages must do), or just the time until the browser is responsible? It "only" benchmarks the rendering, all the overhead the viewer produces is not shown. That is intentional, as we will create our own viewer anyway BTW: PDF.js in FF is typically slower than in Opera / Chrome, all these benchmarks used Opera