[XeTeX] efficiency of placing different graphics formats?

William Adams will.adams at frycomm.com
Mon Sep 29 19:27:11 CEST 2008


On Sep 23, 2008, at 3:07 PM, William Adams wrote:

> When I do the benchmarking, I'll try bitmapped .pdfs (batch conversion
> from PhotoShop) as well.

Okay, here're the results for two consecutive runs:

PNGs
267.780u 11.425s 5:50.27 79.7%	0+0k 0+15io 0pf+0w
267.515u 11.324s 5:52.37 79.1%	0+0k 0+86io 0pf+0w
filesize: 159.7MB

JPEGs
3.804u 4.123s 1:26.34 9.1%	0+0k 0+82io 0pf+0w
3.481u 2.329s 0:32.98 17.5%	0+0k 0+82io 0pf+0w
filesize: 83.1MB

PDFs
3.874u 4.047s 1:20.87 9.7%	0+0k 0+11io 0pf+0w
3.787u 3.897s 1:03.72 12.0%	0+0k 0+12io 0pf+0w
filesize: 157.7MB

(done using time xelatex filename.tex)

There're over 100 graphics for this title, averaging around 10MB each  
as .PNGs (about a dozen get pulled in for each run though), half that  
for .JPEGs, and _twice_ that for .PDFs --- probably I should try  
making .eps files and distilling to get .pdfs, instead of batch  
processing through PhotoShop.

So JPEGs seem to be the big winner time and file-size-wise (though the  
latter may be aiding the former --- I was working off a network drive  
and on the second run it looks like the while set of JPEG graphics  
were cached in memory for the second run)

William

-- 
William Adams
senior graphic designer
Fry Communications




More information about the XeTeX mailing list