Printing speed question
 

Can anyone offer any suggestions to speed up printing? Or to, at least, provide a clue as to why printing may either be abnormally slow or hanging the printer.

I have an HP LaserJet 1200 that sometimes takes forever to complete a print job. There's 72MB (8MB + 64MB added) of RAM in it but any print file that's larger than a few MB will cause the printer to spin its wheels until someone complains and I pause the queue, kill the current job, power cycle the printer, and restart the queue.

Configuration: The print server is running as a plain 'ol lpd/lpr print server (using LPRng). The printer is, as I mentioned above, an HP LaserJet 1200. The PostScript level is `2014.116'. Jean-Serge Gagnon's PS printer test (ppst33.ps) reports an overall rating of 12.1 (with 1.0 = original Apple Laserwriter). Clients are running CUPS (except for one one Windows client) and treating the print server as an lpd/lpr server.

One thing I've noticed is that printing from Okular seems to be less reliable than printing from Acrobat Reader. (So the speedier launch of Okular is wasted.) Printing from `gv' is generally reliable. Printing pure PostScript files from the command line nearly always results in a completed print job depending on what produced the PS file. For example, printing from Firefox mostly works but not always; it looks like FF is producing PS v3.x output, some of which which may be confusing the printer. LibreOffice printing has not been a problem; neither has anything that comes out of LaTeX+dvips.

Thoughts on ways to speed up printing or to make it more reliable -- and by reliable, I mean `doesn't cause the printer to go catatonic? (Short of buying a newer printer, that is.)

TIA...

--
Rick

Have you tried installing and using HPLIP? That's H-P's free Linux printing and imaging software that supports pretty much every printer they've ever made. It builds your CUPS interface for you (you can, of course, use lp too).

Have a look at http://hplipopensource.com/hplip-web/index.html and maybe give it shot; WFM.

Hope this helps some.

I've been using using the HP-supplied PostScript driver that supports the LJ1200. I can't see how using the HP configuration tool is to going to help.

The problem isn't that printing isn't working... unless the printer is fed one of those really large files. Perhaps the problem is not so much the printer and the drivers as the applications that are generating crappy output files. I know one thing that'll cause the printer to choke is when a PDF is a collection of raster scanned images. Those files are huge and that's when printing seems to get either extremely slow or hung. (Makes me really wish there was a port on the printer that I could hang a terminal on to monitor what's going on.) If I'm aware that the file is composed of a series of images -- happily they're not common -- I'll print them 1-2 pages at a time. They still print very slowly but at least breaking the job up into small chunks doesn't cause the printer to hang. There's another memory slot available on the printer. Anyone feel that adding another 64MB would help? I can probably scare up another chunk of RAM amongst my spare parts. (Frankly I'm doubting that it'll help when an 8MB job chokes when there's 72MB available.)

What would be nice for debugging purposes if it would be possible to append a bit of PostScript to each print job (or embed the job in a PS wrapper via a shell script or a header and trailer that could be configured in LPRng) that would print a trailer page showing performance statistics -- processing time, memory used, etc. -- after the job completes. Anyone know of such a thing? (Haven't tried Googling for that yet to see what I can find or dug into my old Adobe PS books to see if the language offers any bits to get that information.)

So, it appears that I'm still looking for a solution.

--
Rick

Well, one of the reasons I suggested a look-see at HPLIP is that I have some enormous image files to print every so often (plus large PDF files with and without embedded images) and my H-P Photosmart C-4680 with USB (which does not have PostScript in it) and my H-P Business Inkjet 2280 tn (Ethernet card and the PostScript engine in it) handle either with ease; the 2280 tn has 32 M of RAM installed. Between CUPS and HPLIP, which provides the print "model," I have no problem printing things like hi-resolution topographic maps (created with software). Prior to HPLIP, I had essentially the same problem with it taking forever and week to print image files and now I don't.

What the heck, it's free and you can always uninstall it if it doesn't work for you.

Hi

+1 on hplip and just to be blunt
It will change your default printing system from LPR to CUPS and cups, is the modern way of printing for most Linux distros.

Your cups configuration page can also change your resolution or DPI settings.
In other words if you don't mind a loss of quality, you can really really speed up printing by printing at less DPI setting
---the printer will drop less ink onto the page so will be much faster.

YMMV

your distro may have multiple ways of installing cups and hplip but once downloaded
--check that your cups daemon is running go into FF and go to
http://localhost:631/
ensure you know your root password
Click on add printer ---NOT search printer
that will bring up the root login dialog box for cups
enter root and your root password
cups will then find your printer if turned on
cups will then offer hplip driver if found etc etc

or you could try
good luck

I also found it to be the cause of some of my printing performance problems, especially when I had a CUPS client -> CUPS server configuration. Things sped up immensely when I replaced CUPS on the server with LPRng.

If by `modern' you mean `what's being shoved down our throats', then, yeah, I agree... it's the modern way. Doesn't mean it's better, IMHO. I do get miffed when I cannot install some packages without some RPM dependency getting in the way when the actual dependency should be `a printing subsystem', not `the CUPS printing subsystem'. Much the same way as when I'm forced to install MySQL by some package when I'm already using PostgreSQL; the requirement should be `an RDBMS' and not a specific one. (Wait until the fun begins when old RPMs that require MySQL start getting used on systems that have migrated to MariaDB.) It's a waste of resources to have both. But I digress...

In my case that would be `toner' and not ink but I get ya. I already have the printer using the toner-saving mode enabled as the default.

I may need it. Once I have a free evening to kill trying out CUPS again. (And, possibly, another to remove it. :^/ )

You guys are obviously having some luck with CUPS so maybe it's worth giving it another chance.

Later...

--
Rick

ahh ok

you may like to peruse the comments herer
http://www.openprinting.org/printer/HP/HP-LaserJet_1200

in particular using cups, someone found the hp drivers slow but gutenprint also supplies a PPD and that was good

YMMV

Incidently online viewing of PPDs show you need the foomatic-rip package as well as cups

pxlmono PPD---needs foomatic-rip as well
http://www.openprinting.org/ppd-o-ma...et_1200&show=1

PS PPD --fooetc
http://www.openprinting.org/ppd-o-ma...et_1200&show=1

I upload a gutenprint to show needs gutenprint
https://www.dropbox.com/s/vi0xa5uw5eit3gt/hp.ppd

and as I don't have a laser jet but a inkjet...giggles this is the best I can show
https://www.dropbox.com/s/wwnm6wd729ag0m8/hp.jpg

I already use B/W or grayscale when available knowing that sending all that color information to the printer is going to increase the size of the print job.

According to the comments/review it appears that I have maxed out the amount of RAM that can be installed. I generated the status page from the printer and it says:
Hmm... what can I do with ``Side 2''? Is this HP's way of describing a double-sided DIMM board configuration? Could I run 136MB in the printer? I haven't worked with these old memory boards in quite a while so I can't recall if such a beast even existed. I haven't seen any documents that say that more than 72MB is supported so I'm guessing that looking for a higher capacity DIMM will be a waste of time. That part of the report is probably just a placeholder for the HP printers that do support two memory boards. That mention of `Side 2' had me hoping, though.

From the review:
That's long been my suspicion. Printing from browsers, in particular, seems to be the biggest culprit with printing from PDF viewers a close second. I rarely have problem printing previously-created PS files using `lpr -l ...' (for example, files created with LaTeX+dvips). But even that is not universally true: printing a complex web page from Firefox to disk (as a PS file) and printing the resulting file hardly ever helps. But I can, at least, print it one page (or a few) at a time to reduce the chances of hanging the printer.

I've done the obvious things -- tweaking applications to send grayscale information to the printer (and not color), reducing the resolution in CUPS (hey... we all lived with 300DPI for years and thought it was good enough), etc. -- and that's likely about as good as I'll be able to get.

That's easy enough to try. The comment about the PCL driver generating pretty fast print results has me thinking that setting up a separate queues using the PCL driver. Something to try out, anyway.

Indeed...

Thanks for the tip(s)...

--
Rick

Seems like the 1200 dates from sometime around 2000/2001 (earlier?)? Kinda a older/slower box with graphics, but not shabby with text; but no speed demon in any case. I have noticed that the PostScript "engine" in the Business Inkjet can take forever and week to process some images that get sent to it; particularly a 66MB PostScript full-color map of Cold War Europe which takes 9.19 s to render with gv on a pretty darn fast 64-bit box with 16G of RAM. That same file takes about 2.5 minutes to render on the printer. On the C4680 Photosmart it renders in 45 seconds. Newer engine, quicker results, I suppose. The Business Inkjet is about 15 years old, the Photosmart is about 3 (and gv is going to a display). Oh, yeah, that same file takes about 10 minutes on the plotter; lots of detail, lots of text, lots of color layers (but an E-size drawing sure is pretty).

Cut down the DPI (you can't really see the difference between 600 dpi and 300 dpi in most circumstances anyway), you go faster.

It's probable that throwing RAM at LJ1200 could help -- but not much, you're limited by the engine. I watch the lights blink on the router sending a big job to the Inkjet, spit some, wait, spit some more, wait, repeat. It's just not going to be a speed demon.

I have found that HPLIP/CUPS is a good pair, that Inkjet is more efficient with that combination than it ever was with lp. The Photosmart with the same combination pumps along at a good rate (and the thing only cost $65!). The Photosmart gets a lot more use nowadays than the Inkjet but the Inkjet has been reliable and useable for a long, long time and has a warm place in my heart (and a lot less twiddling with HPLIP/CUPS, too, set it up and go in about a minute).

Can't hurt, just might help.

ok so some comments on printing PDFs

you may like to look at ghostscript.

the command to run is gs ref http://linux.about.com/library/cmd/b...hostscript.htm
online help http://www.ghostscript.com/doc/9.07/Use.htm

the main issue is finding a "device" friendly enough to use in the commmands to print without using a file manager or web browser etc

Yeah, it's an older printer. If memory serves it is rated for something like 20PPM and it does fly when you send it a plain text document (I write a lot of documentation so I like the text processing speed). Even documents with embedded graphics print pretty quickly though a full-page image does slow things down. (But only for that page.)

Two and a half minutes would be acceptable to me. It's the whole `hours and hours and still not finished' problem that's driving me up the wall. I haven't discovered the threshold above which the print engine seems to go into an infinite loop trying to render the print job. I should start keeping track of that so I'll have an idea of when to expect trouble. Other users never seem to feel the need to inform me of when something hasn't printed. Then I find that someone's print job has been tying up the printer for a day with everyone else's stuff (i.e., `my stuff' :D ) piled up in the queue.

Agreed. I've changed the resolution and done `real world' test prints and can't see the difference unless I get out a magnifying glass. The CUPS test pages do show a marked difference in results depending on the resolution; especially in the color wheel. Not sure how badly that's going to turn out with every day printing.

Well, as I noted, I've maxed out the RAM in the printer and it hasn't really helped. It's not a network problem; the printer is attached to a server so I can't blame this on the network. It's an older system (1GHz, 768MB) but it'll take some real convincing to get me to believe that that combination is not sufficient to keep the printer from reporting that jobs are stalling.

The cost of doing every day printing to an inkjet has put me off. Heck, the cost of printer ink is enough to keep me from doing too much photo processing lately. BTW: According to my CUPS configuration, I'm using the HPLIP driver; at least, I am if I've reading the setup in YaST correctly.

Well, I have a couple of things to check out and reconfigure so I'm going to get on that.

Later...

--
Rick

I don't see an explicit mention of the LaserJet 1200. I seem to recall that the drivers listed in YaST while in the printer/CUPS config mentioned the LJ4 as compatible. Or would the generic `laserjet' be a better device? (I'm assuming that would be a match for all members of the LJ family.)

(And gs still supports the DEC LN03? Those printers would be over 20 years old by now!)

--
Rick