Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
I have used Google, but no joy. Many hits, but none seems to look just like this one.
Printer is direct connect, no other computers on a LAN etc. Printer is connected to parallel port. I am using an IEEE compliant bi-directional cable purchased just two days ago (just in case the old cable might contribute to the problem, no change). No other printer installed.
As an example of print speed (if that is the correct word)
$ ls -l test.txt -rw-rw-r-- 1 jmccarty jmccarty 16 Jan 17 00:44 test.txt
$ cat test.txt This is a test.
$ lpr test.txt
Here are the elapsed times to significant events...
0:00 hit return on command 0:06 printer data light flashing, loaded sheet of paper 0:38 printer actually printed something, not all... at approximately 10 second intervals, printer moved the print head, data light flashing continually 1:58 sheet ejected, with one line on it printed
So, two minutes to print "This is a test." I have to believe that, since the data transfer light was flashing the whole time, something is translating this into a giant postscript program, or into a pure bit map, or something, and transferring a huge amount of data. There must be something wrong in my setup for this printer.
Here is what (Red Hat Symbol)->System Settings->Printing gives for the printer.
Queue Type Queue Type: locally connected Queue Options At start: no banner At end: no banner Top margin: 36 right margin: 36 left margin:36 bottom: 36
cpi 12 lpi 7 page-bottom 86 page-left 57 page-right 57 scaling 100 wrap true Printer Driver DeskJet 870C hpijs Driver Options send FF no send EOT no assume unknown is text no prerender postscript no convert text to ps yes effective filter locale C res,qual,ink,media 300 dpi,Color,Black+Color Cart page size letter model HP DeskJet 870C printout mode normal
I have another queue for that printer which does two-up printing, but I haven't checked it for speed. I just know that this queue is unbearably slow, and I must have set up something wrong.
I tried changing the option "convert text to postscript" to no, but that made no change.
I would like some direction on how to proceed with this.
Thank you for your time in reading this message.
Mike
On 1/17/06, Mike McCarty mike.mccarty@sbcglobal.net wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
If you have the option of testing the printer on another system maybe you can determine if it's the operating or the printer itself that needs fixing.
Klaasjan
Klaasjan Brand wrote:
On 1/17/06, Mike McCarty mike.mccarty@sbcglobal.net wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
If you have the option of testing the printer on another system maybe you can determine if it's the operating or the printer itself that needs fixing.
I was able to do something even better. I was able to take a web page which took 30 minutes per page to print, and subset it to one which prints quite quickly. So I know that it is not just the printer, nor is it the CPU getting bottled up, but something with the configuration and how it handles things.
But I still don't know what the problem is.
Mike
On Tue, 2006-01-17 at 23:56, Mike McCarty wrote:
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
If you have the option of testing the printer on another system maybe you can determine if it's the operating or the printer itself that needs fixing.
I was able to do something even better. I was able to take a web page which took 30 minutes per page to print, and subset it to one which prints quite quickly. So I know that it is not just the printer, nor is it the CPU getting bottled up, but something with the configuration and how it handles things.
But I still don't know what the problem is.
The usual scenario is that the linux app generates postscript, substituting fonts as needed, then for printers that don't accept postscript, it is fed through ghostscript to generate an appropriate bitimage and commands for your printer model, which is then dribbled out the printer port. It sounds like you might be having trouble with several of these steps. One potential problem is that your parallel port isn't generating an interrupt and runs in a really slow polled mode but it has been so long since I've seen that that I've forgotten the solution. The way to avoid the whole mess is to use printers that accept postscript on a network connection.
Klaasjan Brand wrote:
On 1/17/06, Mike McCarty mike.mccarty@sbcglobal.net wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
If you have the option of testing the printer on another system maybe you can determine if it's the operating or the printer itself that needs fixing.
Klaasjan
I now have one .GIF file which, when viewed by Mozilla on my machine, and then printed, takes an inordinate amount of time. I printed it to a PostScript file, and it takes the same length of time to print this with lpr. I ran ps2ps and got a smaller file which is still very slow.
Mike
Mike McCarty wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
If you run the "top" command on your machine while printing, does it show anything using a lot of CPU or memory?
Mogens
Mogens Kjaer wrote:
Mike McCarty wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
If you run the "top" command on your machine while printing, does it show anything using a lot of CPU or memory?
Mogens
It ran 0.7% to 1.3% (max, not often hit) during a test print. That is CPU utilization. So I don't think it is CPU limited.
Mike
Mike McCarty wrote: ...
It ran 0.7% to 1.3% (max, not often hit) during a test print. That is CPU utilization. So I don't think it is CPU limited.
On FC4, one can create a custom device (instead of /dev/lp0), and send the output to /tmp/printer.out (create this file first, with 777 permission). Try this, and see how long it takes to print to the file.
Try to copy the file to /dev/lp0 and see how long that takes.
Mogens
Mogens Kjaer wrote:
Thanks for the help.
Mike McCarty wrote: ...
It ran 0.7% to 1.3% (max, not often hit) during a test print. That is CPU utilization. So I don't think it is CPU limited.
On FC4, one can create a custom device (instead of /dev/lp0),
I'm sorry, but I'm not an experienced UNIX admin. I'd need more information about how to run MAKEDEV. Oh, and I run FC2, not FC4.
and send the output to /tmp/printer.out (create this file first, with 777 permission). Try this, and see how long it takes to print to the file.
Umm, it took not any time at all to create a one line text file. I'm experiencing the problem even with a file with only 16 ASCII characters in it.
Try to copy the file to /dev/lp0 and see how long that takes.
Ok, I tried this file...
$ ls -l test.txt -rw-rw-r-- 1 jmccarty jmccarty 17 Jan 17 03:47 test.txt
$ cat test.txt This is a sheet.
The "data transfer" light on the printer started to flash immediately, and after 16 seconds, the printer grabbed a sheet and printed the line instantly. It LOOKED like it might have been downloading the font, perhaps. It certainly used a different font from the one that lpr used.
But, I did this... I printed a one-page thing from Mozilla to a file in my home directory, and got mozilla.ps
$ ls -l ~/*.ps -rw-rw-r-- 1 jmccarty jmccarty 504768 Jan 17 11:02 home/jmccarty/mozilla.ps
Creation of that file was essentially instantaneous. Copying it to /dev/lp0 caused the printer to spit out one page with the PostScript header on it, and the error light to flash. Pressing resume caused the printer to spit out one blank sheet, and then think it was through.
Thanks again for the help.
Mike
They were only going to hook up to MS SQL servers so the only additional piece they require is:
Part#: 400100-1xx Name: SQL APM Price: 1075 (MSRP)
They need 2 of these. We'll have to get in touch with Chad Whalen about waiving the 20% fee...or maybe getting deeper discounts and passing them along. (i.e. a additional 20% off to cover!)
On Tuesday 17 January 2006 11:08, Mike McCarty wrote:
Mogens Kjaer wrote:
Thanks for the help.
Mike McCarty wrote: ...
It ran 0.7% to 1.3% (max, not often hit) during a test print. That is CPU utilization. So I don't think it is CPU limited.
On FC4, one can create a custom device (instead of /dev/lp0),
I'm sorry, but I'm not an experienced UNIX admin. I'd need more information about how to run MAKEDEV. Oh, and I run FC2, not FC4.
and send the output to /tmp/printer.out (create this file first, with 777 permission). Try this, and see how long it takes to print to the file.
Umm, it took not any time at all to create a one line text file. I'm experiencing the problem even with a file with only 16 ASCII characters in it.
Try to copy the file to /dev/lp0 and see how long that takes.
Ok, I tried this file...
$ ls -l test.txt -rw-rw-r-- 1 jmccarty jmccarty 17 Jan 17 03:47 test.txt
$ cat test.txt This is a sheet.
The "data transfer" light on the printer started to flash immediately, and after 16 seconds, the printer grabbed a sheet and printed the line instantly. It LOOKED like it might have been downloading the font, perhaps. It certainly used a different font from the one that lpr used.
But, I did this... I printed a one-page thing from Mozilla to a file in my home directory, and got mozilla.ps
$ ls -l ~/*.ps -rw-rw-r-- 1 jmccarty jmccarty 504768 Jan 17 11:02 home/jmccarty/mozilla.ps
Creation of that file was essentially instantaneous. Copying it to /dev/lp0 caused the printer to spit out one page with the PostScript header on it, and the error light to flash. Pressing resume caused the printer to spit out one blank sheet, and then think it was through.
Thanks again for the help.
Mike
p="p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} This message made from 100% recycled bits. You have found the bank of Larn. I can explain it for you, but I can't understand it for you. I speak only for myself, and I am unanimous in that!
Here's a copy of /proc/interrupts. Can anyone help me figure out whether the machine is even configured to accept interrupts from the printer? I don't see anything which would lead me to believe that it is.
CPU0 0: 123049771 XT-PIC timer 1: 102313 XT-PIC i8042 2: 0 XT-PIC cascade 3: 0 XT-PIC uhci_hcd 5: 943 XT-PIC Intel 82801DB-ICH4 8: 1 XT-PIC rtc 9: 944 XT-PIC uhci_hcd 10: 0 XT-PIC uhci_hcd 11: 489740 XT-PIC ehci_hcd, eth0 12: 591304 XT-PIC i8042 14: 408191 XT-PIC ide0 15: 1332973 XT-PIC ide1 NMI: 0 ERR: 1
I interpret this as:
timer "i8254" CTC i8042 keyboard controller or ? cascade PIC cascade input uhci_hcd ? 82801DB-ICH4 ? rtc "MC146818" TOD + RAM uhci_hcd ? uhci_hcd ? ehci_hcd ethernet I/F i8042 ? or keyboard controller ide0 ATA I/F ide1 ATA I/F
Mike
Mike McCarty wrote:
Here's a copy of /proc/interrupts. Can anyone help me figure out whether the machine is even configured to accept interrupts from the printer? I don't see anything which would lead me to believe that it is.
You are not using interrupts for the printer port. You may want to run tunelp to turn it on. You probably need:
tunelp /dev/lp0 -i 7
Tunelp can also make other adjustments for you...
Mikkel
Mikkel L. Ellertson wrote:
Mike McCarty wrote:
Here's a copy of /proc/interrupts. Can anyone help me figure out whether the machine is even configured to accept interrupts from the printer? I don't see anything which would lead me to believe that it is.
You are not using interrupts for the printer port. You may want to run tunelp to turn it on. You probably need:
tunelp /dev/lp0 -i 7
Tunelp can also make other adjustments for you...
Mikkel
I read the man page for tunelp, and it looked like just the thing. But...
# tunelp /dev/lp0 -i 7 tunelp: ioctl: Invalid argument /dev/lp0 using IRQ -1
Mike
Mikkel L. Ellertson wrote:
Mike McCarty wrote:
Here's a copy of /proc/interrupts. Can anyone help me figure out whether the machine is even configured to accept interrupts from the printer? I don't see anything which would lead me to believe that it is.
You are not using interrupts for the printer port. You may want to run tunelp to turn it on. You probably need:
tunelp /dev/lp0 -i 7
Tunelp can also make other adjustments for you...
Mikkel
Hmm, I missed this part...
NOTE: This option will have no effect with kernel 2.1.131 or later since the irq is handled by the parport driver. You can change the parport irq for example via /proc/parport/*/irq. Read /usr/src/linux/Documentation/parport.txt for more details on parport.
# uname -r 2.6.10-1.771_FC2
# ls /proc/parport ls: /proc/parport: No such file or directory
# ls /usr/src/linux/Documentation/parport.txt ls: /usr/src/linux/Documentation/parport.txt: No such file or directory
Hmm...
Google turned up a document named parport.txt, but it looks rather old.
I did find this...
# cat /proc/sys/dev/parport/parport0/irq -1
And -1 happens to be what tunelp reports.
I wonder whether putting a 7 in there would help any.
# cat /proc/sys/dev/parport/parport0/autoprobe CLASS:PRINTER; MODEL:DESKJET 870C; MANUFACTURER:HEWLETT-PACKARD; DESCRIPTION:Hewlett-Packard Deskjet 870C; COMMAND SET:PCL,MLC,PML;
which looks correct as much as I understand (command set I dunno)
# cat /proc/sys/dev/parport/parport0/modes PCSPP,TRISTATE,EPP
Hmm. EPP looks ok, I guess. I dunno what PCSPP means.
# cat /proc/sys/dev/parport/parport0/spintime 500
I wonder whether tuning this might speed things up?
Mike
Mike McCarty wrote:
I did find this...
# cat /proc/sys/dev/parport/parport0/irq -1
And -1 happens to be what tunelp reports.
I wonder whether putting a 7 in there would help any.
Well, I have 7 in there, and "tunelp /dev/lp0" returns /dev/lp0 using IRQ 7
But I suspect you will get the same error as with tunelp. What you may want to do is the next time you boot the system, go into the BIOS setup and check the mode, and if it has an IRQ assigned to it.
# cat /proc/sys/dev/parport/parport0/modes PCSPP,TRISTATE,EPP
Hmm. EPP looks ok, I guess. I dunno what PCSPP means.
I forget what it stands for, but I believe that is the mode where the port uses DMA for transfers.
# cat /proc/sys/dev/parport/parport0/spintime 500
I wonder whether tuning this might speed things up?
Well, that is one of the things that tunelp is supposed to be tweak. You may want to use tunelp instead of writing it directly.
Mikkel
Mikkel L. Ellertson wrote:
Mike McCarty wrote:
I did find this...
# cat /proc/sys/dev/parport/parport0/irq -1
And -1 happens to be what tunelp reports.
I wonder whether putting a 7 in there would help any.
Well, I have 7 in there, and "tunelp /dev/lp0" returns /dev/lp0 using IRQ 7
But I suspect you will get the same error as with tunelp. What you may want to do is the next time you boot the system, go into the BIOS setup and check the mode, and if it has an IRQ assigned to it.
I can't modify the "file". (Yes, I was root.)
# cat /proc/sys/dev/parport/parport0/modes PCSPP,TRISTATE,EPP
Hmm. EPP looks ok, I guess. I dunno what PCSPP means.
I forget what it stands for, but I believe that is the mode where the port uses DMA for transfers.
# cat /proc/sys/dev/parport/parport0/spintime 500
I wonder whether tuning this might speed things up?
Well, that is one of the things that tunelp is supposed to be tweak. You may want to use tunelp instead of writing it directly.
It seems that tunelp cannot modify that, either.
BIOS Reports I/O 378, IRQ 7, ECP mode.
Mike
Mike McCarty wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
UPDATE:
I finally bit the bullet, and got out an old adaptor I made with both a male and a female connector on each end, and my 'scope, and looked at the lines. Well, the timing I looked at seemed odd. I saw long periods of time with very brief exchanges (like a few bytes), with the printer printing along, then a furious burst of what looked like normal timing, followed by long periods again. I also saw the BUSY line active for seconds at a time. Hmm.
So, as an experiment, I waited until the comm seemed quiescent, and pulled the cable. The printer continued to print for a few minutes, then started flashing the "error" light, indicating "loss of comm". I plugged in the cable for a second, and the printer was happy again, and I withdrew the cable, and the printer merrily printed for another few minutes. I did this for an entire page, which came out printed perfectly.
Sooo... communication time is not the culprit. For some reason, the printer is very very slow when printing these images.
I wonder what is going on inside there? And how did it print a page one time in less than three minutes?
Perhaps I should investigate the default dpi...
Mike
Mike McCarty wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
Oh frabjous day! Calloo! Callay!
Success at last! Using GNOME I edited the printer driver options to "Prerender Postscript" enabled, and I have now printed a 26 page document (which includes the "test page" I've been printing) start to finish, including queueing time and one paper jam I had to clear in just under six minutes. That's just about four pages a minute. Just the one "test page" was taking over twenty minutes. That's more than 80x speedup.
So, one misconception I had was that this is not a postscript capable printer. Apparently it is. But the postcript interpreting engine is S-L-O-W.
Pre-rendering was the trick.
Thanks all for your time and help.
Where should we make note of this, if at all?
Mike
On 2/8/06, Mike McCarty mike.mccarty@sbcglobal.net wrote:
Mike McCarty wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
Oh frabjous day! Calloo! Callay!
Success at last! Using GNOME I edited the printer driver options to "Prerender Postscript" enabled, and I have now printed a 26 page document (which includes the "test page" I've been printing) start to finish, including queueing time and one paper jam I had to clear in just under six minutes. That's just about four pages a minute. Just the one "test page" was taking over twenty minutes. That's more than 80x speedup.
So, one misconception I had was that this is not a postscript capable printer. Apparently it is. But the postcript interpreting engine is S-L-O-W.
Pre-rendering was the trick.
Thanks all for your time and help.
Where should we make note of this, if at all?
Mike
You might try the CUPS site. I believe they have an FAQ. Also your experience is not unique. Many people with old postscript printers forget that the engine in thoses things run in the 10s of megahertz. Today's desktop/laptop processor is at minimum an order of magnitude faster so it makes more sense to perform the rendering at the host.
On Wed, Feb 08, 2006 at 11:07:49AM -0800, Kam Leo wrote:
On 2/8/06, Mike McCarty mike.mccarty@sbcglobal.net wrote:
Mike McCarty wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
Oh frabjous day! Calloo! Callay!
Success at last! Using GNOME I edited the printer driver options to "Prerender Postscript" enabled, and I have now printed a 26 page document (which includes the "test page" I've been printing) start to finish, including queueing time and one paper jam I had to clear in just under six minutes. That's just about four pages a minute. Just the one "test page" was taking over twenty minutes. That's more than 80x speedup.
So, one misconception I had was that this is not a postscript capable printer. Apparently it is. But the postcript interpreting engine is S-L-O-W.
Pre-rendering was the trick.
Thanks all for your time and help.
Where should we make note of this, if at all?
Mike
You might try the CUPS site. I believe they have an FAQ. Also your experience is not unique. Many people with old postscript printers forget that the engine in thoses things run in the 10s of megahertz. Today's desktop/laptop processor is at minimum an order of magnitude faster so it makes more sense to perform the rendering at the host.
I've got a Lexmark Optra Color 40, Postscript inkjet printer. Its engine is also fairly slow, but nothing like as slow as the OP describes his HP printer to be. Sometimes it takes a while to get a page out of it, for complex pages, so maybe I'll try the 'pre-render' thing too.
Howdy Mike,
On Wed, 2006-02-08 at 12:52 -0600, Mike McCarty wrote:
Mike McCarty wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
Oh frabjous day! Calloo! Callay!
Success at last! Using GNOME I edited the printer driver options to "Prerender Postscript" enabled, and I have now printed a 26 page document (which includes the "test page" I've been printing) start to finish, including queueing time and one paper jam I had to clear in just under six minutes. That's just about four pages a minute. Just the one "test page" was taking over twenty minutes. That's more than 80x speedup.
So, you want to run that past me/us again? I assume you used gedit (gnome editor)? And exactly what path/file did you edit? Inquiring minds want to know ;-)
taharka
Lexington, Kentucky U.S.A.
taharka wrote:
Howdy Mike,
So, you want to run that past me/us again? I assume you used gedit (gnome editor)? And exactly what path/file did you edit? Inquiring minds want to know ;-)
No, I used the "GNOME Print Queue Properties Editor".
At the bottom of my screen, I see an icon of a printer. I clicked on that. It can also be started by clicking [RedHat]->SystemSettings->Printing, which goes directly to "Printer Configuration".
Anyway, using the icon on the "task bar" at the bottom brings up the "GNOME Print Manager". Right clicking on a queue associated with the printer and selecting Properties brings up "Printer Configuration", which shows all queues associated with the printer, associated with that queue. (Hard to understand here. I mean that there is a printer associated with the queue one clicked on to change Properties. The next box shows all queues associated with that printer.)
Right clicking on the queue and selecting "Edit" brings up "Edit a Print Queue". Clicking on the "Driver Options" tab allows one to select or deselect "Prerender Postscript". I clicked it to make a check mark appear there. That had to be done for all queues associated with the printer. Then I clicked "Apply" in the "Printer Configuration" dialogue box.
I hope that was clear.
I use FC2, so YMMV.
Mike
Mike McCarty wrote:
Off topic due to age of FC2?
Anyway, have installed an HP DeskJet 870Cse, and it prints unbearably slowly, like 30 minutes per page.
Well, the one queue works fine, now, but the "two up" queue I created which prints two up now does not. First, it no longer prints "two up" but "one up". And, it is still slow. Interesting. But at least my printer is usable now.
Mike
-
Aaron M. Hirsch -
Fred Smith -
Kam Leo -
Klaasjan Brand -
Les Mikesell -
Mike McCarty -
Mikkel -
Mogens Kjaer -
taharka