| <<<Back 1 day (to 2014/08/12) | 2014/08/13 |
ningu | hi, anyone around? | 01:07.12 |
rayjj | ningu doesn't seem very patient | 01:42.16 |
| (I don't know because the irclogs are in a different timezone from me) | 01:42.50 |
| DARN !!! The timings I did on the Pi WERE NOT AT THE BASELINE 700MHz, apparently. I ran "raspi-config" and set it for 700 and it was SLOWER. Even set for 800 is slightly slower | 02:15.35 |
| but 900 is faster, so I probably was running 800 :-( | 02:15.54 |
| QUESTION FOR ALL: should I re-run at 700 (stock) speed ? | 02:16.29 |
| While awaiting the answer, I will (might as well have it working) | 02:17.15 |
robin_watts_mac | rayjj: It occurs to me, that I made a change to MuPDF just before 1.3 that slowed it down a lot. We did a quick fix to it and released 1.4 | 05:41.11 |
| The bug was to do with the way we handle streams, and could sometimes cause us to decode 1 byte at a time, rather than a bufffers worth (typically 4K) at a time. | 05:42.11 |
mvrhel_laptop | rayjj: mine was set for 700 | 05:42.34 |
robin_watts_mac | We fixed it for flate, but there may be some decoder types out there that haven't been fixed. | 05:42.56 |
mvrhel_laptop | I am currently rebuilding with the latest gs | 05:43.07 |
robin_watts_mac | If you are consistently getting slower times for mupdf than you expect, it may be worth trying a couple of weeks before 1.3. | 05:43.40 |
mvrhel_laptop | found a bug in gsview. printing a document with pages that were landscape does not work so well | 05:43.46 |
robin_watts_mac | If there are problems I will fix them when I get home. | 05:43.49 |
henrys | jogux:yeah I did see that and the cups problems, I searched the web also and found folks asking questions on stack overflow about porting cups to android. Some group from Vietnam had lots of questions that fit the time frame. | 13:14.07 |
jogux | henrys: ah, good :-) hehe | 13:14.50 |
henrys | Iâd like to have our own printing someday so we can report the slime bags | 13:15.18 |
jogux | nods. would be nice | 13:15.57 |
kens | chrisl ping | 13:42.50 |
chrisl | kens: pong | 13:43.29 |
kens | Do you want to take on #695423 ? | 13:47.40 |
chrisl | Yep, I just saw it. I'll deal with it | 13:47.59 |
kens | OK thanks | 13:48.05 |
nemo | kens: heeey | 14:48.10 |
| 13:59 < nemo> Bisecting: 0 revisions left to test after this (roughly 0 steps) | 14:48.21 |
| 13:59 < nemo> [7ef7b0d503fee02da12430ee19a394b23b1bfa6f] pdfwrite - colour conversion, consider sRGB to be RGB | 14:48.24 |
| 13:59 < nemo> I don't get why that would cause http://bugs.ghostscript.com/show_bug.cgi?id=695420 | 14:48.27 |
| kens: rayjj disappeared, was going to send it to him since he filed | 14:48.39 |
| but he keeps referring to you in reference to stuff like this, so I guess this code is your baby | 14:49.04 |
| after dropping to one revision prior to that, all my files converted successfully using his syntax \o/ | 14:49.24 |
mvrhel_laptop | rayjj: so what numbers were you getting with PLRM at 700Mhz? | 19:05.28 |
| I was getting 80 secs to completion with the latest snap shot | 19:05.41 |
henrys | tor8: can you field the questions from the last support email? You can respond directly or send them to me or kens. | 19:18.41 |
rayjj | mvrhel: (for the logs). With gs the PLRM_100 600 dpi I get 69.9 from .pdf and 118.4 from PS. With pcl6 it is 21.9 | 20:29.33 |
| mvrhel: still collecting the monochrome 1-bit and 8-bit | 20:30.08 |
| darn. A typo in the test script, so starting the 8-bit wiped out the 1-bit tests results :-( | 20:33.35 |
mvrhel_laptop | rayjj: you there? | 21:06.22 |
| so I received the board from company c today | 21:06.31 |
| I was a bit surprised | 21:06.37 |
| I don't think it makes any sense for me to duplicate your efforts on this | 21:06.57 |
| henrys: what are your thoughts on this? | 21:07.08 |
henrys | mvrhel_laptop: why did they send you and ray a board? Of course you shouldnât duplicate rayjj âs work | 21:08.45 |
rayjj | mvrhel_laptop: I have no idea why you got a board. Maybe Miles or Scott asked them for it. | 21:11.24 |
| I'm working on porting gs to that board as an "app" (I discussed briefly w/ mvrhel_laptop yesterday on a phone call) -- It runs ThreadX :-( | 21:12.27 |
| AIUI those boards are for sale and cost aboutt $205 | 21:13.02 |
| but I'm not sure the one I was looking at on the internet was the same oen | 21:13.27 |
| BTW, the files I am using for the testing on the Pi are at casper.ghostscript.com:~ray/public/testing | 21:15.58 |
| mvrhel_laptop: when you dusted off your Pi, were you able to get the really fast PLRM times (I think you said you couldn't replicate the 43 seconds) ? | 21:19.10 |
| now that I'm *really* at 700 MHz, I am seeing times generally slower for PDF and faster for PS than your results. | 21:27.02 |
| mvrhel_laptop: and you didn't have a PLRM.ps time included | 21:27.28 |
mvrhel_laptop | rayjj: no I was not able to get the times I had claimed before | 21:32.18 |
rayjj | mvrhel_laptop: :-( | 21:32.28 |
mvrhel_laptop | rayjj: I am getting 80 secs on the 100 page pdf PLRM | 21:32.43 |
| is that what you are getting? | 21:32.47 |
rayjj | mvrhel_laptop: for anything, or just the PLRM you tested | 21:32.51 |
mvrhel_laptop | rayjj: I was just checking the PLRM | 21:33.20 |
rayjj | mvrhel_laptop: well, now that I know mine is set for 700, I _should_ be able to re-test. It's running the PCL mono now | 21:34.15 |
mvrhel_laptop | ok | 21:34.21 |
| I just want to know if we are matching finally | 21:34.31 |
| let me know later what you get | 21:34.38 |
rayjj | mvrhel_laptop: did you update to a current rev? or are you just running the binary you had ? | 21:35.07 |
mvrhel_laptop | current rev | 21:35.16 |
| so should be the same as you | 21:35.23 |
rayjj | I am running what I had in the initial report: commit 5671fc9 | 21:35.53 |
| but there hasn't been any major changes since that | 21:36.12 |
mvrhel_laptop | ok. well I just updated late yestreday | 21:36.34 |
| rayjj: not sure the exact commit | 21:36.42 |
rayjj | mvrhel_laptop: that's why I documented it :-) | 21:38.10 |
| off to lunch. bbiab. | 21:40.34 |
| Forward 1 day (to 2014/08/14)>>> | |