Username:
B
I
U
S
"
url
img
#
code
sup
sub
font
size
color
smiley
embarassed
thumbsup
happy
Huh?
Angry
Roll Eyes
Undecided
Lips Sealed
Kiss
Cry
Grin
Wink
Tongue
Shocked
Cheesy
Smiley
Sad
<- 123 ->
--
--
List results:
Search options:
Use \ before commas in usernames
Oh shit. I gotta check my demos >__> I'm still stuck on 3.015
lag. check my demo for me will you.
for this demo I went to joequake software, lowest resolution, used picmip and mipscale commands.
judging from my demo I´m at medium 1.3 when I start.
Hey Ho Let's Go
If I run JoeQuake in software it crashes the instant I try to run wot.bsp, the same happens to Mandel.

I'm beginning to think I'm jinxed for this contest (slow starttime and not being able to run crappy software resolutions) ;-)
Edit history:
Stubgaard: 2009-07-21 04:40:40 am
Hey Ho Let's Go
I just ran a quick test to see what starttime/finishtime I would get with the different engines by just walking into the exit.

FINISHING TIME:

WinQuake - 0:05.29334
GLquakebjp - 0:05.33312
GLquake - 0:05.38251
JoeQuake-GL - 0:05.39774

The difference between WinQuake and JoeQuake-GL is .1! That's a lifetime for a run like this. Unfortunately JoeQuake Software crashes when I try to run wot.bsp as mentioned in my previous post, so I couldn't test that.

Lag: Can I get you to run the demos through lmpc and see what the actual starttimes are?
Attachment:
Invisible avatar
You don't need to use lmpc, demtool has the same option :P. Let me check it out...
sda loyalist
PlayerEngineTime
Relomywinquake1.31410813
Stubgaardwinquake1.31874371
Relomyjoequake-gl1.31885040
Lag.Comjoequake1.32078755
Lag.Comwinquake1.32390308
Stubgaardglquakebjp1.33320308
Lag.Comjoequake-gl1.33496201
Lodisjoequake1.33526790
Lag.Comd3dquake1.33712661
Rutger?1.34996331
Lag.Comglquake1.37468386
Stubgaardglquake1.38311446
Stubgaardjoequake-gl1.39495251
Ajax?1.50608778

Seems like software beats hardware. By the way, that glquakebjp demo seems to have a ton of demo errors (or at least unrecognised fields) in it. Also, I think I just solved the 'Ann Ahl' mystery!
Invisible avatar
Guess no need for me to post my results  Roll Eyes
Hey Ho Let's Go
Quote from lag:
PlayerEngineTime
Stubgaardwinquake1.31874371
Relomyjoequake(gl?)1.31885040
Lag.Comjoequake1.32078755
Lag.Comwinquake1.32390308
Stubgaardglquakebjp1.33320308
Lag.Comjoequake-gl1.33496201
Lodisjoequake1.33526790
Lag.Comd3dquake1.33712661
Rutger?1.34996331
Lag.Comglquake1.37468386
Stubgaardglquake1.38311446
Stubgaardjoequake-gl1.39495251
Ajax?1.50608778

Seems like software beats hardware. By the way, that glquakebjp demo seems to have a ton of demo errors (or at least unrecognised fields) in it. Also, I think I just solved the 'Ann Ahl' mystery!


Sickening how much difference there is between the engines!

So, care to spill the beans on the Ann Ahl mystery?
Mumma
That's a mighty and weird table. But the differences are so "significant", at least for a short contest like this one. I suppose time varies between restarts of maps as well though?

I can't imagine that Jardrup's engine would produce a single error. I bet it's just that the demo uses a different protocol. Bengt himself has a tool to convert demos from various protocols, called ConvDem.

And do tell who Ann Ahl is! It sounds like a Swedish name (There seems to be 38 persons with that name in Sweden) but I think it's a joke by someone!
sda loyalist
Yeah, the start time varies by a little tiny bit in my retries, .001~.004 in most engines. Demos on say, menk.bsp are likely to have a further .1 penalty or something. You're probably right about the bjp thing. I guess lmpc is too old to recognise this newer format properly. It still parses enough of it to play, at least.

I think I'll let him come forward on his own. Roll Eyes
uh. I may be missing something, but it doesn't really seem valid to compare demos on different clients on different computers.
A faster computer should have an advantage really, and get a faster start time...
sv nice
Quote from lag:
PlayerEngineTime
Stubgaardwinquake1.31874371
Relomyjoequake(gl?)1.31885040
Lag.Comjoequake1.32078755
Lag.Comwinquake1.32390308
Stubgaardglquakebjp1.33320308
Lag.Comjoequake-gl1.33496201
Lodisjoequake1.33526790
Lag.Comd3dquake1.33712661
Rutger?1.34996331
Lag.Comglquake1.37468386
Stubgaardglquake1.38311446
Stubgaardjoequake-gl1.39495251
Ajax?1.50608778

Seems like software beats hardware. By the way, that glquakebjp demo seems to have a ton of demo errors (or at least unrecognised fields) in it. Also, I think I just solved the 'Ann Ahl' mystery!


Mine is JoeQuake-gl.

Here's a demo on WinQuake.  The time isn't bad, but it isn't my best.  I wonder how the starting times compare.
sda loyalist
Predictably, slightly quicker start than Stubgaard's winquake.
Hey Ho Let's Go
dex solved my issue with not being able to run Quake properly in 72 fps. It turns out it gives me a .02 advantage in JoeQuake-GL as opposed to when I was running in 60 fps.
Hey Ho Let's Go
From the contest page:
http://speeddemosarchive.com/quake/contests/shamhell-revisit.html

19 entries, not bad! Even some oldtimers felt obliged to contribute to this contest. The competition was fierce, with Jozsef Szalontai claiming the grand prize, followed closely by Daniel Hansson and Adam Lewandowski. Only a sniff of air seperated top three (and the rest of the field for that matter). Rumors has it that someone recorded a 2.88, but we never received that demo, unfortunately.

It wasn't intentional by any means, but it soon became apparent this contest would turn into a hard-fought duel based on the different Quake engines. It seems WinQuake had the edge by a small margin compared to, say, JoeQuake. We're talking tenths of a second - which normally doesn't sound like a whole lot - but for a short run like this, it's the difference between success and failure. Some would claim a bit of unfairness is involved in this contest - and we agree - so that's why we've divided the tables into three categories based on the finishing time, the duration of the demo, and the start-time (when the demo starts recording). The time that truly counts is the finishing time though, and we've based the results on that.

    * Download all the demos (.zip format): http://speeddemosarchive.com/quake/contests/demos/wot-revisit.zip - 5.2 MB
    * A video of the winning demo in "insane quality" (.mp4 format): http://speeddemosarchive.com/quake/videos/wot_jozsef_iq.mp4 - 2.8 MB
    * High quality video (.mp4 format): http://speeddemosarchive.com/quake/videos/wot_jozsef_hq.mp4 - 1.1 MB
    * ...and low quality video (.mp4 format): http://speeddemosarchive.com/quake/videos/wot_jozsef.mp4 - 287 KB


Results : Finishing Time (exit)

                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   
RANKPLAYERTIME
01.Jozsef Szalontai0:02.9184
02.Daniel Hansson0:02.9222
03.Adam Lewandowski0:02.9439
04.Thomas Bergendorff0:02.9466
05.Pif de Mestre0:02.9508
06.Mathias Thore0:02.9606
07.Thomas Stubgaard0:02.9627
08.Karol Urbanski0:02.9667
09.Paul Davies0:02.9723
10.Mattias Öman0:02.9831
11.Rutger Baks0:02.9833
12.Flavio Quadros0:02.9983
13.Jaakko Alakopsa0:03.0187
14.Rui Neto0:03.0337
15.Mads Peter Stubgaard0:03.1315
16.Laurent Mertens0:03.1476
17.Anders Nordensten0:03.2003
18.Marek Bartoszek0:03.2161
19.Philip Chute0:03.4205



Results : Duration (actual playing time)

RANKPLAYERTIME
01. (---)Jozsef Szalontai0:01.5930
02. (---)Daniel Hansson0:01.6018
03. (12th)Flavio Quadros0:01.6087
04. (---)Thomas Bergendorff0:01.6113
05. (3rd)Adam Lewandowski0:01.6250
06. (5th)Pif de Mestre0:01.6252
07. (6th)Mathias Thore0:01.6262
08. (11th)Rutger Baks0:01.6333
09. (---)Paul Davies0:01.6374
10. (7th)Thomas Stubgaard0:01.6406
11. (13th)Jaakko Alakopsa0:01.6412
12. (8th)Karol Urbanski0:01.6528
13. (10th)Mattias Öman0:01.6665
14. (---)Rui Neto0:01.6817
15. (---)Mads Peter Stubgaard0:01.7390
16. (17th)Anders Nordensten0:01.8144
17. (16th)Laurent Mertens0:01.8198
18. (---)Marek Bartoszek0:01.8292
19. (---)Philip Chute0:02.0007


Results : First-time (time when demo starts recording)

RANKPLAYER    TIME
01. (8th)Karol Urbanski0:01.3138
02. (10th)Mattias Öman0:01.3166
03. (---)Adam Lewandowski0:01.3189
04. (2nd)Daniel Hansson0:01.3204
05. (7th)Thomas Stubgaard0:01.3220
06. (1st)Jozsef Szalontai0:01.3253
07. (5th)Pif de Mestre0:01.3255
08. (16th)Laurent Mertens0:01.3278
09. (4th)Thomas Bergendorff0:01.3352
10. (6th)Mathias Thore0:01.3344
11. (9th)Paul Davies0:01.3349
12. (11th)Rutger Baks0:01.3499
13. (14th)Rui Neto0:01.3520
14. (13th)Jaakko Alakopsa0:01.3775
15. (17th)Anders Nordensten0:01.3859
16. (18th)Marek Bartoszek0:01.3868
17. (12th)Flavio Quadros0:01.3896
18. (15th)Mads Peter Stubgaard0:01.3925
19. (---)Philip Chute0:01.4197


Attachment:
Mumma
Congradurations Jozsef. Good to see you in action again.
thanks Smiley

it's not a comeback though, don't expect new speedruns from me, but entering this contest was quite refreshing, and Stubby also pushed me hard Cheesy
Hey Ho Let's Go
Quote from j0zz1e:
thanks Smiley

it's not a comeback though, don't expect new speedruns from me, but entering this contest was quite refreshing, and Stubby also pushed me hard Cheesy


There's a new contest up in case you hadn't noticed.... ;-)

http://speeddemosarchive.com/quake/contests/155t.html

Time for a new round of the most extreme form of quake speedrunning - a contest in The Stairs of Death by Trinca made by Rui Neto. Your job is clear, run fast or destroy everything. Make monsters feel that you are a great warrior. Eat them as a breakfast. Now go and make sure that you are the best!

  1. This contest will end Midnight CET, Monday, August 17th, 2009.
  2. Send your demos to quake @ speeddemosarchive . com
  3. Name your demo and the accompanying textfile in the format 155t_er/nh_playername.dem/txt. Someone named Rosalie may send in 155t_er_rosalie.dem, for example.
  4. QdQstats must be used to get the exact time.
  5. Two categories are contested this time - Easy Run and Nightmare 100%.
Edit history:
Stubgaard: 2009-08-04 04:48:45 am
Hey Ho Let's Go
I ran some more straight-forward-walking tests with the different Quake engines, this time on the 100m map. The outcome doesn't differ much from those we ran on the contest map, but surprisingly enough an older version of JoeQuake is faster than the newest build. The times listed below are the average finishing times.

ENGINE BUILD FPS TIME
winquake v1.09 - 15.64
joequake-gl v1329 72 FPS 15.66
joequake-gl v1329 60 FPS 15.67
glquakebjp v1.33 - 15.67
joequake-gl v1862 72 FPS 15.72
glquake v0.98 - 15.72
joequake-gl v1862 60 FPS 15.74
I don't think testing it endlessly is the right way. But in any case, I did a bunch of tests myself =P

To me, it just looks like a matter of loading time. The faster the computer, the smaller the map, the earlier the timer starts. I actually checked the exact starting times, like Lag did, of a lot of demos, and got some interesting results.

My demos from up to the end of 07, which were recorded on my older coputer, all started at 1.39999998s. I grabbed a demo from you (Stubgaard) and one from Mandel, all from early 07, and the starting time is that same 1.39999998s. That was the 1329 build. Recent demos show various starting times, above and below that 1.4 mark.

Still, it's not really random. All my demos from my older pc with newer JoeQuake had an aprox time of 1.44167

Lagging my (current, newer) computer and comparing different maps I got different numbers, but they came in packs. A demo in the huge map menk, for instance, gives a starting time of aprox 1.44167! As I lag down the pc, it takes a while before it drops to an even slower start. A demo in a regular sized map will give me 1.41.. with some precision, smaller maps like wot give me 1.389.. and so on. With the game lagged to it's minimum fps, I get a time of 1.700000.. With hfr very low, it's very close to 1.300000...

So, what do we get from all this? The game has its weird ways of loading the game and starting the timer (probably rolling some frames before it really starts; the difference from the numbers I got weren't exactly multiples of 1/72, but whatever). We could keep testing to find out what engines are faster to shave some hundreths of a second, but instead of wasting our time with reverse engineering, it may be wiser to just ask someone who actually understands something (Lag.Com) to look at the code Smiley It may be very engine dependent (soft/opengl, etc), but he should be able to shed some light.

As seen in the contest, my loading times are quite bad, but that never stoped me =D Although it was a weird feeling a couple of days ago, when I missed a record for .07s... Ignorance is bliss?
sda loyalist
I've had a quick look... as most programs do, JoeQuake (and probably the other clients) use time(), i.e. the system clock, to calculate the start time. What I don't understand is why it starts at 1.4 or whatever when most levels take MUCH less than that to load. I guess the easiest thing to assume would be that the timer starts on 1.0 then adds loading time. But I don't know why it starts on 1!
Thunderbolt Forever :)
Quote from Stubgaard:
There's a new contest up in case you hadn't noticed.... ;-)

http://speeddemosarchive.com/quake/contests/155t.html
I've just sent both er and nh demos, please check if they are ok...

Thanks a lot :-)
Thunderbolt Forever :)
http://speeddemosarchive.com/quake/contests/155t.html

Hey people, that's not fair.

I sent 0:10.02146 ER and 0:26.19606 NH, but you listed an old 11.2 ER and no NH at all.

Of course they are slow, but that's my work...
Quake and Sex owns
I´m in vacations :/ and my computer crash... fuckkkkkkkk i had two demos... Sad i didn´t sent then in day 14 because i took my computer with me, but it crashed Sad and now i didn´t sent any... buaaaaaaaaaaaa

:| maybe bext time! i had some nice demos Sad
sv nice
Quote from Droog_Andrey:
http://speeddemosarchive.com/quake/contests/155t.html

Hey people, that's not fair.

I sent 0:10.02146 ER and 0:26.19606 NH, but you listed an old 11.2 ER and no NH at all.

Of course they are slow, but that's my work...


"Some demos were corrupted so they had to be left out of the results. Andrey Kulsha's keyboard demo has to be viewed with Bengt Jardrup's engine."

Maybe the first sentence has to do with your other one.