Zhlt 3.0 Beta

123468

Comments

  • amckernamckern Join Date: 2003-03-03 Member: 14249Members, Constellation Posts: 970 Auto Verified
    yeah your site is geting lot of links in mapping forums, more so becuase its the docs that are missing from some tools

    Beta 4 is almost done, i'm going to complie 2bases2 tonight (my map) and see if it has any errors left over, from then it should be out, how ever would you be willing to do an update on the docs?

    I hope beta 4 is the stable version, and the next one will be 3.0 final, from there I can start working on the source tools

    The lattest Change log is normaly in the /beta folder, but i'll send you one via email when i update it as well

    Adam
    No Frils!
  • The_Real_NemThe_Real_Nem Join Date: 2002-12-16 Member: 10900Members Posts: 53
    I did a little work on RipEnt today. The tool now operates like the other tools (in terms of output) and adds the following options: -chart, -noinfo, -lightdata and -parse. -chart, -noinfo and -lightdata are self explanatory (though I'm, not sure why -lightdata was not in RipEnt (is there a reason for this?)). -parse operates as follows:

    On -export -parse parses the entity data and reformats it for easy reading in any text editor.

    On -import -parse parses the entity data and removes all unnecessary white space. The parsing operation also works as a means to insure that the entity data is well formed, if it isn't an error message is printed to help track down the problem.

    I've attached the new code and I'm pretty confident I haven't messed anything up. The new code is also well commented.

    Nem
  • BulletHeadBulletHead Join Date: 2004-07-22 Member: 30049Members Posts: 2,530
    any ETA yet?

    Will this all work with the HL2 SDK?
    QUOTE (DragonMech @ Jul 9 2005, 09:19 PM)
    QUOTE (Sonic @ Jul 9 2005, 06:49 PM)
    I wish my butcheeks could propel me up flights of stairs are terrifying speeds.

    I sense a custom title right there... :D
  • ReveReve Join Date: 2003-09-23 Member: 21142Members Posts: 64
    amckern: Yes of course I'll keep the docs up to date smile-fix.gif and don't mind doing so whoever is doing the coding (oh, and good luck with your job application!). I'll try and keep it up to date with whatever updates happen with the tools. Actually, I'd appreciate if you can confirm which of the 'severe p15.5 problems' listed on the front page of zhlt.info are fixed in beta 4, and also which versions of the tools they were fixed in. actually, the more detailed the change log you can give me, really the better.

    I'd consider making a 'beta4 testing thread' once beta4 is made, and ask people to test their maps with it. Perhaps worth posting on a few other forums (like countermap) linking to that thread, to keep track of how well it is going. I reckon a thread for each beta release (hopefully not many left before 3.0 gold) will help track any last bugs down.

    Nem: Wow more excellent stuff! Keep up the good work! (and by the way, I love your tools as well smile-fix.gif).
  • amckernamckern Join Date: 2003-03-03 Member: 14249Members, Constellation Posts: 970 Auto Verified
    there is an alpha of beta 4 (strange ?)

    At the moment the alpha has the working rad, and vis, and the beta 2, csg, and bsp, it all works fine

    how ever there are still some bugs that need to be fixed before the beta is out -and i thought i fixed them all, dam you 2bases2!

    The main bugs holding beta 4 back are csg brush errors, and bsp showing leeks,or crashing

    I also found the mapname.rad crash bug, how ever i belive that it its not where it is - part of the on screen print out log code?

    NO 3.0 will not work with source (well have not been tested anyway), if people can complie cs:s maps with these tools, feel free to post, and tell the community

    Adam
    No Frils!
  • ReveReve Join Date: 2003-09-23 Member: 21142Members Posts: 64
    Can I just double check I understood that correctly? Did you mean the beta 4 vis and rad work fine, and the beta 2 csg and bsp work fine (and together can compile a map ok?), but the changes in csg/bsp for beta 4 aren't working yet?

    Or am I getting all confused? confused-fix.gif
  • amckernamckern Join Date: 2003-03-03 Member: 14249Members, Constellation Posts: 970 Auto Verified
    yeah your right, no confusion to be found

    The new tools will be up on the web soon, and the revesion 3 beta 4 source code

    Adam
    No Frils!
  • BulletHeadBulletHead Join Date: 2004-07-22 Member: 30049Members Posts: 2,530
    Pardon me for sounding stupid (tis now 0237 hours) but uh... what is your website?
    QUOTE (DragonMech @ Jul 9 2005, 09:19 PM)
    QUOTE (Sonic @ Jul 9 2005, 06:49 PM)
    I wish my butcheeks could propel me up flights of stairs are terrifying speeds.

    I sense a custom title right there... :D
  • amckernamckern Join Date: 2003-03-03 Member: 14249Members, Constellation Posts: 970 Auto Verified
    http://ammahls.com




    Beta 4 is now out, download dirrect from www.zhlt.tk

    Changelog @ http://downloads.ammahls.com/zhlt/beta/changelog.txt

    QUOTE

    Beta 4

    Package
    Added
    ZHLT.wad This version fixes some issues when using a texture with capital letters
    HalfLife QuArK file (DataHL.qrk (Jan 2001))
    lights.rad (Hammer 3.4)
    wad.cfg (1.7)

    ALL TOOLS
    Started converting functions to assembly, in other words, making each bit of code, as fast as it can go, by making it as Simple as possible

    CSG and BSP
    Cleaned up some tab spacing, so that built in command line options where displayed proper

    RAD
    Fixed
    Crash bug when using Opaue entitys
    Added
    Faster Rad Code by Nem, Improves complie time by about 15-27%

    Ripent
    Added
    New command line switch (-parse), by Nem that Pareses and reformats entity data stripping all non essential formatting, to cause less chance of a bug




    Beta 3

    This version was a test, for including the MSVC.lib code, it did not work, so was removed after a small internal beta test with members of #mapcore, and #svencoop on Gamesurge IRC


    @Reve

    I will be starting a campaing soon, i'll send u an email (via the forums email thing) lets talk about it

    Adam
    No Frils!
  • FragBait0FragBait0 Join Date: 2003-12-16 Member: 24431Members Posts: 58
    Okay built 3 maps with beta 4....and it works great!

    The modified ripent is useful - when i use opt_entdata it used to spit out the charts all the time.

    RAD has a habit of getting several million pagefaults if you use -customshadowwithbounce. This happens in p13 as well.
    This seems to occur during buildvisleafs.
    It's nonfatal but seems to slow RAD down a LOT - buildvisleafs runs merrily till about 40% on this particular map and slows. Giving all the pagefaults while its there.

    Using low chops puts out a nice warning - fair enough.
    But it appears in a most odd way - on the start of the "hlrad v2.5.3 rel.." line.
    Which results in some ugly breaks. Meh.

    One last thing then, could logging be put into ripent?

    Good work guys keep it up!

    *almost hits "Add Reply"
    Oh. Seems the weird shadow bug is around still. sad-fix.gif
    It's not as bad now...
  • MendaspMendasp I touch maps in inappropriate places Valencia, Spain Join Date: 2002-07-05 Member: 884Members, NS1 Playtester, Contributor, Constellation, NS2 Playtester, Squad Five Gold, NS2 Map Tester, Reinforced - Shadow, WC 2013 - Shadow, Community Dev Team Posts: 4,175 Auto Verified
    I'm getting Bad Surface Extents errors on bast and:

    a) Old versions of HLRAD didn't give any problem
    b) Check for problems shows nothing

    >_>
    <_<
  • amckernamckern Join Date: 2003-03-03 Member: 14249Members, Constellation Posts: 970 Auto Verified
    edited November 2004
    sad-fix.gif

    Am I ever gona be able to get a final out?

    Well, this will be the last version before Source is public, but i will work on the tools a little bit more

    I also thought that i had fixed the inverse shadow code, but it must just be my cube map, was a little bit to small

    At least the major issues, of Beta 2 Have been fixed, and tested as fixed

    Also, you might ask why are there beta 2 csg, and bsp?

    Well the complier broke the beta 4 ones, (For some reason), and the beta 4 tools, did not get any code changes from beta 2

    source code updated - http://ammahls.com/zhlt/zhlt30_beta4_rev3-src.rar (184kb)

    Adam
    No Frils!
  • FragBait0FragBait0 Join Date: 2003-12-16 Member: 24431Members Posts: 58
    Humm something is giving me weird red light thrown all over one map...

    Warning: point (120.000 1063.460 80.000) infinite v (-1.#IO -1.#IO -1.#IO)
    Thats a new one.

    Back to p13 RAD. sad-fix.gif

    Again CSG, BSP, VIS are all fine and I'm really loving the modified ripent.
    Wooties @ Nem for that.

    Hey, it all compiles fine now. Yay.
    Cept for netvis of course.
  • ReveReve Join Date: 2003-09-23 Member: 21142Members Posts: 64
    amckern - do you want me to put up the changes for the different beta 4 revisions (rev1 - rev3) on the version history? If you want, you can send me a breakdown, otherwise I can just put up what you've written as the change log (i.e. just the summary as the 'beta 4' version).

    Well, despite some minor setbacks, it looks like you guys have made some great progress on these tools. 'Gold' can't be far off. Keep on hacking smile-fix.gif
  • amckernamckern Join Date: 2003-03-03 Member: 14249Members, Constellation Posts: 970 Auto Verified
    well there has only been one beta 4 made public, the reason for the revisons on the source code is becuase there where major changes to the code at each revesion (Rev 2 was basicly fixing the rad), and rev 3 has nems new stuff

    So no, just place the changes as beta 4

    Thanks for the heads up on the light (is it just the one map?) and also can you please send me the map source (winrar solid archive best compresion) to amckern@yahoo.com, so i can use it as a test map

    @mend - if i had a copy of bast, i would be able to test it the tools with it better then the others, as it seems to be that map has the most errors when useing 3.0

    Adam
    No Frils!
  • FragBait0FragBait0 Join Date: 2003-12-16 Member: 24431Members Posts: 58
    edited November 2004
    amckern: just sent it.

    I did some small tests with VIS tonight:

    38.58s | 3.0 beta4 SSE
    39.75s | 3.0 beta4 STD
    36.92s | 3.0 beta2 STD
    36.63s | 1.7p15
    38.08s | 2.5.3 CB1.7
    49.55s | 2.5.3
    The beta4 SSE is a custom compiled version with SSE turned on - because I could smile-fix.gif

    Obviously these tests shouldn't be taken as a golden bar of vis performance, as they are quite short and I had a bunch of background processes going.
    But it *is* interesting how ZHLT 3 seems to have lost ground - particularly beta 4 to beta 2.
    It's not much but 8% is still 8%

    I didn't change any settings in the compiles and all versions ran without trouble.

    EDIT: formatting of times better
  • amckernamckern Join Date: 2003-03-03 Member: 14249Members, Constellation Posts: 970 Auto Verified
    yeah beta 2 was built with lots of wacky opts turned on, when i built the tools, such as comdat folding, and other stuff

    Beta 4 this time was built with some other functions, namely built for Pentuim pro and above, and desgined to be ran under win 98, though i'll have a look and see if i can switch some of these faster options back on

    I also belive i found where the memory leak is in rad, as floting point is off (This means that each chunck of ram is not fully filed, before the next chunck is started) the downside is, if i enable floting point, rad will take alot longer, at the cost of less ram

    I ask
    "Do you want a slower rad, over less memory useage?"

    Post

    Adam
    No Frils!
  • AnpheusAnpheus Join Date: 2004-09-30 Member: 32013Members Posts: 63
    I'd pick the less RAM version as default, and then have an option for not using it.
  • CageyCagey Ex-Unknown Worlds Programmer Join Date: 2002-11-15 Member: 8829Members, Retired Developer, NS1 Playtester, Constellation Posts: 1,751
    QUOTE (amckern @ Nov 16 2004, 02:03 PM)
    I also belive i found where the memory leak is in rad, as floting point is off (This means that each chunck of ram is not fully filed, before the next chunck is started) the downside is, if i enable floting point, rad will take alot longer, at the cost of less ram

    I ask
    "Do you want a slower rad, over less memory useage?"

    Are you talking about floating-point consistency? What switch do you think is causing the memory leak?

    The people with the worst RAD performance are those who run out of physical memory and are stuck with a chugging hard drive. For those people, improving memory efficiency will improve speed.

    I also don't like the idea of hurting program stability (which a memory leak can do over time) as a tradeoff for speed. There are better ways to optimize (algorithm changes have the most potential, followed by code optimizations like nem's recent work or FASTMATH) that don't bargain qualiity for performance.
    XP-Cagey

    Recommended Reading: NS Mapping Guidelines | Mapping Forum FAQ
    Nostalgia: Power Cells Thread
  • ReveReve Join Date: 2003-09-23 Member: 21142Members Posts: 64
    Will having this floating point turned off affect people building maps using QuArK? The QuArK editor uses floating points throughout.

    Do you have approximate statistics for the differences in cpu time required/ram required?

    I'll update the ZHLT version history tonight if I get the chance. Sorry have been a little busy.

    @XP-Cagey - are you planning to revive xp-cagey.com?
  • ReveReve Join Date: 2003-09-23 Member: 21142Members Posts: 64
    FragBait0 - regarding your Entdata optimizing tool, is it getting stable enough such that it could be included with ZHLT 3? Are there any large dependencies (runtimes, dlls) that would be too big to bundle?

    amckern - I was thinking, this would be quite a nice addition to the 'shiny new ZHLT 3 tools' factor in releasing this. Assuming the tool is getting mature enough, and everyone involved doesn't mind. What do you (and FragBait0, and everyone else) think?

    Perhaps all of the tools other than the csg/bsp/vis/rad (i.e. ripent, opt_plns, netvis if it's working, entdata tool, etc) should be stuck in a subfolder like 'Advanced tools', to help cut down on the confusion about what they all do.
  • ReveReve Join Date: 2003-09-23 Member: 21142Members Posts: 64
    Hi all,

    I've updated the docs locally but my host appears down temporarily, so I'll have to wait till it's working again till I can upload the changes (and so everyone can see them).

    Just to let you know, I've added a 'download zhlt' section linking to zhlt.tk and also to historical versions of the tools.

    Work silly host, work! asrifle.gif

    On a different note, can someone here ask a mod to unsticky the old 'p series' thread and sticky this one in its place?
  • AnpheusAnpheus Join Date: 2004-09-30 Member: 32013Members Posts: 63
    First of all, please edit your post to include new information rather than double, and triple posting.

    Second, keep the p tool page stickied, these tools arent fully stable. And many people, including myself, prefer to use the old tools due to stability... oh, and lack of bugs.
  • FragBait0FragBait0 Join Date: 2003-12-16 Member: 24431Members Posts: 58
    QUOTE (Reve @ Nov 17 2004, 09:19 AM)
    FragBait0 - regarding your Entdata optimizing tool, is it getting stable enough such that it could be included with ZHLT 3? Are there any large dependencies (runtimes, dlls) that would be too big to bundle?

    amckern - I was thinking, this would be quite a nice addition to the 'shiny new ZHLT 3 tools' factor in releasing this. Assuming the tool is getting mature enough, and everyone involved doesn't mind. What do you (and FragBait0, and everyone else) think?

    Perhaps all of the tools other than the csg/bsp/vis/rad (i.e. ripent, opt_plns, netvis if it's working, entdata tool, etc) should be stuck in a subfolder like 'Advanced tools', to help cut down on the confusion about what they all do.

    I'll second the "advanced tools" folder.

    I don't see why not.

    It's not VERY large...
    490kb uncompressed, 439kb RARed. It dosen't gain much because I use UPX on it.

    And I'd be happy to see it with the tools. Might need to modify the doc though to fit in. smile-fix.gif
  • ReveReve Join Date: 2003-09-23 Member: 21142Members Posts: 64
    edited November 2004
    amckern - I've just realised that you've included the DataHL.qrk file in the ZHLT distribution. This file is now bundled along with QuArK (along with many others), and so you don't need to include it. The one you are using (you say it's 2001) is also rather old, and so is probably worse than the one that you get when you download QuArK. Best to just remove it from the ZHLT toolset I think.

    Any more feedback on the opt_entdata/Advanced tools folder idea?
  • ReveReve Join Date: 2003-09-23 Member: 21142Members Posts: 64
    Changes I mentioned earlier are now live on the docs site.
  • amckernamckern Join Date: 2003-03-03 Member: 14249Members, Constellation Posts: 970 Auto Verified
    CODE

    if PostCount < 1000 SpamPost++;
    break;


    tounge.gif

    I am still learning cpp, and as such I understand the princables, but cant implemnt the changes very well

    QUOTE

    floating-point consistency

    Yes thats what I mean

    Netvis wont be complied with a new timestamp, but the next version will have the 1.7 netvis, i forgot about adding it back

    I have also reserected the Primary color bug, that was in p12 (the blue light bug) This time it comes in 3 flavors, and any map can see any or all of these, red, blue, and green colors coming from a sky brush

    If you are to use a brush based entity though the light will be the right color, i belive that this bug is becuase of the code changes in the opaue crash bug

    Reve - Please update the main page, the rad is now faster, due to the updated shadow code, and the new code by nem, also Nem is not on the credits page under Ryan Greeg, or Nemissis (Typo? )

    Adam
    No Frils!
  • ReveReve Join Date: 2003-09-23 Member: 21142Members Posts: 64
    Ok I've updated the docs with those notes. I've also added name anchors to the command reference, so if someone wants to link to a particular reference (e.g. hlrad -full), just link to the anchor #full smile-fix.gif
  • AnpheusAnpheus Join Date: 2004-09-30 Member: 32013Members Posts: 63
    Can somebody explain what the 'softlight' hack is? Its listed but no information is given on what the hell it does.
  • FragBait0FragBait0 Join Date: 2003-12-16 Member: 24431Members Posts: 58
    edited January 2005
    Reve, you brought up including opt_entdata with ZHLT3....

    I just thought I'd mention I've updated it - now it's a fair bit smaller so including it won't make the final package too big. I hope smile-fix.gif

    EDIT: And I don't know what the softlight hack is sposed to be exactly either smile-fix.gif
    Also, I think that this is a little...confusing:
    QUOTE
    To customise this shadow, add this key. The value represents how opaque the brush based entity should be treated in hlrad, on a scale from 1.0 to 0.0 (where 1.0 is fully opaque, casting no shadow, and 0.0 being fully transparent, casting a full shadow)

    If the entity is casting a shadow wouldn't that mean that, at least to RAD, it's opaque?
Sign In or Register to comment.