3DHardware.net scoops up the Videologic PowerVR Neon 250 (aka PVR Second/Next Generation). It goes without saying that this card kind of missed the boat, but still offers outstanding performance and quality (as seen in the Dreamcast), and offers a peek at Series 3. Too bad they missed out on Q3Test Phase4, but they still did a fps bit anyway:

We all know that the Timedemo is broken and does not give an accurate FPS, but still it does give an indication of where the performance is headed. The Neon 250 was run on the very first version of Q3Test due to the fact that there were still some issues concerning the compatibility with the newer Q3Test versions. All tests were run in 1024x768 with the highest possible detail level, tri-linear filtering and 32bpp source and rendering. Using cg_drawfps 1, the average FPS after running around the entire Test Level 1 a few times seemed around 40FPS. Minimum being around 25FPS and maximum weighing in around the 60FPS area. Not bad at all and very splendid considering the fact that the Neon 250 only runs at a clockspeed of 125MHz for both the core and memory. 32bpp doesn’t seem to be a big problem for the Neon 250, much due to the fact that there is no need for a Z-buffer. This also saves up memory meaning that the 16MBs of onboard memory can be used much more effectively for textures.

The guys at Omni Group have finished their v1.08 OSX port of Q3Test for those of you using the OS in question. Also while on the Mac tip the 3dfx folks have released some new Voodoo3 drivers.

Seeing as that QuakeCon is bearing down upon us like a very fat man with a gas problem, the QCon site has been updated with information on the official tournament. Here's some quick info:

• Maps will be Q3Test1 and Q3Tourney
• 9 Rounds of play; the first two are 8 player free for alls, and then you have to 1on1
• You'll play on a AMD K7 550 with 128MB of RAM and a ATI Rage Pro
• You get nifty stuff if you get 1st place :)
• You need to sign the wavier:
  plain text: http://www.quakecon.org/QC99legal.txt" rel="nofollow" target="_blank" class="external
  MS Word 97: http://www.quakecon.org/QC99legal.doc" rel="nofollow" target="_blank" class="external

Just wanted to let you guys that may be having problems know that you have to delete the old test in order to get it to work (as I just found out). It *seems* you can keep your configs, but I'm not sure.

Also, you'll prolly want to nuke 'baseq3' if you used that for the previous tests.

How do you know if you're having these problems? If the game tells you that your pak0.pak file is corrupt. Oh, and this info *is* in the readme.

Since we're all having to download another 30MB+ Q3Test, the boys at id felt sorry for us and added a new player model, Sarge. Graeme Devine tells us that all you have to do is type "model sarge" at the console and you're golden.

[update - Reorganized a bit by Jack]
Well, it looks like the .plan update from John Carmack last Sunday wasnt too far off. Version 1.08 of Q3Test has been released. The coolest thing is the player control setup system, as well as the addition of a new player model, the Sarge. Already played it a bit and the new system for players waiting in line on 1on1 servers seems to work great. Make sure and check out the Quake3 messageboard while you're at it. Waiting for that download, Immy has an realaudio file letting you know that we feel your pain waiting for those releases to appear on the web.

Win32 Mirrors
id Software, Barrysworld, Gamespy1, Gamespy2, Stomped.. (Thanks Redwood)

Mac Mirrors
Stomped, id Software, CDROM.COM, Blue's News

Related .plan updates
I slapped together all the 4 previous .plan updates from Carmack for you to read through with a complete list of changes. 
Check out update1, update2, update3, and update4.  
Graeme Devine also made a .plan update to celebrate the occasion.

Shots
Here are shots of the Sarge: [shot1] [shot2] [shot3] [shot4] [shot5] [shot6]
The new menu: [shot1] [shot2] [shot3] [shot4]. 

Some of the changes
Here's a list of some of the changes, the rest can be found above

* fixed marks fading properly in fog volumes
* show weapon in fov >90, adjusting position down as needed
* update scoreboard info while at intermission
* print "waiting to play" for tourney spectators
* fixed tied rank with 0 score and spectators
* return to roaming spectator when a followed client quits or spectates
* release windows cursor when running windowed and the console is down
* log all client transitions, item pickups, and kills
* changed joystick axis to act just like arrow keys so they can be bound in the controls menu for strafing. Yes, this does remove slow walking from joystick movement,but it makes everything a lot cleaner.
* fix up PantherXL trackball support
* tournement queuing of spectators to enter the game
* track wins and losses as long as you stay on a tourney server
* spectators are now in fly mode instead of noclip, and use teleporters
* pass serverTime instead of msec for command timing, prevents timescale cheating
* fixed rcon

Just a quick reminder for you guys, if you remember this post I made last week then you should already be on the lookout for the new Q3Test v1.08 which should be coming out some time tonight. I'll keep a lookout and let you guys know the minute it hits the net with a list of mirror pages for you to download it from.
update: Oh well looks like that .plan update from Carmack didnt come true. Guess uhm.. oops? Sorry about that.

The 3dfx NDA has been released and the big news? The T-Buffer technology which will be featured in that Voodoo4 card that people are just now starting to talk so much about. Also known as "Napalm" Marcus was down in California checking out the technology and has written up an article for you guys to read. I just slapped it into the Q3Test article template from earlier but I havent had a chance to change the images yet. Go ahead and read the article and I'll have the accompanying images in there within a couple of hours (It's a big download)

Basically, the T-Buffer boosts eye candy content in games. The first and foremost effect of the T-Buffer is real-time full-scene anti-aliasing. The easiest way to describe aliasing are the "jaggies" or "stair casing" that one can see on the edges of polygons. Usually these are the most apparent at lower resolutions (and/or larger displays). Increasing the resolution reduces the "artifact" somewhat, but aliasing is still noticeable even at high resolutions like 1600x1200.

You can also check out the Voodoomag, VoodooExtreme,   CombatSim, and GameSpot articles.

John Carmack has made another late night .plan update with word that a new version of Q3Test should be hitting the net some time next Monday for download. All of the changes mentioned in his previous .plan updates as well as the missing Mac release will be included in that release. Oh yeah, and can you believe they are gonna be making a Rocky VI?

Viper3D offers up a new Q3Test tweak guide for those of you still not getting the framerate you deserve when playing Quake3.

Thresh over at FiringSquad takes a look at the v107 Q3Test and gives his opinions on the latest version. You can also have a look at the article that I put out just 48hours before the big release.

The GameSpy alternative PingTool v2.5 has been released. Nifty stuff for this new release is support for those "fun names" in Q3Test, and some bug fixes for WinNT users who were getting crashes. Here is a link to the release notes. I especially enjoy it when a new PingTool is released because Jack gets his panties in a wad

The Viper3D page has an in depth look at Q3Test scores using all the different video settings on a p2-450 w/TNT2 card.

Something I figured out over the past couple of days is that Q3Test demos are in fact not broken, however the multi-binding system to quickly record demos doesnt work. You now have to record demos like this:

g_syncronousclients 1
record mycooldemo
g_syncronousclients 0
stoprecord <to stop demo>

a|citizen goes that extra mile on the Team Abuse page with Q3 SMP info. His results are very impressive. He has posted a performance monitor screenie that exhibits what goes on with the CPUs with r_smp 0 and r_smp 1 during a timedemo. With SMP disabled, you notice that one CPU skyrockets and hangs around the maximum usage, but with SMP enabled, the workload is shared. One CPU is still workin' it a little more, but you can visibly see the difference in timedemo lengths.

3DGaming.net slaps up a review of Q3Test 1.07 for those who still aren't "sure".

Graeme Devine does a quick update to his .plan file regarding the Mac version of Q3Test (the same platform that had it couple weeks before the others)

Mac Users! We're not ignoring you, we're just having a slight technical problem with the Macintosh build right now. BTW, I'm in love with that new iBook! When they ship that's going to be /the/ portable.

In addition to the 1.07 filter that I released yesterday, Deathincarnate took it a step futher and sent this set of GameSpy filters which will seperate servers for you into the teamplay, 1on1, and FFA modes making it easier to find servers that suit your mood depending on the position of the moon. Also for those of you using PingTool Blue has posted a 1.07 filter for that utl.

Since it was past bed time for a lot of you guys last night when Quake3 v1.07 came out I figured I will go ahead and put it at the top of the news today along with a compiled list of stuff that you might wanna play with. Firstly Win32 as well as Linux versions have been released. You should also have a look at Graeme Devine's .plan update talking about feedback from the players. On top of that, I put together a nifty new GameSpy v1.07 filter for you guys who just dont enjoy making your own. The guys on the Quake3 messageboard are on top of things as always, check there for the latest tips. Finally, Zoid made this posting for you guys running servers:

It appears the dedicated server won't load if "set dedicated 0" is in your q3config.cfg file in the demoq3 dir. Delete that file, when +set dedicated 1 will work again.

/// Zoid.

citizen over at Abuse sends word that he has put together some new benchmarks of a dual CPU machine using the latest SMP code in the v1.07 Q3Test release that came out last night. Here are the results he got in low res:

resolution	r_smp 0	r_smp 1
640x480	88.1 fps	110.5 fps
800x600	87.6 fps	105.3 fps
1024x768	82.4 fps	90.3 fps