Voodoo Extreme heeft een review gepost van de Asus V3800TNT2. Geen UltraTNT2 kaart (dat wordt de V3800 Deluxe), maar wel met een 3D bril:

The stereoscopic glasses are quite cool. They were engineered in house by the Asus folks and it is apparent that they have done a good job. The glasses are somewhat heavy, but they provide a more immersive gaming experience. The glasses seem to make the game “jump” out at you, making the game have more depth and therefore more realistic… You can get more into the game anyways… Not all games worked with these glassess though. With the shipping drivers you should have no problems.

Er zitten verder nog een paar benchmarks en image quality (16 vs 32-bit) vergelijkingen bij. Overklokken hebben ze ook gedaan (max 158/160).

De fusie tussen 3Dfx en STB is rond. AGN 3D heeft de press release:

Under the terms of the merger first announced Dec. 14, 1998, shareholders of STB Systems will receive 0.65 of a share of 3dfx Interactive common stock for each share of STB Systems common stock. Based on the number of shares of STB Systems common stock outstanding, 3dfx Interactive expects to issue approximately 8,270,000 shares of 3dfx Interactive common stock in connection with the acquisition of STB Systems, representing approximately 34 percent of the outstanding shares of 3dfx Interactive common stock after giving effect to such issuance.

"With the merger behind us, 3dfx will accelerate its efforts to deliver powerful new technologies for games, entertainment, and more," said Greg Ballard, president and chief executive officer of 3dfx Interactive. "We will aggressively pursue every opportunity to lead this market as the most innovative manufacturer of 3D graphics technology."

Ook Ars heeft een stukkie text getikt over de (on)betrouwbaarheid van Tom's Q3Test benchmarks. Moet je deze page checken. De timedemo benchmarks in Q3Test zouden onbetrouwbaar zijn. Brian Hook (aka Grandmaster B) van id Software zegt het zelf, maar Tom meent dat hij het beter weet dan de programmeurs zelf... (toch wel knap ):

Brian Hook said that 'timedemo' doesn't work ...

Brian Hook, or 'Grandmaster B', as he likes to call himself very modestly, has said a whole lot in the past, some of it was true, other stuff wasn't. 'Timedemo' DOES work if you do it right, and Brian is unfortunately incorrect. Sorry Brian, but even 'Grandmaster B' is not perfect ...

Frg van HardOCP heeft een stuk getypt over Tom en z'n Hardware Guide. Tom zou te snel conclusies trekken op basis van pre-production hardware en pre-production software. Ook zou hij anti-3Dfx zijn en de binding met z'n totaals totaal verloren hebben:

There are a lot of good hardware reviewers out there and I dont want you to think that they are getting beat to press on this one. I simply think a lot of those guys, and like myself, they want to be sure that what they serve up to the peeps is as close to the truth as possible. I feel as though I OWE you a solid amount of information when reporting to you. In my opinion, that is not happening at Tom's Hardware.

If you want premature results with premature drivers, Tom is your man! If I have come off as a total idiot, please let me know. I have been wrong before but I don't think I am this time. I will put the soapbox away now.....as always, this is just one man's opinion.

Check de complete zeepkist op HardOCP.com.

Voor het geval je er niet genoeg van kunt krijgen: AGN 3D heeft twee screenshots gepost van een scene op de Voodoo2 en TNT. Check 't hier.

Errug heftig nieuws bij AGN 3D: Creative gaat een Glide wraper bij z'n TNT en TNT2 kaarten leveren. Daar zal 3Dfx blij van worden (-> rechtszaak?). Check deze press release:

SINGAPORE – May 13, 1999 – Creative Technology, Ltd. (NASDAQ: CREAF), the renowned innovator of market-leading multimedia solutions for personal entertainment, today announced a beta version of Unified™, a new technology for use in conjunction with its award-winning line of Graphics Blaster RIVA TNT graphics accelerators. This new technology allows the limited number of remaining "GLIDE only" legacy games to run on Creative's Graphics Blaster RIVA TNT line of products. Subsequent releases of Unified are expected to support the entire line of Creative's upcoming graphics cards.

The Unified technology announced today consists of two parts: (1) a software layer that translates GLIDE™ calls into the corresponding commands in Microsoft's broadly supported Direct3D API and, (2) a specification of extensions that go beyond Glide's current limited capabilities to support the more advanced hardware architecture offered by Creative's Graphics Blaster RIVA TNT. Unified will permit those game developers who wish to utilize Glide to take advantage of the advanced rendering features already available in Direct3D but not currently supported by Glide.

Planned future extensions of Unified include 24/32-bit rendering, stencil buffering; 24/32-bit z-buffering, support for textures larger than 256x256, and AGP texturing. Documentation for these proposed extensions is available to the game development community for review, and can be downloaded from the Creative Zone at: http://www.soundblaster.com/unified.

"We are pleased to be able to offer our existing installed Graphics Blaster RIVA TNT customer base and game developers this new technology," said Steve Mosher, vice president, graphics business for Creative. "While research shows that only a small number of "Glide-only" games still exist today, our customers have invested in these titles and want to play them on advanced 3-D accelerators including Creative's Graphics Blaster RIVA TNT."

Owners of Creative's highly-acclaimed Graphics Blaster RIVA TNT cards interested in participating in the beta test program for Unified can download the software at http://www.creative.com/unified. Game developers who do not currently own a Graphics Blaster RIVA TNT card are encouraged to contact Creative's Developer Relations team in the United States at 1-408-428-2345 for information on our Developer Relations programs.

Deze videokaart van Guillemot is ook al sinds een week in Nederland verkrijgbaar. Als het goed is heb ik hem overmorgen ook in mijn systeem . Het is een TNT2 kaart met 16MB.

Which brings us to the Maxi Gamer Xentor. One of the first TNT2-based cards available, it's basically a plain-vanilla 125MHz TNT2 (the memory is clocked at 166MHz). It includes AGP 2x support, 16MB of RAM, and a 300MHz RAMDAC that can drive displays up to 1920x1400 in 32bit color, or 2048x1536 in 16bit color. Refresh rates vary from a maximum of 240Hz at low resolutions up to 60Hz at the highest, which is far more than most users' monitors can handle. If all of this sounds just like a regular TNT-based card, only faster, it should. That's essentially what the Maxi Gamer Xentor is.

De review is niet echt groot of uitgebreid, maar hier kan je hem vinden.

Aan TNT2 reviews geen tekort de laatste tijd. Firingsquad heeft er ook weer eentje, nu van de Hercules Dynamite TNT2 Ultra.

To make a long story short, after much trial and error (and endless reboots), it was determined by the powers that be that our card would clock no higher than 190/230. To me, this is incredible, given that no special steps were taken to cool the card. I'm willing to wager that Hercules has got one of the fastest TNT2 Ultras, if not the fastest! We were assured by Hercules that the sample given to us was not a "golden card" that was a faster card intended to impress hardware reviewers like us. If that's the case, more power to Hercules, as they have some fast RAM and faster cards on their hands.

Wow! 190/230! Dat horen we graag

Lees snel de rest!

Linux kernel versie 2.2.8 is uit, en op te halen bij Kernel.org of een van de mirrors (maar die hadden hem nog niet de laatste keer dat ik checkte).

Deze versie bevat voornamelijk bugfixes & updates bevat. Op Cutting Edge kun je lezen wat er allemaal veranderd en verbeterd is.

3dfx heeft nieuwe Quake 3 compatible drivers voor V2 en V3 kaarten gereleased.
zie hier

Red Hat Linux versie 6 schijnt een paar undocumented features te hebben, waaronder een Blue Screen of Death Simulator:

By default, the bsodsim program activates when the user hits the virtually unused SysRq key (this is customizable), causing the system to switch to a character cell console to display a ficticious Blue Screen. Red Hat hails the bsodsim program as the "boss key" for the Linux world. "Many old DOS games had a boss key, which caused the program to switch to a DOS shell or a benign looking screenshot when the boss walked by," a Red Hat engineer explained." [break]Het lijkt mij een goede grap, maar misschien klopt het wel. Ik heb het zelf nog niet kunnen testen. Bij het volgende stukje tekst ging ik twijfelen:[/break] The bsodsim program doesn't stop at just showing a simulated error message. If the boss doesn't walk away, the worker can continue the illusion by hitting CTRL-ALT-DEL, which causes a simulated reboot. After showing the usual boot messages, bsodsim will run a simulated SCANDISK program indefinitely. The boss won't be able to tell the difference. If the boss continues to hang around, the worker can say, "SCANDISK is really taking a long time... maybe we should upgrade our computers.

Hehe... Als ik (als root) op CTRL-ALT-DEL druk krijg ik ook een "simuated reboot", overeenkomstig met een shutdown -r now.

Probeer het eens, en laat het hier weten.

Hitachi had al een SCSI-2 uitvoering van een DVD-RAM drive, maar heeft er nu ook een IDE uitvoering van, de GF-1000 DVD-RAM. Hier is een stukje uit het persbericht:

For rewritable data storage, the GF-1000 drive accommodates single and double-sided DVD-RAM media with capacities of 2.6GB and 5.2 GB, respectively. Additionally, the drive can read all currently available 120-cm optical disc formats, including single and multilayer DVD-ROM, DVD-R and DVD-Video, and all prerecorded and writable CDs including CD-ROM, CD-R/RW, Photo, CD Audio and Video CD. The GF-1000 has a DVD-RAM data transfer rate of 1.38 MB/s, and a data access time of 200 ms. The ATAPI interface supports a burst transfer rate of 16.6 MB/s (PIO Mode 4 and DMA Mode 2). Access and transfer rates vary when reading other media types.

Het volledige persbericht kan je hier vinden.
De prijs komt op $750.

Ook Anand heeft z'n review over de Diamond Stealth III ook online gezet. Deze kaart bevat de toch wel veelbelovende Savage 4 chipset. Hier is een stukkie ervan:

When it comes down to Savage4 boards, the Diamond S540 will probably end up being one of the best you'll see. It boasts Diamond's quality, attention to detail, and driver support, something very few manufacturers can equate their reputations with. Unfortunately the S540 does have its drawbacks, first of all, all of the drawbacks of the Savage4 are, naturally, drawbacks of the S540 as well (see AnandTech's Savage4 Review) however the biggest problem AnandTech had with the S540 was the price. Diamond stated that the S540 would retail for $129.95, so you can expect the board to sell for as low as $100 at first, dropping significantly after the initial launch.

Goede kaart, maar te duur ? Chech het zelf maar...

Ik zag op Voodoo Extreme dat er weer eens een nieuwe versie van WinAMP is. Weinig bijzonders zo te zien:

- WMA support fixed

- Vastly better DirectSound output

- Other bugfixes

Je kunt 'm bij Nullsoft neer gaan halen (584Kb).

Planet Riva heeft 16 Quake3 screenshots met 'volumetrische shaduwen' (wattes?) gepost. Ik ga er ff bruut 2 jatten :

News.com heeft een artikel gebouwd over de het toekomstige tekort aan IP adressen. Nu zo ongeveer elke zichzelf respecterende broodrooster een eigen IP adres verlangt zullen we binnenkort een 'probleempje' tegenkomen. Hier een stuk uit het artikel:

In a way the potential shortage of IP addresses is most analogous to the shortage of phone numbers that came about with the advent of fax machines and cellular phones, which has spurred the addition of new area codes.

And the perceived scarcity of addresses is just the beginning. As more computers connect to the Net, the databases that map the numbers are growing larger and becoming unwieldy. The ever-increasing size of the network's so-called routing tables has some Net programmers worried.

"There's going to be a point when machines can't handle the size," said Kim Hubbard, president of the American Registry for Internet Numbers, which is responsible for allocating and assigning IP addresses in the Americas.

Although there is hope that a new standard, IP version 6 (IPV6), could help alleviate both problems, the timeline for a rollout is sketchy--estimates range from the next 5 to 25 years. That's why many in the Net addressing trenches agree that allocation of these precious resources must meet strict guidelines.