Brent Kelby (een huurling van Tom) heeft het tweede deel van z'n Intel Whitney / i810 review gepost. Er zitten benchmarks bij van i810, SiS 620 en i440ZX mobo's. Fiets naar Tom's Hardware Guide voor de info.

Tom is wakker: hij zegt als eerste timedemo benchmarks en screenshots van de Q3A Win32 Test te hebben. Op dit moment heeft hij scores van een Hercules Dynamite TNT2 op 180/220 en 175/200MHz, Asus V3800 op 150/183 en een Voodoo3 - 3500:

It's pretty obvious that TNT2 is considerably faster than Voodoo3 if you look at this, but those are only the first results!!! Voodoo3 2000 and 3000 will follow as well as TNT2 as 125/150 MHz!!! ed.

Wat betreft de screenshots: ik kan niet echt veel verschil ontdekken in de TNT2 en V3 beeldkwaliteit. Kijk zelf bij THG.

Ik zag bij Storage Review dat IBM een nieuwe serie derde generatie 10.000rpm SCSI drives aangekondigd heeft. Check deze info uit de press release:

IBM today announced two new high-performance, extremely reliable 10,000 RPM hard drive models, the 36-gigabyte (GB) Ultrastar 36ZX, and the Ultrastar 18LZX, available in 18 and 9 GB capacities.

Ultrastar 18LZX has the fastest seek time (4.9 msec) of any hard drive. Seek time is how fast users can find information stored on their hard drives.

"With these new Ultrastar drives, IBM has once again mastered the art of high performance while improving reliability with several new features and technologies," said M.S. Bhat, IBM Server Hard Drive business line manager.

Als het een beetje mee gaat zitten dan zijn de nieuwe schijven volgende maand te koop (dus als je geld teveel hebt...)

Upgrade Center heeft een review gepost van het 1200 dollar dure Kryotech Cool K5-III 500 systeem:

This was the single most powerful alternative CPU systems I've ever had the priviledge to test.  My regular system is based on an AMD K6-2 400MHz chip and this was noticeably faster in almost every task.  The stability was rock solid throughout all my playing and testing.  I ran a lot of looping benchmarks for days at a time and didn't have it hang once, the stability was awesome.  Very solid in gaming as well as in office applications.

Helaas hebben ze niet de moeite genomen om te testen hoe het ding zich boven de 500Mhz gedraagd. Nu is het weer zo'n standaard Kryotech review ("performance goed, prijs way te hoog").

Ook bji Sharky Extreme hebben ze de nieuwe Celeron 466 gereviewt. Wat betreft overklokbaarheid komen ze tot minder positieve conclusies dan iXBT (daar hadden ze een 466 die moeiteloos naar de 525 ging):

hat 7x multiplier is tough to work with as it is, and another trend is developing with these new PPGA processors we've seen lately: No overclockability.

It's not due to any engineering feat on Intel's part (we've been assured by employees in the know) but for whatever reason both the PPGA Celeron 433 and PPGA Celeron 466 CPUs we've tested are completely unstable and virtually unworkable when overclocked to any degree above their stock levels.

At least with the 433 we were able to stabilize it at a speed of 488MHz (6.5 x 75MHz) which ran well during our endurance testing. But today's PPGA 466 was an absolute beast as it failed each time we tried to utilize it at any front side bus speed above its stock level of 66MHz.

Check de review bij Sharky Extreme.

HW Upgrade heeft een artikel gepost waarin de OpenGL performance van de Permedia II, TNT, Rage128 en TicketToRide4 chips vergeleken wordt. De kaarten werden gebenchmarkt met o.a. 3DMax, Lightwave en AutoCAD. Hier heb je een stuk:

TNT shows the highest fill-rate values, especially at 32bit color depth. Permedia 2 obtains best performances in Polygon rate at 16 bit, because of the superior results in the MCAD test. In the simulation TNT's high values are obtained because of the numerous textures and 3D functions used, as fogging and bilinear filtering. Ticket2Ride4 seems not to performs too well with texture filtering and failed the tests on which its use was requested. Permedia 2 is the best solution on all the applications full of wireframe visualization, such as projecting, and TNT in visual tasks.

iXBT heeft een review van de Celeron 466 gepost en zoals we inmiddels van iXBT gewend zijn is het weer een behoorlijk compleet artikel geworden. Over de Celeron valt inmiddels weinig bezijzonders meer te vertellen, behalve dan z'n overklokbaarheid. De C466 van iXBT ging zonder moeite naar de 525Mhz, maar daarboven was het niets:

However, all our attempts to its further overclocking ended up in a total failure. As we have expected Intel Celeron 466 had a locked clock multiplier equal to 7x. And when the FSB frequency was set to 83MHz the processor started but ceased working in a minute or so. And we could see the same thing with the core voltage either set to standard 2.0V or increased up to 2.1V, 2.2V and 2.3V. So, unfortunately, we have to admit that Intel Celeron 466 processor can be overclocked only to 525MHz.

There is actually another fact proving this sad conclusion: Intel is planning to switch over to a new 0.18 micron technology, when working with the frequencies exceeding 500MHz. Well, it looks as if the resources are almost exhausted. However, let's be fair: 525MHz is not that bad, actually. If overclocked up to this level this processor can easily compete with Intel Pentium II 450, which has a 100MHz FSB frequency.

Beyond3D heeft een artikel geschreven over CPU dependency vs fillrate. Er zit een leuk grafiekje bij met op de x-as de resolutie (in Mpixels) en op de y-as de (final) fillrate. Van de geteste kaarten doet de Videologic Neon250 het erg goed. Handige info => check Beyond3D.

Riva3D heeft 8 screenshots gepost van Quake3 op de TNT, met 32-bit kleurdiepte. De lichteffecten zijn erg vet.

Check Riva3D.com voor de pics.

John Carmack van ID Software heeft weer eens zijn .plan geupdate. Daarin laat hij deze keer weten dat je 100 dollar kunt winnen als je de Q3Test server crashtest.idsoftware.com kunt laten hangen of crashen.

Dit zou moeten voorkomen dat er weer hetzelfde gebeurt als toen Quake 2 uitkwam toen publieke Q2 servers aangevallen werden.

Er is overigens al één persoon die de 100 dollar in z'n zak mag steken.

Final Fantasy 8, alweer het achtste deel uit de vooral in Japan erg bekende Final Fantasy series komt volgens een press release van Squaresoft ook uit op de PC.

Hier alvast een gedeelte uit de press release:

COSTA MESA, CA, May 10, 1999 - Square Electronic Arts L.L.C. (Square Electronic Arts), the exclusive publisher of all Square Soft® products in North America, today announced that FINAL FANTASY® VIII is currently in development for the PC. The title, which will be published by Square Electronic Arts in North America, is scheduled to ship in Winter 1999. FINAL FANTASY VIII for the PC will be showcased this week during the Electronic Entertainment Expo (E3) in Los Angeles.

Het lijkt erop dat VIA een smerige truck heeft uitgehaald met de MVP3 chipset, die bij een hoop gebruikers problemen veroorzaken met de 2x AGP mode van bijvoorbeeld de TNT en de G200. Volgens een diagnose programma van Matrox bestaat de chipset voor de helft uit een oud deel, de VP3, die inderdaad instabiel is met AGP 2x. Check deze info.

I COULDN'T EVEN BELIEVE IT ?!!!! It is actually a half VP3 device and half MVP3 device. According to the datasheet of VIA Apollo MVP3 North Bridge, the VT82C598MVP contains two device. Device 0 is the Host Bridge and the device ID is 0598h. Device 1 is the PCI-to-PCI bridge and the device ID is 8598h. Meanwhile, the datasheet of VIA Apollo VP3 North Bridge, the VT82C597 / VT82C597AT, stated that the chip also contains a Host Bridge with device ID 0597h and a PCI-to-PCI Bridge with device ID 8597h. The Host Bridge contains Cache control, DRAM control, PCI Bus Control, GART/TLB control and AGP Control. The official VIA chipsets FAQ clearly stated that Apollo VP3 chipsets does not work reliably in AGP 2x mode.

Moet je deze verkoopcijfers van de Voodoo3 zien (afkomstig uit een PR van 3Dfx)!:

According to PC Data, which measured retail sales of the entire video card market for the first three weeks of April, the Voodoo3 3000 AGP was the No. 1 selling board in terms of both unit shipments and revenues. The Voodoo3 2000 PCI was ranked second, and the Voodoo3 2000 AGP was ranked third. In fact, the Voodoo3 line outsold the nearest competing product by 450 percent in terms of unit shipments, and an astounding 1,200 percent in terms of revenues. With the launch of the Vooodoo3 line, 3dfx also announced that it has shipped its 10 millionth graphics product

Ars heeft een Motherboard Buying Guide bij elkaar gefixt. Het is nogal basic spul (welke CPU moet in welke plank enzo). Naast een berg algemene info hebben ze er een paar lessen mobo-geschiedenis in verwerkt (486-ers en dat soort fossielen). Kijk maar ff bij Ars Technica voor de info.

Hier heb je een lijstje met mirrors (schaamteloos geript van Voodoo Extreme):

3DFiles
ftp://ftp.cdrom.com/pub/idgames/idstuff/quake3/win32/q3test_1_05.exe http://www.3dgamers.com/3daction/Games/quake3arena/q3test-win32.html ftp://209.1.115.11/idsoftware/q3arena/q3test_1_05.exe http://www.bluesnews.com/files/idstuff/quake3/win32/q3test_1_05.shtml http://www.planetquake.com/dl/dl.asp?quake3/q3test_1_05.exe http://download.ign.com/pc/q3test_1_05.exe ftp://ftp.stomped.com/pub/redwood/q3a/win32/q3test_1_05.exe ftp://qftp1.heat.net/pub/win32/q3test_1_05.exe ftp://ftp.golsyd.net.au/golsyd-public-ftp/quake3/win32/client/q3test_1_05.exe ftp://ftp.barrysworld.com/quake3/test/win32/q3test_1_05.exe ftp://neville2.barrysworld.com/quake3/test/win32/q3test_1_05.exe ftp://chips.barrysworld.com/quake3/test/win32/q3test_1_05.exe ftp://216.81.151.27/quake3/q3test_1_05.exe ftp://ftp.stomped.com/pub/redwood/q3a/win32/q3test_1_05.exe http://www.quake3planet.de/files/q3test/win32/q3test_1_05.exe ftp://ftp.q3arena.com/pub/q3arena/q3test/win32/q3test_1_05.exe

Tom is er na 8 maanden achter gekomen dat je een Celeron na modificatie (of bij de socket370 selderij middels een riser card met dual jumper) in SMP mode kunt draaien! Echt waar!

High end dual processing was never as cheap as now with the Celeron. You can take most CPU converter boards and modify them or go for the new Asus or MSI types in the first place. If you want to run 100 MHz right now, you should stick with the Asus S370, which enables you to increase the CPU's core-voltage, thus making it perfect for overclocking. I've not experienced any problem so far at 100 MHz FSB, but it's never wrong to be careful! Of course you should take a closer look at the motherboard's BIOS; make sure that is not older than 3-4 months and supports the Celeron CPU family, including the right micro code!!!

Applaus voor Tom. En om de pret compleet te maken heeft-ie zelfs benchmarks gepost...

Operation Sports heeft een emmer nieuwe screenshots op z'n site gedropt. Hier heb je de info:

We spent the afternoon racing around and taking more screens of NFS today. We added 32 new ones! These are some great shots of Hot Pursuit Mode at night, in the rain. We also added a bunch of menu screens to show you some of the new options. These HAVE to be seen to be believed. We were nice enough to label the new ones with today's date so you don't have to click through the old ones again. Start drooling by clicking here. This game is so awesome and it's only 65% done!

En de screenshots zijn hier.

De Quake3 Test 1.05 voor Linux en de Mac kunnen inmiddels neergehaald worden:

- Linux v.105 download sites
- Mac v1.05 download sites

...maar Win32 ...

AGN Hardware heeft een review van de TEAC 6x24 CD-Recorder gepost. Hier heb je wat info over ding:

The TEAC 6x24 is a Fast SCSI-2 (50 pin) like most other SCSI CD-R drives. It offers a 2MB buffer to help quell under-buffer write errors, a 150ms access time (which is 20ms faster that the Plextor UltraPlex 4/12 and 8/20) and a 6x burn speed. What really stands out is the 24x CD read speed which certainly puts the TEAC model above the Plextor as far as normal CD operation goes. With features like these, the TEAC 6x24 seems to sit firmly in between the Plextor 4/12 and 8/20 CD-R drives. With the same cache as the 4/12, an intermediate burn speed of 6x, and CD reads faster than the 8/20 it's easy to see why you should consider the TEAC drive an alternative to the Plextor solutions.

Ars Technica heeft een reactie gepost op de BX roundup van Anand. Welke methodes gebruikt Anand om te bepalen wat stabiel is?

However, I must admit, I am a bit perplexed by how one arrives at a stability rating for a motherboard. In my own testing sessions, I've seen systems crash when running a test suite like Winstone even though they would rarely crash otherwise, even in very intensive use. Is Winstone a good test of stability? The purportedly less-than-stellar stability of Abit motherboards hasn't bitten me in hours and hours of intensive use on my BH6 (nor on my old IT5H). Running Windows NT with a massively overclocked Celeron, I only need reboot to update drivers or to pop into Linux or something. The Blue Screen of Death hasn't visited for months on end. If this motherboard, as a piece of hardware, has a stability problem, it's news to me.

Of course, I'm sure it'd crash more often in Win98, but it that the fault of the hardware? The OS? Or, in the case of relative judgements of stability versus competing BX-based motherboards, does a board's reputation for stability or instability have to do with the black art of PnP interaction between the board and the OS? If it's this magical interaction with the PnP parts of Win9x that's causing a board to trip, where should the blame lie for stability problems? Or is that not it at all? Does the manufacture quality of these things vary widely? Or are semi-successful attempts at overclocking--something Abit in particular encourages--the source of stability woes?

Daar ben ik het dus helemaal mee eens. Met Winstone kun je sowieso geen stabiliteit testen omdat Winstone wazug is. Het ding fokt regelmatig zodanig op dat je 'm compleet overnieuw moet installeren, komt echt niet omdat m'n BX6 labiel zou zijn (Win98 wel)...

Ik kwam via BXBoards bij dit artikel van de Los Angeles Times. Het gaat over games, verslaving enzo, waar we allemaal last van hebben

"Playing is so addictive these days because you're playing online against other humans," Mulvihill said. "It's not the same as it used to be where you'd buy a game, play through the levels, and after a few hours it would be over. Online, the game never ends."

With fast Internet connections and better networking software, online gaming has grown more intense--featuring ultra-realistic 3-D graphics, real-time duels and team play. And that, in turn, has made today's games so compelling that many players have trouble breaking away--even at work.

No one knows how many game addicts jeopardize their jobs or severely erode their productivity, but playing games at work is experiencing an explosion of popularity. And as a result, companies are taking action--from written policies on usage to outright bans on games ranging from ultra-violent "Quake" to the seductive "Tomb Raider" series to the passive "Solitaire."

Op zich gezeik, maar ook wel interessant. Fiets er zelf heen.

Volgens dit bericht bij AVault zal Eidos op de E3 beurs een nieuwe Tombraider presenteren:

Speculation that Eidos Interactive will be premiering the next game in the Tomb Raider series at E3 continues to mount. It has been announced that Adrian Smith, one of Core Design’s original designers for the first game, will be at E3 discussing the series. In addition, Eidos recently signed up a new real-life model to portray the series’ heroine, Lara Croft, and are priming her for an E3 debut. Also, last March, a Core representative told the Adrenaline Vault that an announcement regarding the next title in the series would be made "around E3." Need some paint to slap the announcement on Los Angeles billboards? Then refer to various Eidos press releases, including financial statements, that mention the game, including a nod from the company’s CEO, Ian Livingstone. It appears as though Eidos is not going to let their hottest property rest, and given that Core has mentioned that a fresh engine would be developed for the next sequel, she could be getting a whole new appearance.

Q3 Test voor windoos is er bijna. Check deze .plan update van Graeme Devine:

5/10/99

IMPORTANT!

You must completely nuke your old Q3TEST install before installing the update. The config files from the old build are BAD voodoo.

You can also type "exec democycle" to watch a bunch of id people play the map.

5/10/99

Q3Test 1.05 is DONE!

New Linux/Mac versions are going up right now, and we're finishing up the installer for the Win32 version.

If all goes well, the Win32 version will follow in a few hours (few == six or so).

Die postings was om 17:54, dus de binaries zijn kunnen ieder moment komen.