Ik werd door JW (?) getipt dat Graficall de Diamond MonsterSound MX300 (+Unreal) in voorraad heeft. De prijs is FL199,-.

Inmiddels zit ik nog steeds (al 3 weken ) te wachten op m'n MX300 van CompUSA ('per ongeluk' voor $49,95 geprijst). Ik heb weinig vertrouwen dat die ooit nog de polder zal bereiken...

Anand heeft wat info weten los te peuteren over de nieuwe 433 en 466MHz Celerons. Geruchten gingen dat de C433, die 18 maart gereleased zal worden, gebakken zou worden op 0,18micron. Volgens de info van Anand is dit niet waar :

I promised you all last night that I'd find out for sure about the 0.18 micron Celeron 433/466 deal. I talked to Intel today and the official word from them is that the upcoming Celeron 433/466 processors are not 0.18 micron chips and they will be based on the same 0.25 micron Mendocino core. I was originally told that they would be based on the same core, but I needed to confirm, we won't see 0.18 micron chips for a little while now...but Intel definitely has some killer plans for the future with 0.18 micron chips ;)

Klote dus. Er gingen al wilde voorspellingen rond dat de nieuwe 0,18micron Celerons wellicht 650Mhz zouden kunnen halen. Dat kunnen we dus wel vergeten...

Eric van Fastgraphics heeft een review gepost van de 16Meg ATI Xpert 128. Er zitten wat handige benchmarks bij waar de ATI het opneemt tegen de Dynamite TNT.

De kaart die Eric getest heeft maakt gebruik van een nieuwe revisie chips die het minder heet krijgen:

A few weeks ago when I reviewed the first "final version" of the Rage Fury I added some comments on the heat issue which has been discussed on other websites as well. The chip revision on the Fury board is a A21, and ATI has recently sent me an Expert 128 board, also featuring the Rage 128 GL chipset. This board has the newer A22 revision of the chipset which doesn't add anything to the performance, but which has a revised power circuit which causes it to get less hot.

De review vind je hier.

The Register heeft een nieuwsbericht met info over Intel Merced en Win64. Er zijn wat probleempjes...

The problem is that compilers produced for Merced do not perform inherent parallelism in the codelines and Intel engineers have now gone back to the drawing board.

The situation between the two companies has become so bad that Intel engineers working on site have stopped checking Win64 code.

Just before Christmas, we reported that Microsoft engineers had succeeded in booting Win64 on the Alpha platform.

Compiler performance, our sources maintain, is out by a factor of five and this is likely to lead to a delay in the 64 bit Intel chip.

Meanwhile, Merced is threatened by Compaq's Alpha EV68 platform, which will be a .18 micron part. And Merced's die size is likely to be three times bigger than this particular rev of Alpha.

Twentieth Century Fox heeft een press release lost gelaten met info over de Wing Commander movie:

Twentieth Century Fox will release the science-fiction adventure film ``Wing Commander'' in theaters nationwide on March 12, it was announced Monday by Tom Sherak, chairman of Twentieth Domestic Film Group.

``Wing Commander'' combines state-of-the-art science-fiction special effects and settings with the story and creative elements of a grand adventure film. Based on the best-selling, legendary computer- game series of the same name, the film follows the exploits of rebellious rookie space pilot Christopher Blair, played by Freddie Prinze Jr. (``She's All That'').

Carrying an encoded message about an invasion from an alien race, Blair, his fellow maverick pilot Maniac (Matthew Lillard of ``Scream'') and their beautiful superior officer, Deveraux (Saffron Burrows of ``Circle of Friends''), mobilize to evade and ultimately repel the alien attack.

Tcheky Karyo, David Suchet, David Warner, Ginny Holder and Jurgen Prochnow also star.

Chris Roberts, the creator of the ``Wing Commander'' computer games, makes his feature-film directing debut. The film is produced by Todd Moyer. The screen story and screenplay are by Kevin Droney, based on a story and characters created by Roberts.

Freak! heeft een review van de ATI Xpert 128 in elkaar gedraaid. Hier heb je wat spul over de performance van het ding in vergelijking met een TNT:

As you can see, the Rage128's driver immaturity is clearly apparent. In almost all of the benchmarks, the TNT edges the Rage128 in every category. Though the TNT scores aren't killing the Rage128, its drivers have a ways off before they can perform at the same level. 32-bit performance has been the only boon to the Rage128's 3D performance. Unlike the TNT, the Rage128 suffers much less from the increase from 16- to 32-bit.

Ook Thresh's Firingsquad heeft wat benchmarks met de nieuwe nVidia Detonator drivers bij elkaar gegooid. Net zoals bij de benchmarks van AGN Hardware laat vooral 3DMark99 een groot verschil zien. Firingsuquad heeft de nieuwe drivers ook op een PIII getest, maar die toont bar weinig verschil met de performance op een PII:

n 3DMark as well, there is no perceivable performance difference from moving to the new drivers on a P3 vs. a P2. With KNI optimizations, you would expect the P3 to experience a greater percentage gain, roughly 15-25% higher than that of an equivalent P2. However, seeing that this is not the case, two possibilities present themselves. One, the new drivers do not take advantage of KNI, or two, 3DMark's execution time is not holed up in the display driver, reducing the impact of such optimizations. There's also a question of support for KNI in the operating system, which was supposed to be implemented in DirectX6.1. Again, if anyone finds significant performance gains on a P3 as opposed to a P2, please let us know.

Gamer's Platfrom heeft een interview met AMD gepost. Het ding heeft ook wat goeds over over de multiprocessor mogelijkheden van de K7:

Q: Will K7 run multiple processors more efficiently than Intel's line of CPUs?

The AMD-K7 processor uses the Alpha EV6 bus architecture , which is designed to support high-performance, scalable multiprocessing. The K7 will leverage the Alpha server infrastructure that exists in the marketplace. The number of processors is not limited by the AMD-K7 processor design but rather by chipset implementation. System designs based on 2, 4, 8 or more AMD-K7 processors are possible.

The K7 was designed to operate in both single and multi-processor environments. Specific performance benchmarks for the processor, in either a single or multi-processor implementation, will not be made public until the official launch of the K7 during the second quarter of 1999. But, the EV6 bus is designed to provide greater microprocessor bandwidth than a P6 bus. You can take a look at the K7 presentation from the 1998 Microprocessor Forum for additional info on K7.

Voodoo Extreme heeft wat info over de Quantum3D "The Brick", een plank met 8 Voodoo 2's! Klinkt lekker:

"The Brick" you mentioned in your voodoo extreme article is actually 4 Single planar Obsidian2 200SBi boards with a hardware 4:1 pixel averaging subsystem linking them together. This feeds into an I/O card for 1024x768 full scene hardware antialiased output at 60Hz. This is housed in our 6U rack mounted Heavy Metal Mercury Graphics Workstation. Total, that's 8 voodoo2 chipsets running in parallel, each with 16MB of texture memory and 8MB of frame buffer. It looks as good as SGI's Onyx Infinite reality....but cheaper.

AGNHardware heeft wat zitten benchmarken met de nieuwe nVidia Detonator 1.09 TNT drivers. nVidia belooft een performance boost van 30%, maar dat halen de drivers overduidelijk niet. Hier wat geblah:

Well, were not quite talking 30%, but the new drivers do offer an improvement.   From the scores I have seen thus far, the Detonator drivers are good for between 8 and 13% FPS increase.  This isn't too shabby for a free driver upgrade.  The down side is that even 13% is only 4 FPS or so at 1024x768.  It may not seem like much, but we will take any improvements we can get.

De AGN HW benchmarks vind je hier.

Legions.com heeft een editorial genaamd 'Where Intel is Going Today'. Hier heb je wat info uit de intro:

It looks as though the empire is crumbling. High prices, price wars, and significant competition are taking their toll, and if things continue at this rate, the empire won't rule for much longer. Who am I talking about? No, not Microsoft, but the other half of Wintel: Intel.

Over the years, Intel has become so accustomed to telling the computer industry what they want that they seem to have lost the ability to listen. Now, for the first time in recent memory, Intel is facing some significant competition on all fronts. But instead of fighting back, they are carrying on as they always have - putting out products that are decent, although not inspired, and assuming that people will buy them because the box says "Intel Inside." Ignoring the competition has worked so far, but will it work now? Or is this the beginning of Intel's downward slide?

Gamedesign.net heeft wat zitten chatten met Bobby Pavlock, Daikatana level designer bij ION Storm. Dit gebabbel heeft geresulteerd in een interview, waarvan je hier een plak krijgt:

What would you consider the most difficult aspect of level design?

I would say the most difficult aspect of level design, for me, is the planning stages - trying to come up with something fresh and new, that has never been done before. There are SO many levels, both professional and user-made nowadays, that almost everything has been thought of by now, that is, that can be done with the technology available at this point.

Hardware Central heeft een review van de Hercules Terminator Beast Supercharged in elkaar gefabriekt. Hier heb je wat stuff over het verschil tussen de normale Savage3D en de supergecharge chippies:

The biggest change with the SuperCharged Beast is that it is overclocked to a 120Mhz core and memory speed. Because the clock and memory speeds are entangled in the Savage3D architecture, Hercules had to go with SGRAM to ensure the memory would be able to handle the higher clock speed. So how did Hercules get the speeds up to 120Mhz? To ensure overclockability, Hercules screened their stock for Savage3D chips that could handle the overclocked speed and used these chips in the SuperCharged Beasts. In addition, Hercules also decided to make use of the TV-out feature of the Savage3D and enabled it in the Beast SuperCharged. After testing the original Beast and finding it a poor performer in comparison with the TNT, we were quite surprised at the performance difference the overclocking produced. Despite the low cost of the new SuperCharged Beast, Hercules' newest "wonder kid" incorporates some great features and puts up some impressive numbers.

Ars Technica heeft de Gainward CARDExpert TNT onder handen genomen. Waarempel zit er ook nog es een TNT op het ding. De 3D performance van de TNT moge inmiddels bekend zijn. Hier daarom maar ff wat info over de beeldkwaliteit:

Many flames have been exchanged online about video signal quality on nVidia cards versus cards from traditional 2D leaders like Matrox, and perhaps justifiably so.  Complaints about signal quality generally come from those using really high resolutions, like 1280x1024 and 1600x1200.  I run my Riva 128 card at 1152x864 at 100Hz refresh on a 17" aperture grill-style display, and even using a five-pronged BNC input, the Riva's video output produces perceptible ghosting.  The Gainward TNT, however, is significantly better--not perfect, but close enough.  On virtually all of the latest-generation video chipsets, the RAMDAC, which produces the video signal, is integrated as part of a single-chip design.  Hence, video signal quality will vary less from one manufacturer's implementation to another than in the past.  Still, the Gainward implementation appears solid.

De review vind je hier.

TotalPC heeft een artikel over Cooling in elkaar gedraaid. Hier heb je de intro:

Having had time lately to check out some of the articles on other computer sites, I’ve noticed overclocking articles popping up right and left.  “How to get the most bang for your buck”, “Overclocking: Is it right for you” and variations of the two seem to be the only news that us “computer journalists” have been reporting about, which is understandable; who doesn’t get a little thrill every time they “beat the system” with overclocking?  I mean, surely, who in their right mind would go about paying the full amount for a chip that’s most likely going to be obsolete within the year?  Anyway, the time is drawing near for me to upgrade my CPU, and seeing as how my meager salary probably won’t foot the cost of any of the newer Intel models coming out, I decided to go with the Celeron 300A; cheap, common, and reliably overclockable to 450 mhz.  Of course I would need a fan; everyone knows that the cooler you get an overclocked system, the better.  It’s practically a mantra repeated by every site out there… hell, even one site claims that by brining their computer out into snow that they managed to attain a whopping 600 mhz with a 400 mhz Celeron chip!  Yet as I was investigating the potential cooling of my new system, I realized that I had never heard anyone put down a solid, concrete reason as to why colder CPUs can run faster.  The sheer concept that a colder chip can run faster than a warmer one is almost laughably ludicrous if you think about it! (Not too hard…I’m trying to make a point here).  So I did my research; checking out every heat-releated website; every electricity-related website; every overclocking website I could feast my hungry little eyes upon, and what follows, is the compiled reasons of why your computer runs faster, colder.

Andy van BXBoards heeft een Socket370 motherboard roundup in elkaar geplant. Vergeleken worden de Soyo SY6ILA, Epox EP-LX370B, Soyo SY-6IZA, Freetech P6F100 en Abit BM6 (Jippie ).

Hier een stukje over de BM6:

The Abit BM6 is a full featured board, and frankly I'd quite like a Slot 1 version! The BX chipset, and 5/2 PCI / ISA combination means that future proofing is assured. Nevertheless I would like to see less bus speeds above 100Mhz and more speeds below 100Mhz, such as 95Mhz and 88Mhz for example. The voltage tweaks show that this design offers potential, and the BM6 was the only board that would even POST at 550. Given that the Celeron 366 will do 458/83 on all boards tested without the need for voltage tweaking, some more flexibility in bus speed selection at the lower ranges could squeeze a little extra juice out of this CPU.

Alex van Kaam heeft een nieuwe versie van Motherboard Monitor op z'n site geparkeerd. Hier wat info over nieuwe features:

- Support for the new Winbond chips W83782D and W83783S

- As some users requested you can now place an alarm if your resources drop below a certain level

- Option to disable splash screen (yup I finally gave in)

- Some new Chip ID routines

- Some new Voltage routines

Het handige moboproggie heeft nu ook support voor de BX6-II (die van de nieuwe Winbond chip gebruik maakt).