Hardwarelab.de heeft een aantal Quake2, Quake3, 3DMark 2000, Povray en Cinema 3D benchmarks gepost die de performance verschillen tussen een Athlon met 1/2 en 1/3 L2 cache divider laten zien. In de game benchmarks blijken de 700 (1/2) en 750 (1/3) vrijwel gelijk te presteren, terwijl de L2 cache snelheid een minder grote invloed blijkt te hebben in de rendering benchmarks:

For the used games and graphic card the L2 cache speed is not so important. I'm sure this will change when using faster graphic cards (GeForce DDR is comming!). In 3DMark2000 the Athlon 700 1/2 acts like a Athlon 750 1/3, and in some benchmarks it is faster. In renderbenchmarks the L2 cache speed is important. But also higher clockrates mean shorter rendertimes - the 750 1/3 is faster than the 700 1/2.

In short: For the gameplayer, with 'normal' graphic cards there is no need for more than 700MHz when playing todays games. For rendering, go and get the fastest CPU on the market, time is money. If you are using you computer mostly for 'normal' things like surfing the internet or Office stuff, you dont need more than 500MHz, because thats real overkill.

Forbes bericht over geruchten dat S3 mogelijk z'n graphics devisie laat overnemen door VIA Technologies. S3 zou zich meer willen toeleggen op de winstgevende communicatie en entertainment divisies:

Sources familiar with S3 said the company is in talks with another chipmaker. One of the names mentioned is VIA Technologies, the Taiwanese chipset maker that bought microprocessor makers Cyrix and Centaur earlier this year. Silicon Valley sources said that another Asian board and peripheral maker might be looking closely at S3's graphics assets. Both VIA and S3 declined to comment.

Rumors suggest that S3 might hive off its graphics chip business and focus on high growth markets such as digital communications and digital entertainment--two high-growth markets.

In S3's digital future, graphic chips are likely to become less important and it makes good sense for the company to divest these low-yielding assets. VIA could be an ideal buyer for S3's graphics assets. S3 and VIA announced a joint venture in November, tentatively dubbed S3-VIA. Together, they aim to develop highly integrated chips, which will combine graphics and core logic chipset functions and target the high-volume budget PC and notebook PC market.

VIA, at the time of announcement, had a $14 million investment in S3 and was going to invest another $11 million in the graphics chipmaker. S3 was also going to invest $25 million in VIA by December-end. A $25 million investment by VIA represents nearly 3.5% of S3's total equity.

Op mcse.nl kwam ik het volgende bericht tegen (tja, 't is mischien wat oud, maar wel interresant) over windows:

Microsoft geeft toe dat de volgende versie van zijn besturingssysteem, codenaam Millennium, wellicht toch niet de laatste van de 9x-lijn zal zijn. Het bedrijf werkt al sinds de zomer van 1997 aan een consumentenplatform op basis van de Windows NT-kern.

Het hele bericht staat hier, dat is een besloten site, waar je je voor moet aanmelden.

MarcelJ schrijft: "Dinsdag 21 december 1999 en dit is de eerste update van CloneCD van deze week.":

• Version 1.0.0.9 introduced "recalculation of sub channel data during read. This improves readability of the created disk" This was (and still is) a good thing. However, there is a major problem: A lot of CDs made for Game Consoles *require* invalid checksums in the SubChannel Data! After a lot of investigation, the following solution is introduced in this version:
• Recalculation of sub channel data during read removed.
• Recalculation of sub channel data during write added.
• Added option to disable recalculation of sub channel data during write. If you have titles, that worked, when copied with CloneCD 1.0.0.8 or earlier, but failed with later versions, you should check the "Don't repair SubChannel Data Box". Please note, that a lot of CD-Writers do not allow illegal SubChannel Data to be written. If you have such a Writer, the box is disabled. Only CD-Writers that support writing of "CD+G" allow this.
• Increased SubChannel buffer.
• Added MIN:SEC:FRA values to the "Shall I overburn?" Dialog.
• Enhanced reading with Toshiba SD-M1212 and SD-M1302.
• Changed Write Setup. Incompatible writers are now detected during scanning of the SCSI/IDE bus.
• Enhanced the communication with the ASPI Layer. This fixes the "SenseKey 0" problem during write.

Download CloneCD 1.0.2.2 hier.

The man himself Alex "Sharky" Ross heeft een review op zijn site geknikkerd van AMD's Athlon 800MHz. Net als de vorige reviews van gisteren komt Sharky tot de conclusie dat de snelheidswinst relatief klein is vanwege de tragere off-die L2 cache, zoals ook al het geval was bij de 750MHz versie. Hieronder wat uit zijn review:

According to their older roadmaps, AMD had originally planned their transition from .25-micron to .18-micron to be a little different. The off-die L2 cache was due to be scrapped in favor of full-speed on-die cache for the 750MHz. But the 800, just like the 750 before it, still sports a similar set up of the off-die L2 cache, which has to be slowed down in order for the Athlon to be able to cope at higher clock speeds.

How much longer will AMD continue to release Athlon's with off-die L2 cache is a good question and the answer lies within the architecture's limitations. For maximum bandwidth, a good solution is to implement two full-speed parallel caches (the L1 and L2) for the CPU to access in tandem. AMD's Comdex presentation clearly shows that their intentions are along these lines via a 1 to 2MB on-die L2 full-speed cache (16-way associative L2 cache) implementation. Once AMD achieves this, Intel's full speed cache advantage (not to mention Intel's 256-bit datapath with a six-cycle latency) will be put the test.

If AMD's current plans time-out well, you'll likely see .18-micron Athlons with on-die cache, copper interconnects , with VIA KX-133 and PC-133 harboring AGP 4X support (won't it be great to put those GeForce's to the test?) by Q2 of 2000. FAB 30 in Dresden and its much talked-about copper interconnect wafer fabrication facility holds the key to that. AMD formed a strategic alliance with Motorola's Semiconductor Products Sector and licensed its current copper interconnect technology and High Performance Logic Process (HiPerMos) for this purpose. If we were to wager money on it, we'd expect to see Athlon's approaching the 1GHz mark by Q3 via their soon to be released K75 'Thunderbird' core (due out in Q2 2000).

Het is dus nog eventjes wachten voordat we kunnen spelen met Athlons met on-die L2 full-speed cache . Bedankt Remko voor de link.

DOS stuurde ons het volgende bericht van ZDNet.nl over een nieuw macro virus dat effect heeft op windows machine's, een stukje:

Het virus heeft verschillende effecten, met als ergste het formatteren van de harde schijf. Bij initiële infectie verzendt Pri zichzelf naar vijftig adressen uit het Outlook-adresboek. Maar daar eindigt het niet. Doel van het virus is het aantasten van 'templates' of sjablonen. Daarna zal het alle documenten gecreëerd met een sjabloon aanvallen, en als de computerklok een tijd aangeeft waarbij de twee getallen van de minuten gelijk zijn (bijvoorbeeld 9:11of 18:33), worden geopende documenten gevuld met tien geometrische figuren in toevallig gekozen kleuren.

Het hele bericht vind je hier.

HardOCP wist een recente Intel Roadmap te machtigen. Onderstaande screenshots laten o.a. zien wat Intel van plan is met de FC-PGA PIII's (de pikken tegen april '00 95% van de produktiecapaciteit). Verder staat in het tweede kwartaal van 2000 een 850 en 866MHz model gepland, gevolgd door de PIII-933 in Q3 en de 1 GHz Willamette in Q4.

SLaSH! melde ons het volgende berichtje:

BODEGRAVEN (ANP) - Onbekenden hebben maandagavond op een bedrijventerrein in Bodegraven een vrachtauto met een grote partij laptop-computers buitgemaakt. Daarbij gijzelden de drie mannen de 20-jarige chauffeur uit Nieuwerbrug. Later op de avond ketenden ze hem vast aan een hek in Amsterdam Osdorp

Iemand een laptop kopen? :-)

De OpperKoe is vannacht tijdens de dagelijkse stats update weer eens krank gegaan. David McNett (aka Nugget) schrijft in z'n .plan update:

During tonight's stats run things got extremely unhealthy during the creation of the rc5-64 overall email rankings table. Sybase consumed all it's locks and several processes became deadlocked and refused to die gracefully. (Heck, they wouldn't even die screaming).

We spent a few hours poking and prodding, trying to convince the blocking process to go away to no avail. Unfortunately, the only recourse at that point was to down the whole server in a rather indelicate manner. (not quite a kill-9, but pretty close).

Right now we're sifting through the data now that the server is back online. We're trying to determine the scope of the damage (if any) and get a feel for the situation.

More info as the situation progresses...

In de tussentijd kun je natuurlijk mooi DPC gaan joinen, voor het geval je dat nog niet gedaan hebt (niet-DPC leden zijn geen goeie tweakers en mogen - euh moeten - gediscrimineerd worden ).

AMD heeft nieuwe IDE busmaster drivers voor de AMD 756 / Irongate chipset gereleased:

***** Version 1.24 *****

Updated Windows NT installation to accommodate a greater selection of languages

Fixed Windows NT issue of VideoCD not playing in CD-ROMs running in DMA mode

Fixed Windows 95 issue of Slave CD-ROMs not playing Audio CDs - Not WHQL certified

Versie 1.24 kun je hier downloaden.

Leuk dingetje bij Ace's Hardware...

Today, Ace's reveals one of the most hidden mysteries of the Athlon. Yes, This picture is REAL! AMD Engineers have placed a small emblem on the 6th layer metal mask of each .18 µ Athlon: A Gun and the state of texas. We were told that you will find this on each .18µ Athlon (K75 die) coming from Austin.

Notice the direction of the gun: Albuquerque, Phoenix, Oregon... INTEL! . Heh, didn't know that the "CPU war" was getting so real...

It was anonymously (To that unknown person: thank you very much!!) dropped off in my mailbox with the the word "K75" on it.

De demo van Dagoth Moor Zoological Gardens is uit en kan neergehaalt worden bij 3DFiles. DMZG is een door WXP (the Whole Experience) ontwikkelde benchmark tool met o.a. T&L en Cube Environment Bump Mapping support. Ook prettig is dat een GeForce 256 op prijs wordt gesteld (toevallig heb ik er daar nog twee van liggen ):

Dagoth Moor Zoological Gardens is a technology demo and benchmarking tool developed by WXP to show off the latest hardware accelerated features of NVIDIA's GeForce 256 accelerator cards (hardware transform and lighting, cubic environment maps, etc.). It is NOT a game, or a demo of a game.

DMZG requires DirectX 7 and a GeForce 256-based accelerator card. It will probably run on other video cards, but you may experience visual glitches (especially on Voodoo3 cards). WXP has only approved DMZG for use on GeForce-based accelerator cards.

DMZG is a brief glimpse into the world you will see when "Experience", our first person strategic action shooter, ships late next year. We hope you enjoy what you see and hope this raises the bar of what gameplayers expect from the next generation of PC games.

Happy exploring!

Some of the new additions to the engine are:

• GPU accelerated geometry transformation
• GPU accelerated vertex lighting
• GPU accelerated cubic environment maps
• Fading detail maps

Each of these GPU features offloads work from the CPU to the GPU, which frees up the CPU for better physics, AI, and game code. By implementing these optimizations WXP has been able to push the DMZG poly counts from 10,000 polys per frame to over 80,000 on the same CPU. The level of detail this affords is astonishing.

Zie 3D Files voor de download locaties.

JanKees heeft z'n review van de Viper II Z200 online gezet. Hier een prikkie uit de conclusie:

I am impressed by the Quake3 benchmarks, the OpenGL ICD has been optimized for this. OpenGL in general was OK, but it is possible that other OpenGL games are not as fast as Quake3. Quake2 was not extremely fast at lower resolutions, but my CPU might be a bottleneck here. 8 bit stencil buffers were not available in Q3.

D3D performance and Quality leaves much to be desired, but the newest driver showed some progress. The fillrate and rendering speed were much too low with the latest driver.

2D in general was good, although the speed was maybe not as good as it should be compared to the Competition. I can't judge that. Other reviewers reported that fonts can loose their "tail", but I did not notice this. Probably because I don't have Internet Explorer 5 installed ? The Savage4 has the same problem for several months, it seems hard to fix.

Red Hat heeft bekend gemaakt dat het in het derde kwartaal van dit jaar, eindigend in november, een verliest heeft geboek van 3,6 miljoen dollar:

The company, which is the world's leading distributor of Linux, an alternative operating system to Microsoft's Windows, posted a net loss of $3.6 million, or 5 cents per share for the third quarter ended in November, down from a profit of $120,000 in the year-earlier fiscal third quarter.

The latest quarter's results were a penny worse than the 4 cents per share expected by the three securities analysts surveyed by First Call/Thomson Financial, which compiles brokerage estimates.

However, revenue for the quarter was up 24 percent to $5.4 million from $3.3 million a year ago.

Aandelenverkopers lijken zich, gezien het onveranderde koersverloop van de Red Hat aandelen, weinig aan te trekken van het nieuws:

Via HardOCP kwam ik op deze wazige Finse site, met plaatjes van een homemade watercooling systeem:

Ik kwam via HardOCP bij deze pagina van No-skill.net, waar men een leuke therapie voorstelt voor mensen die getraumatiseerd zijn geraakt door hun baggertrage dialup verbinding:

Het principe achter deze Destructus Automobilus behandelmethode zagen we eerder gedemonstreerd worden op een iMac keyboard .

Overclock.com heeft een review van het AOC Iceberg peltier heatsinkje neer gelegd. Conclusie: het ding suckt, prestaties zijn slechter als die van een normale heatsink/fan combo:

Stopping by AOC's booth at comdex, this unit caught my eye. They were selling them right at the show, and I had to have one. I took my trusty p3-450 chip(sl35d) with hopes of hitting 600mhz. The highest I could get with regular cooling was 558mhz. Incidentally, my regular cooler is an AOC dual fan job. Installation was easy, it just snapped on after applying some heatsink compound. I plugged it into my abit bh6 board and set the bus to 133mhz. It ran for a few minutes, and then crashed. This is exactly what happened when I tried it with the regular cooler. At this point, I dropped the speed back down to 124mhz, and all ran fine. Maybe my chip just won't run 600mhz? So now the iceberg is running the chip at the same speed as the regular cooler... I did notice that this unit puts out a lot of heat. I would not recommend using it in a case without additional cooling.

Maarten (t.net / Fokzine webserver admin) stuurde me richting deze aankondiging van FreeBSD.org, waarin bekend wordt gemaakt dat FreeBSD 3.4 is gereleased. FreeBSD is een erg populair webserver OS en wordt o.a. op onze eigen server gebruikt:

Just in under the wire for the current millenium, I'm happy to announce the availability of FreeBSD 3.4-RELEASE, the very latest in 3.x-STABLE branch technology. Following the release of FreeBSD 3.3 in Sept, 1999, a lot of new features have been added, many bugs were fixed and even more important security issues were dealt with. Please see the release notes for more information.

FreeBSD 3.4-RELEASE is available at ftp.FreeBSD.org and various FTP mirror sites throughout the world. It can also be ordered on CD from The FreeBSD Mall, from where it will be shipping soon on a 4 CD set containing installation bits for x86 architecture, as well a lot of other material of general interest to programmers and end-users alike (3.4-RELEASE for the Alpha architecture is available from the FTP site and will also be available on CDROM several weeks after the x86 product is released).

Voor meer info over download mirrors kun je hier terecht.

Be Inc. heeft aangekondigd dat het een licentie overeenkomst heeft afgesloten met Compaq, dat nu de mogelijkheid heeft om Be's Stinger OS op z'n internet appliances te installeren. Klinkt positief voor Be, hoewel er geen definitieve voorwaarden afgesproken lijken te zijn:

Be Incorporated (Nasdaq: BEOS) today announced that it has entered into a licensing agreement with Compaq Computer Corporation (NYSE: CPQ) that allows Compaq to pre-install and distribute Stinger, Be's software platform for Internet appliances, on Compaq's Internet appliance computing devices. The parties are currently working together on joint development and sales initiatives in connection with this license agreement. The agreement is not exclusive for either party and it does not contain any minimum purchase commitments on the part of Compaq.

Stinger is Be's software solution designed for the creation of appliances that deliver information and entertainment over the web. Based on BeOS, Stinger is fully customizable, offers a complete browser and supports popular streaming audio and video standards.

In navolging van GamePC deed ook CT News onderzoek naar de prestaties van de 8/4/32 EIDE burner. De resultaten zijn wederom positief:

Burning 12 600-650 meg CDs (Using Adaptec Easy CD Creator): average deviation from theoretical burn speed - 2% slower

Not bad at all! This actually lived up to its advertised speed. Well Plextor has certainly done it again. They've made a fast drive that is also reliable (in the test I've copied 12 CDs, and none of them bad copy). The 4 MB buffer is extremely important to prevent cds from turning into coasters.

De Ausies van Overclockers Australia hebben een fix in elkaar geknutseld voor FIC SD11 planken die last hebben van spontane reboots:

A few people have complained of the instability of the FIC SD-11 Athlon motherboard. We were contacted by John Gatt, Sales and Marketing Manager (Australia) for FIC. He sent along word that there is a known issue with the board which can be easily remedied if you take it back to your dealer. If you have any trouble getting your dealer to sort it out, print this page and take it with you, or contact John directly. Here's the deal:

Daar moet dus ergens een condensator tussen gefrut worden. Meer info vind je op deze page.

Plextor maakt weer eens duidelijk dat het king-of-the-hill is als het op CD-Rewriters aan komt, want volgens GamePC is de 8-speed Plexwriter 8/4/32 de beste fikker die je voor een non-SCSI systeem in huis kunt halen:

So basically this drive is the best CD-R alternative you can get to buying a SCSI-based drive. It's just as fast as the SCSI drives we tested before, but the speed differences aren't enough to see a noticeable difference. If you need super low CPU utilization, SCSI is the obvious choice, but if you can stand to wait ten minutes while you're burning a CD, this looks like the best CD-R you can buy. It's not too expensive, easy to setup, and just plain fast. Besides the high CPU utilization, which is relevant with all EIDE CD-R's, we couldn't find any problems with this drive, which is a tough thing to come upon. Once again, another incredible drive from Plextor, who would've thought?

Uit de benchmarks wordt maar weer eens duidelijk hoe extreem het verschil in CPU utilization tussen een IDE en SCSI brander is: de 8/4/32 EIDE pikt maximaal 81%, terwijl een 8/2/20 SCSI niet boven de 7% komt. Met m'n eigen Teac 58S SCSI merk ik eigenlijk helemaal niets van enige CPU belasting .

Extreme3D heeft een korte review van de Benwin BW2000 flat panels online gezet:

The speakers themselves are very impressive. They seem to deliver more chimes and sounds that you normally wouldn't hear with a regular pair of speakers. To even enhance this more, Benwin included a 3D Surround Sound button located on the subwoofer. All my mp3's and gaming sounded much better with this. Sometimes, however, the 3D surround sound can mess with a few sum of mp3's. It makes it sound all muddled or out of tone. This happens with only some of the mp3's and most of which are low quality or not popular.

Het volk van Digital Web 3D kreeg een interessant mailtje van Matrox, over AGP4x support in de G400. Het blijkt dat de G400 in principe al vanaf het begin met AGP 4x overweg kon, maar dat support van de AGP 4x mode pas sinds de nieuwere '4A' revisies gegarandeerd wordt:

When we launched G400 in May 1999, 4x motherboards did not exist. Although our part was able to work in 4x transfer mode we felt that before the Camino chip set was launched, it would be unfair to our customers to guarantee 4x operation since there are many factors that influence operation speed that are out of our control. Matrox did however ensure 4x compatibility which means that all Matrox G400 graphics boards will work seamlessly in 2x or 4x systems. Since Camino has launched, we have been able to extensively test our boards with AGP 4x. We've identified the G400 boards that support AGP 4x transfer rates with an additional '4A' in the part number: G4+MDH4A32GB/40.

In plaats van een simpele review heeft Got Apex de zaken heftiger aangepakt. Een onschuldige Annihilator Pro werd van een grotere heatsink en extra geheugen cooling voorzien, wat resulteerde in een maximale core/mem clock van 140/350MHz en een 12% hogere 3DMark 2000 score:

Well now comes the question; "To get an Annihilator Pro or not to get one". Put it this way if you do not need any added toys such as are found in the Hercules Prophet DDR-DVI or ASUS V6800 Deluxe, Video-out, Tv-in etc.. Be very good this year and maybe Santa will surprise you. As to whether you should overclock this card or not. I say … OF COURSE!!!!! But only if you get a new fan/sink as the Geforce cards put out a lot of heat. Just adding the Geforce to my rig upped my CPU and Ambient temp 2 Degrees Celsius. I'd say an extra 10-12 fps in Q3 is worth adding another fan. Now2 that I have this new graphic tyrant you may be wondering what will become of Hercules TNT2 Ultra….

Lees de review voor info & bechmarks.

Gezien de fanatieke toepassing van S3 chips in de videokaarten van Number Nine was het al duidelijk dat die twee een redelijke close verhouding hadden. Nu wordt de tent dus helemaal overgekocht door S3. Number Nine maakte ooit een aantal heftige high-end kaarten, maar is met de opkomst van de 3D accelerators vrij snel afgestorven tot een anonieme OEM supplier:

In a move to strengthen the competitive position of its Multimedia Division, S3® Incorporated (Nasdaq:SIII - news) announced today that it has begun the process to acquire substantially all of the assets of Number Nine Visual Technology Corporation.

The acquisition of Number Nine, a contract supplier of S3-based products to IBM, will allow S3 to consolidate its graphics business with IBM into a single source distribution model, while also adding highly skilled hardware and software engineering resources to S3's existing teams.

``Over the past year, we have been working with IBM to build a long-term business relationship based on a supplier agreement with Number Nine,'' said Ken Potashner, chief executive officer for S3 Incorporated. ``We expect that our acquisition of Number Nine's assets will enable us to streamline this business model and further strengthen our direct relationship with IBM.''

De patsers bij Hardware Central hebben een review van het Abit WB6 slot-1 plankje op hun site geplakt. De WB6 is gebaseerd op de Intel i810E, een chipset waarmee de hardcore tweaker elke vorm van contact wenst te vermijden. Leuk bord voor el cheapo typemachines?

While the graphics subsystem won’t be suitable for newer, more demanding games, it’s definitely sufficient to get your work done.

Arguably the best trait of the WB6, and integrated boards in general, is its plug-and-go nature. Simply plug in a processor, some RAM and a hard disk, and you have a fully-functional system for a great price. However, consider that using the sound and AMR capabilities are both fairly CPU-intensive tasks with this system, so you may wish to invest in a more powerful processor than the Celery 300 used here (consider at least 400+MHz), to prevent slowdowns from becoming an issue.

All in all, an excellent little board that should prove popular among OEMs. Its features, ease of use and low price tag will make it a popular choice in budget systems.

The Tech Zone heeft een review geschreven van de Abit BE6-II. Deze opvolger van de BE6 wordt geleverd met het verbeterde Softmenu II, die een enorme sloot van 121 busspeed settings heeft:

The number of bus speeds to chose from just blows any other overclocker's motherboard away. You can chose from 66, 68, 75, 83-200Mhz in 1Mhz steps. The only thing that would make it better would be if you can go from 66 to 200Mhz in 1Mhz steps. Bus speeds below 83Mhz uses a 1/2 PCI divider. Above that there is a choice between 1/3 or 1/4 PCI dividers. AGP divider can be either 1 to 1 or 2/3.

The Abit BE6-II is every bit as stable as the old BE6. With the Celeron 366 I was even able to squeeze out an extra 5Mhz over what the BE6 could do. My Celeron 366 can run at 605Mhz on the old BE6 using the 110Mhz bus but would never run stable on the 112Mhz. The BE6-II, with its ability to go up 1Mhz at a time was able to run the little Celeron at the 111Mhz bus, giving me a stable 610Mhz system.

I didn't get much of a chance to play with the PIII-500 on the BE6-II but 635Mhz (127 x 5) is no problem. I think it can go higher but I ran out of time. Still 635Mhz is 15Mhz better than what the old BE6 can do. You just got to love this 1Mhz stepping! There is no better way to find the limit of your CPU. Abit's hold as "king of overclocker's board" seems very safe for now.

In een wanhopige poging om de aandacht van ons tweakers te verkrijgen werpt Soyo in ene keer maar liefst 10 nieuwe plankjes op de markt. Interessant zijn vooral de SY-K7AIA (slot-A met AMD Irongate chipset), SY-6ICA (slot-1, Camino met 2xRIMM) en SY-6VCA (VIA Apollo Pro 133). Hier een paar happen uit de press release:

Dec. 20, 1999-- Soyo, Inc. rings in the new year with 10 new motherboards and a new IEEE-1394 Firewire adapter to jump start year 2000 sales to system integrators and endusers building Camino, Athlon, consumer PCs as well as Internet appliances. All new Soyo motherboards support next generation CPUs, chipsets, and peripherals with features such as 100/133 system bus support, 4X AGP support, PCI expansion slots, Universal Serial Bus (USB) ports, and UltraDMA/33 support and integrated multimedia, and integrated networking and home PNA controllers on some models. Most models also include AMR slots for installing low-cost Audio/Modem Riser cards. [break] SY-6ICA / SY-7ICAD (slot-1, Intel i820, 2xRIMM/2xDIMM): [/break] Soyo will offer two new Camino motherboards for Pentium II/III systems, which feature the new Intel 820 chipset. The SY-6ICA ($139 U.S.) and the SY-7ICAD ($139 U.S.) both include Intel's Camino chipset, Pentium II/III CPU support, 100/133 MHz FSB, 4X AGP graphics, and AC 97 Codec Software Audio Solution with an ATA 66/Audio Modem Riser Slot in an ATX form factor. Both boards are expandable to 512MBs RAM. The Soyo SY-6ICA has two RamBus RIMMs slots, while the SY-7ICAD features two DIMMs. [break] SY-K7AIA (slot-A, AMD Irongate, 5xPCI, 2xISA): [/break] Power users also have of Soyo's SY-K7AIA, which is designed to maximize AMD's Athlon chip. The SY-K7AIA ($145 U.S.) pushes the microprocessor and chipset envelope to a new level with its Single Slot-A socket for AMD's Athlon CPU, AMD-750 chipset, and 200MHz system interface. The ATX form factor board supports all Slot-A AMD CPU's from 500MHz and above and is expandable with 5 PCI and 2 ISA slots. [break] SY-6VCA / SY-6VCA (slot-1 / socket370, VIA Apollo Pro 133, Virtual Channel PC133 SDRAM support): [/break] VIA Technologies' Apollo PRO133 chipset with 4x AGP graphics support is the base for Soyo's new SY-6VCA ($105 U.S.) and SY-7VCM ($99 U.S.) motherboards. The SY-6VCA supports Intel Pentium II/III CPUs, including the new 733MHz Coppermine CPU, while the SY-7VCM supports Intel's Socket 370 CPUs. Both boards also accommodate PC-133/VCM to 1.5GB, and feature an ATA-66/AC-97 audio/modem riser slot [break] SY-6IWM / SY-6IWM/L (slot-1, Intel 810/810E, Aureal 8810 audio, micro-ATX): [/break] Intel's ``Whitney'' 810/810E chipsets and Pentium II/III CPUs drive Soyo's Micro ATX footprint SY-6IWM ($115 U.S.) and SY-6IWM/L ($85 U.S.) motherboards. The 810 chipset lowers the cost per system because it eliminates the need for a sound card, video card as well as a modem while not compromising on performance. Soyo has enhanced its Whitney-based boards with an Aureal 8810 hardware audio controller, and embedded direct AGP with 4MB display cache on the SY-6IWM. [break] SY-6SWM / SY-7SWM / SY-5SSM Pro / SY-7SWF (slot-1, socket370 en socket7 met SiS chipsets): [/break] Soyo will offer four new motherboards for consumer entertainment centers, Internet machines, and business designer desktops. The boards will be based on SiS' 630 and 540 chipsets. The SY-6SWM (Pentium II/III CPU support, $155 U.S.), SY-7SWM (Intel Socket 370), SY-5SSM Pro (Super Socket 7, $79 U.S.), and SY-7SWF (Intel Socket 370, $105 U.S.) will all feature integrated 128-bit 2D/3D Ultra-AGP video and will be LAN/Internet-ready with an integrated Ethernet and 1Mbps Home PNA Controller on the modem port. Home PNA support allows users to be simultaneously connected to the Internet and to a PC in another location. Memory support includes PC-133 SDRAM/VCM to 1GB. Each board will also include a ATA-66/ AC-97 Audio/ Modem Riser Slot and MicroATX form factor, except the SY-7SWF, which uses the Flex ATX form factor.

The SY-7SWF is Soyo's answer for designer desktops and consumer gadget enthusiasts who do not want to compromise style for performance. The SY-7SWF provides power, functionality, and TV/Internet-readiness. With its Flex ATX form factor this sleek new motherboard will allow designers to get creative with their enclosures and develop smaller and more innovative systems for consumer entertainment centers, Internet machines, and business designer desktops with flat panel displays. [break] SY-1394 (Firewire PCI controller): [/break] Soyo has also announced a ``firewire'' IEEE-1394 host adapter for high-speed file transfers to 400 Mb/sec. Soyo's SY-1394 ($50 U.S.) is an ultra speed PCI Bus card, with single chip 1394a host controller with integrated OHCI and 3 IEEE 1394 ports which can act as a multi-function adapter. Each port may be used for peer-to-peer networking or as controller for 1394-ready DVD set top boxes, digital cameras, camcorders, printers, scanners, and others. This host adapter will enable all existing systems to be IEEE 1394 ready.

Ars Technica heeft een artikeltje rond de SCSI interface gebouwd. Met de huidige SCSI-3 (160Mb/s) en aankomende SCSI-4 (brute 320Meg/sec) standaard lijkt SCSI voorlopig voor de Tweaker-Met-Teveel-Geld een blije toekomst tegemoet te gaan:

SCSI is an ever evolving technology. Recent additions to the SCSI world have included Fibre Channel SCSI and IEEE 1394 (Firewire). There are even specs out now for SCSI 4 (with a 320MB/s throughput) and SCSI 5 (weighing in at 640MB/s!!). SCSI, with the advent of SCSI 3, has gone from a monolithic standard that took a long time to "upgrade" to more of a topology map with different pieces making up the whole.

There are papers constantly being published that propose new specs and new ways to use the technology. A good listing of these papers can be found at the SCSI Trade Association's website as well as the T10 organization's website. These are also great places to look to see what is coming down the pipe for SCSI.

SCSI is also trying to become more friendly to programmers by adopting the Common Access Model (CAM) to allow for a layer of abstraction between the massive SCSI command set and a programmer who just wants to get a program to work correctly with SCSI technology. With the new model for changes (introduced with SCSI 3) allowing SCSI to quickly adapt to the ever-increasing performance needs of the market today, its history of high performance with low system resource utilization and its backwards compatibility, SCSI promises to not only have a future in technology, but a bright one at that.

Verder in het artikel een erg handig tabelletje waarin de specs van de diverse SCSI standaard op een rijtje worden gezet.

De DRAM prijzen zijn de afgelopen weken door de toegenomen vraag in verband met eindejaars verkopen licht gestegen, meldt AsiaBizTech. Verwacht wordt dat de prijzen begin volgend jaar weer zullen dalen:

December 15, 1999 (TOKYO) -- The spot price for DRAMs kept its downward trend, but the 30-day rolling average price as of Nov. 26, 1999 began an upward trend.

[...] The spot price for 64MB DIMMs in the 30-day rolling average price has reverted to a rising trend. This is because demand for them has increased as the year-end sales battle nears. However, the supply shortage of CPUs and other parts is restraining PC shipments, and memory prices will not fluctuate before the new year. It is feared that the memory price will fall after the new year.

Area Contract price Week-on-week comparison North America US$11.06 +2.27% Europe US$11.31 +1.06% Asia US$10.57 +2.30%

Volgens geruchten bij JC en c't zal de Spitfire (low-end 'Celeron' versie van de Athlon) met 64Kb low-latency L2 cache uitgevoerd worden:

I had heard this in the form of an anonymous tip a week or so ago, but didn't move on it. You could probably assume that this one will be true, based on that and the fact that we're getting a sort of confirmation from c't: They suggest that the low end version of the Athlon, Spitfire, will be packaged with a quaint 64KB L2 cache on the processor die. Interesting strategy ... it appears that there really will be a performance difference between AMD's low and high end parts.

It's going to be interesting to see how this stacks up. If I understand Johan's explanation of "exclusive cache" well enough, this means we're going to have a K7 with 192KB of low latency cache (Spitfire) against a P6 with 128KB of low, but slightly higher than the Spitfire's, latency cache (Celeron, or "Coppermine-128"). I'd imagine that this processor will perform somewhere between the Celeron and the Coppermine in performance, which is probably good but slightly annoying in that I was hoping for an ubercheap ultrapowerful option, and it looks like the Spitfire won't be zoomy enough for me, I'll have to save up extra for something with a little more oomph. [break] Voor de volledigheid de info uit dat artikel van c't: [/break] Wie bereits gemeldet plant AMD im ersten Quartal nächsten Jahres eine 800-MHz-Version des Athlon auf den Markt zu bringen. Die Hürde von 1 GHz will man im 2. Quartal 2000 überspringen. Bis dahin dürften auch die ersten Muster des Thunderbird auftauchen. Hierbei handelt es sich um eine Athlon-CPU, die wie Intels Coppermine über einen integrierten Full-Speed-L2-Cache von 256 KByte verfügt. Als "Spitfire" mit nur 64 KByte L2-Cache soll der Athlon dann auch den Consumer-Markt erobern. Für Server hat AMD den "Mustang" vorgesehen, der mit bis zu 2 MByte integriertem Full-Speed-L2-Cache aufwarten soll.

Haagsche Harrie schrijft: "Eigenlijk niets nieuws, maar uit dit interview van 3D Wars blijkt wel dat we op de korte termijn gaan Athlon bordje van Abit hoeven te verwachten":

Many people think Abit should spend more of their R&D on creating a nice Athlon Motherboard instead of entering the video card market. How do you feel about that?

What a truly excellent question. This is not an issue of us diverting resources from the development of a Slot A motherboard. In terms of pure R&D expertise, for us to make an Athlon board is not beyond our capabilities or resources. The question is rather the cooperation we can expect to receive. Many of the Athlon board designs (like most board designs, including Intel's) are based on what are known as "reference" designs. AMD provides these basic designs to mainboard companies and gives them assistance in coming up with the final product. When we approached AMD in Taiwan, they decided to work with other manufacturers first, and we were unable to get the assistance that we required to come out with a timely Athlon board.

The other option is that the manufacturer can start from scratch, which takes far more time and resources. And even then, it is not just a simple matter of transferring all of the graphics cards guys to another project and telling them to make an Athlon board instead of graphics cards. You have to build up a team with the right people with the expertise and skill that you need. We have been building our graphics card team now for a year.

So to answer your question, the graphics card guys are working on great graphics cards, which will be available soon. As for Athlon, well, we'll just have to wait and see.

Verder in het interview een aantal vragen over de voor overklokkers erg interessante videokaarten die Abit vorige maand tijdens de Comdex aankondigde.

Dit wisten we natuurlijk al veel langer, maar nu heeft ook AOL/Compuserve tijdens een hoorzitting van de Opta laten weten dat het de groei van internet belemmert ziet worden door de hoge tarieven van KPN (thanks mnda voor de link):

Volgens Simon Hampton van AOL Europe staan de hoge tarieven die KPN Telecom in rekening brengt voor lokaal telefoonverkeer, de groei van internet in de weg. Dit meldt de Telegraaf.

"Wat zijn een paar tientjes per maand voor je provider als je een telefoonrekening krijgt van honderden guldens?", aldus Hampton. Hij stelt dat de telefoontikken uiteindelijk helemaal moeten verdwijnen. In plaats van voor de tikken zou maandelijks abonnementsgeld betaald moeten worden. "Ouders houden nu hun kinderen weg bij de computer uit angst voor hoge telefoonrekeningen. Op deze manier staat KPN de ontwikkeling van het internet en e-commerce in de weg."