Tja, als je de mensheid dan toch al zoveel geld afgetroggeld hebt kun je het net zo goed weer weg gaan geven, loop je ook nog kans dat mensen je aardig gaan vinden...

Bill Gates has whacked in another $5 billion (or it could be $6 billion -reports vary) into his very own home-grown charity.

This is the biggest gift yet made by a living donor. But Gates has a couple of billion dollars to go before his charitable foundation overtakes The Wellcome Trust, the world's largest, with $19.1 billion worth of assets.

The foundation stakes are hotting up, with Bill and Melinda merging the William H Gates and the Gates Learning foundations to form, well, the Bill and Melinda Gates Foundation. Combined assets are $17 billion.

The charitable focus remains on healthcare, and technology in education and libraries. One idea is to create a market for a malaria vaccine that hasn't been developed yet. Shades of Windows 2000, perhaps?

Een leuke post bij The Register: volgens K-Bench, een of andere wazige Koreaanse site, is de Intel i820 wonderchipset niet compatible met de TNT2, G400 en Savage4 als deze in AGP4x mode worden gedraaid. Ouch!

According to the report, which may be found here, Taiwanese mobo manufacturers say there is a problem between AGP 4.x and a number of existing graphics boards.

Boards affected include TNT2, the Matrox G400 and the Savage 4, when running in AGP 4x mode, says K-Bench.

The Camino chipset has represented something of a weight round Intel's neck for the whole of this year.

If true, the difficulties will compound the situation for the Taiwanese firms, which have already been placed on severe allocations by the Great Satan of Chipsets (Chipsetzilla).

Last week, we reported that the Taiwanese firms were also fed up with implementations of the 810 chipset, forcing them to consider options including chipsets from Via and SiS.

The Register post wel vaker onzin, maar dit slaat echt alles. Eén van hun lezertjes beweerd dat Intel en Microsoft collectief ten strijde zijn getrokken tegen het kwaad der overklokkers: een PIII zou onder WinNT met Service Pack 5 niet meer overklokbaar zijn. Lees en huiver :

Chip giant Intel has warned since the beginning of this year that to overclock its processors is verboten.

A reader describing himself as 'Firestormer' wrote to us with the following tale: "You might find this interesting, but if I'm correct, Intel and MS have finally come up with a way to defeat overclocking totally, at least under NT4 Service Pack 5 and with the Pentium III. The sorry tale is as follows.

"I've dabbled from time to time with FSB adjustments on my Tyan Tsunami-based system, originally tweaking my Pentium II 350 to see if I could get another 50MHz or so out of it. My NT4 SP5 barfed at 112MHz bus, but I was able to get it running at 103MHz bus, and the speed difference showed up accordingly under NT Diagnostics as well as H-Oda's WCPUclk util. The performance boost was there, but at the time another 10MHz didn't seem to give enough of a boost to be worth while.

"Recently, however, I upgraded this system from the P2-350 to a P3-450, and then tried to tinker with overclocking via the FSB yet again (enough is never enough where speed is concerned ). BIOS, check; post screen, check; DOS level utils, check... all reported ~500MHz correctly (I'd tried 112MHz bus again).

"However, upon getting into NT, I ran my benchmarks and found that the system performed identically to before the overclock. Upon further analysis, I found that all of my speed reporting utils still said that the CPU was running AT EXACTLY STOCK. Nothing I tried could get the chip to run under NT4 SP5 at anything other than stock and only stock speed.

"This changed, yet again, when I manually replaced HAL.DLL (you know, the lowest level part of NT there is) with the HAL.DLL.SOFTEX that is included with SP5 so I could make the system shut off at NT shutdown. Earlier this evening, I decided to experiment with the FSB overclocking yet again, determined to find out why the overclock failed, only to find that it worked this time.

"This has lead me to one and only one conclusion: the HAL.DLL in NT4 SP5 is capable of reading the processor serial number/ID tag and is using that information to ensure that the CPU will only run at stock speed under NT. And considering that there are utilities available which do similar things (see H-Oda's SoftFSB), this seems the most likely, if somewhat disturbing, explanation.

"Make of it what you will, but this is one tidbit of information I'd like very much to see verified (if possible) and if found correct spread far and wide across the Net."

The Tech Zone heeft een review van de Creative 3DBlaster Savage4 neer gemikt. Vergeleken met het G400 en TNT2 Ultra geweld draait de Savage4 Pro chip van de 3DBlaster op een nogal magere 110/125Mhz. De jongens van The Tech Zone gooien er een stevige dosis Viagra tegenaan en kijken in hoeverre overclocking de lage performance kan veranderen:

The savage4 PRO chips which are in use in the 3dblaster are certified only for 110Mhz. So far only Diamond and Elsa are releasing PRO+ based products, which means that in all likelihood some of those extra PRO+'s might've been sold as PRO's as well. It's true the percentage must be low but the chance is still there; just remember it's more or less down to luck.

Luck was definitely on my side though, as I managed to take my 3D Blaster's core speed up to a staggering 155Mhz! (From 110Mhz!) Sadly the memory on my card seemed to be an exact 8ns SDRAM spec part, which meant that I couldn't take the memory more than 135Mhz without seeing visual artifacts. When I ran the card at 155 there was no problem with stability at all. The second 3D Blaster that I tested had much better memory, which could be taken to 150, the problem now was the core which refused to function at anything higher than 143!...Still at least I could get 143/143 without much of a problem.

In some of my benchmarking I noticed that in order to get a proper linear in fillrate in proportion to the core speed, the core speed must be 0.88 of the memory clock which is how both diamond and creative have designed their specs (Creative 3DB 110/125=0.88 and Diamond Stealth 125/143=0.88). That said if I were to set my core speed to 155, in order to get maximum fillrate I would need my memory to be 175Mhz! Which would be equal to the lower spec TNT2's. I am surprised that nobody has decided to attempt making such a product because it would definitely sell as it will make games very playable at 1024x768 (not that it is not playable now, but anything under 50fps just doesn't seem fluid anymore).

Storage Review (wie anders) heeft een artikel in elkaar gedraaid waarin ze de performance van de nieuwe ATA/66 interface vergelijken met die van de oude ATA/33 standaard. Is ATA/66 werkelijk sneller, of gaat het hier om een opgeblazen marketing hype?

Is ATA-66 a sham? Absolutely not. ATA-33 first shipped over two years ago in the form of the Quantum Fireball ST. Seagate's Barracuda ATA ST328040A, a drive due to ship in the near future, can push 28 MB/sec of data. ATA drives are coming very close to ATA-33's limits. Introducing ATA-66 drives and controllers at this time simply allows the industry and the community to place next-generation infrastructure in place before drives that truly need it arrive. It's advanced planning, folks, something that will benefit us down the road.

Okee, zoiets wisten we al langer ...

Panders van Ars Technica heeft een artikel gebakken waarin hij uitleg geeft over de bugfixes in NT4 SP5:

Which patches should be applied? What is the best order in which to apply them? Does order even matter anymore? That's where this guide comes into play. I hope to show you some guidelines in keeping up with all of the hotfixes (regardless of environment complexity) - the ones you need, the ones you can do without, the order in which to apply them, alternative methods of installing them, and so on.

Wellicht interessant voor (dualerende) NT gebruikers -> Ars Technica.

Bij AGN 3D hebben ze een artikeltje in elkaar gedrukt waarin ze uitleggen hoe je je CPU kunt 'lappen'. Hmm ja, klinkt nogal wazig. Wat ze bedoelen is dat je de bovenlaag van je CPU behuizing met (fijn) schuurpapier wegschuurt, waardoor het oppervlak vlakker wordt en de warmte geleiding wordt verbeterd:

Once the CPU shows signs of wear in the center then switch to the next highest grade of paper. The first part of the process will take a while about 20-25minutes depending on how bad the CPU is and what grit paper you started with. At this point some people may ask why can’t I use coarser grit of paper to start with? The reason I don’t recommend that is the coarser the grit you start with the more likely the edges of the CPU will become beveled. At the beginning of the sanding process the paper will “grab” the CPU there and rock the unit to the side first and remove more material than you want as your sanding it down.

De step-by-step 'lapping' guide vind je hier.

Jeremy van AGN 3D heeft een interview gepost met een ventje van AMD, waarin een eind weggebabbeld wordt over de Athlon. Hier bijv. een vraag over de nogal hoge prijs van Athlon moederplanken:

. Speaking of love for Intel... The price of the processors are actually quite impressive, truly lower then Intel's prices for the same speed processors. But at the same time, the motherboards that are listed on Pricewatch and others are featuring rather massive priced motherboards that are over $200 for their price. This release of higher priced motherboards, evens out the price when compared to a processor and motherboard for Intel. Will this higher level of prices for motherboards with Athlon support be something that lasts, or do you think $125 motherboards will finally show up on the market over the next month or so?

I can't really offer a comment on specific pricing for motherboards since those products offered by other companies. They set their own prices and conduct their own business. Their boards are based off an AMD reference design, and we do work very closely with all of our infrastructure partners, so of course we are keenly aware of the issues surrounding their pricing. Any time a new platform is brought online, the pricing models are not based on sustained volume production. I wouldn't expect you will see the same pricing as time goes on, our partners ramp production on their Slot A motherboards and begin to engineer board designs of their own.

PC Paradox heeft een review online gezet van de Elsa Erazor III. Dit is een non-Ultra TNT2 videokaart met 32MB geheugen.

What really took us by surprise was the excellent power and performance which we got from such a cheap card. Despite its non-ultra status, the ERAZOR III pulled in some impressive benchmarks and garnered an even more impressive array of options for the consumer to choose and customize their ELSA experience. From the 3D Revelator to the multitude of video in and out options, the ELSA could come prepared for almost any occasion. In it's most naked form, the ERAZOR III with no add-ons is the perfect card for a gamer with a tight wallet. Weighing in at around $135 for the 32 Meg TNT2 ERAZOR III, it is almost impossible not to buy it. If you are worried a lot about performance, never fear. When overclocked just a small amount, the ERAZOR III continues to close the gap between itself and its Ultra brethren. And if you are bent on making the most of your wise purchase, a small fan sitting upon the ERAZOR III heatsink can take you to Ultra speeds and beyond!

It is amazing what ELSA has done. They have created their own niche market using only the vanilla TNT2 design, by expanding and creating many choices for those of us who not only want the fastest graphics card on the market but a little functionality too. Being the cheapest 32 meg TNT2 on the market (non OEM) and 2nd only to the Guillemot Maxi Gamer Xentor 16 meg and maybe the Hercules Dynamite 16, the ERAZOR III is aggressively priced and deserves a high ranking for value. So far our only complaint are the ELSA drivers which despite their stability come no where close to the performance that the Nvidia reference drivers permit. (Note: we have not found one manufacturer which has stable and fast drivers other than Canopus). And finally, we commend ELSA for adding the 20 game "preview" bundle which allows gamer new on the 3D scene to get rolling and start checking out software which their card can use.

Anand van AnandTech heeft een review gepost van de A-Trend ATC-6254M BX moederplank. Dit is een normaal BX plankie, behalve dan dat A-Trend er een Voodoo3 2000 ingebakken heeft. Naja, in ieder geval beter dan die geïntegreerde ATI troep die je meestal tegenkomt. Anyway, hier heb je een hap uit de review, check de rest hier.

The setup and configuration of the 6254M is essentially jumperless as only the clock multipliers are controlled on the motherboard (via a dip switch block), with all the FSB settings able to be manipulated in the Award BIOS chipset features setup. The available FSB settings are obviously geared towards towards the overclocking population, the clock generator provides support for the following settings: 66 / 75 / 78 / 81 / 83 / 90 / 95 / 100 105 / 110 / 112 / 113 / 115 / 117 / 118 / 120 122 / 124 / 126 / 133 / 135 / 137 / 138 140 / 142 / 144 / 150 and 155MHz. The mouthful of settings rival those found on the newer ABIT boards which are definitely overclocking oriented.

Hardware monitoring on the 6254M is provided courtesy of the Winbond 83781D chip located on the opposite side of the SC242 connector. The 83781D provides for the monitoring of up to three temperatures including the temperature of the chip itself with the other two temperatures supplied by the included thermistor headers on the board. Two of the three fan headers are also monitored by the chip, and A-Trend conveniently bundled a CD-ROM that includes the Intel LANDesk Client Manager software for system monitoring as well as the Intel BM IDE drivers and all patches for the BX chipset. A second CD is provided with all of the required drivers for the 3dfx Voodoo3, however for the most up to date drivers you'll want to pay a visit to www.3dfx.com.

The performance of the board, along with the on-board Voodoo3 is on-par with that of your standard BX board + Voodoo3 2000 combo so there are no worries there. As far as stability is concerned, the board exhibited average stability in our endurance tests and although it wasn't a top quality server-like solution, it is safe to say that most users will be fine with it.

De Athlon moederborden reviews beginnen nu langzaam op gang te komen. Bij Sharky Extreme vind je er nu een van de Gigabyte GA-7IX. Hier heb je wat uit de intro, klik hier voor de rest.

Although AMD manufactured and produced the mainboards that were equipped in each of the Athlon evaluation machines used by the press for reviews last week, they have no intention of continuing this process and instead have recruited six mainboard manufacturing partners to produce Athlon boards.

FIC, Gigabyte, ASUS, Biostar, MSI, and GVC have all committed to supplying Slot-A compatible mainboards for the Athlon CPU, and each will likely debut their first boards over the next two months. Although these boards will differ in some minute ways, they each will utilize AMD's internally developed 750 mainboard AGP 2X core logic chipset.

Eventually ALi, VIA, and SiS will take over production of the Athlon compatible mainboard core logic chipsets, and will likely be introducing more advanced features than the capable AMD-750 chipset includes now, such as AGP 4X and 133MHz SDRAM support.

When will these new chipsets arrive on the market mounted on Athlon mainboards? After returning home from a trip to Taiwan just last week, Sharky Extreme staffers indicated that leading Athlon mainboard producers expected the introduction dates for these "second generation" Athlon mainboards would not be before late December of 1999. For now, the AMD-750 chipset will be utilized in each of the Athlon supporting mainboards being produced by the six vendors we mentioned previously.

NV10.net blijft interessant: veel mensen vragen zich af over de NV10 ondersteuning heeft voor bumpmapping. Het antwoord blijkt enige dagen gegeven te zijn in het Siggraph '99 artikel van Gamasutra:

Another demo showed some new technology that may be included when this board ships (according to Nvidia's personnel, before Christmas). It showed an OpenGL logo texture mapped onto a quadrilateral, and the application of some nice bump-mapping. The way they achieve this is quite interesting: to get per-pixel, good looking bump-mapping, Nvidia plans to store per-pixel surface normals as an RGB texture map which encodes the XYZ triplets. Then a new texture blending mode will be added to OpenGL's existing multi-texturing capabilities. This new mode will dot-product two vectors (the normal stored in the texture and a light vector), and modulate (multiply) underlying layers according to the result. This way, Nvidia can provide a great way to evaluate the classic Lambert formula for diffuse lighting (which, in case you're interested, is I diffuse = Color surface * Color light * k diffuse * (N·L)).

Kreeg weer een handige tip van Sake: ome Tom heeft de multipliers van z'n Athlon zodanig verbouwd dat het ding nu stabiel op 750Mhz loopt. Tom is zoals gewoonlijk weer geheimzinnig over de methode die hij gebruikt heeft om de multiplier te flippen, maar het lijkt erop dat het hij dit op dezelfde manier heeft gedaan als de Japanners uit deze nieuwsposting. Hier heb je twee snippers uit dat artikel van Tom:

It was a very pleasant surprise to find out that our Athlons ran at up to 750 MHz without any additional means of cooling. This shows that the Kryotech-system could possibly run even faster. Athlon ran as stable as a rock, and when I say stable, I really mean it! I am not talking of crazy overclocking and calling a system stable that needs to be rebooted twice daily. Athlon performed really stable and any software I threw against it ran without any crashes for days.

[...] I needed to change Athlon's core voltage a bit to run it at all the different clock speeds. You may think that I'm talking of higher voltages, but as a matter of fact the opposite was the case. Athlon would not run at 500 or 550 MHz unless I lowered the core voltage down to 1.45 V. On the other side, I only had to raise the core voltage from the default 1.6 V to 1.65 V and Athlon worked beautifully at 700 and 750 MHz. 800 MHz were left to the super-cooled Kryotech-Athlon, since I didn't want to tamper with my Athlon at higher voltages than 1.7 V and this was obviously not enough to run it at 800 MHz without Kryotech's cooling.

Voor benchmarks en meer spul over Tommies amazing Athlon adventures moet je hierheen fietsen.

Ik lees bij AGN Hardware dat Creative Labs nu ook kant-en-klare PC's gaat verpatsen. Het ding heet 'Blaster PC' en zit vol gestouwd met allerlei Creative produkten:

Since roll out of the successful Sound Blaster products, Creative Labs has tried its luck in other computing area, including graphics, DVD/CD-ROM, and modems. It's only natural that a company would bundle all these into a multimedia PC package of its own. This is exactly what Creative Labs has done.

Creative Labs's first personal computer -- Blaster PC -- is a joint project between Creative Labs and Empac Computer, a company based in Ontario, Canada. "Customers can build a PC to order through the Blaster PC website, but all systems must be purchased through a local dealer..." Price of Blaster PC will start at $999 CDN; Blaster PC will be available in Canada EXCLUSIVELY starting in mid-September.

The most prominent feature of Blaster PC is the volume control knob in the front of the PC. "The sliding door at the bottom of the front panel, the user has access to USB, audio input/output, joystick, MIDI, SPDIF video and optical ports and also one firewire port.

Blaster PC performed remarkably well in Computer Paper's (CP) lab test. The machine is equipped with a 500Mhz Pentium III chip, 128MB of RAM, a 13GB Western Digital hard drive running Windows 98. On the multimedia side, Blaster PC is packed with 16MB TNT AGP Graphics Blaster Riva TNT card, Sound Blaster Live!, a Creative Labs 56kbps modem and a Dxr3 MPEG decoder card with a 5x DVD ROM. Forgot to mention Blaster PC also has a FM tuner. According to (CP) tests, BAPCo's Sysmark 98 recorded an overall rating of 197 (195 for Office Productivity and 200 for Content Creation). Playing DVD is smooth as silk.

NV10.net heeft een (super brakke) engelse vertaling gepost van het chinese Voodoo Gaming interview met Elsa. Op zich zit er weinig nieuwe info in de vertaling, afgezien dan van de twee punten die gisteren al bekend werden (release Elsa NV10 in oktober, 64Mb geheugen standaard). Hier heb je een stukje over de Elsa drivers:

Q3: Ken: What's the difference between Elsa graphics card and other brand's?

A3: Alex: We are very strict with the materials of our card, especially on the written of Drivers. We spend much of our forces and completely optmize for Elsa card. It's obverious that use public version driver can't compare with our driver. So Elsa's graphics card always be delivered later than other brand's card, just for providing our users have a more stable and efficient graphics card.

Koen van Hardwareinfo.net heeft nog een Hercules Dynamite TNT2 Ultra te pakken weten te krijgen. Erg triest dat Hercules ermee opgehouden is, want de Dynamite TNT2 was absoluut een topper:

Alle benchmarks bewijzen dat de Hercules Dynamite TNT2 Ultra absoluut de snelste videokaart van dit moment is. Op de een of andere manier is het Hercules gelukt om een TNT2 videokaart te bouwen die stabiel is op 175/200 en de prestaties zijn dan ook verbluffend. De videokaart is soms meer dan 25% sneller dan een standaard TNT2 Ultra. Wat drivers betreft is Hercules absoluut ook een van de betere. Het enige nadeel dat ik kon vinden is het gebrek van een goede softwarebundel.

Check de review voor meer info (benchmarks, CPU scaling e.d.) over deze kaart.

Andy van BXBoards heeft een review van de Abit BE6 op z'n site geparkeerd. Verdere uitleg lijkt me niet meer nodig bij dit bord...

An excellent motherboard let down by one or two niggles. I have had a BE6 since early July and it has been pretty hassle free for me. Nevertheless, the integration of the Highpoint controller will cause problems for some people and IRQ's are a scarce resource on this board. Clashes are inevitable and can cause real problems under non-PNP O/S like Windows NT4 which is our primary test O/S here.

OK, lets look at the positives and there are many! Quite simply this is the Best Slot1 mainboard Abit have made, and that means that it is one of the best Slot1 mainboards ever made. Quite a complement but with more mainboards stealing Abit's overclocking thunder, they have again stolen a march with the introduction of UDMA/66, one of the most requested functions for new mainboards if my mail bag is to be believed.

Lost Circuits heeft een review van de FIC F11 neer gekwakt. Deze Athlon plank is gebouwd rondom AMD's eigen 751 en 756 North/South bridge. Verder heeft-ie 3 DIMM slots, 5 PCI gaten en een optioneel ISA slot:

FIC has created a good, solid board that, without any doubt will be an excellent platform for any kind of Athlon CPU one might lay eyes on. Because of the extremely spacious layout there is relatively little concern about heat issues. In fact, even the voltage regulators that, on the Fester B3 showed slight discoloration (heat spots) at the end of the testing period remained spotless. There are several reasons why essentially the same hardware can behave in different ways. One, of course, is that the original Athlon testing was conducted under climatic conditions that were as bad as one could possibly imagine. The second reason lies in the CPU itself. The power consumption of the Athlon rises almost exponentially with the clock speed and thus a delta of 50 MHz can make a significant difference particularly for the voltage regulators.

Hot Hardware heeft een stel 133MHz HSDRAM DIMMen van Enhanced Memory Systems op de testplank gelegd. Hier heb je een stukje uit de review:

I have seen a lot of memory modules and from my perspective, the stability and performance of Enhanced Memory Systems' PC133 SDRAM Module is second to none at this point in time. Other sites have tested this product vrs. other manufacturers like Samsung, who is noted for their high quality "-G8" chips. The outcome was the same. Nothing matched the HSDRAM from EMS with respect to stability.

Having said that, real world application performance will be impacted only by a small margin as a result of this higher speed memory. The real benefit this technology brings you is stability in extreme overclocked systems of today's technology along with some additional headroom for the next generation of bus frequencies with the latest chip set and processor technologies. At this point in time, the EMS HSDRAM is ONE of the fastest modules you can buy. There are other PC133 module makers out there and we will be taking a look at their offerings as well in order to level the playing field

Joe van Overclockers.com heeft een vergelijking gepost van de Alpha socket 7 en GlobalWin FDP32 koel monstors. De Alpha cooler wint in zijn test:

The Global Win FDP32 is the first serious challenger to the Alpha PFH 6035. It is ugly, large and inelegant compared to the Alpha, but it is a very close-performing second. With the potential to increase performance with some tinkering, I would rate the Global an interesting alternative to the Alpha.

Nou heb ik net zo'n FDP32 besteld (in de vorm van de 3dfxCool BigMoFo-REX) en eigenlijk vind ik deze uitslag niet zo leuk . Gelukkig (voor mij) wordt de GlobalWin in reviews van The Tech Zone en BXBoards beter bevonden dan de Alpha.

Ik kwam bij Slashdot een interessante editorial tegen over de vermeende fragmentatie van Linux, veroorzaakt door de toenemende betrokkenheid van commerciële bedrijven. De schrijver van het artikel is het daar niet mee eens en zegt dat Linux sterker is dan ooit:

What we have now have going is a virtuous circle -- as each of the old-line Unix outfits joins the Linux crowd, the gravity it exerts on the others grows stronger. The Monterey and Tru-64 development efforts, the last-gasp attempts to produce competitive closed Unixes, can't even muster convincing majorities of support inside the vendors backing them; both IBM and Compaq are investing heavily in Linux.

Linux fragmenting? No way. Instead, it's cheerfully absorbing its competition. And the fact that it is `absorbing' rather than `destroying' is key; vendors are belatedly figuring out that the value proposition in the OS business doesn't really depend on code secrecy at all, but instead hinges on smarts and service and features and responsiveness.

Het complete artikel vind je hier.