March 1 - 8, 1999
Archive

• 3DMark 99 MAX (Win9X/NT - 17.7 MB)

 

Seeing the TNT and TNT2 perform side by side, it was clear who outperformed who. The TNT 2 equipped system ran the timedemo around 9 seconds faster than the TNT equipped one. Keep in mind that we were never shown actual FPS so we can't only give you an approximation of the performance. In our testlabs, under the same conditions, with normal drivers (not Detonator drivers), our TNT performed 24.5 FPS in demo1.dm2. The 689 frames contained in that demo took 28.1 seconds to run through with the TNT. So taking this in consideration, and that the TNT 2 equipped system finished that demo around 9 seconds faster than the TNT, we calculate the FPS to be: 689 / 19.1 = ~36 FPS.

Now for the ULTIMATE THRASH TEST. Yes friends, we are going to run 3DMark99 at the same time as writing a CD. The only other CD-writer I've tested that can do that is the Yamaha 4260 SCSI. My Wearnes CDR-622 knocks out the moment I double click the 3DMark99 icon. Okay here goes! Put in a VideoCD in my Sony 24X and a blank CDR in the Yamaha. Start the writing process and start 3DMark99. Holy SH*T! It's WRITING! I noticed that 3DMark was slower. Every second or so, it would pause. But hey! It WROTE! 11 minutes later the Yamaha spat the disc out. Whoopee! I watched the VCD (Ghost in The Shell) and yep, it was FINE. We can attribute the 4416E's success in handling interruptions and multitasking to the massive 2MB buffer. This means that if the IDE bus is being accessed by something else, the buffer can still supply the writer with constant data. Not bad. Not bad at all. Check out the 3DMark scores here.

 

• AOpen PA2000 Driver (Win9X - version 3.01.01 - 1.49 MB)
• AOpen PA2000 Driver (WinNT - version 3.01.01 - 2.81 MB)

 

• Aureal A3D Reference Driver (Win9X/NT - version 2.05 - 251KB)
• Aureal A3D 2.0 SDK (Win9X/NT)

 

I have released MBM 4 - Beta 1, this is as Beta as you can get it so I recomend you don't d/l it unless you know what your doing. It will work fine on Windows 9.X but you need to have 3.x installed if you want to use it on NT.

This is Beta so please send me feedback of bugs, for now I have no use for messages saying it won't work on my PC, MBM 4 does not support more boards then 3.8.

• Motherboard Monitor (Win9X/NT - version 4.0 Beta 1)

 

Once again, Gigabyte is showing its knowledge with this superbly designed product and no doubts in our mind about the reason why this manufacturer has survived since so many years. The GA-6BXE motherboard is a real fine piece of equipment and even if we wouldn't recommend this motherboard to the Overclockers fanatics, I sure would recommend it for most of the peoples wishing to built a strong system with one of the best motherboard available on the market nowadays.

Don't think the Zoro is for you? Then try Microsoft's other game pad offer, the Zulu. The Zulu also incorporates many of the same features as the Zoro, but its unique and sharp design really makes it stand out against its competitors. Zulu's great innovation is its "Perceptive Control" which is best equated to the feel and use of a mouse, something that many PC gamers know a lot about. Imagine, Quakers may be able to use the Zulu to enhance the first person multiplayer shooter to new levels. This innovation frees gamers from the typical mouse + keyboard setup that really can get clumsy at times. The Zulu also offers 16 programmable buttons and the SideWinder game controller software, which, by the way, has certain game customizations that allow for better game play.

 

There is one tremendous bonus to a USB version of this controller:  There is virtually no performance hit when 4 of them are used simultaneously.  Previously you needed Gravis' specially designed 4-way hub to connect 4 GamePad Pros to a single game port, and even then performance was not as tight as it should have been.  With a USB version, the connection between the devices and the PC is fast enough not to effect performance when several are used at once.

 

The first thing I noticed about the wheel was the excellent grip and how comfortable it was. It certainly feels a lot more realistic than the Microsoft wheel. Although thick and heavy, it moves smooth too. The generic plastic-like feel is completely eliminated. The wheel is not as powerful as the others I have tested. It is much smoother and not as jerky. Instead of pulling you along, you actually control the game. It's fun to be able to control the game for once instead of vice versa.

 

This board was VERY stable at default voltage when overclocking our P2-333 SL2TV step code chip to 500 Mhz. This is a processor that needs a voltage boost depending on the motherboard used to overclock. Not so with the P6F99! There are no options for CPU voltage adjustment on this board, but we didn't need it. The board ran all our tests, without crashing even once, at default CPU voltage. In a word, robust!

Now let's get down to how this thing actually performs, one word comes to mind, speedy. This drive writes data faster than any CD-Recordable out on the market, the fastest currently is 8x. Clocking in at around 9x, this drive writes at 1.38 MB/s, which may not sound like a lot, but in the world of optical media recording it's top of the line. Compared to the other leading rewritable optical technology, CD-ReWritable, I found the DVD-RAM to be less system reliant, and took no time at all to erase/format data. Whereas CD-RW (Which I'm no fan of) can take hours to format a single 650MB disc, the Flex took only 2-3 seconds to format each side of the DVD-RAM disc, each side holding 2.6 GB. Normal CD performance is rated at around 20x, not too shabby.

 

Remember the days when the Zip Drive 100mb was enough for you to tote your programs and applications around on 1 or 2 Zip disks and when the speed of the parallel was good enough? Well no more of that. The 100mb zip cartridges are getting small fast and the parallel port is just slow. Want a Zip drive with over 200mb of storage and speeds that will blow away your IDE drive ? Well heres my solution to you, the Zip Drive 250mb SCSI. Sporting its new SCSI-2 interface, speeds have increase in a excess of 50% over the original Zip100 SCSI, now if it's that much of an increase in the SCSI model think of how much faster it is over the Parallel model. Unfortunately the Parallel model of the Zip 250 has the same performance of the Zip100.

 

WOW!!! COOL! ABIT works wonders! ABIT BX6 Rev 2.0 left far behind not only our former favorite Chaintech 6BTM but also its current opponent ASUS P2B Rev 1.10. We have to admit that ABIT finally succeeded in designing the quickest mainboard. Congratulations! Though we cannot guarantee that this triumph will last long. Their main competitor – ASUS - keeps on introducing new revisions of P2B, and as for Chaintech, they haven’t yet offered anything new.

 

• PowerStrip (Win9X/NT - version 2.40 - 640 KB)

 

Instead of just normal metal, 3Dfx Cools hddHO uses aluminum to hold the drive, the same type of aluminum you'd see on a CPU heatsink. The aluminum used to absorb the heat even has a thermal resistance of 0.46. On top of using this awesome metal, the hddHO uses three fans to blow air over your drive. The three ball-bearing fans are pretty quiet as well. Most fans are loud, and you can hear them in another room, which is not the case with the hddHO.

 

Very impressive scores, considering that the TNT can only get about half the numbers in Turok 2 D3D. But what about the heavy duty tests like WinBench 99 2D/3D and 3DMark 99? Obviously the V3 doesn't support 32 bit color, so comparing the 16 bit performance to 32 bit TNT performance just isn't feasible. But here's the detailed scoop on the V3 in 16 bit color:

 

As mentioned above by Steve Schick, 32-bit rendering does have an enormous overhead on both the processor and the graphics card. This overhead usually results in frame rates that are decidedly sub par for enjoyable gameplay. In order to fight the enormous strain brought about by 32-bit rendering and keep frame rates high, ATI has optimized their chip extensively for 32-bit use. The end result is a chip that can provide near 16-bit level performance at the 32-bit setting. Here is a comparison against the 32-bit performance of some of their competitors when using 32-bit textures. (The TNT was only using 24-bit)

 

• Diamond Monster Fusion Driver (Win9X - version 2.13 - 3.5 MB)

 

New stuff in v1.1.0

• AMD CPU support
• Savage3D support
• ATi Rage 128 Support
• Warnings!!
• Various code tweaks. Memory footprint is down, memory management is better, tighter loops where possible...etc.

• FastCard Install Program (Win9X - version 1.1.0 - 612 KB)
• FastCard Core Program (Win9X - version 1.1.0 - 95 KB)

"I had the opportunity to speak with Kevin Wagner from AMD today, and here's his fill in for all you AMD/TNT fans out there. AMD and NVIDIA are vigorously engaged in optimizing the TNT drivers for both D3D and OpenGL. The initial timetable for release will be in two to three months, but of course, unforeseen development delays may move that back a bit. "

 

The SoundBlaster Live! uses the EMU10K audio processor, which is touted as having 1000MIPS (Million Instructions Per Second) of audio processing power. Creative Labs claims that roughly half of the EMU10K’s performance is currently utilized with the latest SoundBlaster Live! drivers. Being a sound DSP, it can be easily updated to new industry standards. Additionally, new features can be added by Creative Labs as they see fit. To take advantage of the benefits of using a powerful DSP, Creative Labs has created a "Live!ware" promotional campaign to promote driver updates, which will add increased functionality to the SoundBlaster Live!. Two main feature updates since the SoundBlaster Live!’s release, were the addition of support for 32 concurrent 3D hardware sound streams, and 512-voice MIDI support. I commend Creative Labs for continually adding features to the SoundBlaster Live!. The end-user’s experience is really improved when a given product keeps getting better and better.

 

One of the things that should be reiterated is that this board is based on the 440LX chipset, which doesn't use a "Front-side bus" frequency. This means that overclocking to 100Mhz (while guaranteeing the stability of your PCI/AGP peripherals) isn't possible, but as 100Mhz isn't an option in the bus speed field, it's a moot point. Just remember to look elsewhere if you're set on getting a Celeron 300 to pump up to 450.

 

There has been a lot of confusion about something called the CPU scaling of graphics chipsets. Basically what this scaling thing is all about is the increase in framerate which you get with a certain chipset when you move to a faster CPU. So when a given chipset shows much better performance on a Pentium II 400 than on a 233, then we say that it scales up well... Chipsets that don't seem to benefit much from a faster CPU are said not to scale well. So what's the confusion?

• Logitech Internet keyboard review
• Kenwood's CL-701 3-piece speaker set review

 

Kryotech claims that Vapor Phase Refrigeration is 50 times as effective as traditional forced air cooling (your standard heatsink/fan combo cooling device) and 5 times as effective as forced liquid cooling. Using Vapor Phase Refrigeration, Kryotech has been able to allow their products to run at levels around -40 degrees Celsius, an achievement which opens new doors for desktop computing opportunities, the biggest of which happens to be overclocking.

 

At it's current pricing schedule, the socket 370 version of the Celeron 400A, is roughly $20.00 cheaper than AMD's K6-2 400MHz.  As you can see, there is very little difference in performance.  We will be covering a number of socket 370 mainboards over the next few months but as it stands now pricing of these system boards is similar to their super7 couterparts.  As I mentioned earlier, the only downside to the PPGA 370 version of the Celeron A seems to be very restrictive future upgrade path but other than that, the platform seems extremely stable and capable.   While it may seem like I'm selling out, I nevertheless am happy to give full approval to both the processor and the platform and look forward to more extensive testing in the near future.  Stay tuned...

 

Configurability, options and price.  Those three elements of the BM6 make it a great buy if you are looking to upgrade to the Celeron PPGA line of processors. I wasn't able to test any FSB settings above 75MHz due to the multiplier lock of x6 on the 400MHz Celeron, but I'm fairly certain that it is stable up to the maximum setting of 133MHz like the BX6 II.  What's great about the current Celeron line is that you can still use your old PC66 RAM making this motherboard coupled with a Celeron a very low cost solution vs. the Pentium II & III which don't have nearly the price/performance point as of yet.

 

One of my favourite sites, Hard OCP has posted part 1 of their Card Cooler review.  From what I see, this thing reduces the temperature of the graphics chip big time!  Here is a bit from the review:

OK, back to the main story-line here.  With us using the less than 100MHz bus speeds, we of course increase the PCI and AGP bus on most boards.   Some boards have BIOS settings that can alter this to some degree, but not much.   When we increase the PCI and AGP bus we are running all our devices harder.   But with most of us OC'ers, we probably already have been tweakin on the freakin Vid Card too and now we have OC'd the damn thing not only directly, but indirectly also.  We increase the id card's memory MHz, we increase the core MHz, then for fun, lets kick the bus in the ass and get the whole mutha crankin hard enuff to jump start a smallblock Chevy. 

 

I noticed over at 3DFiles.com, that the Japenese Canopus site has released some new drivers for the Canopus Spectra 2500 card.  I can't make any sense of the page (because its in Japenese), but if you can head over there and download the drivers.

It's been a while, but Brian Galm has released new versions of his popular tweaking utilities. There are a ton of updates to each utlity, so head over to his page to check out all the details.  You can grab the utilities below:

• Required Controls (Win9X - 1.96mb)
• TweakIt (Win9X - version 4.0.0 - 149 kb)
• DirectControl (Win9X - version 3.0.0 - 106 kb)
• Matrox Overclock (Win9X - version 3.0.0 - 167 kb)

 

The computer industry literally changes month to month, and luckily Sharky Extreme is there to keep up with it.   They provide us with a monthly buyers guide to update us on what the best hardware is and which parts can give you the best bang for your buck.  You can find the February Buyers Guide here.

Avault has done a review of Turtle Beach's Montego II OEM audio card which is based on Aureal's Vortex 2 chip. If I were you, I rather go with a MX300 or Storm Platinum, especially considering they are the same price.

• Hercules Dynamite TNT Driver (Win9X - version 0.59.10109 - 2.24 MB)

 

"Coppermine" as it's code-named by Intel, will truly be a step forward in CPU processing power that no one will easily match, yet alone surpass. Offering the impressive specs mentioned above, along with a 600MHz core speed, Coppermine is the "next great thing" from Intel.

As if that wasn't enough, Sharky Extreme has learned that Coppermine's specification sheet now includes the addition of 256Kb of full speed L2 cache to the chip's core. Combine that with the increased amount of 64Kb of L1 cache, and you're talking about a 15 to 20% performance improvement versus a current Pentium 3 running at the same frequency level. The off-die half-speed L2 cache that currently resides on the P2 and P3's PCB wafers will disappear when the integration of the full speed 256Kb is begun with Coppermine's introduction.

• Abit BH6 Beta BIOS
• Abit BX6-2 Beta BIOS

 

The image is roughly equivalent to that of a 21-inch CRT display, yet the TFT8000 weighs just 20 pounds, in contrast to the 70 or so pounds a CRT weighs. Compaq has targeted the TFT8000 primarily at financial markets where space is at a premium; the monitor is also intended for executive and high-profile applications. Compaq provides a one-year warranty on labor and a three-year warranty on parts, including the backlight.

 

The first glance at the C300 shows that Diamond did their homework and found out exactly what the majority of users want in a motherboard - a well built board with lots of expansion that is also easy to setup. The C300 delivers in all those categories.

 

Intel now has 4 Processors in the production line; the Celeron (two formats, PPGA and SEPP), the Pentium II, the Xeon, and the Pentium III. It's getting a little crowded and it looks like Intel is about to do some Spring cleaning. It will probably be announced all over the web in the next few days, but it looks like the Pentium II is headed for Clearance City by the beginning of April. Expect massive price drops approaching 50% on the PII 400! That's nothing to sneeze at.

Our distributors are listing the Intel Pentium II with an arrival date of April 4th at just about half the price they are selling for right now. This is great news, becuase it puts the retail Pentium II 400 at the price range of the Celeron 400 right now! Oooooh. So for those of you considering upgrading to a PII 400 or 450, we suggest you wait just one more month, and you'll save a couple hundred bucks! We expect to offer BX6-2 upgrade kits based upon this processor, overclocked to 448, starting in early April for around $330! Stay tuned for more info. Kits made with the 450 overclocked to 504 would cost about $100 more.

 

I was really surprised by how well the headphones actually worked in A3D games like Half-Life, Wargasm and Sin and even Direct Sound/Direct Sound 3D titles like Starcraft and Delta Force allowed me to feel the force. Some may feel the force-feedback effects are cheesy, others will be wowed by it, but for an additional $20 you're getting a nice set of padded headphones for gaming or listening to music, which isn't a bad deal at all.

• Voodoo2/Banshee MiniGL Driver (Win9X/NT - version 1.47 - 148 KB)

 

• 3DFX Banshee Reference Driver (Win98 - 1.02.03 RC1)
• 3DFX Banshee Reference Driver (WinNT - 1.02.01 RC1)
• 3DStudio MAX 2.5 Compatibility Patch

 

• ELSA Erazor II Driver (Win9X - version 0301-0020 - 2.16 MB)

 

Being an Intel AGP 4X validation partner, S3 became the first company to announce an AGP 4X part. The Savage4 continues where the Savage3D left off with dual texturing pipelines, increased bandwidth, 32 MB framebuffer and AGP 4X support. There will be 2 chips in the lineup: Savage4 PRO and GT. Creative,Diamond Multimedia, ELSA, Gainward andNumber Nine have signaled their intention to build Savage4-based cards. With Savage4's support for 8-bit stencil buffer, Creative will want to implement real-time shadow effects.

 

The combination of the Gainward 6IBA and the Cardexpert 740 is a very stable combination. It's a fact that Intel made the I740 to work with a Pentium II and from what I've heard, the I740 seems more stable on a BX motherboard. At regular system speeds there was no need for extra cooling to prevent "freezing". While overclocked we added an extra fan to the CPU and a fan on the Cardexpert just to be safe.

 

It's a reality. K6-III silicon is now being fabbed in Texas. This chip comes with bold claims of performance equal to that of a Pentium III at the same clock speed. Can AMD deliver? Nine out of the top ten OEM's think so. Do you? In the K6-III, AMD touts a TriLevel cache design, 3DNow! Instructions (which has recently gained significant industry support), and an improved CXT core (first used in the K6-2). More with that later.

 

The problems AnandTech experienced with the original Savage3D seemed to be completely absent from the quick 4 hour experience Diamond let us have in the lab, definitely a promising sign. It seems as if S3's design has finally matured to the point where a release makes sense, and if Diamond is willing to put their name behind the product, it should be a noticeable improvement over S3's last attempt at a rise to 3D power.

 

According to this information, AMD will shift the manufacturing process of the K6-III from .25 micron to .18 micron in the second half of 1999, but has not yet decided whether to perform that miracle on the K6-2, as yet.

The physical limit of the K6-III's .25 frequency is 550MHz and they already have 550MHz running on this process using a simulator.

AMD will ship its mobile K6-III in the first half of this year but do not have a plan to reduce the consumption power, which will be 2.4 volts.

 

One feature that really makes the RAGE 128 stand out from the crowd is its hardware DVD decoding features. While it doesn’t offer full hardware DVD decoding, it has built in motion compensation and IDCT (Inverse Discreet Cosine Transform). The RAGE 128’s hardware features, combined with a software DVD decoder, allow for smooth DVD playback. Compared to the run-of-the-mill software DVD decoder, the RAGE 128 benefits because of its extra hardware DVD decoding features. With the RAGE 128, more CPU time will be available when DVD movies are being played, allowing for some multitasking (depending on your CPU speed). Unfortunately, for this preview, I was unable to completely test the DVD performance of the RAGE 128, but look for the DVD results in the full review when the RAGE 128 hits the market.

 

• Guillemot MAXI Gamer Phoenix Driver (Win9X - version 1.02.02 - 2.7 MB)
• Guillemot MAXI Gamer Phoenix Driver (WinNT - version 1.02.02 - 1.8 MB)
• Guillemot MAXI Gamer Phoenix BIOS Update (BIOS 1.10 - 366 KB)

Netscape's Netcenter has an article about the code-named Coppermine version of Intel's Pentium III CPU.  This is in fact the chip that should have been called the Pentium III as it has a redesigned core and is based on the 0.18 micron process.  Check out this tidbit from the article:

Coppermine will replace the 512K secondary cache that sits alongside the processor with an integrated 256K that's integrated. The shift is possible because Coppermine will be Intel's first chip made on the more advanced 0.18-micron manufacturing process, which allows Intel to squeeze more transistors onto a single chip.

While smaller, the integrated cache will boost performance approximately 5 to 10 percent, even without speeding up the processor, said Brookwood.

 

The Fast Network Kit comes with everything you need to get started on a 10/100 Mbps network.  In the box you will find 2 PCI 10/100Mbps Ethernet adapter cards, a 5 port 10/100Mbps HUB, 2-25' twisted pair cables, and a power brick for the Hub.    There was also a drivers diskette and an installation manual.  if you need to add more that 5 machines, you can easily upgrade with another Hub, which will very easily connect to the first.

While this may all seem a bit intimidating to you network newbies, I can assure you that you can go from the store to the game in about 10 minutes (not including driving time =)

 

It's important to remember that cards supporting EAX won't use Live's reverb engine and the quality of each reverb engine will be different. The CPU hit will depend on the reverb engine and the power of soundcard using it. It's really no point supporting EAX unless you can do it well and with a reasonable CPU hit. So the problem with supporting EAX is the creation of a good and efficient reverb algorithm.

 

…..And work it does. We saw several benchmarks running live, including Quake2, Incoming, Forsaken, and 3DMark99. First we witnessed a single Banshee card running through the paces, and in each case it performed as we're used to seeing it perform. When SLI was engaged, the image quality remained the same, but performance scaled upwards dramatically. We saw gains of 40%+ in most of the marks, and image quality was both consistent and satisfying. This is a similar boost that we've witnessed when adding a second Voodoo2 into a system and engaging SLI.

 

You guessed it, the K7s floating point unit not only has the power to match Intel's alone, but with the help of the extra cache and bus speed, it should beat it by more than a fair margin. I guess Intel decided they didn't need to innovate anymore, and smaller fabs would do the work for them… Oops. That and the fact that AMD has struck deals to distribute with nine of the top ten PC manufacturers, the most recent being Gateway. Although Dell has not given in yet, I'm not sure Michael [Dell] could resist all that power.

We can’t fault Promise for the lack of a measurable performance increase. The current selection of Ultra ATA/66 drives are just not fast enough to really utilize the new found speed. With a street price of $40-$50 the Ultra66 could be a very affordable speed boost when the faster drives appear. The only drawback is that it uses a precious PCI slot. Assuming Promise straightens out the compatibility problems we found, we would strongly recommend it for those wanting to upgrade to Ultra ATA/66. We plan to contact Promise and will update this review pending their answer.