   |
|||
|
|||
|
:
News
: Archives
|
|
|
|
|
|
|||
|
|
|||
|
|
|||
|
May 30, 2005 Happy Memorial Day! - Josh As you are enjoying this extended weekend, remember those who have given all so we may live as nicely as we do. I know that I will be doing a LOT of housework today as we have an appraisal on Wednesday! Hopefully the cleaning experience today will not cause my wife and I to have a divorce or something even more nasty! Cleaning is always much more interesting when you have a 2.5 year old boy running around causing havoc. That kid is a wrecking machine, no wonder I don't have any hair anymore! Anandtech has posted some of the first official looks at the G70 product from NVIDIA. You can read his entire analysis here. He says some very surprising things, and it appears as though A1 silicon was available as early as February. It seems that NVIDIA has been quite busy with this product, much busier than many probably expected. HardwareZone also takes a look at a nonfunctioning G70 card that they somehow got their hands on. This sample is one of the A1 silicon boards, and it sports a single slot cooler. The die size on that chip is tremendous! That is not all bad, as it means more surface area for heat transfer. Apparently quite a few of the Taiwanese partners are openly talking about the G70, so I figure the reports will start flowing in tomorrow once the show opens. On the ATI side there are already a lot of pictures of their Crossfire Multi-VPU setup. This requires a special Master card with a special connector which will attach to the DVI output of the Slave card. This does require an external dongle, unlike NVIDIA's SLI. Which one is better? I guess we will find out once the reviews start pouring in. I think that the competition is good, as NVIDIA SLI prices will go down (as they plan to cut the AMD SLI chipsets in price in the second half of this year). It will also probably push NVIDIA into more driver releases to support more and more titles of products out of the box (without having to create a profile yourself using the Coolbits 8 registry edit). There is a definite market out there for these high end products, and with the introduction of ATI's products, we will see better prices and greater availability of the product. That is all for now. Hopefully after I get the cleaning and painting done, I can edit an article and get it posted. Wouldn't that be nice? May 26, 2005 Thoughts on Confusion - Josh Well, my article on the G70 was really just an exercise in speculation. While I don't have any concrete's on the G70 (and I haven't been briefed yet about it) I am probably wrong about it being 90 nm. While Marv did make the statement that "virtually all (large) parts will be 90 nm from here on out", Dave Baumann from Beyond3D made a very pertinent comment. Basically, from Marv's perspective, the costs of the G70 tape-out and production happened late last quarter. What he was saying is true for him, but not true for us. While internally NVIDIA had already finished the G70 and sent it for production, we still haven't seen it "introduced". What we can deduce from his presentation though is that TSMC's 90 nm Low-K process is quite healthy, and they are very happy with the direction that it is taking. Marv repeatedly said that the people over at TSMC did an excellent job with the process, and NVIDIA couldn't be happier with the results. This is quite a turnaround from a couple of years ago when NVIDIA was one of the first TSMC customers to release a large part on their new 130 nm process. In between some design library changes that happened earlier that year, and NVIDIA's overall design of the NV30 (which as we all know, was not a great design, and was really hard to effectively fabricate), the 130 nm node had a rough start. Now, something that is not widely known is that TSMC's 130 nm Low-K process was only used in one of their Fabs, and they were very production constrained with the parts made on that process. The most notable are nearly all of ATI's parts at one time. Last year pretty much every single major ATI graphics chip was using 130 nm Low-K, and as we know from the extreme lack of higher end parts, TSMC was not able to supply the demand. This shouldn't be a problem with 90 nm. As I said, there are three separate Fabs that are producing 90 nm Low-K parts. Apparently TSMC sees this as their next major product offering, and I think most of TSMC's large customers will happily use this process. While the jump to 90 nm is complex (much more so than the jump to 130 nm), the engineers at TSMC have done an excellent job in releasing this process. ATI is already receiving production quality R500 silicon, which is good because the release of the X-Box 360 is not all that far away (in terms of production schedules). I bet by early September that Microsoft will be in full production of the X-Box 360 and will be stocking up for the Late November release. Just as the 130 nm product from TSMC and IBM were the process Du Jour for the past two years, 90 nm is going to be around for quite some time. While Intel and AMD will be transitioning to 65 nm next year, the graphics industry won't be following that trend anytime soon. The jump to 65 nm has some real nasty hurdles, and it really takes a company the size of Intel or the cooperation of two companies like IBM and AMD, to be able to make that jump successfully in such a short period of time. Intel has been in full 90 nm production since December of 2003, and AMD started full 90 nm production almost a year ago. So we are seeing a full 2 year stretch between the major node transitions for the big companies. For companies like TSMC and UMC, who really have only started full 90 nm production, we can expect them to service their clients with this primary process for at least until 2007, at that point we will see their transition to 65 nm. Now, TSMC will probably do a halfway shrink (80 nm I believe) just as they have done with all their other major nodes- 180/150 nm, 130/110 nm. Nobody outside of the confines of TSMC knows exactly when they will transition to this "budget" process, but I would expect another year and a half before we see anything. So, basically it looks as if I was onto something when I wrote this little article about Slowing Down the Process Migration last year. I personally think process technology is fascinating. It just boggles the mind what they are able to do. When it comes to electrical engineering, I am very much a "rub two sticks together to make fire" kind of guy. When I play a 3D game, or write an article on my computer, I am honestly just amazed that it all works. And it works very well. This really is a fascinating industry to watch. Intel Releases 820D and 670 Processors - Josh Well, the NDA's are lifted, and we are starting to see the reviews of these new processor products from Intel. The 820D is the dual core Pentium 4 running at 2.8 GHz, while the 670 is the 3.8 GHz 2 MB L2 Pentium 4. Budget enthusiasts are looking at the 820D with great interest, and the pricing on this product is actually quite impressive. These also cover the release of the i945 chipset. Here are a couple of reviews so far: May 22, 2005 Mass Confusion About G70 and Other Upcoming NVIDIA Products- Josh Usually the press tends to avoid many of the "technology conferences" given by financial organizations, but every once in a while a gem is uncovered in these presentations. Once such gem was given by Marv Burkett, CFO of NVIDIA at the recent J.P. Morgan technology conference. After going over some of the financials and the company outlook, he was asked a very good question about NVIDIA's move to 90 nm. His response opened a floodgate. I cover his comments, and their ramifications to NVIDIA's product lineup that will start to be unveiled shortly. Here is a quick quote:
You can find the rest of the article here. May 20, 2005 What a Month! - Josh Well, after a very successful April, real life has intervened again and has destroyed my article plans for May. We are going through a nasty refinance at the moment, and much of our time is being spent at our daily jobs and fixing up the house (painting, cleaning, yardwork, etc.). This has left little time to compose articles based on testing that I have already done. That is unfortunate, because I have almost finished testing on the SLI package NVIDIA sent (I have some interesting findings that I haven't seen widely published as of yet), I have finished testing of Sapphire's new Ultimate X800 XL (this one overclocked very well), and a MSI K8N Neo 4 Platinum motherboard. So, I have plenty to write about, but I just haven't had the chance to actually sit down and bang these out! Also, I need to edit Scott's review on an Antec case, because he has already submitted everything to me (I just need to get it looked over and posted!). Well, this weekend is going to be more housework, but I will see if I can't find the time on Sunday to post Scott's article, and maybe get a bit of writing done on my own. We can only hope! The unfortunate thing about this time of year is all of the technology "outbreaks" that we have seen in the past three days with the major console introductions from Sony, Microsoft, and Nintendo. The technologies in those products are actually quite astounding, and very impressive. One area that has been glossed over is their audio capabilities. I know SiS did a lot of work on the Southbridge/Audio portion of the X-Box 360, but as to what it can do and support... we have no idea. Sony is a bit different, as they haven't released anything about their audio portion. I am very, very curious though if NVIDIA will also be handling those duties for the PS3. It would make sense if they did, as they have the talent and engineering staff to do it right! I will have to ask my friends at NVIDIA and see what they have to say. If you are curious, we are having a nice discussion on the graphics portions of the consoles that have been showcased at E3 this year. So please sign up and add your comments! May 18, 2005 They Finally Leaked It! - Josh NVIDIA has had the wraps on their upcoming product, and hardly anyone has heard anything concrete about the next gen of products from that company. Now, much is known about the ATI R520 that will be unveiled sometime this summer, but the speculation has been rampant about NVIDIA's new part. It seems that my basic ideas about the G70 are matching up with reality! SH SOTN was kind enough to alert me to what looks to be official marketing materials for the next generation of NVIDIA cards. The product that has been leaked is the GeForce 7800 GTX. It is a 24 pixel pipeline design with 10 vertex shaders made on TSMC's 110 nm process. I had originally thought that the clockspeed would be around 450 MHz to 500 MHz, but for this SKU the clock is at 430 MHz. Now, note that this is the GTX, and not an Ultra branded product. If NVIDIA releases an UltraX (or something like that) then we can expect the higher clock speeds. It will of course be PCI-E, feature a 256 bit memory bus, and probably have 512 MB of memory (though the fast GDDR-3 is not out yet, so we can still expect to see the memory clocked at a pretty standard 525 MHz, which is 1050 MHz effective). It will continue to have the 64 bit filtering and blending (FP16) for advanced rendering such as HDR. One area that catches my eye is the little bit that mentions "Intellisample 4.0". Now, the current 6800's have Intellisample 3.0, so hopefully NVIDIA has totally revamped their AA and AF unit. Hopefully this is the case, and not a simple change without adding a significant amount of quality. I would certainly like to see NVIDIA match ATI with their programmable multi-sample unit which can handle up to 6 samples per pixel (currently NVIDIA can only handle 4 samples per pixel without resorting to super-sampling). Now, these specs might in fact be faked, and it is entirely possible that it is the case. I am not sure, but it certainly smacks of what you would expect with this upcoming generation of products. Jen-Hsun Huang of NVIDIA has hinted that we might see a peek of the next gen of products at E3 (which is going on at this moment). Again, this is not official information, and it could very well be a hoax. However, it just seems somewhat true to me, and there are nothing in those specs that are outrageous or appears to be inaccurate. I guess only time will tell! May 17, 2005 Sony, Microsoft, NVIDIA, ATI, and the Future of 3D Graphics - Josh E3 is here, and Sony and Microsoft are showcasing their next generation consoles. It just so happens that the two 3D graphics giants have squared off against each other this round, as Sony has procured NVIDIA's help to develop their graphics chip, while ATI has already finished and produced the X-Box 360 graphics chip based on the R500. The RSX, which is the graphics portion of the Sony PS3 is being developed by NVIDIA, and from what I understand, the chip has not taped out yet. I would imagine that the design will be finished by the end of this summer, and production will start before the end of the year. This deal will be much different from NVIDIA's X-Box contract, in that Sony paid NVIDIA to design the part, and then bought that design from them. NVIDIA will receive royalties from Sony for the use of its design, but NVIDIA will not be producing the end product and selling it to Sony. This allows Sony a lot more flexibility in manufacturing the part, and while NVIDIA's revenue will not be as high as it could have been if they had a contract as they did with Microsoft originally, there should be a lot fewer headaches with this setup (and NVIDIA won't have to cover legal fees from price disputes). The next consumer level chip from NVIDIA is rumored to be the G70, but from what I gather it is not based on the RSX technology. The G70 looks to be an extension of the NV4x architecture, mainly that it can handle SM 3.0 better, and will feature faster clock speeds and more physical pipelines (both pixel and vertex shader pipes). The RSX is a step further, but it still looks to be based on a highly revised NV4x technology base. It is a FP32 based chip with separate pixel and vertex shaders, but there will be significantly more of them than in current designs. The RSX is over 300 million transistors in size, and will run at 550 MHz. Since the PS3 is a very small form factor product, I would imagine that heat is not a huge issue with this design. This bodes very well for future NVIDIA products based on 90 nm technology! The current target for the single chip design is that it can put out 136 shader operations per cycle, as compared to the current 6800 chips that can do 53 ops per cycle. NVIDIA is touting this as being far more powerful than two 6800 Ultras in SLI, and from all indications I would guess that this is true. The only question that still remains is if NVIDIA will be able to utilize the technologies in its RSX chip for some of their own future products, or does Sony actually own much of that technology? ATI is in a bit of a different situation in that they have actually finished the X-Box 360 design and it is up and running in developer machines. The demos that Microsoft showed off were probably true renders of upcoming games, as compared to what Sony and NVIDIA showed off (which are most likely "approximations" of what the games will look like on the final product). Microsoft started shipping X-Box 360 developer hardware earlier this year, and they were comprised of 2 x PowerPC processors with a R500 board installed. These developers have had quite a few months with hardware on hand, and there will be quite a few titles available at launch (which is supposed to be around the Holiday Season this year- which probably means late November). ATI and Microsoft are heading in a slightly different direction than Sony and NVIDIA. The R500 will be the first unified shader product to hit the consumer market. This simply means that the physical pipelines can handle both pixel and vertex shader operations, depending on hardware resources and what the application wants. Microsoft has been pushing for a unified model for some time, but many in the industry do not think that it is time for that kind of jump. I have seen many of the pros and cons of a unified architecture, and I would have to agree that at this time the technology and design processes are just not in place to have a unified product, yet still perform well. ATI obviously agrees with this at the moment, as the upcoming R520 will not feature a unified architecture. Unified will come eventually to the PC, but just not this round. NVIDIA is so far concentrating on the separate vertex and pixel shaders, as they say the transistor budget at this time will not really allow an effective unified design. Since the X-Box is a console, Microsoft was not held by the constraints of the PC desktop market, so they could choose to implement a unified architecture. I think currently a unified design would lag behind a traditional design when considering the use of transistors and the overall efficiency of the separate pixel and vertex shader pipelines. Both products are undeniably impressive, and the consumer will win no matter what. I believe we are now reaching a stage that both hardware and software design will deliver content that was previously only seen in the movies. The hardware will be able to push the visual envelope to a great degree, but the upcoming technologies and tools from the software developers will enable incredible visuals. Developers will be able to utilize these technologies and tools to create amazing environments with the same amount of manpower that it took to develop older titles. Unreal 3 is an excellent example of this. Not only does it feature some amazing technologies, but Epic was able to bundle the product with some outstanding tools that gives these developers more time to create the environments and apply artistic and realistic touches to a title without having to spend their entire budget on programmers to make it happen. The tools being written will allow more artists to be hired, and hopefully will allow a greater focus on gameplay and environment. Looking at the demos that were shown, we are about to take that big step into truly cinematic computing. While Far Cry was one of the first shots across the bow, and Doom 3 and Half Life 2 quickly followed, by early 2006 we will have the hardware and software to see some truly awe inspiring visuals come off of our TV and computer screens. I am quite tempted to start saving my pennies for both consoles when they eventually hit the market! Not only will the hardware be impressive, but the convergence of so many entertainment appliances into one (or two) boxes is a very stimulating idea. MP3 player, HD-DVD, DVD-Audio, photos, and nearly any other media you can think of will be playable on these boxes (well, HD-DVD doesn't appear to be supported by X-Box 360... but it isn't set in stone yet). Definitely fascinating times we are living in! May 12, 2005 AMD is of course making some big splashes these days, and the Athlon 64 X2 release was the biggest event this spring it seems. AMD will really push the X2 to the productivity market, but it will continue to present the single core chips to the gamers and enthusiasts out there. The FX-55 will continue its dominance as the highest speed processor out there until the release of the FX-57 in the Fall (or so I hear). The X2s are supposed to be available in June, but only through a few select partners. By August these parts should be widely available to users. Now, currently these X2 parts are all aimed at the socket 939 platform, and none will show up on the Socket 754. By early next year though, AMD is planning another socket migration. This is an area where some people are really unhappy. In the past AMD has been great at keeping the socket infrastructure consistent, we can look at Socket 7, Slot A, and the very long lived Socket 462. Since the Athlon 64 has been released, on the desktop side we have already gone through the Socket 754 and now are into the Socket 939. The Opteron has held steady at 940 pins, but that is going to change as well. Socket S1 is intended for the mobile sector, and it looks to be a very small version of Socket 754. This will be low profile, and aimed primarily at the low power/mobile sector. Socket M2 is the desktop replacement for Socket 939, and it has a total of 940 pins (though it is not pin compatible with the Opteron Socket 940). AMD is planning on using this socket with DDR-2, though there has been a lot of discussion that AMD may skip DDR-2 altogether (though I doubt DDR-3 will be ready in volume by Spring of 2006). This supposedly will give AMD the legs to continue to offer dual core Athlon 64's with plenty of power and bandwidth. In many ways, Socket 939 was a bit overdesigned and could handle some pretty hefty power loads, but it is just not smart to retain the same socket yet transition to DDR-2 (I can foresee many fried chips with a DDR memory controller trying to work with DDR-2). The other socket that AMD is going to release is the Socket F, which is comprised of 1207 pins for the Opteron. This will also feature DDR-2, but one has to question why so many pins when the current dual core Opterons work fine on Socket 940? I think there are two distinct possibilities here. The first is that faster dual core chips will consume more power, so more power and ground pins need to be included for future products. However, with the current 90 nm dual core products, we won't see them scale much higher in speed for the next year, and that is nothing that the current socket can't handle. When AMD makes the jump to 65 nm, those power requirements will go down. So, why so many extra pins? The first explanation is that AMD will be going with a quad core product for the server market when 65 nm is reached. The quad core at 65 nm would be about the same die size as a dual core at 90 nm (this is pure speculation, but we can approximate looking at how the design has changed in the past). However, quad cores running at 2.6 GHz+ will eat up a lot of juice, no matter what process is being used. But does this truly account for the greater amount of pins? This leads to the second spot of speculation. Will AMD release the first CPU to introduce a 256 bit memory bus for DDR-2? If AMD is in fact going for a quad core, then the current 128 bit memory controller will be very bandwidth limited, even with high speed DDR-2 at its disposal. By making a 256 bit memory bus with DDR-2 800, then a dual core or quad core Athlon 64 core would be happily fed with bandwidth, all of which features the low latency of an integrated memory controller. This bit of speculation leads to other big questions, such as will DDR-2 receive a new pinout (can you imagine a 400+ pin DDR-2 module)? Or will AMD simply require these high end machines to utilize 4 DIMMS to enable the full 256 bit path? While this is a bit far out, I think this is very possible. Would DDR-2 and extra power and grounds really take up all 267 extra pins? I think that another 128 data pins for the memory, plus the extra 139 pins for power and ground would fit quite nicely into that 267 pin budget. If this is true, then the server people must be drooling gallons by now. Imagine a dual socket motherboard with 2 x 256 bit memory paths which can host quad core processors. 8 logical processors in two sockets, with 12.8 GB/sec of bandwidth per socket (assuming this is DDR-2 800). That would be pretty amazing, not to mention very cost and space efficient. The best part of this speculation is that it can physically be done, and there are few barriers standing in AMD's way to implement such a design when they migrate to the 65 nm node. Looking at AMD's prices for their top of the line dual core Opteron products, I can only imagine how expensive a quad core would be... Far Cry Goes 64 Bit - Josh Oh geeze, more AMD news. The nice folks at Crytek, with the help of AMD, have released their 64 bit version of Far Cry. Not only does the game run at 64 bits, but it also has quite a bit of new content as well as graphical enhancements. This of course will only work with a 64 bit processor and a copy of Windows XP 64. You can find out more info here. G70 at Computex? - Josh Digitimes is reporting that the G70 from NVIDIA will be showcased at Computex. This certainly looks to be legitimate, and I don't doubt that NVIDIA has working silicon that they can show off successfully. Volume deliveries are not expected until the second half of Q3 though. From all indications it appears as though ATI may hold off on their R520 release, though they have working products that they are shipping to partners and game developers (I noticed in a version change log on a game that says "Fixed rendering problems with SM 3.0 on ATI products"). It also appears very certain that the G70 will be 110 nm, though for a time I had seriously wondered if that would be the case. There has been a lot of misinformation spread about this product, and if you look around you often will see conflicting information. Overall the card will indeed consume more power than the current products, but it will not break the 150 watt envelope that PCI-E can deliver. G70 will also probably be introduced to the AGP infrastructure, so it will still need to fall under those power guidelines. Cooling will probably be very similar to the current 6800 Ultra, but most likely slightly beefier. I certainly can't wait til the end of May! May 9, 2005 AMD Previews Athlon 64 X2 Processors - Josh In a bit of a surprise move, AMD sent out X2 4200 and X2 4800 processors to reviewers even though the X2 processors will not be available until June. The X2 4200 runs at 2.2 GHz with 512 K L2 cache per core, while the high end 4800 runs at 2.4 GHz with a full 1 MB of L2 per core. The usual suspects have these processors in their hands, and the results are damn impressive. The dual core parts are based on the same technology found in the Venice/San Diego cores, with all the optimizations and features that those exhibit. SSE-3, better memory support, better prefetching, and a host of other additions make this a very speedy core. In many of the single thread tests the X2 4800 nearly matches the performance of the FX-55 (Clawhammer, 2.6 GHz, 1 MB L2), and of course destroys everything else when multi-threaded applications are used. Many of the reviews have it set against Intel's dual core offerings, and the X2 series just seems to spank them around for the most part. AMD has certainly been on a roll since the release of the original Athlon 64 3200+, and they haven't slowed down one bit. AMD has been planning on moving their products into the dual core realm, and their current products reflect that fact very well as compared to Intel's dual core products. It seems that Intel has been in a "damage control" situation since their release of the Prescott core, and the way they implemented their dual core strategy really reflects that. The Intel design, while still impressive, is not nearly as integrated as the AMD product. These processors will not be available until June, and only then through a limited number of system integrators. General availability is expected to show up around a July/August timeframe (though some are suggesting that it will be later than that). The dual core product has a die size of around 199 mm square, which is around the size of the original 1MB L2 Clawhammer. AMD will continue to produce their latest Venice and San Diego cores for quite some time, as AMD does not plan to transition fully to dual core until next year. AMD will essentially have to wait until Fab 36 is complete with a running 65 nm process before they could really afford to offer dual core to the masses. For the time being, the single core Athlon 64's will be the mainstream choice for processors. One aspect that many reviewers have brought up is the "creamy smoothness" of dual core/Hyperthreading in daily work. I have been using an Athlon 64 3700+ for the past 8 months or so, and I have rarely run into any issues with my work and this computer. For those that are constantly doing many things at once (ripping MP3's, DiVX, etc.) while doing other work, then the dual core/Hyperthreading aspect may be important. For my work (email, web browsing, html editing, photo editing, etc.) I tend not to utilize more than one heavy application at a time. And when I game, I definitely do not have anything running in the background (because I enjoy being in control of things and supervising the work that goes on). So, before getting sucked into the whole "I have to have dual core/Hyperthreading or else my computing experience will be ruined" think about your usage of the computer, and what you expect out of your processors. Since good performance in single thread applications and gaming are my two big needs, I found that the single core Athlon 64 has been a good bet as compared to going to a dual core machine or using an Intel processor. Again, not everyone has the same needs as myself, so mileage will vary. Still, it is great that AMD does have a consumer level dual core product that will compete with Intel in a feature for feature race. Hopefully one day I will experience the bliss that the X2 4800 is supposed to imbue upon a user! Then perhaps I can give a better idea of how usage can work with a dual core product. Here are a couple of reviews to tide you over: May 6, 2005 Exceptionally Happy it is Friday - Josh Wow, what a long week here. Refinances are not fun at all, especially the waiting after all your stuff is in. That is essentially the reason why no new content has been hitting the site, but things will change this next week. I have apportioned off some time to finish a couple of projects, as well as do some editing and writing. So, hopefully starting late Sunday, we can see something new around here! Hexus has the goods on ATI's upcoming MVP technology (their version of NVIDIA's SLI). The expose a couple of new things that nobody has talked about before, and it definitely appears to be legit. Go take a look at it here. Albatron is showing off a unique cooling solution for their GeForce 6800 product, and for those wishing for something a little different, as well as pretty quiet, should take a good look. After reviewing their 6200 TC product, I have been very impressed with what Albatron can do in terms of their quality. Very solid stuff that I can see so far. BFG, another favorite company of mine, is giving away some really nice little prizes for the next two months. The first is a very impressive GeForce 6800 Ultra 512 MB card (worth $999 in retail). They will also be giving away the BFG GeForce 6600 GT OC and their nForce 4 motherboard. Several of these will be given away, and all of the gory details can be found at BFGGaming.com. These can be had if one decides to play CS:S on the BFG servers, or participates in Folding for the BFG Team. My friend "Battlemaster" runs these servers, and he does an awfully nice job at it. That is all the really interesting stuff in the mailbox, so have a great weekend! May 4, 2005 512 MB Cards for the Masses? - Josh It seems that today is the day that the NDA's are lifted on the 512 MB X800 XL's from ATI (or more specifically, from Sapphire and its other partners). Some months ago ATI showed off the 512 MB X850 XT PE's, but then told us that the card would never be available in retail. It was more of a PR and engineering exercise than anything else. NVIDIA on the other hand is shipping quite a few 512 MB GeForce 6800 Ultra to OEM partners like Alienware and Falcon Northwest. To keep the video card Cold War brewing, ATI is releasing their midrange X800 XL in 512 MB configurations. In most applications there is very little difference in overall performance between the 256 and 512 MB X800 XL's, with the only exception being Half Life 2 at high res and AA/AF enabled. Even Doom 3 at Ultra Quality is not showing the differences between the cards (though it purports to utilize 500 MB of texture data). Looking at the results from several reviews, I have a dread feeling that ATI's drivers are not adequately taking advantage of that extra memory. Either that, or the X800 XL just doesn't have the horsepower to really take advantage of that extra memory. From a trusted contact of mine, I know that he was seeing up to a 20% increase in overall performance in apps like HL2 and Doom 3 with the 6800 Ultra 512 MB version at higher resolutions an AA/AF enabled. So I'm not really sure what to think. Yes, we will eventually need 512 MB cards for future applications, but is the time right now? Currently both the 6800 Ultras and X800 XLs at 512 MB utilize 16 GDDR-3 1.6 ns chips, and that creates a significant price increase over the standard 256 MB configuration. The 6800 Ultra 512s are expected to be in the $800 to $900 range initially, while the X800 XL 512 MB is expected to retail around $450. So, looking at these prices, is it better to buy a faster card at $450, or to get the bigger memory load? It is a tough question, but my honest answer is to go with the faster card and not worry about having the extra memory. While the extra memory will probably become important in another year or so, we are still seeing 128 MB cards perform quite admirably in many of the latest games (namely the 6600 GT). So, if I had a choice between a X800 XL and a 6800 Ultra... I would probably pick the Ultra and get the extra speed (or even the X800/X850 XT for that matter). Here are a couple of reviews. AMD Processor Stuff - Josh It seems from hints being given from a couple of sites around the web that many of them have the Athlon 64 X2's on the test bench right now. I certainly don't have one, but of course wish I did. Still, I am really busy as is, so when would I have the time to test this new processor? I already have a lot of my table! Still, it is good to see that AMD may be in a good position to release new chip updates on the desktop to keep up with Intel. The new Venice cores are getting the good one-two action at a couple of sites, and the most notable one is the review at Lost Circuits. MS sacrificed several motherboards for this review, and the results are really quite impressive. Not only does he take a look at the overall performance of this new core, but he also takes a very close look at the power and heat properties of the 3800+. This is a must read article. Once everything is put into perspective, we see how impressive the Venice core really is for AMD. It is a very power efficient, yet high performing part. It is simply amazing what AMD has done with this core. Marco at Hot Hardware also has a review that is a good read as well.
If you have found this article interesting or a great help, please donate to this site.
Copyright 1999-2005 PenStar Systems, LLC. |
|||
