Slashdot Mirror
Dual GPU graphics solution from ATi?
Graphics Guru writes "Last week TweakTown posted an exclusive picture of the ATi Radeon 8500 MAXX with believable accompanying information also regarding the highly anticipated ATi R300. 3DChipset is today reporting that they have confirmation that the 8500 MAXX is indeed real and is due to be shipped fairly soon. Here's what someone from ATi told them: "The ATI Radeon 8500 Maxx is for real and the card is already in full production and about to be shipped soon. ATi has finally nailed certain issues with the dual chip. Final testings have been done and you should here noise from ATi regarding this offering." You decide if it is real or not, a solid dual GPU solution would surely rock the industry to massive proportions!"
You decide if it is real or not, a solid dual GPU solution would surely rock the industry to massive proportions ...as soon as the next version of the drivers come out, I presume. This is an ATI product, remember.
--saint
I wonder if this online petition had any effect...
Do you even lift?
These aren't the 'roids you're looking for.
It will have less features than the R300, but probably more overall power because of the two chips. I.e., it will probably be a great solution for today's games, but not next year's.
And the dual chip approach will not be cheap.
With current Graphics Card technology the Memory bandwith is becoming less of an issue compared to older cards and the GPU tends to be the bottleneck. This can really open up the doors to some killer crazy cards....
This GPU plus those 8500 OpenGL drivers being paid for by the weather channel should make a kick ass system for Doom III.
2 GPUs: that means 3D viewing. One for red and one for blue. I better find them goggles.
DNA is the ultimate spaghetti code.
Comment removed based on user account deletion
The technology is cool, but I'm afraid of the impact this could have on the Linux world. Drivers for this monster are no doubt going to be part of Windows XP from the get-go, but it'll probably take the XFree86 team a few months to get a Linux driver out.
I think if Slashdot wants to do the Linux world a favor, then it should stop playing up the new great hardware, and instead focus on the reliable, established tech that is well-supported in Linux. Otherwise, we are just driving people to Windows.
Karma: Good (despite my invention of the Karma: sig)
UntiL ATi makes their own *nix drivers, im stickin to Nvidia. Those 3rd party drivers for ATi are pretty shabby and not worth anybody's time.
-=Crazy stuff happens w/ the Bong and me.=-
l33t...
a solid dual GPU solution would surely rock the industry to massive proportions!
Which industry would that be? The gaming industry is slowing down as far as graphics go. Mark my words, there is going to be a shift soon from graphic intensive to gameplay innovation. People don't want games to be any prettier (or don't notice much of a difference). Notice how the mod community is getting bigger and better? Its cause they take the graphics engines and add innovation.
I'm rambling, but I think that these new video cards aren't going to be this big explosion that they were in the past. Sure they are big and powerful, but people aren't going to fork over the cash to get this one when they can get a good GeForce2 that can play their games just as well.
Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
This means scientists using GPU's for in silico experiments don't need a second computer anymore.
They can do the duplo with the second GPU.
Right, like all those dual CPU desktops I see.
--
E_NOSIG
Errr, install Linux, maybe?
The way to solve a Windows98 problem is to remove Windows98. Problem solved.
is the quad version coming out?
Great Linux Site
"Optimization, shmoptimization! Just cram a second GPU in there and we'll be fine."
I really wish people would just stop coming out with new hardware for a couple of years, so that we can all save a few upgrade bucks, and the software industry can get their act together, and start writing clean, well optimized, stable programs, instead of trying to always catch up to the bleeding edge that nobody really asked the hardware companies to push.
Ñ'
This has been around for a while on Wildcat cards - you know, the ones that cost many thousands of dollars. If you're talking about the "industry", though, you can't leave them out.
Here's a URL:
I would be interested to now how this architecture would work.
...
Would the piplines be split between each processor? or do they share the entire pipeline architecture?
Or do they have differently pipelines for each GPU?
All this questions....
"The large print giveth, and the small print taketh away" -Tom Waits
The Nvidia Geforce 4 4200 generally runs most things about 50% faster than a radeon 8500 and costs less. It's definitely the best value gamer card on the market.
_ 42 00.shtml
Here's a good article with some benchmarks on this great value card.
http://www.hothardware.com/hh_files/S&V/abit_ti
This is a nice concept card, but it's not going to put ATI on top.
money for moogles
---
I support spreading santorum
ATI and Apple have been chummy lately, releasing updated drivers every few months. The release date for this has been reported on other sites(The Inquirer, I think??) as July 18th, 2002. The day after his-Steveness' keynote. Coincidence? Probably not.
Ummm... ATI had nothing to do with the voodoo 5, you know that, right? Anyway, the ATI driver support is supposedly improving. Maybe if they concentrated on making solid drivers about figuring how to make things look faster in Quake3, they would have a better reputation on the market.
Ñ'
But is it going to be Cg compatible?
Maybe the next computer animated movie will be real time rendered in the theatre by a computer using this card.
1) ATI has a history of lying about driver support. The Rage Fury MAXX was released around the same time as Windows2000. I, like many other people, believed ATI when they said that they would develop drivers for the platform, that they were just around the corner, so that buying this card over the already-Win2K compatable NV cards would not be a waste of money. HOWEVER, ATI left the Maxx buyers high and dry when they announced, several months later, that they were incapable of producing Win2K drivers and had given up trying.
2) They consistently LAG the cards produced by NVIdiea by wide margins and, while trying to stuff a bunch of useless features down dev's throats, lack the freedom of a programable GPU that the Geforce3 (and higher) cards offer... the cards that game developers are all optimizing their hottest new titles for.
Want a little evidence of how badly ATI stinks? Tom's hardware did a great job testing all of the cards in one huge benchmark here.
If you were paying attention to John Carmack you'd know what cards are already bringing it to life. Specifically the Geforce4 TI 4600, currently appears to be the Quake God's best reccomendation, though Radeon's 8500 series is rumored to be what was running the Doom III preview at the recent E3.
Reading over Carmack's finger, one can see that he is currently having the most success with NVidia drivers and the card will therefore not only be fast, but should support every feature he builds in.
I'm currently building my own system that is focused on Doom III. My best reccomendation for a graphics card is by Gainward; the GeForce 4 PowerPack! Ultra/750 XP Golden Sample. It retails for right around $350.00. However (as far as I know), it still holds the best benchmarks of the GeForce4 series.
There is something to be said about Carmack's opinion on ATI hardware (which is good overall), but for the sake of assuring your computer will be running Doom III perfectly I must stand by the opinion that the GeForce4 TI 4600 is what you need...
...that allow XFree86 drivers to be built and work potentially as well (or better) as their proprietary couterparts (or even better, if they work on Free Software drivers), I'll gladly buy this card. If not, oh well, my Matrox Millenium G200 still has pretty good 2D, and 3D is just about a tad slower than geforce's Free drivers (not the proprietary ones), so its a win-win situation... they sell one more card, and I finally can enjoy good and decent 3D :)
Cheers
about to be shipped soon
Not that I know anything personally, but when's MacWorld? Oh, next week? Hm...
- It's true
- It is a viable product
- reliable drivers become available
- people buy them
How much wasted GPU cycles will there be? I mean, even die-hard gamers don't do it 24/7/365...So if a few thousand get out there in desktops that are only taxing these cards a few hours a day, how long until somebody writes a distributed processing app like seti@home or The Great Internet Mersenne Prime Search to run on these GPUs?I mean, if people can trick the TCP stack into doing distributed math, they can certainly trick these GPUs into doing it to...
Yeah, they said this about the final Kiss tour as well.
hear hear,
ati drivers are bollocks. as well as matrox.
i'm glad someone points this out everytime there is a story about these jokers releasing some whiz bang hardware that probably wont work cuz their drivers are poop!
-
uberfag, and proud of it.
I have a Rage Fury Maxx which I use under W98 with no problems. I also run it under SuSE Linux (although the Linux driver only uses one GPU, I think). You can forget about trying to use the card under Win2K, NT, etc.
What's the solution then?
Gas chambers?
I *just* purchased an nVidia GeForce4 Ti 4200 a few days ago! I hate that one lemma that states that the new hardware you want will not come out until you've paid top-dollar for what's currently on the market.
Why bother.
Isn't this throw-more-chips-at-the-problem, require-more-power-produce-more-heat-slow-down-inn ovation-in-everything-but-raw-speed idea a good part of what killed 3dfx?
is competition good, or is duplication of effort bad?
I know this was a troll, but there is a driver for raedon cards.
Sure it is not specific to the latest card in this slashdot article, but I'm sure the drivers will be made to work with the new Raedon soon enough.
I only saw one comment that came near this so far. If everyone here will recall 3dfx's parting shot before they left the industry was a dual gpu card called the Voodoo 5 5500 and they were working on a quad card also. With nvidia and ati being so competitive in the field it pushed out a good company from the market place which slowed down the innovation in the industry. Before the Voodoo 5500 came out there was the wildcat cards, they are still going for outrageous money for high end CAD systems. They had the ability to tie together multiple cards/GPUs and split up the video processing power. The only real drawback was you had to be running Windows NT4 no other drivers were ever released for them.
I say all of that just to come back to the point who really needs more than 60 fps anyway? Your eyes can't tell the difference much over the 40 - 50 range. So who really cares if your new GeForce4 can do 120fps in quake. My old voodoo 5500 does 40+ in very nice color.
I read a lot of people complaining about ATI. I think people need to put a little perspective on things. NVidia came out of the blue and used their superior 3D chipsets to grab the mainstream Video market. ATI's response was slow at first, but is really gaining steam. I've got a Mobility M4 in my laptop with great new OpenGL drivers. In my home PC I've got a Radeon 8500 LE that runs 99.9% of all the DirectX games. In the case of the former, my 2D performance was the biggest factor. In the latter, the price gap for comarative performance was a joke. $99 for the ATI Radeon 8500 LE (NewEgg.com) vs $180 for a Nvidia GF4 4400. NVidia is now using their market dominance to bleed the market (a familiar strategy that eventually backfires). Not to mention the beautiful All-In-Wonder I bought for my parent's computer that has the best MultiMedia and TV Tuner I've seen (DScaler ain't bad but can't seem to pick up as many stations).
For all you GNU/Linux junkies, ATI has been much more forthcoming in information for developing XF86 drivers than NVidia(proprietary binary only).
"God fights on the side with the best artillery." - Napoleon, Marshal of France - speaking truth to power
when they get the drivers straightened out, I'll think about buying one.
Here's a picture of it off the forum:
http://www.jeffwilhelm.com/files/r250.jpg
Might wanna hold on to that cash :) Carmack himself said that by the time Doom III is out there will be a new generation of cards that are much more suited to running it.
Ñ'
Hey Cat,
r o/ ragefurymaxxdrivers.html
The rage maxx Fury had a 98 driver, in fact im running it on my only 98 box in the house. Did you mean a 2000 driver? That doesnt exist, YET.
The driver released was 95/98/me only
Get your fact straight before posting.
In Fact heres the link,
http://www.ati.com/support/products/pc/rage128p
I'm currently building my own system that is focused on Doom III.
Why?
If you wait until the game is released you can buy the same hardware for half the price.
I got a radeon 8500 right now and I regret the day I ever bought it. No 3d acceleration support in linux. Until they fix that the next card I buy will be a nvidia. Anyone know when ATI plans to make drivers for their cards for linux?
"ATi can eat a dick with their crappy cards."
Dude, an 8500 can stand up to a geforce 3 ti500, which is what it was built to do.
Plus, if ATi wasnt around, can you even imagine what Nvidia could do with their prices, they could easily charge, say, $800 for a high end card, because we would pay it, because we would have no (decent) alternatives.
If only they had 1/10th the driver support for linux that Nvidia has, I might actually care.
I'm hoping that multiple GPUs will mean the ability to smoothly handle multiple GL contexts. This is the main thing that still separates consumer level cards from the professional quality ones, like the various 3DLabs offerings. If you've ever tried to run multiple OpenGL applications at once, or possibly multiple D3D apps as well, you've probably suffered severe 3D slowdown due to the overhead required by the context switching.
Is it me or sometime soon are we going to have to plug our computers into the 220 socket where the washing machine used to be?
;o)
Besides, who needs clean clothes when your getting 200 fps
Hadn't constantly crashed my Win2k box when I bought it for $600, I might not be so bitter.
But when you write a driver and refuse to run a machine with it for more than an hour, and then, worse than that, ship the product and try to sell it for $600 upon release, you do get a bad name.
ATI deserves every flame they get until my radeon supports VfW without an ungodly amount of hacks. And video capture is the absolute least amount of the problems with the driver that shipped (the fact that your DVD support is gone if you lose/ruin your driver disc would be number 2 on the list).
ATI can keep their crappy products. Of course, now I've switched over to Linux, I'm starting to buy their products again (looks like third party drivers written without full specs of ATIs cards are more stable than ATIs own -- who'd-a-thunk-it?).
>ATI has been much more forthcoming in information for developing XF86 drivers
Which would explain why third party X11 drivers are better than their windows drivers. Man, you have to have one really poor set of coders to be beat out on the quality of the drivers for your product by people hacking out code as a hobby.
If you could be told what you can see or read, then it follows that you could be told what to say or think - BoC
Does building a system for a game that doesn't come out for at least a year seem a bit misguided and overeager?
What happens if in that year Carmack changes the game engine so that it doesn't work well with your setup? Also is this system going to just sit around until then?
IMHO I'd wait at lease until the game came out. If you need a new system, build it for current games, then upgrade.
BASIC HTML
What's ATI doing to keep TWO of these in one box from overheating?
Says the RIAA: When you EQ, you're stealing bass!
I too would really like to see free, open-source Linux drivers for ATI graphics cards. I would buy new ATI cards like the one being discussed here, if/when such Linux drivers are available.
I too paid $$ for an xpert 2000 card w/tv. ATI claimed to have open source r128 drivers, so I bought it. Unfortunately, they lied. The open source drivers don't work with all r128 based cards, there are well known problems with 2d+3d acceleration enabled at the same time that cause screen flicker and corruption.
Don't buy ATI, if you want 3D under linux get a geforce card.
how can you see it work for yourself, if you don't trust them, i mean i personally don't buy anything that i don't trust.....
Here is a hint: wait
Doom3 is still a ways off. The present top of the line cards will run doom3 like the tnt2/voodoo3 ran quake3
Judging by how hot my Xpert128 runs, I think we can expect advertising copy along these lines:
The ATI Radeon 8500 Maxx: Your High-End Desktop Graphics and Affordable Home Heating Solution from ATI!
"You decide if it is real or not, a solid dual GPU solution would surely rock the industry to massive proportions!"
;)
(emphasis added)
When did Slashdot start hiring cheezy '80s Hair Metal band rejects to post stories to the front page?
It's "News for Nerds", not "News for Mullet-Sporting Losers Who Can't Get Over Their High School Glory Days".
Flamebait? Maybe a little.
Yeah ATI drivers blow real hard, but Matrox cards work real nice with XFree86 for 2D work.
Try NetBSD... safe,straightforward,useful.
Why settle for just 2 GPUs, here is the Lightning-2 - "A High-Performance Display Subsystem for PC Clusters" - http://graphics.stanford.edu/papers/lightning2/.
I recently heard one of the researchers give a talk about WireGL and found it very fascinating. If one has a GigE dedicated network, then a graphics cluster like this would be perfect for supporting graphics over the network for multiple PCs.
Innovation is worth more than graphics, people don't play games for good graphics alone. I am not saying the quality of the expirience isn't important, but you need the content to lure you into spending your free time *ahem* and all other time *ahem* to fully enjoy the expirience.
Enough with the graphics, and lets get some new ideas.
It's all good.
For LCD panels to talk straight digital.
and end up getting nailed hard by nVidia? I've only been playing "hardcore" with computers the last few years, but it seems I remember a card from 3dfx that had something along the lines of 4 processors and still couldn't touch nVidia's performance. If this works and doesn't cost $$$$$$ more, great, but I have a feeling this is a grasp to move ahead in a small margin area, while those in the mainstream are going to just pass it by.
Knightfall
UBC: An Efficient Unified I/O and
Memory Caching Subsystem for NetBSD
Chuck Silvers
The NetBSD Project
chuq@chuq.com, http://www.netbsd.org/
Abstract
This paper introduces UBC ("Unified Buffer Cache"), a design for unifying the filesystem and virtual memory caches of file data, thereby providing increased system performance. In this paper we discuss both the traditional BSD caching interfaces and new UBC interfaces, concentrating on the design decisions that were made as the design progressed. We also discuss the designs used by other operating systems to solve the same problems that UBC solves, with emphasis on the practical implications of the differences between these designs. This project is still in progress, and once completed will be part of a future release of NetBSD.
Introduction
Modern operating systems allow filesystem data to be accessed using two mechanisms: memory mapping, and I/O system calls such as read() and write(). In traditional UNIX-like operating systems, memory mapping requests are handled by the virtual memory subsystem while I/O calls are handled by the I/O subsystem. Traditionally these two subsystems were developed separately and were not tightly integrated. For example, in the NetBSD operating system[1], the VM subsystem ("UVM"[2]) and I/O subsystem each have their own data caching mechanisms that operate semi-independently of each other. This lack of integration leads to inefficient overall system performance and a lack of flexibility. To achieve good performance it is important for the virtual memory and I/O subsystems to be highly integrated. This integration is the function of UBC.
Background
In order to understand the improvements made in UBC, it is important to first understand how things work without UBC. First, some terms:
A pool of memory allocated during system startup which is dedicated to caching filesystem data and is managed by special-purpose routines.
The memory is organized into "buffers," which are variable-sized chunks of file data that are mapped to kernel virtual addresses as long as they retain their identity.
The portion of available system memory which is used for cached file data and is managed by the VM system.
The amount of memory used by the page cache can vary from nearly 0% to nearly 100% of the physical memory that isn't locked into some other use.
The kernel abstraction which represents a file.
Most manipulation of vnodes and their associated data are performed via "VOPs" (short for "vnode operations").
The major interfaces for accessing file data are:
The read() system call reads data from disk into the kernel's cache if necessary, then copies data from the kernel's cached copy to the application's address space. The write() system call moves data the opposite direction, copying from the application's address space into the kernel's cache and eventually writing the data from the cache to disk. These interfaces can be implemented using either the buffer cache or the page cache to store the data in the kernel.
The mmap() system call gives the application direct memory-mapped access to the kernel's page cache data. File data is read into the page cache lazily as processes attempt to access the mappings created with mmap() and generate page faults.
In NetBSD without UBC, read() and write() are implemented using the buffer cache. The read() system call reads file data into a buffer cache buffer and then copies it to the application. The mmap() system call, however, has to use the page cache to store its data since the buffer cache memory is not managed by the VM system and thus not cannot be mapped into an application address space. Therefore the file data in the buffer cache is copied into page cache pages, which are then used to satisfy page faults on the application mappings. To write modified data in page cache pages back to disk, the new version is copied back to the buffer cache and from there is written to disk. Figure1 shows the flow of data between the disk and the application with a traditional buffer cache.
This double-caching of data is a major source of inefficiency. Having two copies of file data means that twice as much memory is used, which reduces the amount of memory available for applications. Copying the data back and forth between the buffer cache and the page cache wastes CPU cycles, clobbers CPU caches and is generally bad for performance. Having two copies of the data also allows the possibility that the two copies will become inconsistent, which can lead to application problems which are difficult to debug.
The use of the buffer cache for large amounts of data is generally bad, since the static sizing of the buffer cache means that the buffer cache is often either too small (resulting in excessive cache misses), or too large (resulting in too little memory left for other uses).
The buffer cache also has the limitation that cached data must always be mapped into kernel virtual space, which puts an additional artificial limit on the amount of data which can be cached since modern hardware can easily have more RAM than kernel virtual address space.
To solve these problems, many operating systems have changed their usage of the page cache and the buffer cache. Each system has its own variation, so we will describe UBC first and then some other popular operating systems.
So what is UBC anyway?
UBC is a new subsystem which solves the problems with the two-cache model. In the UBC model, we store file data in the page cache for both read()/write() and mmap() accesses. File data is read directly into the page cache without going through the buffer cache by creating two new VOPs which return page cache pages with the desired data, calling into the device driver to read the data from disk if necessary. Since page cache pages aren't always mapped, we created a new mechanism for providing temporary mappings of page cache pages, which is used by read() and write() while copying the file data to the application's address space. Figure2 shows the changed data flow with UBC.
Figure 1: NetBSD before UBC.Figure 2: NetBSD with UBC.
UBC introduces these new interfaces:
These new VOPs are provided by the filesystems to allow the VM system to request ranges of pages to be read into memory from disk or written from memory back to disk. VOP_GETPAGES() must allocate pages from the VM system for data which is not already cached and then initiate device I/O operations to read all the disk blocks which contain the data for those pages. VOP_PUTPAGES() must initiate device I/Os to write dirty pages back to disk.
These functions allocate and free temporary mappings of page cache file data. These are the page cache equivalents of the buffer cache functions getblk() and brelse()[3]. These temporary mappings are not wired, but they are cached to speed repeated access to the same file. The selection of which virtual addresses to use for these temporary mappings is important on hardware which has a virtually-addressed CPU data cache, so the addresses are carefully chosen to be correctly aligned with the preferred addresses for user file mappings, so that both kinds of mappings can be present at the same time without creating coherency problems in the CPU cache. It is still possible for applications to create unaligned file mappings, but if the application lets the operating system choose the mapping address then all mappings will always be aligned.
This is a UVM pager which handles page faults on the mappings created by ubc_alloc(). (A UVM pager is an abstraction which embodies knowledge of page-fault resolution and other VM data management. See the UVM paper[2] for more information on pagers.) Since its only purpose is to handle those page faults, the only action performed by ubc_pager is to call the new VOP_GETPAGES() operation to get pages as needed to resolve the faults.
In addition to these new interfaces, several changes were made to the existing UVM design to fix problems which were glossed over in the original design.
Previously in UVM, vnodes and uvm_objects were not interchangeable, and in fact several fields were duplicated and maintained separately in each. These duplicate fields were combined. At this time there's still a bit of extra initialization the first time a struct vnode is used as a struct uvm_object, but that will be removed eventually.
Previously UVM only supported 32-bit offsets into uvm_objects, which meant that data could only be stored in the page cache for the first 4 GB of a file. This wasn't much of a problem before since the number of programs which wanted to access file offsets past 4 GB via mmap() was small, but now that read() and write() also use the page cache interfaces to access data, we had to support 64-bit uvm_object offsets in order to continue to allow any access to file offsets past 4 GB.
What do other operating systems do?
The problems addressed by UBC have been around for a long time, ever since memory-mapped access to files was first introduced in the late 1980's. Most UNIX-like operating systems have addressed this issue one way or another, but there are considerable differences in how they go about it.
The first operating system to address these problems was SunOS[4,5], and UBC is largely modeled after this design. The main differences in the design of the SunOS cache and UBC result from the differences between the SunOS VM system and UVM. Since UVM's pager abstraction and SunOS's segment-driver abstraction are similar, this didn't change the design much at all.
When work on UBC first began over two years ago, the other design that we examined was that of FreeBSD[6], which had also already dealt with this problem. The model in FreeBSD was to keep the same buffer cache interfaces to access file data, but to use page cache pages as the memory for a buffer's data rather than memory from a separate pool. The result is that the same physical page is accessed to retrieve a given range of the file regardless of whether the access is made via the buffer cache interface or the page cache interface. This had the advantage that filesystems did not need to be changed in order to take benefit from the changes. However, the glue to do the translation between the interfaces was just as complicated as the glue in the SunOS design and failed to address certain deadlock problems (such as an application calling write() with a buffer which was a memory mapping of the same file being written to), so we chose the SunOS approach over this one.
The approach taken by Linux[7] (as of kernel version 2.3.44, the latest version at the time this paper was written) is actually fairly similar to the SunOS design also. File data is stored only in the page cache. Temporary mappings of page cache pages to support read() and write() usually aren't needed since Linux usually maps all of physical memory into the kernel's virtual address space all the time. One interesting twist that Linux adds is that the device block numbers where a page is stored on disk are cached with the page in the form of a list of buffer_head structures. When a modified page is to be written back to disk, the I/O requests can be sent to the device driver right away, without needing to read any indirect blocks to determine where the page's data should be written.
The last of the operating systems we examined, HP-UX, takes a completely different stance on the issue of how to cache filesystem data. HP-UX continues to store file data in both the buffer cache and the page cache, though it does avoid the extra of copying of data that is present in pre-UBC NetBSD by reading data from disk directly into the page cache. The reasoning behind this is apparently that most files are only accessed by either read()/write() or mmap(), but not both, so as long as both mechanisms perform well individually, there's no need to redesign HP-UX just to fix the coherency issue. There is some attempt made to avoid incoherency between the two caches, but locking constraints prevent this from being completely effective.
There are other operating systems which have implemented a unified cache (eg. Compaq's Tru64 UNIX and IBM's AIX), but we were unable to find information on the design of these operating systems for comparison.
Performance
Since UBC is unfortunately not yet finished, a detailed performance analysis would be premature. However, we have made some simple comparisons just to see where we stand. The hardware used for this test was a 333MHz Pentium II with 64MB of RAM and a 12GB IDE disk. The operations performed were a series of "dd" commands designed to expose the behaviour of sequential reads and writes. We create a 1GB file (which is much larger than the physical memory available for caching), then overwrite this file to see the speed at which the data modifications caused by the write() are flushed to disk without the overhead of allocating blocks to the file. Then we read back the entire file to get an idea of how fast the filesystem can get data from the disk. Finally, we read the first 50MB of the file (which should fit entirely in physical memory) several times to determine the speed of access to cached data. See Table 1 for the results of these tests.
Table 1: UBC performance comparison. Experiment Run Time (seconds) Input Output Size NetBSD NetBSD FreeBSD Linux 1.4.2 with UBC 3.4 2.2.12-20smp raw deviceThe great disparity in the results of the first four tests on the three non-UBC operating systems is due to differences in performance of their IDE disk drivers. All of the operating systems tested except NetBSD with UBC do sequential buffered reads from a large file at the same speed as reads from the raw device, so all we can really say from this is that the other caching designs don't add any noticable overhead. For reads, the UBC system is not yet running at device speed, so there's still room for improvement. Further analysis is required to determine the cause of the slowdown.
UBC obviously needs much improvement in the area of write performance. This is partly due to UVM not being very efficient about flushing modified pages when memory is low and partly because the filesystem code currently doesn't trigger any asynchronous writes to disk during a big sequence of writes, so the writes to disk are all started by the inefficient UVM code. We've been concentrating on read performance so far, so this poor write performance is not surprising.
The interesting part of this test series is the set of tests where we read the same 50MB file five times. This clearly shows the benefit of the increased memory available for caching in the UBC system over NetBSD without UBC. In NetBSD 1.4.2, all five reads occured at the speed of the device, whereas in all the other systems the reads were completed at memory speed after several runs. We have no explanation for why FreeBSD and Linux didn't complete the second 50MB read at memory speed, or why Linux didn't complete even the third read at memory speed.
Conclusion
In this paper we introduced UBC, a improved design for filesystem and virtual memory caching in NetBSD. This design includes many improvements over the previous design used in NetBSD by:
Availability
This work will be part of a future release of NetBSD once it is completed. Until then, the source code is available in the "chs-ubc2" branch in the NetBSD CVS tree, accessible via anonymous CVS. See http://www.netbsd.org/Sites/net.html for details.
This being a work-in-progress, there is naturally much more work to do! Planned work includes:
Acknowledgments
We would like to thank everyone who helped review drafts of this paper. Special thanks to Chuck Cranor!
Bibliography
1 The NetBSD Project.
The NetBSD Operating System.
See http://www.netbsd.org/ for more information.
2 C. Cranor and G. Parulkar.
The UVM Virtual Memory System.
In Proceedings of the 1999 USENIX Technical Conference, June 1999.
3 Marice J. Bach.
The Design of the UNIX Operating System.
Prentice Hall, February 1987.
4 J. Moran, R. Gingell and W. Shannon.
Virtual Memory Architecture in SunOS.
In Proceedings of USENIX Summer Conference, pages 81-94. USENIX, June 1987.
5 J. Moran.
SunOS Virtual Memory Implementation.
In Proceedings of the Spring 1988 European UNIX Users Group Conference, April 1988.
6 The FreeBSD Project.
The FreeBSD Operating System.
See http://www.freebsd.org/ for more information.
7 L. Torvalds, et al.
The Linux Operating System.
See http://www.linux.org/ for more information.
About this document ...
UBC: An Efficient Unified I/O and
Memory Caching Subsystem for NetBSD
This document was generated using the LaTeX2HTML translator Version 98.1p1 release (March 2nd, 1998)
Copyright 1993, 1994, 1995, 1996, 1997, Nikos Drakos, Computer Based Learning Unit, University of Leeds.
The command line arguments were:
latex2html -split 0 -no_navigation ubc_usenix.
The translation was initiated by Chuck Cranor on 2000-04-25
Chuck Cranor
2000-04-25
On the other hand, if one could run Distributed.net on the 2 GPUs, I'd be at the top of the OGR list! :-)
ACHTUNG! Das computermachine ist nicht fuer gefingerpoken und mittengrabben. Ist nicht fuer gewerken bei das dumpkopfen.
Comment removed based on user account deletion
That picture looks like a photoshop job ..
..
..
.. although some of the stuff has been switched around this is a pretty amateur job.
8 500-64m b.jpg
For one the Heatsink fans are exactly the same - right down to the positioning of the fan fins
For two the wires from one of the fans are not casting a shadow
In fact if you do a Google image search you can find the original image
see original here:
http://home.earthlink.net/~doniteli/radeon
not really -- Voodoo 2 had "dual" setups and pretty much everyone ran without a hitch;
i think it's not the "dual GPU" that's hard to write drivers for -- it's the way that they implemented the hardware that make it dual -- V2 had SLI - simple, effective, real -- you interleave the lines; but ATI with their rage fury decided that it wasn't "good enough" and just *had* to come up with something else (i think they split the screen top-bottom with some details given to load sharing in case the top-bottom had different number of polygons (read: all the games); THAT will screw up any driver-writing process.
if they stick with SLI this time maybe it will get better
My life in the land of the rising sun.
There's been a lot of chatting about ATi vs NVidia over on Shacknews.com recently about this kinda stuff. Especially since the Carmack came out and said that he thought the R300 was gonna be more capable, from a hardware standpoint, even if ATi's drivers didn't do it justice. If ATi decides to hunker down and make their drivers as good as their hardware, I think they'll do just fine. But if not, then Nvidia will retain its crown, since their drivers seem to work better for the masses than ATi's do.
As far as all this power going unused -- I doubt it. I think at some point game companies _will_ have to focus on better gameplay, instead of simply rehashing the same game concepts and upgrading the graphics engines. But I think that graphics will play a large part in that, too.
A Ti4600 may be more suited for future titles, but you're forgetting that you can run current titles with more eyecandy. NWN with all graphics effects maxed out at 1600x1200 looks pretty damn good.
You know what the best card out there is? It might surprise you. At a low end price tag of ~$800 and up to nearly $6500 the Voodoo 5 6000 is one of the best cards out there. The price tag is that high because it is, unsurprisingly, a collectors item.
I've seen this card work. I runs fast and it looks gorgeous.
You know what the Parhelia tried to do? Fragmented AA? Voodoo could've torn that up years ago. The V5 6000 did 8x Full screen AA. Fast. At 1024. It's amazingly gorgeous.
Think about it. This card is 2-3 years old. The architecture is what matters. Not the amount of GPUs. The GeForce4 4600 can't even consider 8xAA. The V5 6K does, and it does it well. On 128M of SDRAM. I'd still maybe take the 4600 over the V5 6K. But it would be a hard decision. The 4600 with it's DDR memory and GPU can handle some things better. Some. Not all.
This card just proves that it really doesn't matter how much RAM or how many GPUs a card has. It's in the way the card is built. There aren't many cards I'd take over the V5 6K. If I could get one, for myself to keep, I'd pull me Geforce 3 out in a heartbeat. The GPU isn't a factor here. The RAM (DDR over SDR) isn't a factor. The V5 6K is just that well built, even 3 years later.
There. I've said my piece. After seeing the V5 in action, I don't care to get the least bit excited about the "latest greatest" graphics cards ever again.
If I get the choice of adding $200 dollars for a new gfx card or a new cpu to permanently increase the graphics capabilities or computing capabilities I'd way rather pay that than have to pay programmers to optimize every piece of software using it. I want programmers to be busy making high-quality, stable content and gameplay capabilities (though the actual quality depends on the gfx / sfx / storyline / modelling / texturing / whatever staff), not trying to squeeze out the last 5fps / assembly optimizations to make it playable on a bigger marked (= slower systems).
Which is not an excuse to make *unessecerrily* bloated code. But to me I'd be a lot happier if it's featurerich and stable rather than fast. Usually you only get to pick at most two out of three, at least on a sane budget
Kjella
Live today, because you never know what tomorrow brings
How does this fit in with r300. From what I have seen a single r300 should be able to out perform this dual r200 monster. r300 should be coming out soon since doom3 ran on an r300 card at e3 so it should be coming out soon. So where does this leave this card. Assuming r300 will be released in the usual ati septemper timeframe and this card is released today, that will leave this card as the fastest for what two months. Why bother with the dual r200.
To me this is a internal project that will never be released. Good practice though. Maybe they will get a dual r300 out soon after its release.
It is a fake; that's been discussed ad nauseum in the forum.
I typically don't follow trends set by stupid people, besides how can you judge the quality of a item if it is owned by a stupid person, they prolly don't know what fps even means!
Comment removed based on user account deletion
It has been interesting to see all the tricks needed to keep up with the GPU performance growth curve. The growth curve is so much steeper than the CPU growth curve (double every 18 months) because graphics processors can take advantage of more parallelism.
But the parallelism has mostly just happened on chip for consumer GPU's. It will be interesting to see how well it works between chips. This sounds awfully similar to UNC's PixelFlow, which used object-parallelism (GPU-parallelism) and image composition to construct the final image. And between chips parallelism is more feasible than the between-computers solution of WireGL.
MOD THIS UP...
After looking at the image and checking it out in photoshop, I too see the obvious photoshoppery. The connectors and slot cover are from another photo but there are glaringly obvious similarities between this one and this (fake) one
There is no spork.
I've been an avid PC gamer for years, and a am huge fan of Doom. But I've also been far from able to keep up with the pricey trends of the newest best thing. Would I appreciate running Doom III "perfectly"? Certainly, but until a wad of cash falls from the sky, it isn't happening.
As far as games go, my GeForce 2 can handle all the games being released today, and that's fine by me. If a game comes out that isn't backward compatible (even though it might suffer a bit in the fancy graphics), there's no point in playing.
And besides, I'm looking forward to enjoying the game play. If Doom III scares the pants off of me like they claim, but my computer has lackluster texture and lighting, I'm still going to enjoy every minute of it -- and to me, that means the game is running perfectly.
Comment removed based on user account deletion
> UntiL ATi makes their own *nix drivers, im stickin to Nvidia.
In ATI's defense, unlike nVidia (who are strictly proprietary), ATI do make the chipset details available so anyone can write open source drivers for whatever esteoric OS they happen to be using - there's more OSs than just Windows and Linux, you know!
Of course, it would be nice if ATI released both specs and drivers, but IMHO, it is better in the long term for open source OSs if the specs are released.
the picture's been photoshopped. You can see where someone copied and pasted the first fan to the upper right. (theres a little piece of the red wire still in the image) Plus they put fan chips between between two memory chips. No way that pic's real. The card could be coming out, but that pic is definitely fake. (DeJa Vu with Voodoo 5) Read the forums from the original link to find out.
:)
Check it out
"Electric Relaxation" - ATCQ
- Bwana
Comment removed based on user account deletion
"Errr, install Linux, maybe?
The way to solve a Windows98 problem is to remove Windows98. Problem solved."
Yeah, Linux is pretty damn stable when it has no games to run on it.
ATI's dual proc cards used Alternate Frame Rendering
GPU1: renders a frame
GPU2: renders a frame - GPU1: Displays frame it just rendered
GPU1: Renders a frame - GPU2: Displays frame it just rendered
etc.
ATI *cannot* do SLI because it's patented, and nVidia bought 3dfx's patents on SLI among other things. This is why ATI used another method--their methos was just as simple and effective; each GPU renders alternate frames. In fact, it has fewer theoretical problems when doing alternate frames, than Voodoos did with rendering alternate scanlines. The scanline approach produced occasional tearing or shimmering effects, though it's very rare.
The problem with the Rage Fury MAXX has to do not with the method of interleaving or the presence of GPUs, but rather the method ATI chose to bridge the two chips, which isn't permitted in the NT AGP code. ATI couldn't find a way around it, so they abandoned the card. Sad, since it was a nice performer under Win9x...
However, many other implementations do 2 graphics chips right with NT support, such as the Voodoo 5 5500 and the high-end multi-chip Quantum3D boards.
So, ATI could easily do a Radeon MAXX part with WinXP support, since they know what mistakes not to make in silicon this time around...
Chasing Amy
(We all chase Amy...)
"The more corrupt the state, the more numerous the laws"-Tacitus
I'm a little surprise parent post got modded down as 'Flamebait'. His comment was appropriate. There's lots of comments floating around about bad driver support for ATI products. I can personally testify to that. The Radeon 8500 I used to have in my box wouldn't draw the titlebar properly in Windows. Instead, it'd draw it transparently. (Only the text was still there.) That made Windows a BITCH to use.
Nvidia, so far, is the king of drivers. They have a unified driver set. (only one download) They are stable. (I've used 3 different Geforce cards and they work wonderfully.) And they're constantly revised. ATI and every other video card company could learn a lot from Nvidia.
Please mod up parent post. He may not have the most interesting point, but it certainly was not worthy of a 'flamebait -1' moderation.
I heard that the E3 DoomIII demo was using an R300, not an 8500. I think VIA had an R300 running to demo its AGP8X chipset because it was one of the only stable AGP8X cards available. I could be wrong, but I'm pretty sure I read this on at least one fairly reputable site.
ATI is a typical example of how much more mature hardware manufacturing is and how pathetic software "manufacturing" is. Their drivers barely ever work. They create and update drivers when they want to, not when there are problems.
I would venture to say that this new graphics card is technically great. At the same time, I'm practically positive that ATI will not produce working drivers for this card.
OS SUPPORT
* Windows 95/98/ME/2000/XP/NT
So much for that card.
You're right. I'm also building it sooner rather than later because my current computer is a PII 450 Mhz and a Diamon Viper 770 (TNT2) video card... I'm feeling the upgrade burn more than most gamers these days.
The only thing that's allowed me to last this long on this remarkably slow system is fast hard drive communication with a SCSI controller native to the motherboard. Every other PII 450 I've seen has shit its guts out on the software I run today (photoshop 6, Aliens vs Predator 2, etc.).
I always build in favor of a long lifespan, if I wait for the next generation of video cards (as Carmack suggested) I'll be in a situation where I can't buy the top of the line for that generation of hardware because the top of the line won't exist yet.
Just another case of unique solutions to unique needs you'll find with almost everyone building their own computer...
http://www.jeffwilhelm.com/files/r250.jpg
I apologize in advance for the AC post, but being as my company has a working relationship with ATI, blah, blah, etc etc.
On to why I think this is fake:
1. Look at the heatsink/fans. From the picture, it looks like they are using different model fans for the different GPU's, looking at the position at which the power wires are coming from. Being that I am a board designer, I can tell you that this would never happen, in order to keep the bill of materials down.
2. On the very bottom right of the card, under the last SRAM chip, there is a small device (regulator?) that looks like its overlapping the edge of the board. This would never pass board layout verification, because there are certain clearances you need to observe when laying out pcb's.
3. It looks like the lower GPU is violating the AGP spec for connector keepouts. I'm not sure on this, as I dont have the AGP design guide handy, but that GPU looks like it's positioned extremely low.
4. Silkscreen for some of the parts further down the board (compare some of the electrolytic can & SRAM silkscreens) seems to be conspicuously absent.
5. Look at the ATI symbol silkscreen. Right above it is a fiducial (these are used during assembly, as a way for the machine doing the assembly to calibrate it's position to the board), and part of a silkscreen that looks exactly like the assembly guide for the SRAMs! This is the thing that to me stands out the most as being doctored.
"Why would anyone need more than 640k of memory?".
Think about it. This card is 2-3 years old. The architecture is what matters. Not the amount of GPUs.
Again, this card had 4 processors!
It sort of had 128M of RAM. It actually has 32MB of RAM per processor. So, all the latest games that use up more than 32MB of RAM in texture / geometry caching will run really slowly on the V5. Also, for those that don't remember, this was the card that you had to plug into the wall separately from the computer.
Don't get me wrong, I've used the V5 5500 (2 GPU version), and it was really cool at the time. But I'll take a GF4 any day of the week over any voodoo you offer me (unless of course I can sell it at the collector's item price :)
Dan
Well, well...and 640k ought to be enough for anybody!
Until I can get fully emersive stereo 3D raytraced photorealism at >60 Fps similar to the look of Exile with realisic ocean waves crashing on the shore I will continue supporting upgrade after upgrade of graphics solutions.
You probably still have a 3DFx Voodoo 1 right?
What I find amusing about this discussion is that the same people who slam Microsoft's proprietary cutthroat tactics and support a ,for the most part, inferior product jump on the NVidia bandwagon every chance they get. Or, as the GNU gurus so often put it, you don't like the drivers? Write better ones!
I also happen to think that people get way too hung up over the 2% of the time they are playing games when the other 98% ofthe time they would be better off with a ATI or Matrox card.
Not to mention the times I was hosed by crappy NVidia drivers in Linux until NVidia got their crap working better.
"God fights on the side with the best artillery." - Napoleon, Marshal of France - speaking truth to power
Can you tell the difference between the real world and the simulated virtual world, in terms of visual fidelity? I sure can, and as long as there is a noticable difference, the graphics industry will rocket on.
...
I predict the curve won't break until realtime computer graphics are far more convincing than the computer graphics we see in movies today.
That will be a short while
it looks like a gimp job to me
My voodoo5 can still run most all of the modern games, Warcraft 3, Return to Castle Wolfenstein, etc. I have only a couple of problems exiting games now and then, but that is only because the drivers have not been updated (understandably) as of late for these games.
If ati wasn't around, nvidia would go two ways, start charging huge prices and there would be a new company starting that will kill them or they would stay about the same and do what they love.
____
So by that logic, where's your microsoft killer?
So will ATI actually make this MAXX product work in the Windows NT/2000/XP line of products? Or will they pull that windows 98 only crap that shafted alot of previous MAXX owners when windows 2000 was released?
I've owned a MAXX product that didn't work in NT/2000/XP. I've had a USB ATI TV wonder that wouldn't work in XP for about a year, and i've had drivers on many other ATI products that either sucked or didn't work at all.
ATI has a long way to go to get my confidence back. Getting screwed three times without vaseline makes one a little apprehensive.
-ted
Don't really care about Windows or Linux, but, will FreeBSD drivers be available? I can only use cards which run on FreeBSD, my favorite operating system, which already has a stable VM.
I'm really not into the a-fan-for-everything movement going on. Motherboards, video cards, etc. The only things I really want in my PC with the fan are the CPU and the power supply, and if I could do without those, it would be even better. ATI would have a winner in my eyes if it could use two lower-clocked fanless chips together to deliver performance on-par with the rest of the one chip cards. Driver issues or not (I'm on an All-In-Wonder 128 right now... don't even get me started), it would definitely get a buy consideration from me.
Also, if you look at the big silver capacitor on the bottom right, you can see the black fin mark where the GPU's heatsink overlaps the Capacitor in the original picture.
;)
Also, hold a piece a paper up against the top edge of the board itself. You can see where the two pieces don't form a straight line.
Don't steal. The government hates competition.
You obviously have not used a dual cpu system. The fact remains that the only way to truly run to threads concurrently is to run them on seperate cpus. Otherwise, they just run on the one cpu a little bit at a time to simulate true multiprocessing.
My 2x 300mhz PII is more responsive than a 1x 600mhz system. It's not twice as fast, but it can do twice as much in the same time.
There are times I wish I'd opted in for a dual board on my AMD 1900+ setup now for that reason.
Don't suppose you've taken a look at the other side of the card... the connectors on the backplane aren't the same between those two pictures, so they've obviously done something a bit differently.
AA performance basically comes down to memory bandwidth. Yes, the Voodoo 5 6000 did have a stupidly high bandwidth (11-12Gb/s), which still just about beats Nvidia's Ti4600 10.4Gb/s
It has little to do with the number of GPUs you've got. The Voodoo 5 probably had to have several just to keep up with the bandwidth it had.
So what was so great about the Voodoo 5 6000? they put a huge amount of bandwidth into a card when it just wasn't economically viable. I'm sure that nVidia and ATI probably both had internal test setups that could equal it, but they both had the sense not to try and make a commercial product out of it until the cost of fast RAM came down.
As far as I am aware the V5 6000 didn't have any particularly special AA tricks, which nVidia seem to have now (compare Geforce3 AA performance with GeForce4 Ti...) so I'd imagine that the Ti4600 would beat the V5 nowadays, on 4x AA at least. Shame they don't have a higher AA mode, but with the next gen of games coming out, you wouldn't be able to afford it anyway, even with e Ti4600 or a V5 6000.
Of course, the V5 had no pixel or vertex shaders (which is gonna hurt image quality) and no hardware T&L. As the majority of current games are still CPU-bound, that's gonna hurt the Voodoo 5.
Do I wait for the above mentioned cards to come out or buy a top of the line GeForce 4 now?
In a nutshell, GPUs are outdated. Remeber back in April 99, when nvidia announced the Geforce256, the first chip shipping with hardware transformation and lighting. They called this a gpu, since then, various chipmanufactors have extended the "GPU" instructionsets to include stufflike AA/Filters/texture comp./etc etc etc.
But, the successor for thiese kinda have already been created. 3Dlabs have introduced the VPU, which is bacially a GPU with a codeable instructionset.
oh wow, its just 3dfx all over again. hopefully the junky ATI will just die, and surrender its market share to a company that actually makes good products.
Take a look at Duke Nukem Forever.
They are trying to perfect both, in the mean time, everyone else will have passed them by.
Visit the Arcade Restoration Workshop @ http://www.arcaderestoration.com
"Final testings have been done and you should here noise from ATi regarding this offering." Here? Come on guys, you can do better than that.
Oh really, a Geforce 4? whoa, what cave have i been living in? the radeon 8500 wasn't made to be faster than a geforce 4, it was made to compete with a geforce 3, which it can do, basically, using your argument, i can say that every 3d card ever made before geforce 4 was total crap, no, they're just crap compared to todays standards, besides, thats a totally unfair comparison, geforce 4 is newer and thus we should expect it be faster.
Damn,
Just bought a Radeon 8500LE 128.
Ah, well. At least I got it cheap. I suspect they won't be giving these things away.
Out.
Comment removed based on user account deletion
YAY.. glad it was a canadian company to do it first
now can they get their ver 3.04b34 rev 9967 drivers out before everyone else does it too and takes over the market?
a steganography distributed client?
This is a HOAX. The official word from ATI is, "We have no current intentions of making this or similar boards."
Check out hardocp and look at their blurb, with the same
photoshopped rendition of the card sent to them last week. Notice the jab pointed in Slashdots direction? Hilarious!
[H]ardOCP says it's a hoax, got a statement from Rubeena Hussein of ATi.
"We have no current intentions of making this or similar boards."
that most game coders are lazy. This thing is running dual GPU, so it probably can't run in AGP mode and take advantage of all the memory bandwidth that no one take advantage of.
Check it out here
I am disrespectful to dirt! Can you see that I am serious?!
HardOCP http://www.hardocp.com/ has
Christ on a bicycle.. when I was a young'un we called them thar things Graphics Cards.
Down with Corporate Doublespeak!
HEY! Get a clue! The issue isn't whether ATi can compete speedwise! The issue is whether or not ATi will ever produce decent drivers. NVidia has a few things going for it. They make great hardware AND they make the industries best drivers.
"a solid dual GPU solution would surely rock the industry to massive proportions"
Agh! Marketing Splooge, attacking from 3 o'clock!
Dual "GPU" configurations have been around since early '98, when the Voodoo2 came out. Sure, 3Dfx called it SLI, but it was essentially two 3D cards working as one - and someone (Quantum3D?) made single cards with dual Voodoo2's on them. Not to mention the Voodoo5 which had 4 GPUs on it.
I remember seeing someone (could have been Quantum3D again) who was promising a 16-GPU version of the Voodoo5 for mass $$$.
Multiple GPU's is nothing new, and it's definitely not going to "shake the gaming industry to it's core only on PAY PER VIEEEEEEEEEW....."
Go to www.rage3d.com to get this one cleared up. If you actually want to know what is going on within ATI, then go there. This post is somewhat misinformed, to say the least.
deul gpu's? dude, they don't fight
Some numbers for you.
5 .html
5 83&p =9
Q3A, 1600x1200x32 bit (no FSAA)
GF4 Ti4600 : 160.6 fps
V5 6000 : 58.7 fps
Expect almost a linear scaling for FSAA. Note that at 4x, the GF4 would be pushing out around 40 fps. The 6k? About 13. At 8x? Let's be generous, and call it 8. Yes, the machines being tested are very different (a 1.3ghz Athlon vs. an 800mhz P3), but at those resolutions, you're very close to being 100% CPU bound.
I admire the meaningless iconoclasm that would lead one to tout an evolutionary dead-end like the 6000 as the be-all end-all of video cards, but in the future, you would be better served by appealing to the Voodoo's superior blast capacity, or the "warmth" of its image, rather than trying to make a technical argument without even the slimmests of legs to stand on.
Best,
'jfb
Links:
V5 6k benchmarking: http://www.voodooextreme.com/hw/previews/v5_6000/
GF4 numbers: All over, but I used these:
http://www.anandtech.com/showdoc.html?i=1
To spur "enterprise Linux," Big Bang, the distributed two-phase commit.
Why is the photo that was supposedly leaked by someone there so obviously photochopped? Where's the official hype (surely they don't plan, officially, on releasing the product onto a totally unsuspecting market? They've officially told various people about their upcoming R300 product, why would they, officially, keep this so secret)? Where in ATi's lineup does this Radeon 8500 MAXX fit? Above the single GPU Radeon 8500, sure. But wouldn't it steal sales from their upcoming R300 based product (reportedly called the Radeon 9700)? Sure, a Radeon 8500 MAXX won't have DirectX 9 compliance, but there's no DirectX 9 games out yet, nor will there be any that *require* it (notice I didn't say can't take advantage of DirectX 9 features) for some time.
Yes, the geek in me thinks "Dual GPU Radeon card. Sweet!". But the realist in me thinks "Well, the ONLY "proof" of it we've seen are unconfirmed leaks, and a badly photochopped photo of a product that ATi already have in full production.
Ahuh. I'll believe it when I see it, in person.
The heatsink fan stickers are different. The connectors are in a different order. The COPYRIGHT string has a different date (2001 vs. 2002 -- no jpg artifacting suggesting a blur and redo).
The board layouts are very similar, and it's likely that someone did take a stock Radeon picture and mod it a bit, but those are still different boards entirely.
--
Internet Explorer (n): Another bug -- that is, a feature that can't be turned off -- in Windows.
This has been determined to be a hoax. Slashdot get's fooled more and more each day.
The new trend of computer tech is going for dual, then quadrupo, and so forth.. i guess the in like 5 years of time, we need to choose the mother board and as well as a video board.. just insert the GPU into the computer.. (for higher performance, go for the new QuadDR 5Gb RAM 2GB FSB video board.. =P)
The rage fury maxx runs on linux using the rage128 drivers. Even the framebuffer driver will work.
:)
DGA, on the other hand, will absolutely not work, so if you need that, you're screwed.
I had one of these cards, and frankly, it was a pain in the ass. To get it working properly you need two monitor settings in your XF86Config file (for the same monitor) and do a few other odd things.
Now I run it on a system without X or framebuffer drivers as a scrap server. Does the job well enough
Those who can't do, teach. Those who can't teach either, do tech support.
Please excuse me while I run home and get rid of the stack of linux games that I obviously don't have.
That should be easy, they're all on your RedHat CD!
I don't know what country you're in, but in Canada we plug our washing machines in 120V outlets, and our dryers go in 240V.