Slashdot Mirror
AMD Cuttin' Deals, Releases 800 Mhz Athlon
MatriXOracle writes "AMD seems to be on fire lately. According to this C|Net article, HP will be including K6-2's in new portable models, and is considering the Athlon for desktop use. Meanwhile, Gateway is blaming its disappointing earnings on supply (or lack thereof) of Intel chips, and will start selling systems with AMD chips very soon. Finally, an 800MHz Athlon is being released today. "
A small thing to be sure but it's given me a warm glow about AMD
Rich
Have they really had these since November? I wouldn't be surprized if you're right but if you have evidence I'd be interested in seeing it. Even some good juicy rumours would be cool.
You're certainly right about them holding back product to keep profits up. I'm sure they could release some 1Ghz parts now, but the yields wouldn't be great and it would kill the market for lower speed chips (where the yield is decent).
Some have said they should release their fastest stuff right away to grab market share, but that would be costly. Intel (or Microsoft) can to that - and it's not a bad strategy - but when you've been losing money for so long you need to worry about short term profit or your investors will lose confidence.
It seems to me they're playing it just about right. If the demand for Athlons suddenly took off they probably wouldn't be able to supply them. This way the supply is probably in line with the demand, Intel is looking like a poor cousin, and everybody is happy.
It is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail. - Abraham Maslow
Intel announces their faster CPUs during the holidays and AMD announces _availability_ of their 800 MHz CPU in the first week back from holidays.
Is it just me, or does it seem like they had these things sitting in the warehouse waiting for an Intel announcement. I bet if Intel had announced 850 MHz AMD would have matched that too.
It is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail. - Abraham Maslow
The reason today's software sucks is quite simple.
Programmers are becoming more stupid:
Today a lot of software is written in VisualBasic. 10 years ago - C was dominant. 20 years ago assembly was dominant. When you lower the barriers to entry, you get more stupid participants.
Users are becoming more stupid:
20 years ago a computer user would have to be pretty smart. Now, any old moron uses a computer, so the software is mostly designed for morons
(not necessary, but true).
People are becoming more stupid:
This is more complicated, and only applies to 1st world. It's caused by automation making it possible for people to think less, along with the physical addictiveness of mindless ritual (iso9000 etc) leading to reduced cognitive capacity (or morons, to use the correct scientific term).
http://rareformnewmedia.com/
For intelligent devices (such as disk drives with their own OS & drivers), see: Commodore PET, IEEE 488, Commodore 4040 and 8080 drive units, instructions for daisy-chaining two Commodore drive units, sending the drives commands, and unplugging the main computer. The disk drives would continue to operate, sending files to the printer, or copying from one drive to the next, without difficulty.
There would be absolutely NOTHING to stop a manufacturer putting a 80486, 8 megs of RAM, and some ROMs loaded with the Linux kernel, a HD driver and some basic filesystem drivers onto a disk drive. It wouldn't cost significantly more than is already spent on disk controllers, would accelerate performance enormously, and not take resources away from the main CPU.
The advantage? Devices which are intelligent can do their work LOCALLY. You don't need to ferry signals half-way across the motherboard to do simple calculations, and ferry the results all the way back, for each block read or written to. That is STUPID and SLOW!
Why have disk drives with printer drivers? Easy. Where are all your files for printing? In memory? Nope! On disk. If they go through any kind of spooler, they will be on the disk drive. But, if they're on the disk drive already, why drag them all the way into memory to process? Why not just process them locally, where they are, and keep the main processor out of it? Saves time, moving all that data around, and if your disk drive is intelligent, there's precicely no overhead in doing it this way. (If anything, it reduces overhead, as you've cut out transport time, and you can move the spooler to the disk drive too, reducing your main CPU load still further.)
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Here's a link that has a good list of Athlon motherboards that have been announced, preannounced, or are available.
Slot A
the more convinced I am that Intel has taken aim at it's foot and is loading a 10 gauge. AMD is in good position to take a big lead. Intel has wonderful parts no one can have, and AMD has good parts you can actually buy. Granted AMD has a problem with the cache speed and all, but at least you can buy em and sell them.
Me too. I've got an Athlon 500 at home, which, with the TNT2 Ultra I've got, feels so incredibly fast at every game I've tried...
And I'm buying two more. Doing my part to support them.
I don't understand why people would be buyin Intel chips - okay, Athlon motherboards are a bit more expensive and as far as I can tell, none of them support AGP 4X, but AGP 4X hasn't proved to be a big win anyway (and not many Intel boards support it). And after you pay for the Intel chip, the cheaper board doesn't win you anything.
I've got an Asus K7M, btw, which is a great board. Overclocking options right in the bios - woohoo!
(Oh, and check out this article, which basically says that you can add about 5% to all the benchmarks you've seen if you buy a newer motherboard or are lucky enough to have the newer version of the chipset that supports Super Bypass).
Perhaps my understanding of the problem isn't complete. The Athlon will support MP, and the EV6 will support MP... There's just no chipset for it yet, correct? I wonder why AMD isn't cracking on that. Seems like the Athlon would be a great Intel-killer if they got some MP motherboards out there.
--
A host is a host from coast to coast...
A host is a host from coast to coast...
Unless it's down, or slow, or fails to POST!
I ordered one for my boss last July, via web entry and it was a nightmare - after weeks of hearing nothing got on the phone and was passed from one dept to another, eventually back to sq. 1 - the old 'runaround'. A month later we got the unit, a solo laptop, and just yesterday I dropped it off at the 'Country Store' for servicing the display and a shift key that thinks it's a tab. "When she gets it on the service bench...." I think they're the usual big company more interested in a media image of being P.C. than selling quality PC's. Quoth the admin, Nevermore.
Boojum
try { do() || do_not(); } catch (JediException err) { yoda(err); }
The biggest bottleneck is in three parts:
1. RAM size. People today should be running at bare minimum 64 MB, with 128 MB of RAM preferred. Even Linux users who run graphical environments like KDE or Gnome will benefit from 128 MB of RAM.
2. Get a MUCH faster hard drive. While SCSI is the preferable solution, it's still quite expensive to implement, mostly due to the higher cost of SCSI host adapter and UW/UW2-SCSI hard drives. If your operating system can take advantage of the PIIX4E Intel chipset implmentation, you can do bus-mastering on ATA-33/66 IDE hard drives, which lifts a big load off the CPU; in that case, a good 7200 RPM hard drive is a must.
3. Get a faster graphics card. If the new graphics card has drivers supported by the operating system, then you will get faster redraws everywhere (even beyond the 3-D graphics acceleration so much touted nowadays).
Raymond in Mountain View, CA
I couldn't agree more. The last computer I bought (a couple of years ago at this point) uses an AMD chip, and I have no complaints. AMD has established itself as a "real" competitor to Intel, which has to be good from the consumers point of view. (At least, if you buy into the whole free market idea. *grin*)
Anyone have some benchmark numbers yet? I'd be interested in seeing how fast this baby can fly. Maybe I'll think about picking up a new machine soon.
My UID is the product of 2 primes.
This way he got something delivered to him in a few days ready to go with a warranty and customer support. He has had 0 problems with the thing for the last 3 months.
Two months ago I and a co-worker received two PC's from Gateway. They arrived slickly packages, booted up fine and performed like champs. Until I fdisked and attempted to install first Windows NT and then Linux. No luck with either - after many tries I determined the 20 gig harddisks must be broken in some way. Over the course of 3 calls to gateway support, each time being forced to wait about 40 minutes on hold to a machine, I was told by a Gateway techie that I'd have to go download a driver to boot NT, and that, for Linux, well, sorry. When I suggested that I had never had much luck with harddisks that require a driver located on the hardisk to boot, I was advised to go onto the motherboard and start yanking cables. Gateway's techie blamed the problem on the ultra-66 control, but it turned out to be nothing of the kind.
After hours of frustrated hardware juggling (the Gateway cables are cut exactly to length which sounds like a good idea and looks pretty but is actually stupid because you can't move anything to a new location without getting new cables) the computer still wouldn't take an install of anything. All we had for the effort were the usual skinned knuckles. Suspecting that the Gateway techie had a big problem in the cluefullness department, I went for a surf and learned that drives over 8 gig in size break the bad old bios way of addressing disk sectors, and that a bios extension was required. Gateway shipped those machines without the bios extension which had been commonly available for months. Not only that, but there wasn't even an update to fix the problem on their web page. Those machines only ever worked because they'd be juryrigged with a mysterious Win98-only hack to get around the fact that Gateway was shipping them with an obsolete bios.
The machines went back. We were out the cost of the shipping, the time, the skinned knuckles and the high blood pressure. I specced out a similar machine and ordered it from a local, medium-sized box builder and had no problems with it at all.
Life's a bitch but somebody's gotta do it.
SCSI drives cost way more than IDE drives. Do you really think it costs that much more to make them? It doesn't.
The drive manufacturers like having IDE around. They could produce SCSI drives at nearly the same price but they don't want to. It's better for them to sell IDE to the masses and have expensive SCSI disks around so they can get more money out of the businesses who need servers.
Put simply, it's not a matter of how much the drives cost, it's how much people are willing to pay.
I know this sounds like a big conspiracy (and it is) but it's really not uncommon... For example, Intel selling the Celeron cheaper than the P-II even though they cost the same to produce, or going farther back the 486SX vs. 486DX. Businesses like to have a low-end line where they can make money in volume and a high-end line where they can gouge the people who can afford to be gouged.
I'd say that disks are the biggest bottleneck today, just because memory is still too expensive to be able to fit all your apps and data in RAM. Not that you need to, but everytime the disk light lights up, that's what's eating up your time...
And as for "intellegent" devices... SCSI... it's here... it's been here forever... it's just most people are too much into cost and not into quality. Yep they cost more, but they're generally faster (because they represent the high-end of disk drives) and better constructed (again, better margins means they can spend more on parts to build them). And with me, copying files from one disk to the other has yet to cause my CPU usage to rise more than 4%...
Now we just need the motherboard vendors to jump on the bandwagon. We're off to a decent start but more choices, and easier availability of systems and parts is needed.
-- What you do today will cost you a day of your life.
RDRAM will NOT make your PC go much faster - sad but true! Maybe it will go a little bit faster but definately not more than you will feel from going from a 200Mhz - 550Mhz!
:-(
AGP is not a significant bottleneck as long as your video card has enough onboard RAM, and todays cards has 32MB and the next generation will have 64MB! It's quite enough, and futhermore, more and more games (which are the ones that use all this memory) will use texture compression, which will make the demand even smaller.
RDRAM has a huge bandwith, *but* it's latency "stinks" - it is worse than all the other kinds of memory, and often it is better to have memory with a lower latency than a higher bandwith.
It's sad but the producents of memory hasn't been able to make memory with lower latency - the latency has been going down with a very slow rate
The main thing to faster speed is to have enough RAM - there is a major difference between having to swap to the harddisk and having enough RAM and therefore not use a swapfile!
You're right about the harddisks - they have to be faster - much faster, they are the bottleneck right now.
Yeah, Western Digital is another one where you hear horror stories and glowing praise from different people.
I had great luck. I had a 5 gig go out. I went to their automated RMA page and downloaded their diagnostic software and ran it on the drive. It came back with a defect code that I then entered on the RMA page and they e-mailed an RMA the same day.
I shipped off the drive the next day, within 3 days I checked on the board to see if they had received it, they showed no records of the RMA number. Looking through the support posts, I saw complaints that drives had been sent in weeks ago with no word. Uh oh. I posted a message with my RMA asking what was going on. In an hour a tech posted that their RMA database was offline but he had tracked it by hand and a replacement was due to ship out the next day.
Two days after a brand-new 6 gig was dropped off by Airborne. Sheesh was I happy. Makes you wonder what goes on when you get really good service and then you see others that are in customer service hell with the same company.
Like I said, amazing to me how one customer interacts with a company and has a really bad experience, then the next has an easy time. Flawed systems.
I stick with IBM deskstar drives now, pretty much.
I hear both sides of this all the time...
I've been buying computers from South Dakota since I got a 386/33 back in '90.
Until I started building my own systems a couple years ago, I bought almost all my machines from Gateway and never had any trouble with the boxes themselves (GW2K's customer service, that's another story).
When my Dad went to buy a new machine to replace his old one late last year, I first tried to talk him into letting me build one for him. When he brought me an ad with the P-III/500 on sale at the Gateway Country Store and asked how much it would cost to build a similar machine, I took a step back and went with hime to the store to check it out.
They put together a very nice package for mid to entry-level machines. I could have made something pretty close, but that would have taken much more of my time than I had to space at the time and wouldn't have saved him that much time.
This way he got something delivered to him in a few days ready to go with a warranty and customer support. He has had 0 problems with the thing for the last 3 months.
I can recommend Gateway to anyone looking for a decent beginner to intermediate PC.
You may be surprised to find out that most 500MHz and 550MHz Athlons are actually 600MHz or 650MHz chips clocked down to meet demand at the "low end". There are reasonably cheap devices (GFDs -- GoldFinger Devices) which allow you to modify the multiplier and voltage of the Athlon, so you could take it upon yourself to simply run your part at a higher clock.
I mean, granted, you should be careful and test it thoroughly, but it is notable that I know many people who've attempted this sort of "overclocking", and none of those whom I've met (who have tried it) have yet failed to run a Athlon-500 at 700MHz, and a substantial fraction of these people run theirs at 750MHz or higher.
Just a suggestion. If you're not too timid and you're willing to take every precaution, then you could avoid spending five hundred or so dollars on an upgrade.
-JC
PS: Since I don't post here often, I should put up a disclaimer or something: Overclock at your own risk -- burnouts happen about one out of every hundred thousand attempts for careful overclockers, so please don't kill or blame me if bad things happen as a result of an overclocking attempt (eg, I take no responsibility).
Try compiling some serious C++ code
Um, where did I say I was talking about the PCI bus?! I was talking about the FSB (which should be pretty obvious, unless your CPU is on the PCI bus.
.sig is appropriate...
My comment was in reply to the following:
>Memory: Ever wondered why the transition from PC100 to PC133 doesnt give any significant performanceadvantage? Any why Celery with PC66 is pretty fast regardless?
which I thought was pretty obvious... oh well - at least your
"It's tough to be bilingual when you get hit in the head."
Hmm.... well, my dad's Cow (486DX-33) is still running just fine (only had a memory upgrade since he got it about what... 7 years ago...
My faithful P-100 Cow is running for a friend of mine now... I lost a hard drive once... after a big power surge. I also know of a few friends who got cows back in 1995 when I got mine - they were really nice boxes then - standard parts, easily upgradable, nice cases. Nothing to worry about.
Another friend has 2 cows, a P-200 and a P-II 450. He had one of the quantum drives die on him in the 200 (after some really weird stuff), but the machine was otherwise sturdy.
If I didn't build my own, I'd probably get another one...
From what I've heard lately though, people are nowhere near as satisfied as I was... sad. Though I personally would never go within 10 feet of a Micron, and used to like the Dell boxes, but lately, they have dissapointed me greatly...
I just want a dual GHz Athlon (133 DDR FSB, of course). Is that too much to ask 8^)
"It's tough to be bilingual when you get hit in the head."
Hmm... my celeron is greatly improved by the bus jump (comparing it at equal clock speeds)...
And my P-MMX 200 is *much* faster at 2.5 * 100 (250MHz) than at 4 * 66 (266MHz)... of course, this CPU is a very limiting factor in and of itself, but the 66->100 jump is pretty significant. The 100 -> 133 is less noticable.
Similar to the argument about 7200 vs. 10k rpm drives.
"Bah!" - Dogbert
"It's tough to be bilingual when you get hit in the head."
>The inability of Alpha NT to run Intel NT binaries
Well, there's always FX!32 (or whatever). Though it is software emu, so it is painfully slow... it does work, tho...
"It's tough to be bilingual when you get hit in the head."
Thanks for the reminder - there was some way to permanently alter the binary to be Alpha-based, and yeah - that did run pretty quick after the initial conversion - but if you are running iNT software on your alpha NT box off of a network share, it tends to not work quite as quickly.
Let's see - 1997 - Intel 200-233ish... Alpha - 533... mmmmmmm... clock speed.
"It's tough to be bilingual when you get hit in the head."
>All benchmarks I saw for celeron/PC66 showed a performance similar to P3/PC100.
Yeah - I've seen some of those, too... I just choose not to believe everything (or anything). All I can tell you is that I tested the C-466/66 against the C-300(@450/100) and the 450 won out everytime by a few percent... didn't blow it away, that's for sure, but it made a difference.
Most of the gaming benchmarks I've seen for the C vs. P-II/!!! have put the Celeron ~= with the P2/3 at the same clockspeed/FSB. Those with the same clockspeed, but the 66MHz FSB on the Celeron show a definite degradation of performance (or lack of enhanced performance, whatever).
Tests like CPUMark, FPUMark and a lot of the "normal everyday" benches aren't as reliant on the higher bus freq as something like Q3/UT, where you are spitting so much data out to the AGP slot...
True that on the PMMX comment... it makes a great router/firewall for the cable modem, though 8^D
"It's tough to be bilingual when you get hit in the head."
Yeah - every company has people on the plus and minus sides...
when that drive died - I called them (3pm sunday afternoon right after Thanksgiving 1996). They said the new one would get to me in 5-7 days (I was not thrilled). Turns out, I had it by 10 am the next morning, and the drive was a 2G to replace my 1G (just as cheap by that point).
I was also treated with a lot of respect when I called, and other times I've called back (as recently as last year when I fried the speakers that shipped with it), I've had only good experiences - less than 8 minutes on hold, people who understand that (since I do sysadmin work) I do do know something (but I'm always willing to try what they ask, too), and when I tell them what I've done they don't say "well, I need you to do that now, anyway".
I also know someone who wouldn't touch them with a ten foot pole, but he owns the only functioning Micron PC I've ever seen (1 of 7 ain't bad, is it?).
$.02
"It's tough to be bilingual when you get hit in the head."
How was I to know the Micro-Star MS-5185 motherboard that the system was built on was flaky as hell? (Micro-Star doesn't even acknowledge that model's existence on their site, BTW.) Confronting Gateway about the problem resulted in them saying, "It's due to that 'Linux' thing you're running." I ended up replacing nearly half the system components. The original processor is still cranking along, though.
Gateway will probably do well selling systems with AMD processors. They just need to make sure the other components of their systems aren't wretched crap.
Give a monkey a brain and he'll swear he's the center of the universe.
Um, that's your frontside bus, not your PCI bus. These are not the same. Your PCI cards do NOT run at 100 MHz.
Trees can't go dancing
So do them a big favor
Pretend dancing stinks!
Also, does this mean we'll see a Kryo CoolAthlon announcement for 1.066 GHz? (upward multiplier is 133%)...drool drool....
Yup, and you can read about it here on Tom's Hardware.
numb
Instead of bumping your RAM speed up to 800 MhZ, why not make your RAM bus 4x larger and capable of moving 256 bits in one cycle?
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
The K6-2 533 costs exactly the same as the 533 Celeron.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Gateway's never really been marketing to our demographic. We like to build our own (Which we can generally do for about 1/2 to 2/3'rds the price for comparable machines.)
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
AMD has a great chance to stomp Intel into the ground when Merced flops -- and it WILL flop. If AMD has the foresight to have their own 64 bit chip in the wings when Merced is released, they could take the market leadership away from Intel more at least through the end of the decade.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
is AGP 4x, PC133 RAM, more than 3 RAM slots, and SMP on K7 MBs. VIA has a chipset they claim is already out to deal with the first three. Of course official Red Hat support would be nice also as opposed to the "it's a to recent technology to be properly tested" on there website.
Of course, given the $100 difference between a K7@700MHz and a K7@750MHz I have fears of what the 800 price will be it, or it's availability (try to find a P3 800 without being a corporate customer to Dell, etc. and you will see what I mean)
Also, does this mean we'll see a Kryo CoolAthlon announcement for 1.066 GHz? (upward multiplier is 133%)...drool drool....
- Sig
Were you using Win 95? There is a Win 95 patch that fixes problems with 350+ Mghtz AMD processors. In fact, the memory errors you say you had sound like exactly one of the symptoms.
_______
I just wish I could c:\format Internet
So why did this happen for video cards, but not for disk drives? Probably because most users (who aren't running high-traffic Web sites at home) aren't complaining about the disk performance. They were complaining about video performance. Also, I'm not sure that NT or Win98 are really flexible enough to accept a plug-in replacement for the standard filesystem drivers, in which case there isn't much motivation for anyone to create such a product. (I'm assuming that a "filesystem accelerator card" would basically be a disk controller with a filesystem driver built in or loaded at boot time, which would then accept high-level filesystem requests from the OS's filesystem driver and take care of all the underlying details, just as is done with video accelerator cards.)
RDRAM performance only gives you a small improvement over systems with PC133 RAM, according to the latest Windows general and game specific benchmark tests at Tom's Hardware. In the windows general test, the point spread between the i820/PC133 RAM and the i840/RDRAM is something like 25 points, with the lowest scoring 321 on the BAPco SYSMark98, and the highest at 343. In the game test, the i840 scores 122 in Q3 Arena (640x480x16), while the i820 scores 111.6. Hardly a mind boggling improvement.
You are not going to notice that difference in a game. But you are going to feel it quite long, hard and wide in your pocket book. A whole lotta spankin' for nothin'. You can get an Athlon 750 for less, and the performance in real time is either superior, or not far enough back behind the Coppermine to justify the cost of the Coppermine/RDRAM combo.
Likewise RDRAM + the Athlon won't mean anything, either. Athlon's support of AGP4x will mean a lot, however.
--- Grow a pair, liberals... stop letting the Republicans bully you!
Kryotech is selling the Super G, a 1000mhz Athlon running at a temperature of -40C.
However, your problem is not the chip speed, but the surrounding bus speed and RAM speed. The chip runs at 1000mhz, but the Front Side Bus is only 200mhz and the RAM runs at 100mhz. YUCK! Talk about your kinks in the hose!
The things you'll want to address before you improve your cpu speed are:
1) RAM speed. You'll theoretically wanna blast open that blockage in the pipeline with 800mhz RDRAM, although it has been shown that this does not deliver anywhere near the astronomical improvement that it is meant to deliver.
2) Cache speed. It runs at half the speed of the cpu - or, almost half. This needs to be rectified ASAP.
3) AGP support. This is very important in games. I predict AGP4x will really unleash the speed demon in every computer.
Is it any wonder that with all these bandwidth limitations, the Super Bypass only yields a 2-5% increase in performance over a non Super Bypass board?
--- Grow a pair, liberals... stop letting the Republicans bully you!
I just upgraded to an Athlon 500 from an old P200MMX. Did I notice that much speed difference? Only in graphic intensive 3d games. And that's because the video card is 10 times better. It doesn't matter how fast the CPU is, there's just too many bottlenecks right now to slow it down. Now, if the Athlon MB came with support for RDRAM (and the cost of RDRAM goes down significantly) and onboard SCSI, along with an AGP 4X, then it might go faster.
I think a lot of Slashdotters tend to forget that there are people outside of Slashdot who buy computers. Your average home user, for example, has never heard of AMD, but ask them about Intel, and they'll tell you about funky television ads.
Ask them what the Intel Pentium Bug is all about, or the F0 0F bug, and they won't have any idea what you're talking about.
Average consumers, and even business consumers, are influenced by advertising, and not by technical performance. I know it's an American tradition (and even more a Linux tradition) to root for the underdog, but remember that Microsoft still has ~85% market share as far as operating systems go. That's due to the same people who would buy and Intel processor because they saw them on TV.
Compaq is going with AMD too - both on their desktops and on their laptops!
Futhermore they are using Firewire instead of USB.
This could be a serious boost for Firewire!
The story is here:
http://www.theregister.co.uk/000103-000001.html
I know this is slightly off topic. I was thinking about doing an "Ask Slashdot" but felt it would be wasteful. I'd like to use an Athlon on my next box and want it to run Linux/Win9x. I've heard there are some problems with some mainboards. Does anyone know for certain which work the best under Linux? I haven't seen anything on the distro sites other than "some mainboards have caused problems with Linux."
Do really dense people warp space more than others?
>We need intelligent devices! Badly! Floppy drives and hard disks with their own memory and processors.
Get a good SCSI subsystem - good cards have nice processors on them, and can have several to many megs of cache. Costs more, but so does everything in your comment...
64 bit or 66MHz PCI (both, preferably) would be a welcome addition, and are not as cost prohibitive as many other solutions. 64-bit 33MHz devices can coexist with 32-bit devices, and still maintain the speed advantage... PCI-X makes this even better - remember PCI was designed as a cost-efficient performance bus. Mostly cheap, sorta fast, reliable. What are you using that presently ships with EISA? Yikes...
Not that your sound card, modem, or even 10BaseT network really taxes PCI all that much... and those should all have processors, bus-mastering, and DMA (damn win-modems!).
If you want a pretty extreme example of what you are mentioning, go look at an AS/400 - separate processors for all of the I/O functions, high speed internal busses between subsytems... and expensive... do you want to spend $1-2k for a decent system, or $80k... it's up to you...
"It's tough to be bilingual when you get hit in the head."
800 mhz...
Since I've stopped playing games other than nibbles, tetris and such gems, the mhz-race is something I watch and giggle somewhat at.
What you really need for ordinary desktop apps is MEMORY (goes for Linux too, only slightly less than in Windows), not more mhz. And even so - Mhz is not everything for a processor, a well designed processor with slightly less Mhz can be faster than a poorly designed Mhz-rich processor.
// Simon, remembers his 1-Mhz 8-bit computer...
The -biggest- bottleneck is the bus. PCI and (E)ISA are all way too slow. I believe there's an extended version of VME, which is a seriously nice bus system. A PC with a VME bus would get a serious performance boost.
The next biggest bottleneck is memory. It's WAAAY too slow, partly (I suspect) due to the large distances between the chips. Large distances mean a lot of time wasted synchronising and/or waiting. That's time better spent on other tasks. Fewer physically larger chips would solve this problem, by reducing the average distances travelled considerably.
Another factor in the slowness is that chip manufacturers prefer to churn out low-cost, low-speed memory in bulk, which forces people to then buy lots more much faster memory for cache. If main memory ran at a decent speed, we wouldn't have that problem in the first place.
A third bottleneck is in the speed of devices. These are ALL controlled by the main processor, even today. =COMMODORE= were doing better than that, in the 70's! We need intelligent devices! Badly! Floppy drives and hard disks with their own memory and processors. Not just a few scraps of cache, but enough to do useful work. If you're copying files from one drive to another, there is NO reason, WHATSOEVER, for the main processor to be involved at all.
The same is true of printing. If you had your disk drives loaded with a basic OS and some print drivers, you wouldn't end up swapping files everywhere on the system just to get them to the printer. Computers should be designed to be efficient, rather than like a Donald Duck cartoon. Cartoons are great, but they aren't really the best place for inspiration for computer design.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)