Actually from what I understand it is those profit numbers that actually made the FCC se mi forcegoogle to IPO in the first place. Apparently there are regulations concerning how much a private company can earn.... shrugs.
Like jhunsake says, you are quite incorrect.
First off, the FCC has nothing to do with Google's line of business. Maybe you're thinking of the SEC.
Secondly, the rule is not about how much profit/income you can make before you have to go public. (UPS was a $20+ billion dollar company prior to going public, with a reasonably healthy profit margin. A *lot* of employees held private stock too.)
What there is a rule about (and I don't remember the specifics off-hand) is that once X number of people hold private stock in the company, you have to start making annual (quarterly?) reports to the SEC. Even if you're not a publically traded company. The cost of preparing these reports is rather high, and not much different then the cost of being a public company and reporting to the SEC anyway. At which point, since you're already doing the legwork, why not go public and try and reap a windfall from the IPO. Or you can choose to just suck up the costs and stay completely private.
(Whether or not that's the point of an IPO is another debate for another time.)
Some of the unluckly ones of us will still be driving exactly the same cars we're driving today. It's only ten years away right?
Unlucky? I have a brand new car, that's completely paid off, so that could be ten years without having a monthly car payment. The car after that could well be a hybrid though, if they can drive the costs down a bit more. I expect that even in the US, gasoline will cost 50% more then today (maybe adjusted for inflation... shrug). That might be enough economic incentive to drive folks towards fuel efficiency over vehicle size.
(Short-term price spikes in gasoline prices can/will be shrugged off, it will take 24+ months of high prices before people will seriously consider switching to something more efficient.)
"BZZT Wrong answer" has to be the most annoying and childish catch-phrase of Slashdotters. It implies enormous intellectual superiority over the original poster, and yet a majority of the time, someone else (as happened here) proves the opposite is true.
Ranks right up there with the show, "Missing Link", doesn't it? (Don't get me wrong, I'm in the same boat as you.)
According to the FAQ, Sunbird gets the date format from your OS. To use the ISO format, go to the Windows Start Menu, then Settings > Control Panel > Regional Options (not Date/time) > Date tab.
Bit of a cop-out. I would prefer to have the option to specify my date-format and over-ride the built-in O/S settings. I might find "dd/mmm" to be a pleasing to the eye format for my task due dates. (There's also the issue that some older software, written by Microsoft, will *break* if you change those regional settings.)
All of my computer needs are pretty much open source now, except for the bloody calander, and now that there is a decent alternative, it doesn't do me much good.
I'm in the same boat, stuck with MSOutlook because Mozilla doesn't:
- sync its calendar/task list with my Palm
- offer pop-up mail alert dialogs as a filter action
- have a way to play a sound as part of a filter action
- allow you to set a task with a due date but no time
And of course, the issue that you have to have the calendar app open in order to receive alerts...
1) Makes incremental backups easier. More of an issue for some users then others, depending on the size of your MBOX files (rsync helps a lot).
2) Some A/V software do nasty things when they find a virus in a MBOX file (as in quarantining the entire MBOX, unless you exclude that directory). With individual files, the A/V scanner would only end up trashing the affected e-mail.
3) Much easier to grep your mail at the command line. You can even use command-line tools to move messages around.
Well, at least Mozilla uses a "standard" format. MBOX ain't perfect, but it does the job well (some mailing list folders that I have are well into the 5-digit post count range).
The problem is that all of the WebDAV documentation is written for intermediate/advanced users or is focused soley on configuration documentation. That generalization applies to all of the documentation that I read a few months ago while trying to figure out SubVersion+WebDAV.
Rsync and OpenSSH are just as bad when it comes to the whole issue. It's very hard to find a high-level view of a lot of these systems, that explains more then "WebDAV allows you to edit documents over port 80". A good high level summary should include information about the security model, where it fits into the infrastructure, and good examples of usage (simple and moderately complex cases).
(Yes, I know this is probably rantish, but I'm not surprised that the original poster found WebDAV confusing.)
Re:Power consumption and price comparison
on
NSLU2 Now More Useful
·
· Score: 3, Informative
Most 5400rpm drives are 6-7W idle and ~10W when seeking. The 7200rpm drives are a bit more power hungry, 7-10W idle and 10-13W seeking. Dunno about 10k RPM drives, but for a low-power fanless server you'd want to stick with cool running 5400rpm drives anyway. (Those numbers are from the manuf websites.)
Here in the northeast US, 10W of power draw costs $0.60/mo. Figure a 25W low-power CPU like a VIA C3, another 10W for the motherboard, plus 2x7.5W for a pair of 250GB 5400rpm drives in RAID1. That *should* clock in at around 50W on average, and maybe 40W if the disks spin down. The NSLU2 draws 10W (max) plus another 15W for the (2) USB hard drives for a total of 25W (being conservative). So the cost savings is around $1/mo.
Unit price for the NSLU2 is $80, plus another $50 for a pair of USB enclosures. Definitely cheaper then building a mini-ITX system (est $300-$400, not including drives).
All that being said, I prefer my toaster-sized mini-ITX linux server.
I've seen rumors that MS is throttling the number of systems that can download SP2 automatically. So only a limited number of systems per day will receive SP2 as a download.
1) Burn a CD with XP SP2 on it at work, a friend's place or wherever
2) Install XP fresh without being connected to the net
3) Install SP2 from the CD next
4) Install everything else
You forgot step #5, make a backup image/snapshot of the system using Ghost or Knoppix. (Usually prior to installing anything other then basic applications.)
I personally don't see a point in the displays that are thin, like plasma and LCD panels. I don't see a point in the display being thinner than the accompanying components are deep, because the components would protrude or you'd have a big air gap behind the display, if the display is flush with the DVD player, amp, etc. So I'd consider a 15" to 20" depth to be pretty comfortable.
Weight.
I'm not overly impressed with the idea of wall mounting a display either, given that the components aren't also wall mounted. However, a thin display does give you the option to wall mount, which could make it easier to re-arrange the room. But a screen that is the same size, but half the weight has some big advantages come moving time.
I still wonder why we can't move away from floppies. I mean we made the switch from 5.25 to 3.5. The only thing I see taking the floppies place right now is the cdburner and there are so many limitations to that media. I've got floppies from the early 90s that I still read/write to....I don't use cds that way.
Why did nothing ever replace the floppy disc?
Because manufacturers got greedy.
Iomega's ZIP, Sony's LS-120, and a bunch of other small sized, 100MB+ capacity discs were all supposed to be "floppy killers". However, due to greed, none of the companies would cross-license or agree to a common standard. Which had the nasty side-effect of keeping prices for both the media and drives high. Drive costs needed to be on the order of $50 with media costs in the $2 range (MD could've been a contender, but Sony is their own worst enemy).
Once CD-R media broke $2/disc (or CD-RW), it no longer made sense and they quickly priced themselves out of the market. Even at the tail end, ZIP disks were $10 or $15 compared to a $2 CD-RW which held 5x or 2x as much. Even better, a CD-RW could often be read in any system with a regular old CD drive.
USB flash drives are probably the only thing that's going to kill of floppies, even though they're slightly more difficult to use.
That's what always utterly mystified my about marketing dweebs:
I think the idea is that there are enough *other* rubes out there, that your lost business can be blamed on "poor market conditions". Plus, if you're a critical-thinking person, you're not really a good consumer anyway (they want impulse buyers of expensive things like cars, boats, hair gel).
I'm not convinced modern machines could even handle this: Bear in mind that IDE buses run at what - 250mbps max?
How would an OS react to suddenly having to catalogue a multi-terabyte disk? By locking, I suspect.
Even today's systems would have zero issues with a 1TB removable media. The only question would be whether the UDF file system can handle discs of that size, or if we'd have to come up with yet another file system.
After all, we already have 300GB drives in common use, and it's not difficult to buy a $1000 RAID0 external firewire drive that's 1TB. (Those are either formatted as HFS, NTFS, or something else.)
Bandwidth might be a bit of an issue, but the capacity of SATA is 150 MB/s (slightly faster then PCI bandwidth, but PCIe raises that limit a lot). So that wouldn't remain a bottleneck for long (or at least it would move around to another part of the system again).
However, correctly handled, the CD-R:s should last quite long. I've a few that are over 10 years old (burned back in 1993 with a single-speed CD-R-drive or something like that). And they still work just fine. I've never had a CD-R fail on me except during the burning phase.
Still, you're depending on luck (or redundant copies).
One of the things I've started doing with any DVDs is putting additional recovery data on the discs. That gets me both an additional verification method (MD5), plus if I detect errors it's possible to repair the damage.
QuickPar will let you specify anything from 0.5% recovery records (maybe lower!) up to 100% recovery records (where half your disc is filled with recovery data). I usually use 5%, which doesn't take much space, is better then nothing, and useful for cases where I don't want to burn a backup copy of the media.
Man, it is such a challenge buying a video card these days. It's really difficult to find the "sweet spot" of price versus performance.
No kidding.
I prefer not to pay more then $200 for a video card, because that's the pain threshold for my wallet. (If a $500 card goes up in smoke, I'd be seriously put out, but a $200 card I would be more reasonable about it going poof.)
Combined with the fact that there are easily two dozen manufs of NVIDIA cards, complete with their own differences (memory speeds, etc.) and completely made-up naming conventions. Very confusing to someone who just wants to figure out what video card is going to be twice as fast as his old, trusty GeForce4 Ti 4600. A lot of the review sites only focus on the latest and greatest, without putting average costs next to the performance scores.
Eventually, after a week of stupid web browsing, I'm either going after the baseline 5900XT or a baseline 6800 (if the prices drop below $200).
It's been a while so I might have the numbers wrong...NT 4 SP4 was issued to fix NTFS which was horribly crippled by NT 4 SP3.
Yep, some of the NT service packs were horrid.
Personally, I'm holding off for a few weeks to see if XP SP2A shows up. I'm in no rush to break systems, and I have other things on my plate to play beta-test monkey this week.
Re:This is not a crappy computer
on
You've Got PC
·
· Score: 1
Someone please tell me why a normal computer user (the type who uses AOL) needs anything faster than a 2GHz Celeron with 256MB of RAM?
You mean besides the fact that WinXP runs like mollasses in January on only 256MB of RAM?
That's one I have with the box, the rest of it looks alright for a basic PC, but only putting 256MB in there is really iffy. They should've sourced a cheaper CPU (even a 1Ghz CPU can work great if it has enough memory). It's gonna swap like crazy if they try to do AOL + e-mail + office all at the same time.
The second complaint is lack of a DVD-ROM. They could've easily dropped one in there if they had gone with a cheaper CPU (same with more memory).
How depressing. I just droped $500 on my 256mb Geforce FX 5900 Ultra back in September and, less than a year later, it's going to be two or three revisions behind? Usually a top model card takes a year to be replaced, let alone be replaced twice or thrice over.
Your cash to blow and you're the one that made the choice.
Personally, my limit is $150-$200 for a video card. Sure, the 6800UltraExtremeSuperDooper gets 3x the scores of my aging but trust GeForce4 Ti4600, but no way am I paying $500 for one.
Instead, I'll upgrade to the GeForce FX 5900 XT, which seems to be a good tradeoff between cost/performance. From the benchmarks that I've seen, it's about 1.5-2.0x faster then my Ti4600, and the cost is easily under $200. Heck, I got a 10% boost in performance by switching back to older NVIDIA drivers, rather then using the latest and greatest for this 2-year old card.
If it makes designing motherboards easier why is there only one Opteron chipset out there and it's made by AMD?
NVIDIA has an nForce3 chipset, VIA has a K8T800 chipset.
I'm sure there are more by now.
From what I recall, AMD always makes the first chipset as a reference board to make sure everything plays together and you can fit everything into an ATX-sized motherboard. They're not in the chipset business, so they get back out of the pool as soon as possible and let VIA/NVIDIA (and others) carry that load.
Remember too, that you can only "barely" get 2GB ram chips and they're VERY expensive... there's nobody with enough money to max the things out...or at least be more than a niche!!
While spot-checking prices... here's basically how it is today. (This is for PC2100, registered ECC, which is the most expensive. Kingston - DDR 266Mhz Cas 2.5 (PC2100) 184Pin Dimm 3.3v Gold)
512MB chips are $112
1GB chips are $260
2GB chips are $600
4GB chips (not Kingston) are $4000
So the 2GB chips at $300/GB are only 15% more expensive then the 1GB chips. So that 16-slot Opteron board could hold 32GB of RAM without breaking the bank. (Dunno if it supports 2GB DIMMs or not.)
And everyone who can afford 2^36 bytes of RAM, raise your hand...
2^36 is only 64GB of RAM (16x). At today's prices (making the very iffy assumption that you can somehow cram 32 2GB chips in), that's only $19100 or so. More likely, you'd have to use the $1000/GB chips which would be $64000.
The problem is not that we can't afford 2^36 bytes of RAM today, it's that we *will* be able to afford 2^36 bytes of RAM in just a few years. Five years ago, 64MB PCs were common, and five years before that (roughly) 4MB PCs were common. In today's world, 1GB machines are common and it's reasonable to assume that in another 5 years that 64GB machines will be quite possible/common for close to the same costs as today.
Servers will naturally get there first (or workstation-class machines). Either using the memory as cache to fill those gigabit or 10 gigabit networks or to hold larger databases completely in memory.
Heck, with a 64GB workstation, you can load an entire uncompressed DVD clip into memory which should make applying effects to the clip faster. (Just like older PCs couldn't load entire sound clips in because there wasn't enough RAM.)
Slashdot Mirror
Actually from what I understand it is those profit numbers that actually made the FCC se mi forcegoogle to IPO in the first place. Apparently there are regulations concerning how much a private company can earn.... shrugs.
Like jhunsake says, you are quite incorrect.
First off, the FCC has nothing to do with Google's line of business. Maybe you're thinking of the SEC.
Secondly, the rule is not about how much profit/income you can make before you have to go public. (UPS was a $20+ billion dollar company prior to going public, with a reasonably healthy profit margin. A *lot* of employees held private stock too.)
What there is a rule about (and I don't remember the specifics off-hand) is that once X number of people hold private stock in the company, you have to start making annual (quarterly?) reports to the SEC. Even if you're not a publically traded company. The cost of preparing these reports is rather high, and not much different then the cost of being a public company and reporting to the SEC anyway. At which point, since you're already doing the legwork, why not go public and try and reap a windfall from the IPO. Or you can choose to just suck up the costs and stay completely private.
(Whether or not that's the point of an IPO is another debate for another time.)
Some of the unluckly ones of us will still be driving exactly the same cars we're driving today. It's only ten years away right?
Unlucky? I have a brand new car, that's completely paid off, so that could be ten years without having a monthly car payment. The car after that could well be a hybrid though, if they can drive the costs down a bit more. I expect that even in the US, gasoline will cost 50% more then today (maybe adjusted for inflation... shrug). That might be enough economic incentive to drive folks towards fuel efficiency over vehicle size.
(Short-term price spikes in gasoline prices can/will be shrugged off, it will take 24+ months of high prices before people will seriously consider switching to something more efficient.)
"BZZT Wrong answer" has to be the most annoying and childish catch-phrase of Slashdotters. It implies enormous intellectual superiority over the original poster, and yet a majority of the time, someone else (as happened here) proves the opposite is true.
Ranks right up there with the show, "Missing Link", doesn't it? (Don't get me wrong, I'm in the same boat as you.)
According to the FAQ, Sunbird gets the date format from your OS. To use the ISO format, go to the Windows Start Menu, then Settings > Control Panel > Regional Options (not Date/time) > Date tab.
Bit of a cop-out. I would prefer to have the option to specify my date-format and over-ride the built-in O/S settings. I might find "dd/mmm" to be a pleasing to the eye format for my task due dates. (There's also the issue that some older software, written by Microsoft, will *break* if you change those regional settings.)
All of my computer needs are pretty much open source now, except for the bloody calander, and now that there is a decent alternative, it doesn't do me much good.
I'm in the same boat, stuck with MSOutlook because Mozilla doesn't:
- sync its calendar/task list with my Palm
- offer pop-up mail alert dialogs as a filter action
- have a way to play a sound as part of a filter action
- allow you to set a task with a due date but no time
And of course, the issue that you have to have the calendar app open in order to receive alerts...
1) Makes incremental backups easier. More of an issue for some users then others, depending on the size of your MBOX files (rsync helps a lot).
2) Some A/V software do nasty things when they find a virus in a MBOX file (as in quarantining the entire MBOX, unless you exclude that directory). With individual files, the A/V scanner would only end up trashing the affected e-mail.
3) Much easier to grep your mail at the command line. You can even use command-line tools to move messages around.
Well, at least Mozilla uses a "standard" format. MBOX ain't perfect, but it does the job well (some mailing list folders that I have are well into the 5-digit post count range).
The problem is that all of the WebDAV documentation is written for intermediate/advanced users or is focused soley on configuration documentation. That generalization applies to all of the documentation that I read a few months ago while trying to figure out SubVersion+WebDAV.
Rsync and OpenSSH are just as bad when it comes to the whole issue. It's very hard to find a high-level view of a lot of these systems, that explains more then "WebDAV allows you to edit documents over port 80". A good high level summary should include information about the security model, where it fits into the infrastructure, and good examples of usage (simple and moderately complex cases).
(Yes, I know this is probably rantish, but I'm not surprised that the original poster found WebDAV confusing.)
Most 5400rpm drives are 6-7W idle and ~10W when seeking. The 7200rpm drives are a bit more power hungry, 7-10W idle and 10-13W seeking. Dunno about 10k RPM drives, but for a low-power fanless server you'd want to stick with cool running 5400rpm drives anyway. (Those numbers are from the manuf websites.)
Here in the northeast US, 10W of power draw costs $0.60/mo. Figure a 25W low-power CPU like a VIA C3, another 10W for the motherboard, plus 2x7.5W for a pair of 250GB 5400rpm drives in RAID1. That *should* clock in at around 50W on average, and maybe 40W if the disks spin down. The NSLU2 draws 10W (max) plus another 15W for the (2) USB hard drives for a total of 25W (being conservative). So the cost savings is around $1/mo.
Unit price for the NSLU2 is $80, plus another $50 for a pair of USB enclosures. Definitely cheaper then building a mini-ITX system (est $300-$400, not including drives).
All that being said, I prefer my toaster-sized mini-ITX linux server.
I've seen rumors that MS is throttling the number of systems that can download SP2 automatically. So only a limited number of systems per day will receive SP2 as a download.
1) Burn a CD with XP SP2 on it at work, a friend's place or wherever
2) Install XP fresh without being connected to the net
3) Install SP2 from the CD next
4) Install everything else
You forgot step #5, make a backup image/snapshot of the system using Ghost or Knoppix. (Usually prior to installing anything other then basic applications.)
I personally don't see a point in the displays that are thin, like plasma and LCD panels. I don't see a point in the display being thinner than the accompanying components are deep, because the components would protrude or you'd have a big air gap behind the display, if the display is flush with the DVD player, amp, etc. So I'd consider a 15" to 20" depth to be pretty comfortable.
Weight.
I'm not overly impressed with the idea of wall mounting a display either, given that the components aren't also wall mounted. However, a thin display does give you the option to wall mount, which could make it easier to re-arrange the room. But a screen that is the same size, but half the weight has some big advantages come moving time.
http://www.pinkroom.biz/owl/minirot13/ First hit on Google for Mozilla ROT13
Useful link, but why is that functionality not built directly into the base product?
(Yet another topic for a bug in Bugzilla.)
I still wonder why we can't move away from floppies. I mean we made the switch from 5.25 to 3.5. The only thing I see taking the floppies place right now is the cdburner and there are so many limitations to that media. I've got floppies from the early 90s that I still read/write to....I don't use cds that way.
Why did nothing ever replace the floppy disc?
Because manufacturers got greedy.
Iomega's ZIP, Sony's LS-120, and a bunch of other small sized, 100MB+ capacity discs were all supposed to be "floppy killers". However, due to greed, none of the companies would cross-license or agree to a common standard. Which had the nasty side-effect of keeping prices for both the media and drives high. Drive costs needed to be on the order of $50 with media costs in the $2 range (MD could've been a contender, but Sony is their own worst enemy).
Once CD-R media broke $2/disc (or CD-RW), it no longer made sense and they quickly priced themselves out of the market. Even at the tail end, ZIP disks were $10 or $15 compared to a $2 CD-RW which held 5x or 2x as much. Even better, a CD-RW could often be read in any system with a regular old CD drive.
USB flash drives are probably the only thing that's going to kill of floppies, even though they're slightly more difficult to use.
That's what always utterly mystified my about marketing dweebs:
I think the idea is that there are enough *other* rubes out there, that your lost business can be blamed on "poor market conditions". Plus, if you're a critical-thinking person, you're not really a good consumer anyway (they want impulse buyers of expensive things like cars, boats, hair gel).
I'm not convinced modern machines could even handle this: Bear in mind that IDE buses run at what - 250mbps max? How would an OS react to suddenly having to catalogue a multi-terabyte disk? By locking, I suspect.
Even today's systems would have zero issues with a 1TB removable media. The only question would be whether the UDF file system can handle discs of that size, or if we'd have to come up with yet another file system.
After all, we already have 300GB drives in common use, and it's not difficult to buy a $1000 RAID0 external firewire drive that's 1TB. (Those are either formatted as HFS, NTFS, or something else.)
Bandwidth might be a bit of an issue, but the capacity of SATA is 150 MB/s (slightly faster then PCI bandwidth, but PCIe raises that limit a lot). So that wouldn't remain a bottleneck for long (or at least it would move around to another part of the system again).
However, correctly handled, the CD-R:s should last quite long. I've a few that are over 10 years old (burned back in 1993 with a single-speed CD-R-drive or something like that). And they still work just fine. I've never had a CD-R fail on me except during the burning phase.
Still, you're depending on luck (or redundant copies).
One of the things I've started doing with any DVDs is putting additional recovery data on the discs. That gets me both an additional verification method (MD5), plus if I detect errors it's possible to repair the damage.
QuickPar will let you specify anything from 0.5% recovery records (maybe lower!) up to 100% recovery records (where half your disc is filled with recovery data). I usually use 5%, which doesn't take much space, is better then nothing, and useful for cases where I don't want to burn a backup copy of the media.
Whoever modded the above as off topic clearly hasn't heard of ROT13. Who doesn't have ROT13 for their webbrowser and mail/usenet client?
Mozilla Mail doesn't have it (neither in the mail portion or the USENET portion).
Man, it is such a challenge buying a video card these days. It's really difficult to find the "sweet spot" of price versus performance.
No kidding.
I prefer not to pay more then $200 for a video card, because that's the pain threshold for my wallet. (If a $500 card goes up in smoke, I'd be seriously put out, but a $200 card I would be more reasonable about it going poof.)
Combined with the fact that there are easily two dozen manufs of NVIDIA cards, complete with their own differences (memory speeds, etc.) and completely made-up naming conventions. Very confusing to someone who just wants to figure out what video card is going to be twice as fast as his old, trusty GeForce4 Ti 4600. A lot of the review sites only focus on the latest and greatest, without putting average costs next to the performance scores.
Eventually, after a week of stupid web browsing, I'm either going after the baseline 5900XT or a baseline 6800 (if the prices drop below $200).
-1 Flaimbait
-2 Astroturf
-3 Spelling Nazi
-4 Fanboy
It's been a while so I might have the numbers wrong...NT 4 SP4 was issued to fix NTFS which was horribly crippled by NT 4 SP3.
Yep, some of the NT service packs were horrid.
Personally, I'm holding off for a few weeks to see if XP SP2A shows up. I'm in no rush to break systems, and I have other things on my plate to play beta-test monkey this week.
Someone please tell me why a normal computer user (the type who uses AOL) needs anything faster than a 2GHz Celeron with 256MB of RAM?
You mean besides the fact that WinXP runs like mollasses in January on only 256MB of RAM?
That's one I have with the box, the rest of it looks alright for a basic PC, but only putting 256MB in there is really iffy. They should've sourced a cheaper CPU (even a 1Ghz CPU can work great if it has enough memory). It's gonna swap like crazy if they try to do AOL + e-mail + office all at the same time.
The second complaint is lack of a DVD-ROM. They could've easily dropped one in there if they had gone with a cheaper CPU (same with more memory).
How depressing. I just droped $500 on my 256mb Geforce FX 5900 Ultra back in September and, less than a year later, it's going to be two or three revisions behind? Usually a top model card takes a year to be replaced, let alone be replaced twice or thrice over.
Your cash to blow and you're the one that made the choice.
Personally, my limit is $150-$200 for a video card. Sure, the 6800UltraExtremeSuperDooper gets 3x the scores of my aging but trust GeForce4 Ti4600, but no way am I paying $500 for one.
Instead, I'll upgrade to the GeForce FX 5900 XT, which seems to be a good tradeoff between cost/performance. From the benchmarks that I've seen, it's about 1.5-2.0x faster then my Ti4600, and the cost is easily under $200. Heck, I got a 10% boost in performance by switching back to older NVIDIA drivers, rather then using the latest and greatest for this 2-year old card.
If it makes designing motherboards easier why is there only one Opteron chipset out there and it's made by AMD?
NVIDIA has an nForce3 chipset, VIA has a K8T800 chipset.
I'm sure there are more by now.
From what I recall, AMD always makes the first chipset as a reference board to make sure everything plays together and you can fit everything into an ATX-sized motherboard. They're not in the chipset business, so they get back out of the pool as soon as possible and let VIA/NVIDIA (and others) carry that load.
Remember too, that you can only "barely" get 2GB ram chips and they're VERY expensive... there's nobody with enough money to max the things out...or at least be more than a niche!!
While spot-checking prices... here's basically how it is today. (This is for PC2100, registered ECC, which is the most expensive. Kingston - DDR 266Mhz Cas 2.5 (PC2100) 184Pin Dimm 3.3v Gold)
512MB chips are $112
1GB chips are $260
2GB chips are $600
4GB chips (not Kingston) are $4000
So the 2GB chips at $300/GB are only 15% more expensive then the 1GB chips. So that 16-slot Opteron board could hold 32GB of RAM without breaking the bank. (Dunno if it supports 2GB DIMMs or not.)
And everyone who can afford 2^36 bytes of RAM, raise your hand...
2^36 is only 64GB of RAM (16x). At today's prices (making the very iffy assumption that you can somehow cram 32 2GB chips in), that's only $19100 or so. More likely, you'd have to use the $1000/GB chips which would be $64000.
The problem is not that we can't afford 2^36 bytes of RAM today, it's that we *will* be able to afford 2^36 bytes of RAM in just a few years. Five years ago, 64MB PCs were common, and five years before that (roughly) 4MB PCs were common. In today's world, 1GB machines are common and it's reasonable to assume that in another 5 years that 64GB machines will be quite possible/common for close to the same costs as today.
Servers will naturally get there first (or workstation-class machines). Either using the memory as cache to fill those gigabit or 10 gigabit networks or to hold larger databases completely in memory.
Heck, with a 64GB workstation, you can load an entire uncompressed DVD clip into memory which should make applying effects to the clip faster. (Just like older PCs couldn't load entire sound clips in because there wasn't enough RAM.)