Not all of us are willing to blow $500+ on a device that doesn't enable us to do anything we couldn't do before.
I need a laptop to get my job done. I need a mobile phone for a variety of reasons. I can't think of a single damn reason why I need a $500 tablet.
Until you get that not all of us are willing to spend $500 on gadget porn, you won't grok why some people think the iPad is a tremendous waste of money and attention.
The "transmission ratio is controlled by the relative speed of the motor driven shaft" technique is exactly what's used by the transmission in the Prius (and every other Toyota, Nissan, and Ford hybrid).
No one 'seeded' me to think we should be hanging senators for their crimes.
If you literally believe that, you're crazy and dangerous. If not, stop egging on the people who are crazy and dangerous.
Saying that you want to vote the bastards out of office is reasonable. Saying that you want to see criminal prosecutions against those who you believe have acted illegally is reasonable. Threatening violence is not.
I can assure you that restore points (shadow copies) DO work, and they have saved my ass more than once. However, Windows can only keep as many shadow copies as you give it room to store. If you fill up your disk, shadow copies WILL be deleted. This is a persistent issue on my Intel X25-m 80GB SSD, which is pretty much constantly near full because of its small size. I usually only get a few days of shadow copies as a result.
That said, neither Linux nor Mac OS X have anything like shadow copies unless you go configure them yourself.
On Linux, you can use a COW filesystem like ZFS or btrfs to get similar functionality, but practically no one uses either on a personal machine since ZFS runs in FUSE on Linux and btrfs is not entirely stable yet. The vast majority of Linux distros today default to ext4, which is a file filesystem but doesn't offer any sort of snapshotting.
On Mac OS X, you would need to configure Time Machine, which is great but requires an external disk or a Time Capsule (or a hacked NAS). I would bet that a substantial portion of Mac users never set up Time Machine simply because they don't have the necessary hardware.
The point is, Windows - even with shadow copies off - is no worse than Linux or Mac OS X out of the box. Shadow copies are not a backup mechanism and they aren't intended to replace one. Obviously if there's a bug here, it's serious and needs to be fixed. But the very nature of shadow copies means that they cannot be a long-term disaster recovery mechanism. If you're relying on shadow copies instead of backing up, you're stupid. Shadow copies offer zero protection against hardware failure, disasters (fire), or malware, even when they work.
That said, shadow copies are one of the best reasons to use Windows 7 (some versions of Vista had them too). No, they're not perfect. But the fact that they're on by default means that millions of users who are too lazy or uninformed to do backups have at least some minimal protection against the most likely (in my opinion) cause of data loss: user error.
Is that clear enough for you? PCs aren't going away, but the traditional PC profit margins are going away, and this will cause a shakeup in the PC manufacturing industry. Apple has, so far, managed to make higher margins than the typical 10%, but how long can they continue this?
Talk to anyone who works in the PC industry and they will tell you that this has already happened. HP doesn't make shit on a PC anymore. They make it up with ancillary revenue like extended warranties, accessories (e.g. printers), and pre-installed trialware.
This isn't new, it's not different, and it's not the end of the PC industry. Apple may not be able to survive in such an environment, but companies like Acer have shown that they can prosper selling ultra-low-cost devices.
I received a notice that I either switches to digital cable (i.e. one free DTV box, plus $5 per month for each additional box), or I will lose all channels above 15.
You get up to 3 free DTAs, and I believe that they are $3/mo beyond that.
how the hell can finding a phone that you recognize as maybe a prototype (not something Jane or Joe can....) offer to bring it back to Apple be a reason to get your home searched?
The second you attempt to sell the phone rather than try to return it to Apple, you're no longer just "finding" a phone. You're selling property that's not yours, which is a felony.
Both Gizmodo and the person who found it damn well knew that this was Apple's property. Gizmodo wouldn't have paid $5000 if it had been anything but a prototype, and the person who found it wouldn't have demanded $5000.
iPhone prototypes are, of course, the property of Apple. The person who lost it didn't own it. The person who found it didn't own it. Gizmodo didn't own it.
The person who found it didn't drive to Apple's headquarters (5 miles away) and return it. He didn't give it to the police. He didn't return it to the bartender so that the Apple employee who lost it could recover it.
He sold it to Gizmodo for $5000. It sounds like he - and Gizmodo - knew exactly what they were dealing with. You can't sell property that's not yours.
Oh, except it was microsoft's operating system, and microsoft's messenger. I don't understand this concept of computing where you can click in "the wrong link". I can click in whatever link I want, and that is not supposed to destroy my computer. I use Pidgin on GNU/Linux. I can click on ANY link that I want. Clicking on the link won't do anything besides opening it on a browser, or asking me to download it
Your attitude of invincibility is both dangerous and stupid. Firefox, like all web browsers, is complex software that has a long history of vulnerabilities. One buffer overflow vulnerability (and Firefox has a history of such vulnerabilities) is enough to run arbitrary code on your system.
Except I sudo su and chmod +x $file and./$file nothing is going to happen.
Not true. The software you use every day almost certainly has security vulnerabilities that may allow code execution. History has shown that determined hackers have little trouble finding one.
But we hear all the time from windows users getting randomly infected with malware by just clicking on a fucking URL, or going to the wrong site, etc.
No, mostly we hear those stories from people who don't know what the hell they're talking about. If you download and run some arbitrary executable, well, yeah, you can get infected. The same could happen if you went and installed a malicious deb/rpm.
Those people who truly *were* infected by "just clicking on a fucking URL" (and not by deliberate acts of stupidity on their part) are victims of software vulnerabilities. And those vulnerabilities exist on every platform.
Oh, except it was microsoft's operating system, and microsoft's messenger.
Neither Microsoft's OS nor their messenger software had anything to do with this hole, although Internet Explorer might. Neither the messenger software nor the OS were vulnerable; the vulnerability was most likely either in the web browser or a plugin like Flash.
What kind of shared host gets on you for 8GB/mo? Almost all of the shard hosting providers I've seen provide at least 50GB/mo.
Maybe you should use a better hosting provider, figure out how to add the correct code to robots.txt, or use Google's webmaster tools (which are quite easy to use).
And, FYI, pretty much any web language can set headers. You don't need to have root access to do it. You don't need to modify Apache configs.
Most blogging, CMS, or forum software already handles this correctly. If you don't want to learn about handling caching or other HTTP issues, you might want to consider building on top of a framework that handles those issues for you.
This isn't a malicious attempt to get you to upgrade to DVR service. It has to do with the fact that the digital cable box you have (Motorola DCT2000 series) has 2MB of flash memory.
The VCR recording feature requires an IR database (that stores the correct power/record codes for each VCR), code to operate the IR blaster, and of course UI and other features. All of this takes space. It may only be a few KB, but Comcast keeps adding features to the DCT2000 boxes and eventually something has to go. The VCR feature is one that isn't particularly popular (it's hard to configure and most people don't even have a VCR anymore), and it takes up more space than many other features, so it gets the axe.
Comcast's guide software (i-guide) is not particularly great, but it's a hell of a lot better than what used to run on the DCT2000. Those boxes are very old at this point, but the i-guide software has given them a reasonable level of functionality for people who don't want HD or a DVR.
If you don't like the change, you are free to do any of the following: - Return the Comcast box and use a video recording device (TiVo, Moxi, Media Center, etc.) that uses a CableCard. Comcast charges $1.50/mo for a CableCard. - Use a recording device or software (Media Center, MythTV, TiVO series 1/2) that supports your cable box with an IR blaster. - Switch to Comcast's DVR.
FYI, Comcast's DVR is $15.99/mo if it's the first box on the account in most areas ($20 if it's an additional box). Conventional boxes are free (first box) or $6 (additional boxes). Some of these rates vary by area, but they're increasingly standardized.
Re:It's not a computer, it's a living-room applian
on
iPad Review
·
· Score: 1
I've never liked being strapped to a desk.
So buy a laptop. They're good for using on things like, well, your lap.
I don't know what system you were running. I was using an AlphaServer ES40; four 667 Alphas with 8gb RAM. It was one of the most reliable systems I've ever used for HPC. There was a rack of intel x86 systems of the same era right next to it - something like 32 Intel Xeon CPUs - and the Alpha made the rack look silly and wasteful. On BLAST, the Alpha ran circles around the intel rack, and it became even more embarrasing for the intel rack when the data sets got larger. That was only one example, though; we found pretty much anything we could get source code for, the Alpha ran better. And that was going up against 1.8ghz Xeons.
So you're comparing an SMP system with a cluster? Anyone who knows HPC can tell you that's not a fair fight.
Everyone knows that Netburst was not competitive with (just about anything) on a clock-for-clock basis. That's not the point. The question is whether Alpha was competitive on a performance per dollar standpoint.
x86 isn't about being the fastest. It's about being fast and dirt cheap. You can get a complete Core i7 2.8GHz system for around $1000, and there's no non-x86 competitor that's even close, even at 4x the price. Even x86 servers are dirt cheap.
x86 isn't a passable architecture at all. What it has going for it, is MONEY. Intel, AMD, and others have dumped tons of money into it to keep it moving along, against all odds.
It's fundamentally irrelevant whether anyone thinks that x86 is "passable" - it's a proven fact. We have 15 years of out-of-order x86 implementations that prove that.
Yeah, you have to handle the brain-dead instruction encodings in the decoder, and you need to emit micro-ops for a bunch of obscure instructions that no one ever uses (to maintain compatibility). You also have to handle the multiple obscure and obsolete memory addressing modes.
But the reality is that no one but engineers gives a crap about this. In a world of 300M+ transistor cores, there just isn't that much overhead to making the CPU compatible. Most of the die space is cache anyway nowadays.
We can't compare what x86 is to what POWER or MIPS or SPARC "would have been" in some speculative world where Intel wasn't the dominant desktop/server CPU manufacturer. There's no magic bullet that can make load-store architectures amazingly fast but that doesn't apply to x86. Almost all of the technology out there can apply equally to a modern x86 CPU.
What sells CPUs is not having a clean and simple ISA. What sells CPUs is performance, power consumption, and, in many cases, compatibility. If having a clean ISA accomplishes those objectives, so much the better. But Intel and AMD have shown that you can make a fast, low-power, compatible x86 CPU and sell it at a very low price. That's what matters.
Indeed, the 21264 is commonly cited as an example of a "modern" out-of-order CPU in architecture courses. It's a multi-issue, out-of-order, speculative design with a multi-level cache and an advanced branch predictor.
The huge amount of instruction-level parallelism (dependent on a very good compiler) really seems like the best way to do things. It's too bad it doesn't work out in practice.
The problem is that Intel didn't come up with this concept (ILP through compiler-scheduled instructions), nor were they the first to try it.
VLIW designs have *always* looked great on paper and *always* sucked in practice. Intel did make a bunch of improvements to VLIW with Itanium, but they should have known that you can't just dump the problem of scheduling on the compiler and pretend it's magically solved at an architectural level.
When I was at the SC08 Cluster Challenge in Austin, there was a team from Dresden. I found a simple way to express the size of the US:
The state we were in (Texas) was almost twice the size of Germany. (696,241 km^2 vs 357,021 km^2). And it's one of four US states (Alaska, Texas, California, and Montana) that are larger than Germany.
The population density of Germany is roughly 20 times that of Texas. And it's not just because there are a lot of rural areas in Texas. There are vast, vast swaths of suburbia with wide streets and mega-stores with huge parking lots.
The upside of this is that it's convenient and comfortable for the car-owner. You get to live in a big house and it's easy (and reasonably quick) to drive wherever you want. Even a good public transit system (like the NYC subway) is *way* slower than driving in suburbia. My brother's commute from Jersey City, NJ to Manhattan is about 5 miles, and it takes just about an hour. I can drive the same 5 miles in under 10 minutes here in Boulder, CO. You'd be nuts to own a car in Manhattan, because it's expensive and really no faster than the subway. But that's as much because driving is slow in dense urban settings as it is because transit is better there.
There is of course a huge cost to all of this: energy. People in the US use about twice as much energy as people in Western Europe. Some of that is sheer stupidity (using less efficient heating, lighting, appliances, etc.). But a lot of that comes down to a simple fact: larger houses mean more energy is required for heating, lighting, and other factors. Larger houses mean you need to be further from services, which means driving greater distances and less transit use.
It's easy, cheap, and convenient to use a lot of energy in the US. Big SUVs, huge houses, and leaving the A/C set on 66 are *nice*. You can argue that people are irresponsible for using a ton of energy, but it's hard to argue that they're irrational. That's what makes decreasing US energy consumption so challenging.
I used an iPad for about 30 minutes today in the Apple Store. It's a giant iPod touch.
I keep hearing how fast the browser is, but I'm not seeing it. Yes, it's faster - compared to the iPhone 3G or even the 3GS. But take, for example, Engadget.
Engadget loads in about 2 seconds on my ThinkPad in Chrome. Not only does it load, but it's fully rendered, ready for smooth scrolling and instant interaction. The iPad takes 20+ seconds to do the same thing, and while the page is loading scrolling is surprisingly slow. Yes, it's smooth, but the checkerboard pattern is everywhere and it takes forever to disappear.
Using the browser, you just can't shake the impression that this thing is an iPhone/iPod Touch. Things load slower, everything is annoyingly scaled, it's impossible to do things like drag/drop, and using multiple pages at once is harder.
Then there are other things, like the fact that you can't use a USB printer/scanner (some of us *do* occasionally print things out), the fact that you can't plug in a USB storage device and copy files around, the fact that it's totally useless as a development environment, or the fact that you can't run multiple apps at the same time.
It's abundantly clear that this *isn't* a revolution. The revolution happened in 2007 when the iPhone came out and made touchscreen technology work. This is an evolution, a bigger and (usually) better version of a product we all know. And while it's clear that the iPhone's technology makes it a great PDA/Phone, it's also abundantly clear that it does a poor job of replacing a PC.
My wife's friend brought by a shiny new $600 netbook, one that actually had a basic non-intel GPU capable of limited video performance (most netbooks fall flat with flash, and can not do H.264 at native screen resolution let alone 720P). It had a 2GHz Atom/arm/whatever it was, and 2GB of RAM. It took more than 3.5 minutes to boot windows 7 to a login screen, and more than 70 seconds after login to open outlook and a web browser. by 5 minutes in, I'll have forgotten why I was booting it up. Technically, it smoked the iPad's specs, but it was compeltely unusable from a concencince/companion device standpoint. $250 more and I'd have gotten a machine capable of playing WoW, running virtual machines, a 13" screen, and the power and performance to edit video and run a full OS, on a 7 hour battery (aka, a White Macbook).
Yeah, and my EEE PC 900HA ran Windows 7 just fine, once I threw 2GB of memory in it. It booted in under a minute and Chrome opens in a couple of seconds. Anecdotes are like that - not reliable.
Netbooks play H.264 at screen resolution fine, and even at 720p. You have to use a decoder that's not brain dead (like libavcodec) and not Flash, though.
And, FYI, there aren't any $600 netbooks, because netbooks are by definition a low-cost computer. The very upper range is around $500.
And, FYI, there's no such thing as an $850 MacBook - the white MacBook is $999. So you're talking $400 more than your expensive "netbook" and $600 more than a very-well-specced EEE PC 1005PE with 2GB of memory.
Netbooks are cheap, small, light, long-battery life (1005PE gets 12+ hours) general-purpose computers. $500 gets you a 16GB iPad, or an EEE PC 1005PE, a 2GB DIMM, and $100 extra to spend on whatever the hell you want to.
And, yeah, the EEE runs Linux. It runs Linux extremely well.
Nearly all popular SSDs are quoted in base-10 (SI) units, including Samsung, Crucial, WD, Seagate, Intel, and a variety of smaller brands including Patriot/Corsair/Kingston.
The one exception I'm aware of is OCZ, which uses base-2 units.
Except when we complain how shitty our telecommunications service is in the United States compared to other nations
Except it's not. I'm in a shitty apartment in Jersey City, NJ right now and I can still get 5Mbps+ HSPA+ (from T-Mobile), ~3MBps HSPA (from AT&T), 50Mbps DOCSIS 3.0 from Comcast, or (soon) FiOS from Verizon.
More power draw than a CPU from the bad old days of Prescott
Prescott at its hottest (Pentium 4 HT 571) was only 115W, which is about the same or (in some cases) vastly less than nearly every mid-range to high-end GPU today.
Radeon 5830 is 175W Radeon 5850 is 151W Radeon 5770 is 108W
Prescott at its hottest actually used less power than some of the current high-end Core i7 CPUs (i7-920 is 130W), although of course that's comparing a 1-core CPU to a vastly faster 4-core CPU.
What's happened is that CPU coolers have gotten much better (thanks in part to heatpipes and larger fins/fans), power supplies have gotten more efficient and larger, and cases are better ventilated. The result is that today a 130W CPU is no big deal, whereas with the Prescott it caused all kinds of thermal nightmares for people building their own PCs (professionally engineered commercial PCs generally fared OK with Prescott).
Still, 250W on a GPU is stupid. Even with modern efficient air cooling, it's hard to keep such a GPU cool without making a ton of noise. Add the crazy power supply requirements (most people are recommending 550W or more, which means $100+ if you want a quality PSU), and it's a pretty big burden. The real problem is that the ATI card is almost as fast, cheaper, and 80 watts cooler. And it's been on the market for 8 months.
Windows Explorer no longer kills network transfers after a failure as of Windows Vista.
Maybe some of the people complaining about Windows should stop using a version thats 9 years old (XP). Red Hat 7.2 isn't particularly great by today's standards either.
Slashdot Mirror
Dear dickhead,
Not all of us are willing to blow $500+ on a device that doesn't enable us to do anything we couldn't do before.
I need a laptop to get my job done. I need a mobile phone for a variety of reasons. I can't think of a single damn reason why I need a $500 tablet.
Until you get that not all of us are willing to spend $500 on gadget porn, you won't grok why some people think the iPad is a tremendous waste of money and attention.
The "transmission ratio is controlled by the relative speed of the motor driven shaft" technique is exactly what's used by the transmission in the Prius (and every other Toyota, Nissan, and Ford hybrid).
If you literally believe that, you're crazy and dangerous. If not, stop egging on the people who are crazy and dangerous.
Saying that you want to vote the bastards out of office is reasonable. Saying that you want to see criminal prosecutions against those who you believe have acted illegally is reasonable. Threatening violence is not.
puny single-core ARM they have specs equivalent to a high end phone
The iPad has a "puny single-core ARM" that's not significantly faster than the 1GHz Snapdragon that is currently in many high end phones.
I can assure you that restore points (shadow copies) DO work, and they have saved my ass more than once. However, Windows can only keep as many shadow copies as you give it room to store. If you fill up your disk, shadow copies WILL be deleted. This is a persistent issue on my Intel X25-m 80GB SSD, which is pretty much constantly near full because of its small size. I usually only get a few days of shadow copies as a result.
That said, neither Linux nor Mac OS X have anything like shadow copies unless you go configure them yourself.
On Linux, you can use a COW filesystem like ZFS or btrfs to get similar functionality, but practically no one uses either on a personal machine since ZFS runs in FUSE on Linux and btrfs is not entirely stable yet. The vast majority of Linux distros today default to ext4, which is a file filesystem but doesn't offer any sort of snapshotting.
On Mac OS X, you would need to configure Time Machine, which is great but requires an external disk or a Time Capsule (or a hacked NAS). I would bet that a substantial portion of Mac users never set up Time Machine simply because they don't have the necessary hardware.
The point is, Windows - even with shadow copies off - is no worse than Linux or Mac OS X out of the box. Shadow copies are not a backup mechanism and they aren't intended to replace one. Obviously if there's a bug here, it's serious and needs to be fixed. But the very nature of shadow copies means that they cannot be a long-term disaster recovery mechanism. If you're relying on shadow copies instead of backing up, you're stupid. Shadow copies offer zero protection against hardware failure, disasters (fire), or malware, even when they work.
That said, shadow copies are one of the best reasons to use Windows 7 (some versions of Vista had them too). No, they're not perfect. But the fact that they're on by default means that millions of users who are too lazy or uninformed to do backups have at least some minimal protection against the most likely (in my opinion) cause of data loss: user error.
Talk to anyone who works in the PC industry and they will tell you that this has already happened. HP doesn't make shit on a PC anymore. They make it up with ancillary revenue like extended warranties, accessories (e.g. printers), and pre-installed trialware.
This isn't new, it's not different, and it's not the end of the PC industry. Apple may not be able to survive in such an environment, but companies like Acer have shown that they can prosper selling ultra-low-cost devices.
Where did you hear that? It's something that I came up with on my own. Obviously someone else could have come up with the same thing.
SIdenote: any time we win in WoW arenas because the other side didn't show up, we call it a "Comcast Win".
The funny thing is that I kept using the term after I switched to Qwest.
You get up to 3 free DTAs, and I believe that they are $3/mo beyond that.
http://digitalnow.comcast.com/About_Digital_Update.aspx
Well, you're wrong about the DTAs.
Antenna? DSL? Satellite? Wireless broadband?
In my area there are two satellite providers, Comcast, and over-the-air for TV.
There is a WiSP, DSL (Qwest) and Comcast for Internet, plus various WWAN providers.
Obviously this isn't as competitive as I would like it to be. But Comcast is almost certainly not your only option.
The second you attempt to sell the phone rather than try to return it to Apple, you're no longer just "finding" a phone. You're selling property that's not yours, which is a felony.
Both Gizmodo and the person who found it damn well knew that this was Apple's property. Gizmodo wouldn't have paid $5000 if it had been anything but a prototype, and the person who found it wouldn't have demanded $5000.
iPhone prototypes are, of course, the property of Apple. The person who lost it didn't own it. The person who found it didn't own it. Gizmodo didn't own it.
The person who found it didn't drive to Apple's headquarters (5 miles away) and return it. He didn't give it to the police. He didn't return it to the bartender so that the Apple employee who lost it could recover it.
He sold it to Gizmodo for $5000. It sounds like he - and Gizmodo - knew exactly what they were dealing with. You can't sell property that's not yours.
939 uses DDR, AM2 uses DDR2.
The only reason that AM2+/AM3 are compatible is that newer AMD CPUs support both DDR2 and DDR3.
Your attitude of invincibility is both dangerous and stupid. Firefox, like all web browsers, is complex software that has a long history of vulnerabilities. One buffer overflow vulnerability (and Firefox has a history of such vulnerabilities) is enough to run arbitrary code on your system.
Not true. The software you use every day almost certainly has security vulnerabilities that may allow code execution. History has shown that determined hackers have little trouble finding one.
No, mostly we hear those stories from people who don't know what the hell they're talking about. If you download and run some arbitrary executable, well, yeah, you can get infected. The same could happen if you went and installed a malicious deb/rpm.
Those people who truly *were* infected by "just clicking on a fucking URL" (and not by deliberate acts of stupidity on their part) are victims of software vulnerabilities. And those vulnerabilities exist on every platform.
Neither Microsoft's OS nor their messenger software had anything to do with this hole, although Internet Explorer might. Neither the messenger software nor the OS were vulnerable; the vulnerability was most likely either in the web browser or a plugin like Flash.
What kind of shared host gets on you for 8GB/mo? Almost all of the shard hosting providers I've seen provide at least 50GB/mo.
Maybe you should use a better hosting provider, figure out how to add the correct code to robots.txt, or use Google's webmaster tools (which are quite easy to use).
And, FYI, pretty much any web language can set headers. You don't need to have root access to do it. You don't need to modify Apache configs.
Most blogging, CMS, or forum software already handles this correctly. If you don't want to learn about handling caching or other HTTP issues, you might want to consider building on top of a framework that handles those issues for you.
This isn't a malicious attempt to get you to upgrade to DVR service. It has to do with the fact that the digital cable box you have (Motorola DCT2000 series) has 2MB of flash memory.
The VCR recording feature requires an IR database (that stores the correct power/record codes for each VCR), code to operate the IR blaster, and of course UI and other features. All of this takes space. It may only be a few KB, but Comcast keeps adding features to the DCT2000 boxes and eventually something has to go. The VCR feature is one that isn't particularly popular (it's hard to configure and most people don't even have a VCR anymore), and it takes up more space than many other features, so it gets the axe.
Comcast's guide software (i-guide) is not particularly great, but it's a hell of a lot better than what used to run on the DCT2000. Those boxes are very old at this point, but the i-guide software has given them a reasonable level of functionality for people who don't want HD or a DVR.
If you don't like the change, you are free to do any of the following:
- Return the Comcast box and use a video recording device (TiVo, Moxi, Media Center, etc.) that uses a CableCard. Comcast charges $1.50/mo for a CableCard.
- Use a recording device or software (Media Center, MythTV, TiVO series 1/2) that supports your cable box with an IR blaster.
- Switch to Comcast's DVR.
FYI, Comcast's DVR is $15.99/mo if it's the first box on the account in most areas ($20 if it's an additional box). Conventional boxes are free (first box) or $6 (additional boxes). Some of these rates vary by area, but they're increasingly standardized.
So buy a laptop. They're good for using on things like, well, your lap.
So you're comparing an SMP system with a cluster? Anyone who knows HPC can tell you that's not a fair fight.
Everyone knows that Netburst was not competitive with (just about anything) on a clock-for-clock basis. That's not the point. The question is whether Alpha was competitive on a performance per dollar standpoint.
x86 isn't about being the fastest. It's about being fast and dirt cheap. You can get a complete Core i7 2.8GHz system for around $1000, and there's no non-x86 competitor that's even close, even at 4x the price. Even x86 servers are dirt cheap.
It's fundamentally irrelevant whether anyone thinks that x86 is "passable" - it's a proven fact. We have 15 years of out-of-order x86 implementations that prove that.
Yeah, you have to handle the brain-dead instruction encodings in the decoder, and you need to emit micro-ops for a bunch of obscure instructions that no one ever uses (to maintain compatibility). You also have to handle the multiple obscure and obsolete memory addressing modes.
But the reality is that no one but engineers gives a crap about this. In a world of 300M+ transistor cores, there just isn't that much overhead to making the CPU compatible. Most of the die space is cache anyway nowadays.
We can't compare what x86 is to what POWER or MIPS or SPARC "would have been" in some speculative world where Intel wasn't the dominant desktop/server CPU manufacturer. There's no magic bullet that can make load-store architectures amazingly fast but that doesn't apply to x86. Almost all of the technology out there can apply equally to a modern x86 CPU.
What sells CPUs is not having a clean and simple ISA. What sells CPUs is performance, power consumption, and, in many cases, compatibility. If having a clean ISA accomplishes those objectives, so much the better. But Intel and AMD have shown that you can make a fast, low-power, compatible x86 CPU and sell it at a very low price. That's what matters.
Indeed, the 21264 is commonly cited as an example of a "modern" out-of-order CPU in architecture courses. It's a multi-issue, out-of-order, speculative design with a multi-level cache and an advanced branch predictor.
The problem is that Intel didn't come up with this concept (ILP through compiler-scheduled instructions), nor were they the first to try it.
VLIW designs have *always* looked great on paper and *always* sucked in practice. Intel did make a bunch of improvements to VLIW with Itanium, but they should have known that you can't just dump the problem of scheduling on the compiler and pretend it's magically solved at an architectural level.
When I was at the SC08 Cluster Challenge in Austin, there was a team from Dresden. I found a simple way to express the size of the US:
The state we were in (Texas) was almost twice the size of Germany. (696,241 km^2 vs 357,021 km^2). And it's one of four US states (Alaska, Texas, California, and Montana) that are larger than Germany.
The population density of Germany is roughly 20 times that of Texas. And it's not just because there are a lot of rural areas in Texas. There are vast, vast swaths of suburbia with wide streets and mega-stores with huge parking lots.
The upside of this is that it's convenient and comfortable for the car-owner. You get to live in a big house and it's easy (and reasonably quick) to drive wherever you want. Even a good public transit system (like the NYC subway) is *way* slower than driving in suburbia. My brother's commute from Jersey City, NJ to Manhattan is about 5 miles, and it takes just about an hour. I can drive the same 5 miles in under 10 minutes here in Boulder, CO. You'd be nuts to own a car in Manhattan, because it's expensive and really no faster than the subway. But that's as much because driving is slow in dense urban settings as it is because transit is better there.
There is of course a huge cost to all of this: energy. People in the US use about twice as much energy as people in Western Europe. Some of that is sheer stupidity (using less efficient heating, lighting, appliances, etc.). But a lot of that comes down to a simple fact: larger houses mean more energy is required for heating, lighting, and other factors. Larger houses mean you need to be further from services, which means driving greater distances and less transit use.
It's easy, cheap, and convenient to use a lot of energy in the US. Big SUVs, huge houses, and leaving the A/C set on 66 are *nice*. You can argue that people are irresponsible for using a ton of energy, but it's hard to argue that they're irrational. That's what makes decreasing US energy consumption so challenging.
I used an iPad for about 30 minutes today in the Apple Store. It's a giant iPod touch.
I keep hearing how fast the browser is, but I'm not seeing it. Yes, it's faster - compared to the iPhone 3G or even the 3GS. But take, for example, Engadget.
Engadget loads in about 2 seconds on my ThinkPad in Chrome. Not only does it load, but it's fully rendered, ready for smooth scrolling and instant interaction. The iPad takes 20+ seconds to do the same thing, and while the page is loading scrolling is surprisingly slow. Yes, it's smooth, but the checkerboard pattern is everywhere and it takes forever to disappear.
Using the browser, you just can't shake the impression that this thing is an iPhone/iPod Touch. Things load slower, everything is annoyingly scaled, it's impossible to do things like drag/drop, and using multiple pages at once is harder.
Then there are other things, like the fact that you can't use a USB printer/scanner (some of us *do* occasionally print things out), the fact that you can't plug in a USB storage device and copy files around, the fact that it's totally useless as a development environment, or the fact that you can't run multiple apps at the same time.
It's abundantly clear that this *isn't* a revolution. The revolution happened in 2007 when the iPhone came out and made touchscreen technology work. This is an evolution, a bigger and (usually) better version of a product we all know. And while it's clear that the iPhone's technology makes it a great PDA/Phone, it's also abundantly clear that it does a poor job of replacing a PC.
Yeah, and my EEE PC 900HA ran Windows 7 just fine, once I threw 2GB of memory in it. It booted in under a minute and Chrome opens in a couple of seconds. Anecdotes are like that - not reliable.
Netbooks play H.264 at screen resolution fine, and even at 720p. You have to use a decoder that's not brain dead (like libavcodec) and not Flash, though.
And, FYI, there aren't any $600 netbooks, because netbooks are by definition a low-cost computer. The very upper range is around $500.
And, FYI, there's no such thing as an $850 MacBook - the white MacBook is $999. So you're talking $400 more than your expensive "netbook" and $600 more than a very-well-specced EEE PC 1005PE with 2GB of memory.
Netbooks are cheap, small, light, long-battery life (1005PE gets 12+ hours) general-purpose computers. $500 gets you a 16GB iPad, or an EEE PC 1005PE, a 2GB DIMM, and $100 extra to spend on whatever the hell you want to.
And, yeah, the EEE runs Linux. It runs Linux extremely well.
Nearly all popular SSDs are quoted in base-10 (SI) units, including Samsung, Crucial, WD, Seagate, Intel, and a variety of smaller brands including Patriot/Corsair/Kingston.
The one exception I'm aware of is OCZ, which uses base-2 units.
So, no, not "for the most part".
Except it's not. I'm in a shitty apartment in Jersey City, NJ right now and I can still get 5Mbps+ HSPA+ (from T-Mobile), ~3MBps HSPA (from AT&T), 50Mbps DOCSIS 3.0 from Comcast, or (soon) FiOS from Verizon.
Prescott at its hottest (Pentium 4 HT 571) was only 115W, which is about the same or (in some cases) vastly less than nearly every mid-range to high-end GPU today.
Radeon 5830 is 175W
Radeon 5850 is 151W
Radeon 5770 is 108W
Prescott at its hottest actually used less power than some of the current high-end Core i7 CPUs (i7-920 is 130W), although of course that's comparing a 1-core CPU to a vastly faster 4-core CPU.
What's happened is that CPU coolers have gotten much better (thanks in part to heatpipes and larger fins/fans), power supplies have gotten more efficient and larger, and cases are better ventilated. The result is that today a 130W CPU is no big deal, whereas with the Prescott it caused all kinds of thermal nightmares for people building their own PCs (professionally engineered commercial PCs generally fared OK with Prescott).
Still, 250W on a GPU is stupid. Even with modern efficient air cooling, it's hard to keep such a GPU cool without making a ton of noise. Add the crazy power supply requirements (most people are recommending 550W or more, which means $100+ if you want a quality PSU), and it's a pretty big burden. The real problem is that the ATI card is almost as fast, cheaper, and 80 watts cooler. And it's been on the market for 8 months.
Windows Explorer no longer kills network transfers after a failure as of Windows Vista.
Maybe some of the people complaining about Windows should stop using a version thats 9 years old (XP). Red Hat 7.2 isn't particularly great by today's standards either.