ECC support is disabled on all non-Xeon chips. There isn't a technical reason, it was purposeful market segmentation.
Agreed, but the price difference isn't really that much for ECC with Intel.
I'm not a fanboi of either company, I just tend to buy the best price/performance at the time. I have a pair of dual-socket Opteron servers that were by far the best deal for the performance at the time, plus they use far less electricity than the Xeons of the same era.
But, I'm replacing them with dual-socket Xeon (Westmere) machines because it's by far the cheaper way to go for the same performance, and Intel has also leapfrogged AMD in lower electricity use.
It's damn cheap for a fast performing ECC workstation.
Right now, motherboard/CPU combos are just about perfectly priced so as to match their performance. So, your $360 parts are just about 75% the performance of $480 worth of Intel gear (Passmark 3931 vs. 5850 for a Core i7-930). Your ability to use ECC is a bonus, though, and I suspect that as standard memory grows to 8-16GB, Intel will have to re-think their ECC support stance.
The reason I'd still buy Intel myself for a desktop is that I can get the $280 CPU for $200 at a local store, which makes it only a 11% price increase for 48% more performance.
The only issue might be that you need a motherboard that supports ECC, but $270 for this one isn't a lot more than the $200 or so you'd pay for a non-server board with equivalent build quality. Unless things have changed drastically since the last time I looked at AMD motherboards, not all of them support ECC, either.
The new Core i5 brings top-of-the-line Nehalem-class performance at a $200 price point. We recently awarded it our Recommended Buy honor after seeing it stand up to more expensive CPUs in games and other demanding apps.
They don't recommend spending more than $200, though.
I'm lucky in that I live near a MicroCenter store, and they are currently selling the i5-750 for $180 and the i7-860 for $200.
Since they both use the same socket, the extra $20 for a lot more performance (133-266MHz, depending on turbo-boost, plus twice the threads and better virtualization) is a no-brainer. Even with the $40 bundle discount on AMD CPU/motherboard from MicroCenter, Intel is still far and away the price/performance leader.
Stop using Azureus for your bit torrent client, and downloading a file will no longer require a fill core with of CPU time.
Over the past month, my torrent VM (running Azureus) has averaged 603MHz CPU usage.
I suspect it's that high because of the fairly high average transfer rate (1.5MB/s disk, 984KB/s network), plus the usual suspects of other background programs (anti-virus, etc.).
Or maybe lists could be identified in a different way? Sometimes Thunderbird shows a "reply to list" button, so how does it find out?
Most list software adds extra headers that can be used identify the actual original recipient (i.e., the list address) and the end recipient, along with a lot of other useful information.
I still use Pegasus Mail (although I'm starting to look for a more modern replacement that has the power and the security), and my rules for filtering "not-directly-to-me" e-mail work pretty well. But, they're still add-on rules, and I don't know of any e-mail client that has similar ones out-of-the-box.
"If *I* think it's crap, but the marketplace thinks it's vastly superior, the only way to resolve this paradox is to assume that the marketplace is profoundly stupid and duped by Apple's svengali-like marketing. Because it couldn't possibly be that I don't have a freaking clue what people want."
I believe the US elections of the past few years show that as long as the average consumer is distracted by shiny toys, nothing else makes any impact on their brain. Apple products are the ultimate in shiny toys, thus, they are wildly successful.
There are many other shiny toys that target people with more money than brains, and most are also wildly successful: almost every heavily advertised movie or video game (regardless of actual quality), "premium" automobiles that are just re-badged versions of cheaper makes, and, of course, casinos.
I think anyone (including/. readers) should be proud if they are not one of the sheep, but rather a thinking human being. Even if you are an Apple fan (which means you likely aren't a big thinker...I kid) it's pretty easy to figure that an iPhone plus some other device (netbook, eReader, etc.) is a better bang for your buck than an iPad.
Compared to any given Netbook product this year, the iPad will outsell it by a wide margin.
If you want to compare fairly, compare one iPad SKU (out of the 5-6 total...I don't keep track of every variation on vendor-locked products with no possibility for permanent expansion) to a single netbook SKU.
Still, I'm sure Apple will sell a gazillion iPads, and many buyers will use them for a few hours each month once the original novelty wears off (and the lack of true features kicks in after similar products hit the market). The rest of us will spend our money wisely.
Top posting makes since because I already read everything up until the reply.
...until you get added to an e-mail chain late in the game, and have to read everything "backwards", and 90% of the volume of the message is signatures and disclaimers that are a waste of time on internal e-mail.
Until Outlook provides tools that have been available in other e-mail packages for years (e.g., support for varying signatures based on recipient, sensible formatting of HTML when viewed as plain text, etc.), it will continue to be an impediment to good communication.
And another favorite: Idiot doing a reply-all to a Mail that was Bcc'ed to him... (with the result that now all the others in Cc: and To: know about him being Bcc'ed...)
Without resorting to rules written by the user, are there any e-mail clients that give the user a very solid notification that some e-mail address that doesn't end up at them is not in the "To:" or "Cc:" fields?
Remember that "To:", "Cc:", and "BCC" (not really a header field) are handled by merely placing the address in the envelope and sending the text of the e-mail, and the text just happens to contain the "To:" and "Cc:" fields.
And on the main point, it is pretty simple: Parody makes fun of the work you are borrowing from (Wierd Al),
You clearly need to pay more attention to the lyrics of "Weird Al".
Although there is some fun poked at the original song, often a third party bears the brunt of the joke:
Ricky is targeted at I Love Lucy more than at Toni Basil (or her song).
Livin' in the Fridge does nothing to comment on the Rolling Stones or their song...it just borrows the music and meter.
Phony Calls also makes no comment on TLC or their song.
The Brady Bunch and Safety Dance...again, just the music and meter in common. Fitting some joke about "hats" in the song would have been easy and made it reference the original more, though.
There are obvious cases where Al is going after the original (Michael Jackson, Coolio, etc.) as well, but the majority that he does aren't that way. Since he pays for rights to use the music, he's in the clear no matter what he is commenting on.
He's probably OK, as he does secure the necessary recording rights for the songs (which requires indirect payment to the songwriters through a clearinghouse).
If he then copies the look of the video, he's OK because he can then separately parody the video safely, since the audio portion is already either licensed or covered by a separate parody. Most of his non-original songs are of this double parody variety.
I suspect the main issue in this case was they didn't pay the fee that allows them to record the song. If you do that, any BMI or ASCAP songwriter has zero say over whether you can record "their" song, or perform it in public.
The problem is here that from what the community at large is saying, while some have tried it, they haven't been happy with the product to continue using it. That means that while the money is being sunk into point 3 above, it's not retaining those users, so much more needs to be spent to get them to try it again.
Worse than that, you have people who use Google for their searches but then once they decide to buy something online they "search" for it using Bing to get the cash-back.
Also I have to point out that the article mistakenly compares paying $10 for hulu (on demand) vs just watching it on "tv for free".
With DVRs capable of holding literally thousands of hours of TV shows at the same quality as Hulu, the only advantage Hulu would have is the older episodes (which is what they are making you pay for).
Otherwise, for any show in reruns, it's merely a matter of time until you could collect the entire series so it was available "on demand".
There are a lot of "text files" that are really interpreted code, like JavaScript.
I really don't want to spend the time whitelisting "safe" text files when there's not that much of a loss of speed for scanning a few writes to those files.
Don't forget that AV software costs you ALL the time when it's installed in "real time" scanning mode.
Most AV software can be configured to only scan on write, instead of on access.
This is a big help for the vast majority of cases (like loading a program, which would require no extra work), but still can be an unnecessary burden if you have a lot of files being written that can't really be virus carriers (e.g., log files).
My crystal ball says few people will be buying standalone graphics cards in 5 years.
Based on the fact that there are games being released today that won't run smoothly (i.e., at least 30fps at all times) with highest levels of detail when using the current fastest standalone card, I suspect your crystal ball is cloudy.
Since these standalone cards can generate upwards of 200 watts of heat, there is no way to put that much computing power into a package that already generates 80-100 watts to do CPU things. As the chips continue to use smaller processes, there will be advances that allow more computing with less heat, but those graphics cards are already at 40nm tech, so they're only about half a generation behind.
Based on the progress over the last few years, I suspect that 5 years from now we will see about 2-3 times the GPU power with about 50-75% of the heat. Even if the CPU portion does as well in technological advance, you'd end up with a 140 watt TDP (at best) for a combined CPU/GPU that doesn't require a standalone card to run most games.
There's actually no need to buy anything to use VMware, as Player is free and allows you to create VMs...it just doesn't have some of the better features (snapshot control, etc.) of VMware Workstation.
VMware Server is also free, and has a lot of great features. For bare-metal, ESXi is also free.
With all that plus VirtualBox, Xen, KVM, various Microsoft offerings, etc., there really is no need to pay anything for a hypervisor.
Granted, it does directly or indirectly stress the fpu, cache, maybe task switching and interrupt handling. However, there are many more things that can go wrong.
Off the top of my head, I can think of a lot of things that specifically need tested that one program probably won't do. For example, you need to verify both 32-bit and 64-bit operations. Prime95 is specifically compiled for one or the other, so would stress less of the "other" version.
There are also a lot of SIMD instructions that need tested. Some are obscure enough that only a few apps would use them.
Then, there's all the instructions that support virtualization. I have found that bad hardware running a hypervisor will fail much more frequently than if it is running a "normal" OS (YMMV).
But, unlike Memtest86+ for RAM, there doesn't appear to be any program that specifically tests all CPU subsystems (registers, cache, instruction execution, etc.).
What about the LATER assignments? One I had for my early course involved writing a ticketmaster-like ticket purchasing system where it would find and temporarily reserve the 'best' seating for a block of tickets. It would release these tickets if they were not purchased. It was a very simple problem (didn't get into multiple users), but I'd be willing to bet that it wouldn't be likely for two students in a group of 100 to come up with the same solution.
In my data structures class, when we started on balanced trees, the prof had an assignment where the input would be dictionary words and we were supposed to sort them using balanced trees as the temporary store. We also were supposed to re-use the last assignment where we used normal binary trees. This was to show that if the input was already sorted, the tree devolved into a list and you hit O(n) instead of the O(log n) that balanced trees should always result in.
As extra credit, he wanted us to avoid the worst case for the binary tree, so I stored the values after applying strrev (or a homemade version thereof). People tried many things (some of them similar), but nobody else used my trick. If you know the domain of your data, you can often improve your algorithm.
To test my theory, I reversed all the strings in/usr/dict/words, sorted it, then reversed them back. What you get looks pretty much random, so my trick shouldn't end up worst case very often.
Slashdot Mirror
ECC support is disabled on all non-Xeon chips. There isn't a technical reason, it was purposeful market segmentation.
Agreed, but the price difference isn't really that much for ECC with Intel.
I'm not a fanboi of either company, I just tend to buy the best price/performance at the time. I have a pair of dual-socket Opteron servers that were by far the best deal for the performance at the time, plus they use far less electricity than the Xeons of the same era.
But, I'm replacing them with dual-socket Xeon (Westmere) machines because it's by far the cheaper way to go for the same performance, and Intel has also leapfrogged AMD in lower electricity use.
It's damn cheap for a fast performing ECC workstation.
Right now, motherboard/CPU combos are just about perfectly priced so as to match their performance. So, your $360 parts are just about 75% the performance of $480 worth of Intel gear (Passmark 3931 vs. 5850 for a Core i7-930). Your ability to use ECC is a bonus, though, and I suspect that as standard memory grows to 8-16GB, Intel will have to re-think their ECC support stance.
The reason I'd still buy Intel myself for a desktop is that I can get the $280 CPU for $200 at a local store, which makes it only a 11% price increase for 48% more performance.
Anandtech managed to get a stable 4.0 GHz overclock with air cooling. It makes an already great deal all that much better in my opinion.
How is a $299 6 core/6 thread chip at 4GHz a better deal than a $199 4 core/8 thread chip that can also be overclocked on air to the same speed, and benchmarks far faster at that point?
To get ECC support from Intel, you need to buy a Xeon, at which point they charge you an extra $800-$1000 for the gates to be enabled.
Boy, when you make up numbers, you really reach deep into your ass, don't you?
Core i7-920 for $280 and the same-socket, indentical spec Xeon W3520 for $310.
The only issue might be that you need a motherboard that supports ECC, but $270 for this one isn't a lot more than the $200 or so you'd pay for a non-server board with equivalent build quality. Unless things have changed drastically since the last time I looked at AMD motherboards, not all of them support ECC, either.
Best gaming CPU for $200:
Core i5-750
The new Core i5 brings top-of-the-line Nehalem-class performance at a $200 price point. We recently awarded it our Recommended Buy honor after seeing it stand up to more expensive CPUs in games and other demanding apps.
They don't recommend spending more than $200, though.
I'm lucky in that I live near a MicroCenter store, and they are currently selling the i5-750 for $180 and the i7-860 for $200.
Since they both use the same socket, the extra $20 for a lot more performance (133-266MHz, depending on turbo-boost, plus twice the threads and better virtualization) is a no-brainer. Even with the $40 bundle discount on AMD CPU/motherboard from MicroCenter, Intel is still far and away the price/performance leader.
Stop using Azureus for your bit torrent client, and downloading a file will no longer require a fill core with of CPU time.
Over the past month, my torrent VM (running Azureus) has averaged 603MHz CPU usage.
I suspect it's that high because of the fairly high average transfer rate (1.5MB/s disk, 984KB/s network), plus the usual suspects of other background programs (anti-virus, etc.).
Or maybe lists could be identified in a different way? Sometimes Thunderbird shows a "reply to list" button, so how does it find out?
Most list software adds extra headers that can be used identify the actual original recipient (i.e., the list address) and the end recipient, along with a lot of other useful information.
I still use Pegasus Mail (although I'm starting to look for a more modern replacement that has the power and the security), and my rules for filtering "not-directly-to-me" e-mail work pretty well. But, they're still add-on rules, and I don't know of any e-mail client that has similar ones out-of-the-box.
"If *I* think it's crap, but the marketplace thinks it's vastly superior, the only way to resolve this paradox is to assume that the marketplace is profoundly stupid and duped by Apple's svengali-like marketing. Because it couldn't possibly be that I don't have a freaking clue what people want."
I believe the US elections of the past few years show that as long as the average consumer is distracted by shiny toys, nothing else makes any impact on their brain. Apple products are the ultimate in shiny toys, thus, they are wildly successful.
There are many other shiny toys that target people with more money than brains, and most are also wildly successful: almost every heavily advertised movie or video game (regardless of actual quality), "premium" automobiles that are just re-badged versions of cheaper makes, and, of course, casinos.
I think anyone (including /. readers) should be proud if they are not one of the sheep, but rather a thinking human being. Even if you are an Apple fan (which means you likely aren't a big thinker...I kid) it's pretty easy to figure that an iPhone plus some other device (netbook, eReader, etc.) is a better bang for your buck than an iPad.
Compared to any given Netbook product this year, the iPad will outsell it by a wide margin.
If you want to compare fairly, compare one iPad SKU (out of the 5-6 total...I don't keep track of every variation on vendor-locked products with no possibility for permanent expansion) to a single netbook SKU.
Still, I'm sure Apple will sell a gazillion iPads, and many buyers will use them for a few hours each month once the original novelty wears off (and the lack of true features kicks in after similar products hit the market). The rest of us will spend our money wisely.
Top posting makes since because I already read everything up until the reply.
...until you get added to an e-mail chain late in the game, and have to read everything "backwards", and 90% of the volume of the message is signatures and disclaimers that are a waste of time on internal e-mail.
Until Outlook provides tools that have been available in other e-mail packages for years (e.g., support for varying signatures based on recipient, sensible formatting of HTML when viewed as plain text, etc.), it will continue to be an impediment to good communication.
Jupiter radiates more heat than it receives from the Sun.
And another favorite: Idiot doing a reply-all to a Mail that was Bcc'ed to him... (with the result that now all the others in Cc: and To: know about him being Bcc'ed...)
Without resorting to rules written by the user, are there any e-mail clients that give the user a very solid notification that some e-mail address that doesn't end up at them is not in the "To:" or "Cc:" fields?
Remember that "To:", "Cc:", and "BCC" (not really a header field) are handled by merely placing the address in the envelope and sending the text of the e-mail, and the text just happens to contain the "To:" and "Cc:" fields.
Weird Al is protected because his works are clearly parody
If you listen to his non-original songs, they are actually mostly "satire" by the definition used by the court in TFA.
The songs often comment on a third party, sometimes to the point that the tune from the original song is the only connection between the two.
Again, he's completely protected because he pays the compulsory licensing fees that allow him to record using the tune.
And on the main point, it is pretty simple: Parody makes fun of the work you are borrowing from (Wierd Al),
You clearly need to pay more attention to the lyrics of "Weird Al".
Although there is some fun poked at the original song, often a third party bears the brunt of the joke:
There are obvious cases where Al is going after the original (Michael Jackson, Coolio, etc.) as well, but the majority that he does aren't that way. Since he pays for rights to use the music, he's in the clear no matter what he is commenting on.
Are you really saying that "All She Wants to do is Dance" and "Boys of Summer" are politically motivated works?
All She Wants to Do Is Dance is pretty obviously a commentary on US diplomacy.
An article about parody videos with music set to lyrics mocking someone, and the first thing you thought of was a comic strip? I immediately thought of Weird Al Yankovic and how screwed he'd be if they're found to be infringing.
He's probably OK, as he does secure the necessary recording rights for the songs (which requires indirect payment to the songwriters through a clearinghouse).
If he then copies the look of the video, he's OK because he can then separately parody the video safely, since the audio portion is already either licensed or covered by a separate parody. Most of his non-original songs are of this double parody variety.
I suspect the main issue in this case was they didn't pay the fee that allows them to record the song. If you do that, any BMI or ASCAP songwriter has zero say over whether you can record "their" song, or perform it in public.
The problem is here that from what the community at large is saying, while some have tried it, they haven't been happy with the product to continue using it. That means that while the money is being sunk into point 3 above, it's not retaining those users, so much more needs to be spent to get them to try it again.
Worse than that, you have people who use Google for their searches but then once they decide to buy something online they "search" for it using Bing to get the cash-back.
Also I have to point out that the article mistakenly compares paying $10 for hulu (on demand) vs just watching it on "tv for free".
With DVRs capable of holding literally thousands of hours of TV shows at the same quality as Hulu, the only advantage Hulu would have is the older episodes (which is what they are making you pay for).
Otherwise, for any show in reruns, it's merely a matter of time until you could collect the entire series so it was available "on demand".
There are a lot of "text files" that are really interpreted code, like JavaScript.
I really don't want to spend the time whitelisting "safe" text files when there's not that much of a loss of speed for scanning a few writes to those files.
WHOOSH!
Don't forget that AV software costs you ALL the time when it's installed in "real time" scanning mode.
Most AV software can be configured to only scan on write, instead of on access.
This is a big help for the vast majority of cases (like loading a program, which would require no extra work), but still can be an unnecessary burden if you have a lot of files being written that can't really be virus carriers (e.g., log files).
My crystal ball says few people will be buying standalone graphics cards in 5 years.
Based on the fact that there are games being released today that won't run smoothly (i.e., at least 30fps at all times) with highest levels of detail when using the current fastest standalone card, I suspect your crystal ball is cloudy.
Since these standalone cards can generate upwards of 200 watts of heat, there is no way to put that much computing power into a package that already generates 80-100 watts to do CPU things. As the chips continue to use smaller processes, there will be advances that allow more computing with less heat, but those graphics cards are already at 40nm tech, so they're only about half a generation behind.
Based on the progress over the last few years, I suspect that 5 years from now we will see about 2-3 times the GPU power with about 50-75% of the heat. Even if the CPU portion does as well in technological advance, you'd end up with a 140 watt TDP (at best) for a combined CPU/GPU that doesn't require a standalone card to run most games.
So what the hell am I supposed to do in the face of this annoying setup?
Stop defying the laws of physics?
<Scotty>Ya cannot change the laws of physics, Jim.</Scotty>
There's actually no need to buy anything to use VMware, as Player is free and allows you to create VMs...it just doesn't have some of the better features (snapshot control, etc.) of VMware Workstation.
VMware Server is also free, and has a lot of great features. For bare-metal, ESXi is also free.
With all that plus VirtualBox, Xen, KVM, various Microsoft offerings, etc., there really is no need to pay anything for a hypervisor.
Granted, it does directly or indirectly stress the fpu, cache, maybe task switching and interrupt handling. However, there are many more things that can go wrong.
Off the top of my head, I can think of a lot of things that specifically need tested that one program probably won't do. For example, you need to verify both 32-bit and 64-bit operations. Prime95 is specifically compiled for one or the other, so would stress less of the "other" version.
There are also a lot of SIMD instructions that need tested. Some are obscure enough that only a few apps would use them.
Then, there's all the instructions that support virtualization. I have found that bad hardware running a hypervisor will fail much more frequently than if it is running a "normal" OS (YMMV).
But, unlike Memtest86+ for RAM, there doesn't appear to be any program that specifically tests all CPU subsystems (registers, cache, instruction execution, etc.).
What about the LATER assignments? One I had for my early course involved writing a ticketmaster-like ticket purchasing system where it would find and temporarily reserve the 'best' seating for a block of tickets. It would release these tickets if they were not purchased. It was a very simple problem (didn't get into multiple users), but I'd be willing to bet that it wouldn't be likely for two students in a group of 100 to come up with the same solution.
In my data structures class, when we started on balanced trees, the prof had an assignment where the input would be dictionary words and we were supposed to sort them using balanced trees as the temporary store. We also were supposed to re-use the last assignment where we used normal binary trees. This was to show that if the input was already sorted, the tree devolved into a list and you hit O(n) instead of the O(log n) that balanced trees should always result in.
As extra credit, he wanted us to avoid the worst case for the binary tree, so I stored the values after applying strrev (or a homemade version thereof). People tried many things (some of them similar), but nobody else used my trick. If you know the domain of your data, you can often improve your algorithm.
To test my theory, I reversed all the strings in /usr/dict/words, sorted it, then reversed them back. What you get looks pretty much random, so my trick shouldn't end up worst case very often.