What's even worse is when your company's IT department uses Windows group policy to lock the resolution of every machine at 1024x768 regardless of what monitor is attached, and also disables right-clicking on the desktop so you can't even think about changing it. That's bad, but what makes it worse is that most of the monitors at work are 22" widescreens with a native resolution of 1680x1050, 1600x900, or 1920x1080 so everything is blurry AND stretched.
AMD *does* push out affordable 4-socket Opteron setups- the Opteron 6000 series CPUs. They are selling those a whole ton less expensive now than they did in the K8 days. The least-expensive Opteron 6000s sell for $266 each and the most-expensive ones are around $1200-1500, compared to starting around $800 each and going on up to close to $3000 for the K8-era 4-way-capable Opterons. Considering a 4-way-capable Intel Xeon still costs close to $2000 and goes on up to near $5000- and is based on two-year-old technology- the Opterons are that great deal you were wishing for.
However on the desktop, Intel has gotten much better in their pricing (i.e. they don't cripple lower-end chips as severely as they used to) and is giving AMD a real run for their money.
No. Computer search algorithms work by taking a large amount of possible matches and then reducing that number by applying filters. If you had a person come in with two days of fever and a cough, the computer would list every condition that has fever and cough as a possibility. It would then recommend a bunch of tests to try and eliminate conditions from the "possible conditions" list until only one remains. If you want to talk about extremely expensive, that's how you get there- go down a list of potential causes of your symptoms and then run tests until you reach the bottom of the list or find your answer.
Also, a computerized search algorithm only works well when it is fed enough accurate and relevant data. Patients frequently have ill-defined, vague symptoms that can change with time. A particular disease can manifest itself quite differently in different patients- the old joke is that a certain set of symptoms being "textbook" is because you will rarely see those exact symptoms for a condition outside of the textbooks! Patients also do not remember their symptoms perfectly and may give inaccurate or untrue information. They also may have multiple different medical conditions that may be playing a role in their symptoms. Lab tests can be erroneous or misleading. A major task of the physician is to determine what information is relevant or irrelevant, accurate or inaccurate, complete or incomplete. Computers can't do this, and will often end up way out in left field because of it. I would hate to see a computer try to handle a hypochondriac, malingerer, or drug seeker- all patients who are deliberately giving untrue information.
Humans aren't perfect, but we are a heck of a lot better in doing the neccessary pattern matching required for practicing medicine than computers. Also, remember that a lot of medicine isn't "House." Most of general medicine is management of known conditions, not diagnosis of new ones. Go ask a family doc or internist how many of their visits are regular follow-up visits for chronic conditions or preventative care visits versus visits for new-onset conditions they have to work up. Many of the new problems are also obvious and the key is the management, not diagnosis. A guy comes in with a 5 cm laceration on his arm. Somebody has blood pressure readings of 150/95 on the last two office visits. A 30-year-old female trying to have a baby for the past 6 months has nausea in the morning, missed her last period, and has a positve home pregnancy test. You don't need an IBM "Watson" to make those diagnoses, but you do need skills to treat them and good follow up.
Medicine is largely heuristics as the physician is matching the symptoms the patient has to a diagnosis, and patients' symptoms are rarely ever "textbook" symptoms. There is a lot of overlap of symptoms and history between many different conditions early on in the illness. People are great at heuristics. Computers are awful at heuristics but are excellent at data storage, algorithmic sorting, and data retrieval. The reason the Jeopardy computer did well is because it was asked to retrieve data. You entered in a question with only one correct answer and it could brute-force search an entire enormous database to arrive at the answer. Medicine is nothing like that unless you want every last test known to man performed on every patient so that the computer can arrive at a brute-force "we have eliminated every possible other diagnosis" kind of an answer. Computers really only aid the physician in doing what computers do best, assisting with information storage and retrieval, such as storing patients' medical histories and allowing the physician to search journals and online databases for information.
Not quite. AMD did create two completely new architectures in the past 10 years, they are Bulldozer and Bobcat. The K8 is a heavily tweaked K7, to the point that a lot of the core logic is similar, although modified for working in amd64 mode. K10/Stars is a decently tweaked K8. Bobcat and Bulldozer are in fact all-new designs from the ground up.
Intel however has been using the same Pentium Pro P6 architecture from 1995 in their current chip lineups. They have invented NO new microarchitectures in the past 10 years. The last truly new architecture Intel has made for x86 is NetBurst in 2000, which flamed out. (pun intended.) Intel used the barely modified P6 architecture in the PII and PIII, and then tweaked it a little for the Pentium M and original Yonah CPUs. They modified it a decent amount for Core 2, but no more than what AMD did to make K7 into K8. Everything Intel has made since the Core 2 has been pretty minor upgrades to the underlying architecture. Your list exemplifies this. Merom and Conroe are the same die; one was for laptop products and the other was for desktop products. Kentsfield was simply two Conroes MCMed on one CPU. Wolfdale is a die-shrunk Conroe with a few additional SIMD instructions. Arrandale never saw the light of day, and Clarkdale is a Wolfdale shrunk to 32 nm with a few more minor SIMD tweaks and the MCH MCMed onto the CPU package. Lynnfield is the "most different" of those chips that you present, only because it incorporates an IMC, L3 cache, and uses the Pentium 4's SMT implementation. But the core block diagram is very similar to the Core 2s and readily identifiable as a P6-based part. The biggest difference between Intel's chips isn't in the microarchitecture, but the platform. Rolling an IGP into a chip doesn't change the microarchitecture, but it certainly changes the platform a lot. Ditto with adding an IMC.
So by your metric, it is Intel that is really lagging in making truly new products as they keep warming over a 1995 CPU design.
You are both right. Most sinus infections are viral, and many patients with viral sinus infections demand that the doctor "do something," which generally means they want antibiotics. Many times they still demand antibiotics even though the doctor explains that antibiotics will not work for their *viral* infection. Thus anybody with sinusitis, be it viral or bacterial, is somebody who "might" be prescribed antibiotics.
An armed populace very well may win a war against the military. Look at what's happening overseas, you have largely illiterate, uneducated guys often with poor vision with Soviet Bloc rifles successfully fighting a war of attrition against the U.S. military. A hundred thousand soldiers with planes and tanks and drones can't kill those guys, how are they expected to take on tens of millions of people armed similarly?
The "AMD processors start on fire if you don't put a heatsink on them" bit hasn't been true in a long time. Yes, early AMD CPUs that needed customer-applied heatsinks did not have any sort of catastropic overheat protection, namely the K6s. However, by the time AMD came back from the SECC sold-with-a-heatsink-attached Slot A CPUs to Socket A CPUs, the platform was supposed to have a catastropic thermal shutdown feature. Not all motherboard vendors actually implemented it though, and that's where the infamous Tom's Hardware Guide video where they roasted an old Duron came from. AMD mandated the catastrophic overheat shutdown feature in K8s and some/all Stars core (K10) and later AMD CPUs can throttle like P4 and later Intel chips when they overheat as well.
The BCG vaccine screws up using the skin test; your body has "seen" the TB antigen in the BCG vaccine and will react to the TB antigen in the skin test. It does NOT screw up the antibody test (Quantiferon), so people who received the BCG vaccine should get the antibody test instead of the skin test. Also, there is a direct test for M. tuberculosis bacteria- it's called a sputum culture. However you won't be able to grow any unless the patient has active TB as latent TB by definition has a negative culture. The skin test and Quantiferon test for latent TB, in the hopes that you catch latent TB, treat it, and kill it before it becomes active (and actively contagious) TB.
No, it increases the cost of prisons. The state has custody of the prisoner and thus is responsible for providing the prisoner's medical care. Prisoners with extremely difficult to kill diseases cost the state a lot of money because they not only have to treat the disease but tie up on officer or two 24/7 to supervise the prisoner in the hospital. The state would love it if the prisoners were perfectly healthy.
Intel: I know, let's try to see just how many features/cores/cache we can fuse off in our dies and different socket combinations to try to make *puts pinky finger to mouth* one MILLION SKUs! Oh, and while we're at it, let's add a FOURTH memory channel, because more is better! Sure, we could get all the bandwidth we need with two DDR3-1866 or -2133 channels and that you really only get about three channels' worth of bandwidth because we have to clock the IMC down to DDR3-1333 with two modules per channel- but we still have FOUR channels! Oh, and we forgot, it's the start of a new quarter so we need to release a new socket. Can't let those socket suppliers get lazy making last quarter's socket design. What, you guys want us to release Sandy Bridge-based Xeon MPs because MP platforms actually need that much bandwidth and core count? We just released the Westmere-based ones a few months ago! Don'tcha know that Xeon MPs run two years behind everything else? Geez, what did you do, wake up yesterday? Next you'll want us to stop crippling our chips, stop using a new socket every other month or something ridiculous like that. Where do you guys get those ideas?
AMD: Based on market analysis, most server applications use primarily integer code and require a lot of bandwidth, memory capacity, and a high core count. We don't have over a hundred billion dollars in market cap to fund several parallel R&D teams to design a specific CPU for every edge use case, so we will design a CPU that is highly modular, has good integer performance (because that's what our research indicated most server apps are), and has a lot of cores. Experience with Intel's HyperThreading is less than stellar with regards to predictable performance, so we will use our CMT approach that leads to better integer performance than HyperThreading but doesn't increase the die size by a huge amount, since we can't afford to make 400-600 mm^2 dies like Intel does to have a lot of physical cores. Oh, and we'll continue to use the existing server platforms out there so our customers can drop-in upgrade and we'll also not change any feature sets in the SKU stack other than the clock speed and number of enabled modules and their associated caches. We do apologize for being "late" with these parts since we usually release server and client at about the same time...
You are a little off.
1. Nehalem introduced LGA1567 alongside LGA1156 and LGA1366. Westmere was never released for LGA1156, Clarkdale was but Clarkdale used a significantly different die than Westmere (two vs. six cores, no on-CPU-die memory controller). They pretty much only shared the 32 nm node and general microarchitecture.
2. Arrandale used Socket G1, not LGA115x.
3. Prescott was the first chip on LGA775, followed by Smithfield, followed by Cedar Mill/Presler. Conroe was the third or fifth generation of LGA775 chip depending how you look at it.
4. Dempsey (Xeon equivalent of Presler) was the first generation of LGA771 chip. Woodcrest (Conroe) came later. You could argue that Dempsey stank and that Woodcrest was the first chip on LGA771 worth buying, and you'd probably be right. LGA771 really didn't catch fire (pun intended) until Clovertown quad-core MCMs because AMD was notoriously late in bringing the Opteron quads to market.
LGA2011 really replaces LGA1567 for the Xeon MPs. Intel's on-again, off-again planned replacement for LGA1366 is actually LGA1356. However, it seems like Intel may somewhat simplify their socket lineup and just have LGA1155 for general uniprocessor platforms and LGA2011 for performance UP and all multprocessor applications.
If you are going to leak something, at least make sure you don't have the test result date be two years before the lab doing the testing operated under that name and offered their services to the general, non-porn-industry public.
Caloric restrictions may keep the weight down, but actual exercise has cardiovascular benefits on top of just not adding more cholesterol plaques to the arteries. Also, weight-bearing exercise increases bone mineral density and makes you less likely to break a hip and end up bedridden when you are old.
No, but which school YOU go to might change as the district uses fewer school buildings after the consolidation. Fewer schools in a given area than before = longer drives for most of that area's students to get to school.
They wouldn't refuse to use your body. A likely use would be as a cadaver for medical and nursing/allied health students to dissect and learn anatomy. Teaching anatomy via dissection requires a large number of cadavers on an annual basis, and there is nothing patentable in doing a dissection.
The only Extreme Editions based on Xeons have been the first P4 Extreme Edition (Gallatin 3.2, 3.4, 3.46) and the Core 2 Extreme 9775. The former was based on the Xeon MP/DP "Gallatin" die with the 2 MB of L3, which never was used for any non-P4EE single-socket CPU. The Core 2 Extreme 9775 was essentially a rebranded Xeon X5482 with an unlocked multiplier, complete down to the dual-CPU capability and LGA771 socket that no other Core 2-branded CPU had. All of the other Extreme Edition CPUs have simply been the fastest desktop or laptop CPU of the line with an unlocked multiplier.
Only the UP models. Xeon DPs have been sold only without a fan since the Xeon 5400 series. Xeon DPs before that could be bought as fanless "tray" CPUs, retail box units with an active-fan heatsink, or retail-box units with a passive HSF. I believe that the Xeon MPs have always been sold as "tray" CPUs without heatsinks since the first socketed P4 Xeon models.
Xeon DP 5500 and 5600 series ship without heatsinks; the previous Xeons did ship with heatsinks. Intel sold separate SKUs with the passive (ended with a -P) and 2U active (ended with an -A) heatsink. I heard they stopped since people frequently ordered the wrong SKU and they had a lot of returns, so they quit offering heatsinks. AMD also used to ship all of their Socket 940 Opterons with heatsinks as well. They stopped with Socket F chips since some board makers used a 4.1" bolt hole spacing and others used the Socket 940 3.5" pitch. Now AMD specifically defined a bolt hole spacing (3.5" for C32, 4.1" for G34) but they ship all CPUs without a heatsink.
The bandwidth between the CPU and memory *has* been increasing pretty noticeably recently. Your "8-core, 3 GHz machine with lots-o-ram and a wide, fast main memory bus" sounds like a dual Xeon X5365, E5450/X5450, or X5472. The 5000X or 5400 chipset used with those CPUs does have a wide memory bus, but that quad-channel DDR2-667 or -800 FB-DIMM bus generally results in a pretty paltry 6-9 GB/sec when it's all said and done, along with significant latency in both the FSB and FB-DIMMs, especially since you have multiple DIMMs per channel ("lots-o-RAM") and two dies per FSB with those quad-core chips. That's beaten by a good margin by a typical dual-channel desktop CPU, let alone a modern server with 6 or 8 DDR3-1333 channels. My dual 8-core 2.0 GHz Opteron 6128 workstation with 8 channels of DDR3-1333 has an actual memory bandwidth north of 50 GB/sec, by comparison.
Slashdot Mirror
What's even worse is when your company's IT department uses Windows group policy to lock the resolution of every machine at 1024x768 regardless of what monitor is attached, and also disables right-clicking on the desktop so you can't even think about changing it. That's bad, but what makes it worse is that most of the monitors at work are 22" widescreens with a native resolution of 1680x1050, 1600x900, or 1920x1080 so everything is blurry AND stretched.
AMD *does* push out affordable 4-socket Opteron setups- the Opteron 6000 series CPUs. They are selling those a whole ton less expensive now than they did in the K8 days. The least-expensive Opteron 6000s sell for $266 each and the most-expensive ones are around $1200-1500, compared to starting around $800 each and going on up to close to $3000 for the K8-era 4-way-capable Opterons. Considering a 4-way-capable Intel Xeon still costs close to $2000 and goes on up to near $5000- and is based on two-year-old technology- the Opterons are that great deal you were wishing for.
However on the desktop, Intel has gotten much better in their pricing (i.e. they don't cripple lower-end chips as severely as they used to) and is giving AMD a real run for their money.
No. Computer search algorithms work by taking a large amount of possible matches and then reducing that number by applying filters. If you had a person come in with two days of fever and a cough, the computer would list every condition that has fever and cough as a possibility. It would then recommend a bunch of tests to try and eliminate conditions from the "possible conditions" list until only one remains. If you want to talk about extremely expensive, that's how you get there- go down a list of potential causes of your symptoms and then run tests until you reach the bottom of the list or find your answer.
Also, a computerized search algorithm only works well when it is fed enough accurate and relevant data. Patients frequently have ill-defined, vague symptoms that can change with time. A particular disease can manifest itself quite differently in different patients- the old joke is that a certain set of symptoms being "textbook" is because you will rarely see those exact symptoms for a condition outside of the textbooks! Patients also do not remember their symptoms perfectly and may give inaccurate or untrue information. They also may have multiple different medical conditions that may be playing a role in their symptoms. Lab tests can be erroneous or misleading. A major task of the physician is to determine what information is relevant or irrelevant, accurate or inaccurate, complete or incomplete. Computers can't do this, and will often end up way out in left field because of it. I would hate to see a computer try to handle a hypochondriac, malingerer, or drug seeker- all patients who are deliberately giving untrue information.
Humans aren't perfect, but we are a heck of a lot better in doing the neccessary pattern matching required for practicing medicine than computers. Also, remember that a lot of medicine isn't "House." Most of general medicine is management of known conditions, not diagnosis of new ones. Go ask a family doc or internist how many of their visits are regular follow-up visits for chronic conditions or preventative care visits versus visits for new-onset conditions they have to work up. Many of the new problems are also obvious and the key is the management, not diagnosis. A guy comes in with a 5 cm laceration on his arm. Somebody has blood pressure readings of 150/95 on the last two office visits. A 30-year-old female trying to have a baby for the past 6 months has nausea in the morning, missed her last period, and has a positve home pregnancy test. You don't need an IBM "Watson" to make those diagnoses, but you do need skills to treat them and good follow up.
Medicine is largely heuristics as the physician is matching the symptoms the patient has to a diagnosis, and patients' symptoms are rarely ever "textbook" symptoms. There is a lot of overlap of symptoms and history between many different conditions early on in the illness. People are great at heuristics. Computers are awful at heuristics but are excellent at data storage, algorithmic sorting, and data retrieval. The reason the Jeopardy computer did well is because it was asked to retrieve data. You entered in a question with only one correct answer and it could brute-force search an entire enormous database to arrive at the answer. Medicine is nothing like that unless you want every last test known to man performed on every patient so that the computer can arrive at a brute-force "we have eliminated every possible other diagnosis" kind of an answer. Computers really only aid the physician in doing what computers do best, assisting with information storage and retrieval, such as storing patients' medical histories and allowing the physician to search journals and online databases for information.
Not quite. AMD did create two completely new architectures in the past 10 years, they are Bulldozer and Bobcat. The K8 is a heavily tweaked K7, to the point that a lot of the core logic is similar, although modified for working in amd64 mode. K10/Stars is a decently tweaked K8. Bobcat and Bulldozer are in fact all-new designs from the ground up.
Intel however has been using the same Pentium Pro P6 architecture from 1995 in their current chip lineups. They have invented NO new microarchitectures in the past 10 years. The last truly new architecture Intel has made for x86 is NetBurst in 2000, which flamed out. (pun intended.) Intel used the barely modified P6 architecture in the PII and PIII, and then tweaked it a little for the Pentium M and original Yonah CPUs. They modified it a decent amount for Core 2, but no more than what AMD did to make K7 into K8. Everything Intel has made since the Core 2 has been pretty minor upgrades to the underlying architecture. Your list exemplifies this. Merom and Conroe are the same die; one was for laptop products and the other was for desktop products. Kentsfield was simply two Conroes MCMed on one CPU. Wolfdale is a die-shrunk Conroe with a few additional SIMD instructions. Arrandale never saw the light of day, and Clarkdale is a Wolfdale shrunk to 32 nm with a few more minor SIMD tweaks and the MCH MCMed onto the CPU package. Lynnfield is the "most different" of those chips that you present, only because it incorporates an IMC, L3 cache, and uses the Pentium 4's SMT implementation. But the core block diagram is very similar to the Core 2s and readily identifiable as a P6-based part. The biggest difference between Intel's chips isn't in the microarchitecture, but the platform. Rolling an IGP into a chip doesn't change the microarchitecture, but it certainly changes the platform a lot. Ditto with adding an IMC.
So by your metric, it is Intel that is really lagging in making truly new products as they keep warming over a 1995 CPU design.
You are both right. Most sinus infections are viral, and many patients with viral sinus infections demand that the doctor "do something," which generally means they want antibiotics. Many times they still demand antibiotics even though the doctor explains that antibiotics will not work for their *viral* infection. Thus anybody with sinusitis, be it viral or bacterial, is somebody who "might" be prescribed antibiotics.
An armed populace very well may win a war against the military. Look at what's happening overseas, you have largely illiterate, uneducated guys often with poor vision with Soviet Bloc rifles successfully fighting a war of attrition against the U.S. military. A hundred thousand soldiers with planes and tanks and drones can't kill those guys, how are they expected to take on tens of millions of people armed similarly?
The "AMD processors start on fire if you don't put a heatsink on them" bit hasn't been true in a long time. Yes, early AMD CPUs that needed customer-applied heatsinks did not have any sort of catastropic overheat protection, namely the K6s. However, by the time AMD came back from the SECC sold-with-a-heatsink-attached Slot A CPUs to Socket A CPUs, the platform was supposed to have a catastropic thermal shutdown feature. Not all motherboard vendors actually implemented it though, and that's where the infamous Tom's Hardware Guide video where they roasted an old Duron came from. AMD mandated the catastrophic overheat shutdown feature in K8s and some/all Stars core (K10) and later AMD CPUs can throttle like P4 and later Intel chips when they overheat as well.
Sick prisoners often go to an actual hospital and thus require 1:1 guarding. Trust me, I work at a hospital and frequently see this.
The BCG vaccine screws up using the skin test; your body has "seen" the TB antigen in the BCG vaccine and will react to the TB antigen in the skin test. It does NOT screw up the antibody test (Quantiferon), so people who received the BCG vaccine should get the antibody test instead of the skin test. Also, there is a direct test for M. tuberculosis bacteria- it's called a sputum culture. However you won't be able to grow any unless the patient has active TB as latent TB by definition has a negative culture. The skin test and Quantiferon test for latent TB, in the hopes that you catch latent TB, treat it, and kill it before it becomes active (and actively contagious) TB.
No, it increases the cost of prisons. The state has custody of the prisoner and thus is responsible for providing the prisoner's medical care. Prisoners with extremely difficult to kill diseases cost the state a lot of money because they not only have to treat the disease but tie up on officer or two 24/7 to supervise the prisoner in the hospital. The state would love it if the prisoners were perfectly healthy.
13 years...let me bet, you're an MD/PhD or you switched specialties during residency several times?
Eh, how about this:
Intel: I know, let's try to see just how many features/cores/cache we can fuse off in our dies and different socket combinations to try to make *puts pinky finger to mouth* one MILLION SKUs! Oh, and while we're at it, let's add a FOURTH memory channel, because more is better! Sure, we could get all the bandwidth we need with two DDR3-1866 or -2133 channels and that you really only get about three channels' worth of bandwidth because we have to clock the IMC down to DDR3-1333 with two modules per channel- but we still have FOUR channels! Oh, and we forgot, it's the start of a new quarter so we need to release a new socket. Can't let those socket suppliers get lazy making last quarter's socket design. What, you guys want us to release Sandy Bridge-based Xeon MPs because MP platforms actually need that much bandwidth and core count? We just released the Westmere-based ones a few months ago! Don'tcha know that Xeon MPs run two years behind everything else? Geez, what did you do, wake up yesterday? Next you'll want us to stop crippling our chips, stop using a new socket every other month or something ridiculous like that. Where do you guys get those ideas?
AMD: Based on market analysis, most server applications use primarily integer code and require a lot of bandwidth, memory capacity, and a high core count. We don't have over a hundred billion dollars in market cap to fund several parallel R&D teams to design a specific CPU for every edge use case, so we will design a CPU that is highly modular, has good integer performance (because that's what our research indicated most server apps are), and has a lot of cores. Experience with Intel's HyperThreading is less than stellar with regards to predictable performance, so we will use our CMT approach that leads to better integer performance than HyperThreading but doesn't increase the die size by a huge amount, since we can't afford to make 400-600 mm^2 dies like Intel does to have a lot of physical cores. Oh, and we'll continue to use the existing server platforms out there so our customers can drop-in upgrade and we'll also not change any feature sets in the SKU stack other than the clock speed and number of enabled modules and their associated caches. We do apologize for being "late" with these parts since we usually release server and client at about the same time...
You are a little off. 1. Nehalem introduced LGA1567 alongside LGA1156 and LGA1366. Westmere was never released for LGA1156, Clarkdale was but Clarkdale used a significantly different die than Westmere (two vs. six cores, no on-CPU-die memory controller). They pretty much only shared the 32 nm node and general microarchitecture. 2. Arrandale used Socket G1, not LGA115x. 3. Prescott was the first chip on LGA775, followed by Smithfield, followed by Cedar Mill/Presler. Conroe was the third or fifth generation of LGA775 chip depending how you look at it. 4. Dempsey (Xeon equivalent of Presler) was the first generation of LGA771 chip. Woodcrest (Conroe) came later. You could argue that Dempsey stank and that Woodcrest was the first chip on LGA771 worth buying, and you'd probably be right. LGA771 really didn't catch fire (pun intended) until Clovertown quad-core MCMs because AMD was notoriously late in bringing the Opteron quads to market.
LGA2011 really replaces LGA1567 for the Xeon MPs. Intel's on-again, off-again planned replacement for LGA1366 is actually LGA1356. However, it seems like Intel may somewhat simplify their socket lineup and just have LGA1155 for general uniprocessor platforms and LGA2011 for performance UP and all multprocessor applications.
If you are going to leak something, at least make sure you don't have the test result date be two years before the lab doing the testing operated under that name and offered their services to the general, non-porn-industry public.
Caloric restrictions may keep the weight down, but actual exercise has cardiovascular benefits on top of just not adding more cholesterol plaques to the arteries. Also, weight-bearing exercise increases bone mineral density and makes you less likely to break a hip and end up bedridden when you are old.
No, but which school YOU go to might change as the district uses fewer school buildings after the consolidation. Fewer schools in a given area than before = longer drives for most of that area's students to get to school.
How about something a little less expensive?
The real question that science can't answer about the universe but one religions do answer is WHY.
They wouldn't refuse to use your body. A likely use would be as a cadaver for medical and nursing/allied health students to dissect and learn anatomy. Teaching anatomy via dissection requires a large number of cadavers on an annual basis, and there is nothing patentable in doing a dissection.
The only Extreme Editions based on Xeons have been the first P4 Extreme Edition (Gallatin 3.2, 3.4, 3.46) and the Core 2 Extreme 9775. The former was based on the Xeon MP/DP "Gallatin" die with the 2 MB of L3, which never was used for any non-P4EE single-socket CPU. The Core 2 Extreme 9775 was essentially a rebranded Xeon X5482 with an unlocked multiplier, complete down to the dual-CPU capability and LGA771 socket that no other Core 2-branded CPU had. All of the other Extreme Edition CPUs have simply been the fastest desktop or laptop CPU of the line with an unlocked multiplier.
Only the UP models. Xeon DPs have been sold only without a fan since the Xeon 5400 series. Xeon DPs before that could be bought as fanless "tray" CPUs, retail box units with an active-fan heatsink, or retail-box units with a passive HSF. I believe that the Xeon MPs have always been sold as "tray" CPUs without heatsinks since the first socketed P4 Xeon models.
Xeon DP 5500 and 5600 series ship without heatsinks; the previous Xeons did ship with heatsinks. Intel sold separate SKUs with the passive (ended with a -P) and 2U active (ended with an -A) heatsink. I heard they stopped since people frequently ordered the wrong SKU and they had a lot of returns, so they quit offering heatsinks. AMD also used to ship all of their Socket 940 Opterons with heatsinks as well. They stopped with Socket F chips since some board makers used a 4.1" bolt hole spacing and others used the Socket 940 3.5" pitch. Now AMD specifically defined a bolt hole spacing (3.5" for C32, 4.1" for G34) but they ship all CPUs without a heatsink.
The bandwidth between the CPU and memory *has* been increasing pretty noticeably recently. Your "8-core, 3 GHz machine with lots-o-ram and a wide, fast main memory bus" sounds like a dual Xeon X5365, E5450/X5450, or X5472. The 5000X or 5400 chipset used with those CPUs does have a wide memory bus, but that quad-channel DDR2-667 or -800 FB-DIMM bus generally results in a pretty paltry 6-9 GB/sec when it's all said and done, along with significant latency in both the FSB and FB-DIMMs, especially since you have multiple DIMMs per channel ("lots-o-RAM") and two dies per FSB with those quad-core chips. That's beaten by a good margin by a typical dual-channel desktop CPU, let alone a modern server with 6 or 8 DDR3-1333 channels. My dual 8-core 2.0 GHz Opteron 6128 workstation with 8 channels of DDR3-1333 has an actual memory bandwidth north of 50 GB/sec, by comparison.