I've been to Fallingwater. It's a disaster. I believe the cantilever is currently propped up while they try to figure out how to fix the mess Mr. Wright made. The ceilings are low, the cauldron is too large to be useful, the floor is uneven.
"The History of Modern Computing" claims that this definition of "computer" didn't change until the 50's or so. I've lent out my copy of that book, and can't check.
In response to the parent post: people studied different things. For instance, "data mining" is something very new that requires computers to be reasonable. Differential equations is something old, but has benefitted from computers on the "applied" side (if you're not a math person, "applied" probably means something different than what you think it means;-). Nearly all of mathematics doesn't need computers in any way, but that will probably change now that the cold war is over and funding of math theory is (has been) in decline. Of course GW might start a new cold war.
Don't worry so much -- the P4 fpu isn't that good. The memory bandwidth with (very expensive) Rambus is nice, but the fpu is nothing to write home about. And don't get me started about the P4's fp exception handling (I have a web page about it).
Outside of memory bandwidth when used with Rambus RAM, I haven't found anything exciting about the P4. The context of my benchmarks is typically numerical statistical algorithms on large datasets.
I've been working about 14 hours per day for three weeks (oops, I'm lying -- I've only been doing 14 hours per day for the last week) on a conference paper due tomorrow at midnight. My only entertainment is during 10 minute typing breaks, and I occasionally steal a little time to post to/. or send an email. I feel like I'm inside my iPAQ by now.;-)
I've worked inside a handful of medium-sized machines, including a couple Microway dual Alpha "rugged racks" and a Compaq ES40 Model II. The 8u (or 9?) ES40 is nicely laid out, but removing the motherboard or messing with drive cables is a pain. The 4u "rugged racks" are a disaster of fans and wires. We've had 4 or 5 fans go out on those, and it takes approximately 45 screws (I counted) and about 60 minutes of fast work to remove and replace a midboard fan.
That opteron case, on the other hand, appears to have plenty of cooling that is easy to reach. I don't see any wires permanently attached to the case. It looks very clean and easy to service, except possibly getting the motherboard out.
At present the government has far less centralized information on you than the private sector, and there are far stricter controls on what the government does with this information. This is reasonable, because the government has executive power (i.e. "they" can arrest you).
TIA is a poorly thought-out and mixed-up program to be sure. But it's not really what you need to worry about. What you need to worry about are the laws governing how the government can use information from the TIA program. In that sense, things have been getting worse ever since September 11th and Congress got all stupid on us.
I expect that Congress and President George W. Bush have inflicted more lasting damage on this country than terrorists ever could.
That said, TIA needs close watch and supervision, so that we can arrest Poindexter (again) if he screws up (again).
I've seen a keypad like this. It was a security device to enter a (clearly over-valued) university computer lab. The keys were lenses that absolutely could not be viewed except for straight-on. Before entering a pass code, you pushed a button to scramble the digits, making key positions irrelevant.
Anyone know who makes such a device? I'm sure I can't afford one (especially having no good reason to own one), just like those awesome $20 ceramic toggle switch protectors the military uses. But I still want to know *where* I could get them, if I wanted (this reminds me of Life of Brian).
We have some old 233MHz Pentiums in our lab, one of which had the cd burner. When transferring Red Hat isos or some such, even ftp couldn't keep up with the network (100Mbit) -- it would hang around 6MB/sec (and using ftp was a pain for various uninteresting reason). Using -c blowfish, I would get around 5MB/sec, which was close enough for me.
I haven't had the slowdown problems you're describing. I don't see exactly the same times, but the difference is rather small. I expect it's a client or server cpu thing. Check if using lighter encryption helps, for instance, compare
scp foo@bar:~/baz.
with
scp -c blowfish foo@bar:~/baz.
If you are using really small files for testing, then the initial handshake alone could swamp your comparison. Try something in the 50MB range, or larger.
You skipped the helmet lecture, so I'll do it for you (no matter how popular).
I've been hit, from the side, by a motorist ignoring right-of-way at a stop sign (she claimed she didn't see me, gave the police 2 different stories in which she did see me, and gave the court yet another story). I had a minor concusion but no other serious injuries. My bicycle had bent forks, bent frame, bent weels, and bent cranks. Without the helmet, I don't know what would have happened to me.
Always ride with a helmet. Always ride with lights.
Note that the least important part of lights is helping you see. The most important part is helping you to *be seen*.
Don't follow the example of moronic bike police and where all black at night. Especially without a light.
When you ride, be *predictable*. If you ride in traffic, which is the law and best thing to do in many circumstances, pretend you're a car. Do car like things. You'll probably be relatively slow, but at least the drivers will know what you'll do next.
If you ride on the sidewalk, act like a pedestrian. Dismount at intersections and walk your bike across the street.
If you ride in bike lanes in most US cities, be aware that you are a prime target for a right-turning car.
I have one question for you. Where? Where did you find 10K SCSI drives for those prices?
-Paul Komarek
Re:No, that's anti-capitalist.
on
NARAS vs. the RIAA
·
· Score: 2, Informative
This is as good of time as any to point out that voting with your wallet is not how changes are made in the US capitalist system. If I bring my small wallet to a fight, and the RIAA brings their war chests, I'm gonna lose. Even if me and my best 1000 friends bring their pathetic wallets, we'll still lose like sword fighters to nuclear weapons.
We have to make our views known with intelligent, clever action. If you think US society is being perverted by corruption in companies and Congress, then *do something* about it -- boycotts aren't enough. You don't have to paint a sign and walk around city streets, though that's one option. It is probably more effective to educate your friends and family directly. Other *do something* options are to fund people who are already doing things, like the EFF or FSF.
I would like to make two notes about the eduction process:
1) People learn best when they control the dialogue, i.e. when they are asking questions. No matter how much you want and try to teach something to someone, they control the learning.
2) You cannot educate unless you are educated on the topic. Read. Think. Rehearse arguments to find your weaknesses, and explore those weaknesses. If you want to know how smart you are, see how simply you can state your position. If you can't state it succinctly, then you don't know your position well enough.
These are things I've learner (learner the wrong way, I might add) in GNU/Linux advocacy over the last 5 years. I've been far more successful in having my voice heard, and changing people's minds, since I learned these things. Never have these lessons been more important than now, with America's apathy towards the "reasoned and wise" actions of our current Congress and administration.
Are you sure that putting (copyrighted) video games in public places constitutes public performance of a copyrighted work? For instance, I've never heard anyone suggest that cybercafes need a public performance license for Quake, or that using Matlab in an interactive demo during public talk (e.g. at a conference) needs a special Matlab license.
I'm guessing that using software does not constitute performance, public or otherwise.
Even the Simpsons have (indirectly) lampooned Edison's stealing of ideas. In this case, Edison was credited with a fancy leaning chair that Homer made.
You are correct, I meant the Archimedes. At least, I meant Archimedes after I'd read enough to find out I'd made a mistake.;-) Thanks for the correction.
I finally learned where the name "ARM" comes from:
StrongArm came from DEC, but isn't it a derivative of the Arm? The Xscale is also a derivative of the Arm, IIRC. And the Arm can be traced back to the BBC Micro, which was evidently constructed for a television series about computers.
Does the LinuxBIOS project even show up on your company's radar? That is, is there any concern that x86 BIOS will be (further) commoditized, and hence decrease BIOS profit margins? Where would your company go next if the bottom fell out of the BIOS market?
As an example of the decreased relevance of the BIOS in modern computing, consider how Linux (the kernel) ignores the BIOS wherever possible and goes straight to the hardware, relegating the BIOS to the boot sequence and (usually buggy) power management.
I guess that's why I responded. The Alpha primary market has been research, though Compaq tried to make noise about its server potential. As a researcher that does a lot of computation, I have no code that ignores half of my cpu. Thats why I like the Alpha, and why our lab group is stunned by the way a 667MHz Alpha performs as well as a P-IV at double or nearly triple the speed.
Big caches, big bandwidth, big memory (which is the real reason we're using them), and well-designed for numerical codes. I'm happy to defend the Alpha for the "corner" of the research market it does best -- numerical and statistical code. The only people I've met who argue performance/dollar for the Itanium got their Itanium for free.
My comment about Intel's level of CPU R&D was meant one of two ways:
1) They do lots but have little new to show for it. 2) They don't bother doing much since their slowly evolving x86 cores seem to earn them a whole lot of money.
Intel obviously has some great engineers holding that stupid x86 arch together. That anyone has ever gotten any speed from that ISA is remarkable. But Intel hasn't come up with any new *and* good designs for quite some time. The Itanic has been hyped for many, many years and has only managed to pull even with existing RISC cores (many of which are less expensive and lower power). As you said, "Itanium 2 is now competitive". Finally, Itanic is the only really new-ish cpu thing Intel has done in recent years.
Furthermore, everything I've read about the Itanic over the last five or so years suggests that HP was responsible for the bulk of Itanic's design. I don't have any percentages, as Intel wouldn't release them if they knew what they were anyway. I've read articles suggesting that Intel gagged HP about their role in the design of "Intel's" chip.
Overall, I'm not impressed by the R&D of Intel, a company whose R&D budget is probably larger than MIPS' gross revenue.
CPU R&D is *really* expensive. That's probably why Intel doesn't seem to do much CPU R&D and sticks mainly to manufacturing (note: Itanic arch came mainly from HP, StrongArm and XScale came from other outside sources). It takes very special companies to support this cost, and there aren't many that want to stay in the CPU R&D business anymore (much less get into it).
A large part of the reason GNU/Linux is doing so well compared to other UNIX and UNIX-like systems today is that it runs well on cheap *commodity* hardware. Nobody is going to be impressed if GNU/Linux runs better than WIntel on really expensive equipment. Not to say that people aren't impressed by GNU/Linux running on the IBM S390 for its own sake. =-)
It's not even clear that people *want* better hardware. After all, better hardware has been losing to cheaper hardware for at least 5 years now. It is suprisingly hard to explain to someone why adding an $1800 (or $5400 full-retail) cpu card to a Compaq ES40 is better than buying a new Athlon XP or P-IV since many of the costs aren't easy to account for. But the sysadmins using the scalable ES40 know just how much easier that machine is to care for (esp. remotely) than a uniprocessor $1800 commodity x86 box.
If I were to believe that everything was as simple as you make it, my conclusion would be that I need an Itanium 2 + Pentium IV + PA-RISC or POWER4 to have an overall better machine than an Alpha Server. Is that what you meant? Incidently, the reason the Alpha 21264 has 3 integer units (providing integer scores that dominate Itanic integer scores) is to keep the 2 fp units fed. I'm not an expert on the Itanic architecture, but I'm led to wonder if the Itanic integer units are capable of providing all of the array indexing and loop-counting chores needed for many floating-point numerical anaylysis algorithms. I'd really like to know an answer to this, since the Itanic appears so unbalanced when looking at SPEC scores alone.
Furthermore, you're comparing prices on processor cards for different systems using full retail price. Have you ever bought this kind of equipment? I know I've never paid full retail price when I have. And you don't need to use Compaq's pricing utility -- just get the model number and search with google like you probably did for th Itanic board. Unfortunately, for the cpu board (KN610-EB) you're describing, there aren't many links. Another problem is that the price you quote for the Alpha board includes additional cpu licenses for Tru64 (for the KN610-EB) or VMS (not sure -- KN610-EC?). I don't see anything at the link you provide which states that the price includes a Windows or HP-UX license (and don't kid yourself, you'll need a license).
It's not clear why you mention the 3MB cache on the Itanium. The EV67 boards (KN610-BA, for instance) for the ES40 Model II (which is old) have 8MB cache, and the EV68 board (KN610-EB, for instance) have 16MB of cache (note that these are the big L3 caches for Alphas, not on-chip like I think the Itanium is). Also, it's not clear to me that the processor boards include the same functional components -- do you know that they do (you already seem to have missed the OS license issue)?
You haven't provided the price for the systems, nor stated their default configurations. This is certainly important when making a comparison. You'll also want to compare memory prices, since ram for these servers sometimes has a special form factor and costs a bundle. Again, never use list prices for any of this stuff, as you can occasionally halve the price with a bit of negotiation. That goes for cars and good office chairs, too.
Last I heard, the important Alpha guys ended up at AMD or in small companies. I remember hearing that some or all of the DEC engineers working on StrongArm refused to move to Intel when Intel aquired that cpu, and that Intel had to move some people from the, geez, i810 or something similarly not StrongArm-ish into the StrongArm project.
At some point, I thought the important Alpha folks went to Nexgen, which was bought (directly or indirectly) by AMD.
I'd really like to have a more detailed and more accurate picture of who went where, but it's bedtime for me. Maybe someone else would like to elaborate.
We bought one of these lite-ons at CompUSA on special about a month ago. They ran out of a slower Buslink drive they had advertised, and sold us a 52x Buslink for $100 instead (it's a rebranded lite-on). It seems to work pretty well so far with the memorex media we've been using, but I can't say we've stressed it much (it's in a test machine).
Slashdot Mirror
I've been to Fallingwater. It's a disaster. I believe the cantilever is currently propped up while they try to figure out how to fix the mess Mr. Wright made. The ceilings are low, the cauldron is too large to be useful, the floor is uneven.
But the location can't be beat!
-Paul Komarek
"The History of Modern Computing" claims that this definition of "computer" didn't change until the 50's or so. I've lent out my copy of that book, and can't check.
;-). Nearly all of mathematics doesn't need computers in any way, but that will probably change now that the cold war is over and funding of math theory is (has been) in decline. Of course GW might start a new cold war.
In response to the parent post: people studied different things. For instance, "data mining" is something very new that requires computers to be reasonable. Differential equations is something old, but has benefitted from computers on the "applied" side (if you're not a math person, "applied" probably means something different than what you think it means
-Paul Komarek
Don't worry so much -- the P4 fpu isn't that good. The memory bandwidth with (very expensive) Rambus is nice, but the fpu is nothing to write home about. And don't get me started about the P4's fp exception handling (I have a web page about it).
Outside of memory bandwidth when used with Rambus RAM, I haven't found anything exciting about the P4. The context of my benchmarks is typically numerical statistical algorithms on large datasets.
-Paul Komarek
Thank-you. I needed the humor.
/. or send an email. I feel like I'm inside my iPAQ by now. ;-)
I've been working about 14 hours per day for three weeks (oops, I'm lying -- I've only been doing 14 hours per day for the last week) on a conference paper due tomorrow at midnight. My only entertainment is during 10 minute typing breaks, and I occasionally steal a little time to post to
-Paul Komarek
I've worked inside a handful of medium-sized machines, including a couple Microway dual Alpha "rugged racks" and a Compaq ES40 Model II. The 8u (or 9?) ES40 is nicely laid out, but removing the motherboard or messing with drive cables is a pain. The 4u "rugged racks" are a disaster of fans and wires. We've had 4 or 5 fans go out on those, and it takes approximately 45 screws (I counted) and about 60 minutes of fast work to remove and replace a midboard fan.
That opteron case, on the other hand, appears to have plenty of cooling that is easy to reach. I don't see any wires permanently attached to the case. It looks very clean and easy to service, except possibly getting the motherboard out.
-Paul Komarek
At present the government has far less centralized information on you than the private sector, and there are far stricter controls on what the government does with this information. This is reasonable, because the government has executive power (i.e. "they" can arrest you).
TIA is a poorly thought-out and mixed-up program to be sure. But it's not really what you need to worry about. What you need to worry about are the laws governing how the government can use information from the TIA program. In that sense, things have been getting worse ever since September 11th and Congress got all stupid on us.
I expect that Congress and President George W. Bush have inflicted more lasting damage on this country than terrorists ever could.
That said, TIA needs close watch and supervision, so that we can arrest Poindexter (again) if he screws up (again).
-Paul Komarek
I've seen a keypad like this. It was a security device to enter a (clearly over-valued) university computer lab. The keys were lenses that absolutely could not be viewed except for straight-on. Before entering a pass code, you pushed a button to scramble the digits, making key positions irrelevant.
Anyone know who makes such a device? I'm sure I can't afford one (especially having no good reason to own one), just like those awesome $20 ceramic toggle switch protectors the military uses. But I still want to know *where* I could get them, if I wanted (this reminds me of Life of Brian).
-Paul Komarek
We have some old 233MHz Pentiums in our lab, one of which had the cd burner. When transferring Red Hat isos or some such, even ftp couldn't keep up with the network (100Mbit) -- it would hang around 6MB/sec (and using ftp was a pain for various uninteresting reason). Using -c blowfish, I would get around 5MB/sec, which was close enough for me.
I'm a little jealous of your 500MHz machine. =-)
-Paul
I haven't had the slowdown problems you're describing. I don't see exactly the same times, but the difference is rather small. I expect it's a client or server cpu thing. Check if using lighter encryption helps, for instance, compare
.
.
scp foo@bar:~/baz
with
scp -c blowfish foo@bar:~/baz
If you are using really small files for testing, then the initial handshake alone could swamp your comparison. Try something in the 50MB range, or larger.
-Paul Komarek
You skipped the helmet lecture, so I'll do it for you (no matter how popular).
I've been hit, from the side, by a motorist ignoring right-of-way at a stop sign (she claimed she didn't see me, gave the police 2 different stories in which she did see me, and gave the court yet another story). I had a minor concusion but no other serious injuries. My bicycle had bent forks, bent frame, bent weels, and bent cranks. Without the helmet, I don't know what would have happened to me.
Always ride with a helmet. Always ride with lights.
Note that the least important part of lights is helping you see. The most important part is helping you to *be seen*.
Don't follow the example of moronic bike police and where all black at night. Especially without a light.
When you ride, be *predictable*. If you ride in traffic, which is the law and best thing to do in many circumstances, pretend you're a car. Do car like things. You'll probably be relatively slow, but at least the drivers will know what you'll do next.
If you ride on the sidewalk, act like a pedestrian. Dismount at intersections and walk your bike across the street.
If you ride in bike lanes in most US cities, be aware that you are a prime target for a right-turning car.
-Paul Komarek
I have one question for you. Where? Where did you find 10K SCSI drives for those prices?
-Paul Komarek
This is as good of time as any to point out that voting with your wallet is not how changes are made in the US capitalist system. If I bring my small wallet to a fight, and the RIAA brings their war chests, I'm gonna lose. Even if me and my best 1000 friends bring their pathetic wallets, we'll still lose like sword fighters to nuclear weapons.
We have to make our views known with intelligent, clever action. If you think US society is being perverted by corruption in companies and Congress, then *do something* about it -- boycotts aren't enough. You don't have to paint a sign and walk around city streets, though that's one option. It is probably more effective to educate your friends and family directly. Other *do something* options are to fund people who are already doing things, like the EFF or FSF.
I would like to make two notes about the eduction process:
1) People learn best when they control the dialogue, i.e. when they are asking questions. No matter how much you want and try to teach something to someone, they control the learning.
2) You cannot educate unless you are educated on the topic. Read. Think. Rehearse arguments to find your weaknesses, and explore those weaknesses. If you want to know how smart you are, see how simply you can state your position. If you can't state it succinctly, then you don't know your position well enough.
These are things I've learner (learner the wrong way, I might add) in GNU/Linux advocacy over the last 5 years. I've been far more successful in having my voice heard, and changing people's minds, since I learned these things. Never have these lessons been more important than now, with America's apathy towards the "reasoned and wise" actions of our current Congress and administration.
-Paul komarek
Are you sure that putting (copyrighted) video games in public places constitutes public performance of a copyrighted work? For instance, I've never heard anyone suggest that cybercafes need a public performance license for Quake, or that using Matlab in an interactive demo during public talk (e.g. at a conference) needs a special Matlab license.
I'm guessing that using software does not constitute performance, public or otherwise.
-Paul Komarek
I knew someone would come to my rescue and correct my weak Simpson's knowledge!
-Paul Komarek
Even the Simpsons have (indirectly) lampooned Edison's stealing of ideas. In this case, Edison was credited with a fancy leaning chair that Homer made.
-Paul Komarek
You are correct, I meant the Archimedes. At least, I meant Archimedes after I'd read enough to find out I'd made a mistake. ;-) Thanks for the correction.
I finally learned where the name "ARM" comes from:
ARM = Acorn RISC Machine
-Paul Komarek
StrongArm came from DEC, but isn't it a derivative of the Arm? The Xscale is also a derivative of the Arm, IIRC. And the Arm can be traced back to the BBC Micro, which was evidently constructed for a television series about computers.
-Paul Komarek
Does the LinuxBIOS project even show up on your company's radar? That is, is there any concern that x86 BIOS will be (further) commoditized, and hence decrease BIOS profit margins? Where would your company go next if the bottom fell out of the BIOS market?
As an example of the decreased relevance of the BIOS in modern computing, consider how Linux (the kernel) ignores the BIOS wherever possible and goes straight to the hardware, relegating the BIOS to the boot sequence and (usually buggy) power management.
-Paul Komarek
I guess that's why I responded. The Alpha primary market has been research, though Compaq tried to make noise about its server potential. As a researcher that does a lot of computation, I have no code that ignores half of my cpu. Thats why I like the Alpha, and why our lab group is stunned by the way a 667MHz Alpha performs as well as a P-IV at double or nearly triple the speed.
Big caches, big bandwidth, big memory (which is the real reason we're using them), and well-designed for numerical codes. I'm happy to defend the Alpha for the "corner" of the research market it does best -- numerical and statistical code. The only people I've met who argue performance/dollar for the Itanium got their Itanium for free.
-Paul Komarek
My comment about Intel's level of CPU R&D was meant one of two ways:
1) They do lots but have little new to show for it.
2) They don't bother doing much since their slowly evolving x86 cores seem to earn them a whole lot of money.
Intel obviously has some great engineers holding that stupid x86 arch together. That anyone has ever gotten any speed from that ISA is remarkable. But Intel hasn't come up with any new *and* good designs for quite some time. The Itanic has been hyped for many, many years and has only managed to pull even with existing RISC cores (many of which are less expensive and lower power). As you said, "Itanium 2 is now competitive". Finally, Itanic is the only really new-ish cpu thing Intel has done in recent years.
Furthermore, everything I've read about the Itanic over the last five or so years suggests that HP was responsible for the bulk of Itanic's design. I don't have any percentages, as Intel wouldn't release them if they knew what they were anyway. I've read articles suggesting that Intel gagged HP about their role in the design of "Intel's" chip.
Overall, I'm not impressed by the R&D of Intel, a company whose R&D budget is probably larger than MIPS' gross revenue.
-Paul Komarek
CPU R&D is *really* expensive. That's probably why Intel doesn't seem to do much CPU R&D and sticks mainly to manufacturing (note: Itanic arch came mainly from HP, StrongArm and XScale came from other outside sources). It takes very special companies to support this cost, and there aren't many that want to stay in the CPU R&D business anymore (much less get into it).
A large part of the reason GNU/Linux is doing so well compared to other UNIX and UNIX-like systems today is that it runs well on cheap *commodity* hardware. Nobody is going to be impressed if GNU/Linux runs better than WIntel on really expensive equipment. Not to say that people aren't impressed by GNU/Linux running on the IBM S390 for its own sake. =-)
It's not even clear that people *want* better hardware. After all, better hardware has been losing to cheaper hardware for at least 5 years now. It is suprisingly hard to explain to someone why adding an $1800 (or $5400 full-retail) cpu card to a Compaq ES40 is better than buying a new Athlon XP or P-IV since many of the costs aren't easy to account for. But the sysadmins using the scalable ES40 know just how much easier that machine is to care for (esp. remotely) than a uniprocessor $1800 commodity x86 box.
-Paul Komarek
-Paul Komarek
If I were to believe that everything was as simple as you make it, my conclusion would be that I need an Itanium 2 + Pentium IV + PA-RISC or POWER4 to have an overall better machine than an Alpha Server. Is that what you meant? Incidently, the reason the Alpha 21264 has 3 integer units (providing integer scores that dominate Itanic integer scores) is to keep the 2 fp units fed. I'm not an expert on the Itanic architecture, but I'm led to wonder if the Itanic integer units are capable of providing all of the array indexing and loop-counting chores needed for many floating-point numerical anaylysis algorithms. I'd really like to know an answer to this, since the Itanic appears so unbalanced when looking at SPEC scores alone.
Furthermore, you're comparing prices on processor cards for different systems using full retail price. Have you ever bought this kind of equipment? I know I've never paid full retail price when I have. And you don't need to use Compaq's pricing utility -- just get the model number and search with google like you probably did for th Itanic board. Unfortunately, for the cpu board (KN610-EB) you're describing, there aren't many links. Another problem is that the price you quote for the Alpha board includes additional cpu licenses for Tru64 (for the KN610-EB) or VMS (not sure -- KN610-EC?). I don't see anything at the link you provide which states that the price includes a Windows or HP-UX license (and don't kid yourself, you'll need a license).
It's not clear why you mention the 3MB cache on the Itanium. The EV67 boards (KN610-BA, for instance) for the ES40 Model II (which is old) have 8MB cache, and the EV68 board (KN610-EB, for instance) have 16MB of cache (note that these are the big L3 caches for Alphas, not on-chip like I think the Itanium is). Also, it's not clear to me that the processor boards include the same functional components -- do you know that they do (you already seem to have missed the OS license issue)?
You haven't provided the price for the systems, nor stated their default configurations. This is certainly important when making a comparison. You'll also want to compare memory prices, since ram for these servers sometimes has a special form factor and costs a bundle. Again, never use list prices for any of this stuff, as you can occasionally halve the price with a bit of negotiation. That goes for cars and good office chairs, too.
-Paul Komarek
Last I heard, the important Alpha guys ended up at AMD or in small companies. I remember hearing that some or all of the DEC engineers working on StrongArm refused to move to Intel when Intel aquired that cpu, and that Intel had to move some people from the, geez, i810 or something similarly not StrongArm-ish into the StrongArm project.
At some point, I thought the important Alpha folks went to Nexgen, which was bought (directly or indirectly) by AMD.
I'd really like to have a more detailed and more accurate picture of who went where, but it's bedtime for me. Maybe someone else would like to elaborate.
-Paul Komarek
And it's because of this point that I'm glad the Bonneville Power Administration exists, despite former president Ronald Reagan's best efforts.
-Paul Komarek
We bought one of these lite-ons at CompUSA on special about a month ago. They ran out of a slower Buslink drive they had advertised, and sold us a 52x Buslink for $100 instead (it's a rebranded lite-on). It seems to work pretty well so far with the memorex media we've been using, but I can't say we've stressed it much (it's in a test machine).
-Paul Komarek