Current versions of Windows CE also run on PowerPC chips, it's likely that this was used as the starting point for the OS. Adding a few features to WinCE to run a gaming console is probably a much better bet than trying to strip down WinXP to get to the same point.
I've seen hundreds of those Nigerian e-mail scams, and I don't think ANY of them have ever been sent using a Hotmail/Yahoo account. The amount of spam coming out of Hotmail, Yahoo and the other big e-mail services is essentially zero.
Take a look at the headers of one of these messages sometime and compare them to a REAL Hotmail/Yahoo message. The return address will say "@hotmail.com" or "@yahoo.com", but these are dead-easy to change (this isn't really even forging here, I use an @yahoo return address for mail I send through my ISP). The Message-ID line is where you can usually tell the difference, even though those are usually forged (pretty much dead-easy to do, but done extremely poorly by 99% most spammers, this is one of the best headers to filter out spam with).
The US makes up only 47.7% of the 13 countries listed on that list, but they make up 56.7% of the total spam. What's more, that is HARDLY a complete list of countries, or even the 13 most prolific internet users (they don't list Canada, the #2 spam-sending country and where there are undoubtably more internet users than Hong Kong, Switzerland or Sweden).
The US probably makes up 30-35% of the all the worlds internet users, but they are sending 56.7% of the worlds spam, that's a significant difference.
To be fair to the US though, it looks like Canada is actually sending more spam per-capita than the US is, so in a sense it's us hosers up in the Great White North that are the worst offenders (not exactly a stat that fills me with national pride).
Re:FPUs of the future? Re:Floating point performan
on
Mini-ITX Clustering
·
· Score: 1
Pico-BTX is still a bit bigger than Mini-ITX and it only solves the cooling issue when used in a fairly standard (albeit small) desktop case.
Has anyone figured out how to use the floating point power in their graphics cards for non-video applicaitons?
Actually yes, some people have started doing floating point in graphics cards. Check out what the GP GPU people are doing in this regard. The latest and greatest GPUs from nVidia and ATI actually offers a LOT of theoretical processing power, significantly more than the latest P4 or Opteron chips (or IBM PPC 970/Apple G5 chips for that matter). They also offer a whole boatload of memory bandwidht, another very important consideration for FP work.
There are some downsides though. First off GPUs are essentially vector processors. This is not necessarily a bad thing (the fastest supercomputer in the world, Earth Simulator, uses vector processors), but some code does not lend itself well to vector processing and other code needs to be reworked. Perhaps a bigger concern is that GPUs only handle single-precision floating point. This is perfectly acceptable for the purpose they were designed for, but they could cause problems for more general pupose FP work which sometimes requires double precision FP. There's also the issue of non-ECC memory on the cards.
Still, they are interesting from an academic point of view. Of course, on the other hand, CPUs have started incorporating some vector processing aspects of their own. SSE/SSE2 and Altivec are designed as vector processing units as well.
The floating point performance of Nehemiah still stinks, it just stinks somewhat less than the FP performance of previous processors. The built-in cryptography units in quite nifty though, albeit for only rather limited purposes.
As for the P4 and AthlonXP chips in Mini-ITX, the issue is primarily to do with cost, not power/heat. Just compare the size of a typical Mini-ITX system to a laptop, and now go see how many mobile P4 and AthlonXP-M laptops there are out there. Some mobile P4 chips have a TDP of up to 70W and they still manage to stick them in a laptop, so sticking a low-power AthlonXP-M with a TDP of only 25W in a Mini-ITX case would be easy, but it would cost money.
The Athlon64 and Opteron dies contain a built-in thermal cutoff point, though it's a bit of a non-issue. Unless you're some complete nob who likes to rip their heatsink off a processor in the middle of running a Quake 3 timedemo, this is NOT a problem.
As for the PCI bus thing, that's ONLY an issue for overclockers. When run at the specified bus speed both AMD and Intel systems have their PCI clock almost exactly at 33.3MHz.
If you're overclocking your server farm, you have WAY more problems than the brand of processor that you're using!
There was no real competition. (it was the Pentium days) There were other processors, but the Pentium pretty much blew them away.
The Intel Inside marketing program started two years before the Pentium came out. At that time AMD was competing very effectively with the 486. So much so that Intel wanted a new marketing campaign to try to bring people back. Even in the early Pentium days AMD continued to compete effectively. Their 5x86 120MHz chips were very competitive with the Pentium 60 and Pentium 66, and even the 75MHz Pentium chips. It wasn't really until '94 or '95 that Intel really started leaving AMD in the dust, mainly because AMD was WAY late at releasing their K5 processor and when it did come out they had so many problems manufacturing it that it was clocked much lower than initially hoped for. Cyrix continued to offer some competition for Intel during this time, but they were plagued by crappy motherboards which gave them a poor reputation (it was a bit of a self-fulfilling prophecy thing: reputation for being cheap crap meant that they were put on cheap crap motherboards which resulted in a poor quality system).
it will be [better] because it is cheaper
And that is somehow an invalid reason for a product to be better?
The problem was never so much the processors as it was the motherboards. Cyrix chips were great on boards that supported them well. Unfortunately the cheap processors were almost always stuck on dirt-cheap motherboards using a POS power supply and every other crappy component people could find. This was especially bad when Cyrix was trying to push technology forward by using 75MHz bus speeds when Intel was only using 66MHz bus speed.
Crappy motherboards continued to plague AMD for ages, even up to today where VIA still hasn't figured out how to write drivers properly. Fortunately for AMD they now have nVidia laying all of the older chipset vendors to shame by doing a better job on their first shot than Intel/VIA/ALi/SiS managed after 10 years.
As for the clock speed thing, AMD just replaced one totally meaningless number (MHz) with another totally meaningless number (model number). It was a termendously successful idea though, boosted their sales a lot because most people like big meaningless numbers better than small meaningless numbers (compensating for something?).
First off, if they port it to damn near any other architecture out there except for PowerPC, then that other platform WILL have this no-execute bit implemented already. x86 is REALLY late to the game with this technology, it already exists in SPARC, HPPA, IA64, Alpha, etc. etc. It's really just x86 and PowerPC that are the two notable exceptions to the rule; they enforced this sort of thing through segmentation (ie it can and has been implemented in regular x86 code, it's just a bit sloppier and requires some kind of ass-backwards hacks to get it to work with the operating system).
Also this no-execute bit is NOT a sure-fire fix. In fact, it's not a fix at all, buffer overflows can definitely still occur, it's just a whole heck of a lot harder to do anything too malicious once that buffer has been overflowed. Basically you just end up with a DoS attack instead of a remote access vulnerbility. Still something that should be fixed though.
So, is this a good idea? Hell yeah! An extra line of defense is ALWAYS a good thing!
Easier still, go to Spam Gourmet and get temporary accounts for them. It'll do pretty much exactly what you listed but you don't need to have your own domain to do it (yeah, I know we're geeks and should all have our own domain, but some of us are poor at the moment!:> ).
Fine, write back to me in 50-70 years when someone might actually need more than 10^19 bytes of memory (and that's assuming we continue with the current rate of doubling memory usage). If we're still using PCs that in any way, shape or form resemble today's PCs, then maybe I'll grant that a 128-bit desktop CPU might have some reason to exist.
Until that time, stick with 64-bit chips.
Re:AMD have been better than Intel for some time..
on
AMD Back in the Black
·
· Score: 1
Err, I've got one of hte first generation nForce boards, and it's performance is every bit as good as any competitors at the time whiel it offered decent integrated graphics (great for 2D stuff, acceptable for limited 3D gaming at the time) and excellent integrated sound. Add to that the fact that even the version 1.0 drivers worked better than ANY drivers I've ever used for a VIA motherboard and I'm happy.
As for Intel "just working", you obviously never suffered through the hell of early revisions of Intel drivers. There were occasions when they were worse than VIA, the only difference being that at least Intel fixed their problems while VIA didn't. I'm thinking specifically of the first set of drivers for the i810 and i820 motherboards as well as the first drivers for the PIIX4 southbridge (430TX and 440LX chipsets). Those were CRAP! Things definitely did NOT "just work", they just didn't work! If you installed the PIIX4 drivers in the wrong order or with the wrong combination of Windows service packs you had to format the drive to get things working properly again.
Intel's motherboard drivers have been generally ok, but they have are hardly a pinnacle of proper functioning. VIA, on the other hand, has been generally bad and always problematic.
All in all, having used all kinds of different systems over the years, I feel quite confident in saying that nVidia has the best drivers of ANY motherboard chipset maker, including Intel.
Of course, if you REALLY want something that "Just Works", buy a Mac.
Not to get picky but the AthlonXP and Athon are essentially the same core [all the way up through Barton].
And by the same token, the PentiumPro, PII and PIII were all the same core as well.
The only real diffs are the cache size [but even the manual notes the limitations of the core with TLB entries]
Err, do you even know what that means? Because ALL processors will have some limitations caused by the TLB, and all should note them if they're doing a decent job of documenting the chip!
They've always lacked in certain areas. For instance, the cache bus is 64-bits which kills some load times
With it's 128KB of L1 cache the Athlon has always been much less dependant on the L2 cache when compared to the P4. Sure, a 128-bit wide L2 cache bus might help a bit, but it's going to make a big difference most of the time, given the chips design. The latency isn't really bad either. Again, lower latency would help a bit, but not much. Both AMD and Intel need to trade off speed for cost in their designs, but since their chips are VERY different they often end up requiring different trade-offs.
The only way to get a 3IPC then is if previously decoded instructions stall and one of the three schedulers is free.
Yeah, and we hardly ever see x86 instructions stall in a 10 stage pipeline... Ohh wait, that happens damn near all the time!
They still have a six cycle multiply [I dunno if they can make this faster...] for the ALU. Even if they dropped this to 4 it would rock.
I don't know the multiply latency of the Athlon off-hand, but the Athlon64 and Opteron have 3-cycle latency in 32-bit mode and 5-cycle latency in 64-bit mode. Of course, instruction latency usually isn't that big of a deal with an out-of-order processor like the Athlon or Pentium4. It's a damn good thing that it isn't for the P4 though, because it has a HUGE latency on integer multiplies (or at least it did until the new Prescott P4s).
They still don't have good idle support [e.g. Barton 3000+ always runs at 2.1Ghz regardless of what it is doing], etc..
At this time, the Athlon64 is the ONLY desktop processor that I'm aware of that runs at different speeds while idle. Intel's P4 doesn't, IBM's PPC 970 (aka Apple G5) doesn't, Sun's UltraSparcs don't, Alphas don't, Itaniumss don't. Before AMD brought their "Cool 'n Quiet" to the desktop this was an almost unheard of solution outside of laptops. I saw "almost" because AMD actually DID implement this technology with some of their desktop AthlonXP chips, but they were only sold in south-east Asia, and mostly just as a test-market kind of thing.
As I'm sure you know, this technology originated in laptops. In fact, not just any laptops, it originated in AMD laptops! AMD is the one that STARTED this whole thing of having the chips run at different speeds while idle as comapred to when they're running full-out (err, actually it was probably started in some embedded chips, but I'm talking about main-stream consumer-grade CPUs here). The K6-2+ was the first chip to support "PowerNow!" as AMD called it. Intel followed up a bit later with "SpeedStep", and then Transmeta came along later still with their "Longrun" technology. In short, dynamic processor speeds is an area where AMD clearly IS being innovative, first in laptops and now on the desktop.
While I hate the P4 pipeline they have kept a huge cache bus
That is true. As mentioned above, the P4 is more dependant on it's L2 cache than the AthlonXP or Athlon64 are, so it's not suprising that the P4 has a 256-bit wide L2 cache bus running at full core clock speed.
their cpus can scale nicely in terms of heat, etc
Not with the Prescott revision of the P4. At 3.2GHz they're already consuming over 100W of power.
[re: you can remove the heatsink of a P4]
Sure, as long as you don't mind your computer freezing up. You can NOT run a P4 without a
The P4 will shut down as well or it will die, die horribly. Tom's video was faked, go read Intel's thermal design guidelines sometime. Hell, you don't even need to do any research to see it was faked, just watch the video and notice how the processor stays at less than 30C when "throttled", even though thermal throttling doesn't kick in until the processor hits 65-70C.
Even if fully throttled a P4 will still consume at least 20-30W of power, and that is WAY more power than a die with no heatsink can dissipate. Ever try running a PentiumMMX processor without a heatsink? They consume only about 10W of power and would overheat in under a minute without a heatsink.
Socket 939 has really no relation at all to socket 940. AMD did did not just remove 1 pin, they completely redesigned the socket, it's only by sheer co-incidence that they ended up being nearly identical in terms of pin-count (ok, maybe not sheer coincidence, they needed to use a socket that is physically the same size).
The socket has been re-layed out for a different purpose, so yet it WILL be so much easier to design boards that you can do it with almost half the number of layers. On the flip-side though, it won't be possible to use more than one processor with this socket type.
No, the previous poster was right, the issue is virtual memory.
You can easily hack more than 4GB of physical memory onto a 32-bit processor, just increase it's address lines, this much isn't even that ugly of a hack (more on that later) and shouldn't hurt performance much. The problem is that you still only get 2 or 3GB of virtual address space (1 or 2GB reserved for the OS). Even if all that space is in physical memory it's still limited to just that much space.
There are some REALLY ugly hacks that let you use more than 3GB of virtual address space on 32-bit processors though, and THATS where things get messy. This gets back into the same sort of segmentation mess that we had back in the 16-bit DOS days, and it's just UGLY! Even when it doesn't hurt your performance by too much, it's still an unnecessary mess that will drive up development and testing costs.
The PS2 has a 32-bit CPU core with 128-bit vector units. The Pentium3 also uses a 32-bit CPU core with 128-bit vector units (SSE), as does the Apple/Motorola G4 chip (with Altivec). There has never been a 128-bit CPU used in ANY gaming console, and I'm only aware of 1 64-bit CPU ever used (Nintendo64).
Of course, the reason for this is that going to more bits makes a CPU SLOWER! All else being equal, a 64-bit is 5-10% slower than a 32-bit CPU, and a 128-bit CPU is 10-20% slower than a 64-bit one. Since games don't need to address more than 4GB of memory, it's totally pointless to use a 64-bit CPU in a gaming console. The only other thing that a 64-bit CPU buys you is an integer range of more than 4 billion, and that's RARELY used outside of cryptopgraphy (how often do you do cryptography on your gaming console?).
Of course, all else usually isn't equal (eg AMD64 adds 8 more general purpose registers and cleans up some cruft when compared to IA32). Also PCs often do need to address more than 4GB of memory (virtual + physical).
PCs do not, however, need to address more than 10^19 bytes of memory, and they definitely don't need more than 10^19 integer range for much of anything, so 128-bit CPUs get you absolutely NO positives but you still would have to deal with the 10-20% performance loss.
Re:Does AMD have anything to compete with Centrino
on
AMD Back in the Black
·
· Score: 1
Intel just recently introduce the "Celeron-M" processor for laptops. It's a cheaped-down version of the "Pentium-M" processor used as part of the Centrino marketing campaign.
Unfortunately Intel disabled many of the most important dynamic power saving features of the Pentium-M when they released the Celeron-M, so it's not as interesting of a chip as it might be.
Re:Does AMD have anything to compete with Centrino
on
AMD Back in the Black
·
· Score: 1
Transmeta? HIGH-END?!?! What the heck have you been smoking? Their Crusoe chips get absolutely owned by the ULV Celeron processors with similar power consumption and lower prices! They are pure bargain-basement stuff! Even their upcoming "Efficeon" chip is only promising 50-80% better performance, which might make them competitive with the bottom-end of the Celeron line in terms of performance, but not much else.
Good power consumption, sure, but at the cost of much lower performance than anything Intel and AMD are designing. Transmeta just has to look out that the world doesn't discover VIA chips, which offer better performance with the same power consumption and they offer it for dirt-cheap!
Re:Does AMD have anything to compete with Centrino
on
AMD Back in the Black
·
· Score: 1
The AthlonXP-M actually doesn't do too bad compared to the Pentium-M ("Centrino" is purely a marketing campaign to sell a Pentium-M processor combined with an Intel chipset and an Intel WiFi chip).
AMD sells a line of AthlonXP-M processors with a TDP (Thermal Design Power, ie kinda-almost maximum power) of 25W, basically the same as the mainstream Pentium-M chips (Intel also sells some ULV Pentium-M chips with a TDP of only 7W). The performance is pretty good in comparison too, especially if you compare the AthlonXP-M 2000+ to the Pentium-M 1.7GHz (top-end of either line). Perhaps even more importantly though, the AthlonXP-M is MUCH cheaper.
The real downside for AMD is that the Pentium-M has slightly better dynamic power management features, so in normal use it consumes more power on average (maybe 20% more, or about a 5-10% difference in battery life).
In the end, the AthlonXP-M can compete VERY effectively with the new Celeron-M. The two chips will be priced similarly but the AMD will offer much better performance and better power consumption (Intel disables the dynamic power management features of the Celeron-M, so AMD comes out on top here). Not quite the high-end market of the Pentium-M, but then again, not everyone wants to spend $1500-$3000 on a notebook.
One word of warning though, AMD also sells an AthlonXP-M with a TDP of 45W as well as "Desktop Replacement" AthlonXP-M chips, and they don't really make much of an effort to differentiate between the three chips. They will all perform the same at the same clock speed, but they'll have different power consumption figures, so you really have to check just WHAT chip you're actually getting. And just so that Intel doesn't feel left out, they do damn-near the exact same thing with their "Mobile Pentium4-M" and "Mobile Pentium4" processors, which have drastically different power consumption figures.
Why can't Intel just stick the Itanic's 64-bit instruction set onto a P4 like how AMD has stuck x86-64 onto the Athlon?
The AMD64 (aka x86-64) instruction set is a natural continuation of the the old IA32 (aka 32-bit x86) instruction set. Pretty much the entire chip can be used for both, it's just a few tweaks here and there. The Itanium's instruction set (IA64, no connection to IA32 other than both came from Intel) is a TOTALLY different beast. It's not even remotely like x86 and would require a completely separate processor core. Now, that's not to say that Intel couldn't put two separate processor cores on the same chip, but then you get into the matter of cost.
Time to through a few numbers out here... The Opteron has a die size of 193mm^2 at a 130nm manufacturing process. This is considered VERY large, especially since the "Northwood" P4 was only 131mm^2 on a similar process. The Itanium2 has a die size that is 374mm^2, ie it's HUGE! If you were to combine a P4 and an Itanium on the same die, it would be well over 400mm^2. The cost for Intel to make such a chip would be VERY high, significantly more than 3 times the cost of making their P4 processors (a larger chip always results in lower yields in addition to taking up more die space). This would eat into their profits by an enormous amount.
Maybe if nothing changes in die sizes by the time Intel starts shipping chips made on a 65nm manufacturing process (late 2005/early 2006) it might be economically feasible. Of course, things have already changed, the new "Prescott" P4 built on a 90nm fab process uses a LOT more transistors (both logic and cache) than the "Northwood".
Intel's model for profitabilty is simple. They make their profit on Xeons, where until recently they have had no competition.
Xeon may be one of Intel's big cash-cows, but it's not their only one. The Pentium-M processor (part of the "Centrino" marketing package) is another good source of revenue, as are the high-end P4 chips. Intel sells a LOT of processors (10M+ every month), and a lot of their higher-end chips make over $100 in profit. That adds up REAL fast.
A friend once told me that the Celeron is priced, "one penny above variable cost,"
The Celeron doesn't make nearly the profit of the P4 and Xeons, but it does still make money. Variable cost for most processors is down in the $25-$50 range for commodity type chips. The bottom-end Celerons sell for not much more than that, but the top-end ones sell for over $100 (though they are definitely NOT worth that much, current Celerons are dog-slow, but that's another story).
In that respect, Intel is a lot like Microsoft. Microsoft makes so much money on Windows and Office that they can afford to lose it everywhere else
That much is definitely true, just have a look at Intel's balance sheet some time. The ONLY sector of the company that makes money is their PC processor division. They make roughly 75% of the companies revenue and about 150% of the profits.
But the IA-64 market is a hard nut to crack, and for a newcomer there's no money to be made there
Yeah, just ask Intel, they still haven't managed to crack that nut! The Itanium line is not making any money and hasn't really cracked into the high-end server market in a big way (Sun and IBM still own it). I haven't seen the numbers for Q4, but in Q3 of last year Intel managed to sell a grand total of 5000 Itanium servers. For comparison, AMD sold 10,000 Opteron servers while the Xeon found it's way into around a 1,000,000 servers (source here).
I'm not sure that it was entirely a case of Joss WANTING it to end after season 5, just that WB canceled the show! Since the show was being canceled, they tried to make it a good series finale.
Of course, partway through the season UPN announced that they would pick it up the following year, so they had to backtrack slightly on their plans and modify others. A lot of the original story line of season 5 ended up being folded into season 6 and season 7 (eg Willow going bad, that was supposed to happen is season 5, but instead they threw in the character of "Doc").
I agree though, Buffy had a good run, if they had tried to continue it on it probably would have gone downhill a lot. Everything seemed to be moving towards a finale and the season ended well.
I don't think that the same is true for Angel though. After season 4 they REALLY changed the story. They got rid of the annoying bitch and the whiny brat... err, I mean Cordilia and Connor, and the other characters are starting to grow into their new roles. Now they're just going to kill it as they get off the ground? It seems to me like the series could have had at least one more really good year in it. Ending it now just seems a bit premature.
I have a rather sneaking suspicion that there IS a bigger story going on here, it's just much more subtle than in any previous season of Buffy or Angel.
For example, everyone seems to have missed the hint/foreshadowing dropped in the last episode ("Why We Fight") that Lindsey may very well not be dead, not to mention little bit with Gunn. I refuse to read spoilers so I really don't know if any of these will have any importance for the rest of the season. I just get the feeling that by the end of the season we'll look back and see lots of little bits leading up to the finale that we ignored the first time aorund.
Surely they're planning on replacing it with a reality TV show, that's what EVERY network is doing. I guess reality TV must be really cheap to produce or something, because it's rapidly taking over ALL programming out there.
Slashdot Mirror
Current versions of Windows CE also run on PowerPC chips, it's likely that this was used as the starting point for the OS. Adding a few features to WinCE to run a gaming console is probably a much better bet than trying to strip down WinXP to get to the same point.
I've seen hundreds of those Nigerian e-mail scams, and I don't think ANY of them have ever been sent using a Hotmail/Yahoo account. The amount of spam coming out of Hotmail, Yahoo and the other big e-mail services is essentially zero.
Take a look at the headers of one of these messages sometime and compare them to a REAL Hotmail/Yahoo message. The return address will say "@hotmail.com" or "@yahoo.com", but these are dead-easy to change (this isn't really even forging here, I use an @yahoo return address for mail I send through my ISP). The Message-ID line is where you can usually tell the difference, even though those are usually forged (pretty much dead-easy to do, but done extremely poorly by 99% most spammers, this is one of the best headers to filter out spam with).
The US makes up only 47.7% of the 13 countries listed on that list, but they make up 56.7% of the total spam. What's more, that is HARDLY a complete list of countries, or even the 13 most prolific internet users (they don't list Canada, the #2 spam-sending country and where there are undoubtably more internet users than Hong Kong, Switzerland or Sweden).
The US probably makes up 30-35% of the all the worlds internet users, but they are sending 56.7% of the worlds spam, that's a significant difference.
To be fair to the US though, it looks like Canada is actually sending more spam per-capita than the US is, so in a sense it's us hosers up in the Great White North that are the worst offenders (not exactly a stat that fills me with national pride).
Pico-BTX is still a bit bigger than Mini-ITX and it only solves the cooling issue when used in a fairly standard (albeit small) desktop case.
Has anyone figured out how to use the floating point power in their graphics cards for non-video applicaitons?
Actually yes, some people have started doing floating point in graphics cards. Check out what the GP GPU people are doing in this regard. The latest and greatest GPUs from nVidia and ATI actually offers a LOT of theoretical processing power, significantly more than the latest P4 or Opteron chips (or IBM PPC 970/Apple G5 chips for that matter). They also offer a whole boatload of memory bandwidht, another very important consideration for FP work.
There are some downsides though. First off GPUs are essentially vector processors. This is not necessarily a bad thing (the fastest supercomputer in the world, Earth Simulator, uses vector processors), but some code does not lend itself well to vector processing and other code needs to be reworked. Perhaps a bigger concern is that GPUs only handle single-precision floating point. This is perfectly acceptable for the purpose they were designed for, but they could cause problems for more general pupose FP work which sometimes requires double precision FP. There's also the issue of non-ECC memory on the cards.
Still, they are interesting from an academic point of view. Of course, on the other hand, CPUs have started incorporating some vector processing aspects of their own. SSE/SSE2 and Altivec are designed as vector processing units as well.
The floating point performance of Nehemiah still stinks, it just stinks somewhat less than the FP performance of previous processors. The built-in cryptography units in quite nifty though, albeit for only rather limited purposes.
As for the P4 and AthlonXP chips in Mini-ITX, the issue is primarily to do with cost, not power/heat. Just compare the size of a typical Mini-ITX system to a laptop, and now go see how many mobile P4 and AthlonXP-M laptops there are out there. Some mobile P4 chips have a TDP of up to 70W and they still manage to stick them in a laptop, so sticking a low-power AthlonXP-M with a TDP of only 25W in a Mini-ITX case would be easy, but it would cost money.
The Athlon64 and Opteron dies contain a built-in thermal cutoff point, though it's a bit of a non-issue. Unless you're some complete nob who likes to rip their heatsink off a processor in the middle of running a Quake 3 timedemo, this is NOT a problem.
As for the PCI bus thing, that's ONLY an issue for overclockers. When run at the specified bus speed both AMD and Intel systems have their PCI clock almost exactly at 33.3MHz.
If you're overclocking your server farm, you have WAY more problems than the brand of processor that you're using!
Do you remember when the "Intel Inside" logo came out?
1991, according to Intel themselves
There was no real competition. (it was the Pentium days) There were other processors, but the Pentium pretty much blew them away.
The Intel Inside marketing program started two years before the Pentium came out. At that time AMD was competing very effectively with the 486. So much so that Intel wanted a new marketing campaign to try to bring people back. Even in the early Pentium days AMD continued to compete effectively. Their 5x86 120MHz chips were very competitive with the Pentium 60 and Pentium 66, and even the 75MHz Pentium chips. It wasn't really until '94 or '95 that Intel really started leaving AMD in the dust, mainly because AMD was WAY late at releasing their K5 processor and when it did come out they had so many problems manufacturing it that it was clocked much lower than initially hoped for. Cyrix continued to offer some competition for Intel during this time, but they were plagued by crappy motherboards which gave them a poor reputation (it was a bit of a self-fulfilling prophecy thing: reputation for being cheap crap meant that they were put on cheap crap motherboards which resulted in a poor quality system).
it will be [better] because it is cheaper
And that is somehow an invalid reason for a product to be better?
The problem was never so much the processors as it was the motherboards. Cyrix chips were great on boards that supported them well. Unfortunately the cheap processors were almost always stuck on dirt-cheap motherboards using a POS power supply and every other crappy component people could find. This was especially bad when Cyrix was trying to push technology forward by using 75MHz bus speeds when Intel was only using 66MHz bus speed.
Crappy motherboards continued to plague AMD for ages, even up to today where VIA still hasn't figured out how to write drivers properly. Fortunately for AMD they now have nVidia laying all of the older chipset vendors to shame by doing a better job on their first shot than Intel/VIA/ALi/SiS managed after 10 years.
As for the clock speed thing, AMD just replaced one totally meaningless number (MHz) with another totally meaningless number (model number). It was a termendously successful idea though, boosted their sales a lot because most people like big meaningless numbers better than small meaningless numbers (compensating for something?).
First off, if they port it to damn near any other architecture out there except for PowerPC, then that other platform WILL have this no-execute bit implemented already. x86 is REALLY late to the game with this technology, it already exists in SPARC, HPPA, IA64, Alpha, etc. etc. It's really just x86 and PowerPC that are the two notable exceptions to the rule; they enforced this sort of thing through segmentation (ie it can and has been implemented in regular x86 code, it's just a bit sloppier and requires some kind of ass-backwards hacks to get it to work with the operating system).
Also this no-execute bit is NOT a sure-fire fix. In fact, it's not a fix at all, buffer overflows can definitely still occur, it's just a whole heck of a lot harder to do anything too malicious once that buffer has been overflowed. Basically you just end up with a DoS attack instead of a remote access vulnerbility. Still something that should be fixed though.
So, is this a good idea? Hell yeah! An extra line of defense is ALWAYS a good thing!
Easier still, go to Spam Gourmet and get temporary accounts for them. It'll do pretty much exactly what you listed but you don't need to have your own domain to do it (yeah, I know we're geeks and should all have our own domain, but some of us are poor at the moment! :> ).
Fine, write back to me in 50-70 years when someone might actually need more than 10^19 bytes of memory (and that's assuming we continue with the current rate of doubling memory usage). If we're still using PCs that in any way, shape or form resemble today's PCs, then maybe I'll grant that a 128-bit desktop CPU might have some reason to exist.
Until that time, stick with 64-bit chips.
Err, I've got one of hte first generation nForce boards, and it's performance is every bit as good as any competitors at the time whiel it offered decent integrated graphics (great for 2D stuff, acceptable for limited 3D gaming at the time) and excellent integrated sound. Add to that the fact that even the version 1.0 drivers worked better than ANY drivers I've ever used for a VIA motherboard and I'm happy.
As for Intel "just working", you obviously never suffered through the hell of early revisions of Intel drivers. There were occasions when they were worse than VIA, the only difference being that at least Intel fixed their problems while VIA didn't. I'm thinking specifically of the first set of drivers for the i810 and i820 motherboards as well as the first drivers for the PIIX4 southbridge (430TX and 440LX chipsets). Those were CRAP! Things definitely did NOT "just work", they just didn't work! If you installed the PIIX4 drivers in the wrong order or with the wrong combination of Windows service packs you had to format the drive to get things working properly again.
Intel's motherboard drivers have been generally ok, but they have are hardly a pinnacle of proper functioning. VIA, on the other hand, has been generally bad and always problematic.
All in all, having used all kinds of different systems over the years, I feel quite confident in saying that nVidia has the best drivers of ANY motherboard chipset maker, including Intel.
Of course, if you REALLY want something that "Just Works", buy a Mac.
Not to get picky but the AthlonXP and Athon are essentially the same core [all the way up through Barton].
And by the same token, the PentiumPro, PII and PIII were all the same core as well.
The only real diffs are the cache size [but even the manual notes the limitations of the core with TLB entries]
Err, do you even know what that means? Because ALL processors will have some limitations caused by the TLB, and all should note them if they're doing a decent job of documenting the chip!
They've always lacked in certain areas. For instance, the cache bus is 64-bits which kills some load times
With it's 128KB of L1 cache the Athlon has always been much less dependant on the L2 cache when compared to the P4. Sure, a 128-bit wide L2 cache bus might help a bit, but it's going to make a big difference most of the time, given the chips design. The latency isn't really bad either. Again, lower latency would help a bit, but not much. Both AMD and Intel need to trade off speed for cost in their designs, but since their chips are VERY different they often end up requiring different trade-offs.
The only way to get a 3IPC then is if previously decoded instructions stall and one of the three schedulers is free.
Yeah, and we hardly ever see x86 instructions stall in a 10 stage pipeline... Ohh wait, that happens damn near all the time!
They still have a six cycle multiply [I dunno if they can make this faster...] for the ALU. Even if they dropped this to 4 it would rock.
I don't know the multiply latency of the Athlon off-hand, but the Athlon64 and Opteron have 3-cycle latency in 32-bit mode and 5-cycle latency in 64-bit mode. Of course, instruction latency usually isn't that big of a deal with an out-of-order processor like the Athlon or Pentium4. It's a damn good thing that it isn't for the P4 though, because it has a HUGE latency on integer multiplies (or at least it did until the new Prescott P4s).
They still don't have good idle support [e.g. Barton 3000+ always runs at 2.1Ghz regardless of what it is doing], etc..
At this time, the Athlon64 is the ONLY desktop processor that I'm aware of that runs at different speeds while idle. Intel's P4 doesn't, IBM's PPC 970 (aka Apple G5) doesn't, Sun's UltraSparcs don't, Alphas don't, Itaniumss don't. Before AMD brought their "Cool 'n Quiet" to the desktop this was an almost unheard of solution outside of laptops. I saw "almost" because AMD actually DID implement this technology with some of their desktop AthlonXP chips, but they were only sold in south-east Asia, and mostly just as a test-market kind of thing.
As I'm sure you know, this technology originated in laptops. In fact, not just any laptops, it originated in AMD laptops! AMD is the one that STARTED this whole thing of having the chips run at different speeds while idle as comapred to when they're running full-out (err, actually it was probably started in some embedded chips, but I'm talking about main-stream consumer-grade CPUs here). The K6-2+ was the first chip to support "PowerNow!" as AMD called it. Intel followed up a bit later with "SpeedStep", and then Transmeta came along later still with their "Longrun" technology. In short, dynamic processor speeds is an area where AMD clearly IS being innovative, first in laptops and now on the desktop.
While I hate the P4 pipeline they have kept a huge cache bus
That is true. As mentioned above, the P4 is more dependant on it's L2 cache than the AthlonXP or Athlon64 are, so it's not suprising that the P4 has a 256-bit wide L2 cache bus running at full core clock speed.
their cpus can scale nicely in terms of heat, etc
Not with the Prescott revision of the P4. At 3.2GHz they're already consuming over 100W of power.
[re: you can remove the heatsink of a P4]
Sure, as long as you don't mind your computer freezing up. You can NOT run a P4 without a
The P4 will shut down as well or it will die, die horribly. Tom's video was faked, go read Intel's thermal design guidelines sometime. Hell, you don't even need to do any research to see it was faked, just watch the video and notice how the processor stays at less than 30C when "throttled", even though thermal throttling doesn't kick in until the processor hits 65-70C.
Even if fully throttled a P4 will still consume at least 20-30W of power, and that is WAY more power than a die with no heatsink can dissipate. Ever try running a PentiumMMX processor without a heatsink? They consume only about 10W of power and would overheat in under a minute without a heatsink.
Socket 939 has really no relation at all to socket 940. AMD did did not just remove 1 pin, they completely redesigned the socket, it's only by sheer co-incidence that they ended up being nearly identical in terms of pin-count (ok, maybe not sheer coincidence, they needed to use a socket that is physically the same size).
The socket has been re-layed out for a different purpose, so yet it WILL be so much easier to design boards that you can do it with almost half the number of layers. On the flip-side though, it won't be possible to use more than one processor with this socket type.
No, the previous poster was right, the issue is virtual memory.
You can easily hack more than 4GB of physical memory onto a 32-bit processor, just increase it's address lines, this much isn't even that ugly of a hack (more on that later) and shouldn't hurt performance much. The problem is that you still only get 2 or 3GB of virtual address space (1 or 2GB reserved for the OS). Even if all that space is in physical memory it's still limited to just that much space.
There are some REALLY ugly hacks that let you use more than 3GB of virtual address space on 32-bit processors though, and THATS where things get messy. This gets back into the same sort of segmentation mess that we had back in the 16-bit DOS days, and it's just UGLY! Even when it doesn't hurt your performance by too much, it's still an unnecessary mess that will drive up development and testing costs.
The Playstation2 does NOT use a 128-bit CPU!!!
The PS2 has a 32-bit CPU core with 128-bit vector units. The Pentium3 also uses a 32-bit CPU core with 128-bit vector units (SSE), as does the Apple/Motorola G4 chip (with Altivec). There has never been a 128-bit CPU used in ANY gaming console, and I'm only aware of 1 64-bit CPU ever used (Nintendo64).
Of course, the reason for this is that going to more bits makes a CPU SLOWER! All else being equal, a 64-bit is 5-10% slower than a 32-bit CPU, and a 128-bit CPU is 10-20% slower than a 64-bit one. Since games don't need to address more than 4GB of memory, it's totally pointless to use a 64-bit CPU in a gaming console. The only other thing that a 64-bit CPU buys you is an integer range of more than 4 billion, and that's RARELY used outside of cryptopgraphy (how often do you do cryptography on your gaming console?).
Of course, all else usually isn't equal (eg AMD64 adds 8 more general purpose registers and cleans up some cruft when compared to IA32). Also PCs often do need to address more than 4GB of memory (virtual + physical).
PCs do not, however, need to address more than 10^19 bytes of memory, and they definitely don't need more than 10^19 integer range for much of anything, so 128-bit CPUs get you absolutely NO positives but you still would have to deal with the 10-20% performance loss.
Intel just recently introduce the "Celeron-M" processor for laptops. It's a cheaped-down version of the "Pentium-M" processor used as part of the Centrino marketing campaign.
Unfortunately Intel disabled many of the most important dynamic power saving features of the Pentium-M when they released the Celeron-M, so it's not as interesting of a chip as it might be.
Transmeta? HIGH-END?!?! What the heck have you been smoking? Their Crusoe chips get absolutely owned by the ULV Celeron processors with similar power consumption and lower prices! They are pure bargain-basement stuff! Even their upcoming "Efficeon" chip is only promising 50-80% better performance, which might make them competitive with the bottom-end of the Celeron line in terms of performance, but not much else.
Good power consumption, sure, but at the cost of much lower performance than anything Intel and AMD are designing. Transmeta just has to look out that the world doesn't discover VIA chips, which offer better performance with the same power consumption and they offer it for dirt-cheap!
The AthlonXP-M actually doesn't do too bad compared to the Pentium-M ("Centrino" is purely a marketing campaign to sell a Pentium-M processor combined with an Intel chipset and an Intel WiFi chip).
AMD sells a line of AthlonXP-M processors with a TDP (Thermal Design Power, ie kinda-almost maximum power) of 25W, basically the same as the mainstream Pentium-M chips (Intel also sells some ULV Pentium-M chips with a TDP of only 7W). The performance is pretty good in comparison too, especially if you compare the AthlonXP-M 2000+ to the Pentium-M 1.7GHz (top-end of either line). Perhaps even more importantly though, the AthlonXP-M is MUCH cheaper.
The real downside for AMD is that the Pentium-M has slightly better dynamic power management features, so in normal use it consumes more power on average (maybe 20% more, or about a 5-10% difference in battery life).
In the end, the AthlonXP-M can compete VERY effectively with the new Celeron-M. The two chips will be priced similarly but the AMD will offer much better performance and better power consumption (Intel disables the dynamic power management features of the Celeron-M, so AMD comes out on top here). Not quite the high-end market of the Pentium-M, but then again, not everyone wants to spend $1500-$3000 on a notebook.
One word of warning though, AMD also sells an AthlonXP-M with a TDP of 45W as well as "Desktop Replacement" AthlonXP-M chips, and they don't really make much of an effort to differentiate between the three chips. They will all perform the same at the same clock speed, but they'll have different power consumption figures, so you really have to check just WHAT chip you're actually getting. And just so that Intel doesn't feel left out, they do damn-near the exact same thing with their "Mobile Pentium4-M" and "Mobile Pentium4" processors, which have drastically different power consumption figures.
Why can't Intel just stick the Itanic's 64-bit instruction set onto a P4 like how AMD has stuck x86-64 onto the Athlon?
The AMD64 (aka x86-64) instruction set is a natural continuation of the the old IA32 (aka 32-bit x86) instruction set. Pretty much the entire chip can be used for both, it's just a few tweaks here and there. The Itanium's instruction set (IA64, no connection to IA32 other than both came from Intel) is a TOTALLY different beast. It's not even remotely like x86 and would require a completely separate processor core. Now, that's not to say that Intel couldn't put two separate processor cores on the same chip, but then you get into the matter of cost.
Time to through a few numbers out here... The Opteron has a die size of 193mm^2 at a 130nm manufacturing process. This is considered VERY large, especially since the "Northwood" P4 was only 131mm^2 on a similar process. The Itanium2 has a die size that is 374mm^2, ie it's HUGE! If you were to combine a P4 and an Itanium on the same die, it would be well over 400mm^2. The cost for Intel to make such a chip would be VERY high, significantly more than 3 times the cost of making their P4 processors (a larger chip always results in lower yields in addition to taking up more die space). This would eat into their profits by an enormous amount.
Maybe if nothing changes in die sizes by the time Intel starts shipping chips made on a 65nm manufacturing process (late 2005/early 2006) it might be economically feasible. Of course, things have already changed, the new "Prescott" P4 built on a 90nm fab process uses a LOT more transistors (both logic and cache) than the "Northwood".
Intel's model for profitabilty is simple. They make their profit on Xeons, where until recently they have had no competition.
Xeon may be one of Intel's big cash-cows, but it's not their only one. The Pentium-M processor (part of the "Centrino" marketing package) is another good source of revenue, as are the high-end P4 chips. Intel sells a LOT of processors (10M+ every month), and a lot of their higher-end chips make over $100 in profit. That adds up REAL fast.
A friend once told me that the Celeron is priced, "one penny above variable cost,"
The Celeron doesn't make nearly the profit of the P4 and Xeons, but it does still make money. Variable cost for most processors is down in the $25-$50 range for commodity type chips. The bottom-end Celerons sell for not much more than that, but the top-end ones sell for over $100 (though they are definitely NOT worth that much, current Celerons are dog-slow, but that's another story).
In that respect, Intel is a lot like Microsoft. Microsoft makes so much money on Windows and Office that they can afford to lose it everywhere else
That much is definitely true, just have a look at Intel's balance sheet some time. The ONLY sector of the company that makes money is their PC processor division. They make roughly 75% of the companies revenue and about 150% of the profits.
But the IA-64 market is a hard nut to crack, and for a newcomer there's no money to be made there
Yeah, just ask Intel, they still haven't managed to crack that nut! The Itanium line is not making any money and hasn't really cracked into the high-end server market in a big way (Sun and IBM still own it). I haven't seen the numbers for Q4, but in Q3 of last year Intel managed to sell a grand total of 5000 Itanium servers. For comparison, AMD sold 10,000 Opteron servers while the Xeon found it's way into around a 1,000,000 servers (source here).
I'm not sure that it was entirely a case of Joss WANTING it to end after season 5, just that WB canceled the show! Since the show was being canceled, they tried to make it a good series finale.
Of course, partway through the season UPN announced that they would pick it up the following year, so they had to backtrack slightly on their plans and modify others. A lot of the original story line of season 5 ended up being folded into season 6 and season 7 (eg Willow going bad, that was supposed to happen is season 5, but instead they threw in the character of "Doc").
I agree though, Buffy had a good run, if they had tried to continue it on it probably would have gone downhill a lot. Everything seemed to be moving towards a finale and the season ended well.
I don't think that the same is true for Angel though. After season 4 they REALLY changed the story. They got rid of the annoying bitch and the whiny brat... err, I mean Cordilia and Connor, and the other characters are starting to grow into their new roles. Now they're just going to kill it as they get off the ground? It seems to me like the series could have had at least one more really good year in it. Ending it now just seems a bit premature.
I have a rather sneaking suspicion that there IS a bigger story going on here, it's just much more subtle than in any previous season of Buffy or Angel.
For example, everyone seems to have missed the hint/foreshadowing dropped in the last episode ("Why We Fight") that Lindsey may very well not be dead, not to mention little bit with Gunn. I refuse to read spoilers so I really don't know if any of these will have any importance for the rest of the season. I just get the feeling that by the end of the season we'll look back and see lots of little bits leading up to the finale that we ignored the first time aorund.
Surely they're planning on replacing it with a reality TV show, that's what EVERY network is doing. I guess reality TV must be really cheap to produce or something, because it's rapidly taking over ALL programming out there.