Uhh, yeah, because Creative Labs in the pinnical of driver quality? Umm.. right!
I haven't used an Audigy, so I can't comment on that line, but every Creative card I have used, from the old SB Pro up to an SBLive! had TERRIBLE drivers! Creative is one of the WORST companies out there for driver quality! They do NOT play well with any other components in the system. The SBLive blatently violated the PCI spec, and despite this being a known problem, Creative still hadn't fixed their drivers a year later when I finally dumped the card!
Yeah, it's terrible how nVidia MADE 3Dfx screw up their entire distribution channels, how they made them buy out STB and try to become a card manufacturer. Absolute horrible how they made 3Dfx deliver all their products a year late (or more) and missing much-needed features. And it was especially bad how nVidia made 3Dfx release crappy drivers (or no drivers) for so long.
Face it, 3Dfx killed themselves, nVidia just moved in to pick up the slack. Even if the whole lawsuit between the companies had gone in 3Dfx's favor it's unlikely that they would have managed to survive long enough to see the results of it. A combination of bad decisions and products that were a day late and a dollar short (but still expensive) killed them, not nVidia.
Try to find a faster RISC chip with lower power consumption than an Athlon64 or P4? I'll give you a hint, you aren't likely to find one. DEC's old Alpha chips were faster than anything around in their day, but they also consumed more power (over 100W in some Alphas, and that was a few years back). IBM's Power4 is also up in the 100W+ range. Intel's VLIW Itanium consumes up to 107W.
Heck, even the IBM PowerPC 970, that lots of people point to as being a cool-running chip in their Powermac G5s is in the same ballpark as AMD's Athlon64/Opteron. It's listed as 48W typical power consumption at 1.8GHz. Add about 10% to get the typical power consumption at 2.0GHz and another 20-40% to get maximum power consumption and it's right up there with the 60-80W maximum power of the P4 or the Athlon64.
Honestly there really isn't a difference betwen RISC and CISC these days. CISC chips look like RISC chips on the inside and the instruction set of RISC chips has mostly become more complicated. All of the chips read instructions from memory and execute different operations on the inside, just like how it's done on x86 chips. The biggest remaining difference is that x86 chips have only 8 visible general purpose registers (16 GPRs in the case of AMD64 chips) while most RISC chips have 32 GPRs and VLIW chips like Itanium have a whole lot. Of course, once you factor in rename registers, even that difference isn't all that huge.
FWIW AMD lists the maximum power consumption of their Desktop Replacement Athlon64 at 81W, and the Mobile Athlon64 as 62W. Both are pretty high, though the Athlon64/Opteron core seems to have a pretty wide power range (ie typical power is quite a bit lower than the maximum power listed). I'm not sure whether this machine uses the Desktop Replacement Athlon64 or the Mobile Athlon64, but I believe that it's the latter (AMD only lists a 3200+ speed grade for the DTR chip, and the laptop comes with a 3000+ speed grade).
Certainly not the sort of chip that you can stuff into a super-small laptop, but about the same as it's main competitor. Intel's Mobile P4 processors have a Thermal Design Power (almost-sorta-kinda maximum power) of up to 70W.
In both cases these machines are really targetting the desktop repalcement crowd. They're designed for people who need to move their computer from one place to another every once in a while but rarely need to use them on the road.
Some of what you say has some validity, but it ignores the fact that sound is digitized at a much higher sampling rate and higher bit rate than a CD. Then you just throw some digital filters at it to get it down to the CD specifications.
The result is that you actually DO get an exact reproduction of the analog signal frequency (except maybe for the first and last half-second of the record) and very nearly an exact replication of the amplitude of the signal (at the low-end of the volume scale it's not quite as exact as it should be due to the use of a linear codec for a very logrithmic problem).
Think of this as being like scanning the image at 32-bit colour and very high resolution, but then cropping it down to 16-bit colour and a lower resolution. The resulting image almost always looks much better than if you scanned it initially at that same 16-bit colour and low resolution.
The "2800+" IS the speed. No, it's not the clock speed that the chip runs at, but that really doesn't make it any more or less relevant to the speed.
Intel decided to market their processors using one totally meaningless measure of performance, the clock speed.
AMD decided to market their processors using a different totally meaningless measure of performance, the model numbers.
The problem with the chips is not so much sleezy marketing as with are ridiculous focus on clock speed. That's like buying a car solely based on the displacement of the engine. Sure, a 3.0L engine might produce more power than a 2.5L engine, but that certainly isn't always the case, and it sure doesn't tell you much else about the car. But if the public bought cars based ONLY on the displacement of the engine like they buy PC's based ONLY on the clock speed of the processor, you better believe that companies would bring out "3.0+" engines with 2.5L of displacement.
In any case, AMD's marketing numbers have been extremely successful for the company. They almost instantly increased AMD's profit when they first brought them out years ago. Argue all you like, but the fact is that the computer buying public is, by and large, generally uniformed and buys systems based on two numbers alone, clock speed/model number of the processor and the price.
That's the official story, and AMD is sticking to it!
Of course, in reality the model numbers are in direct comparison to the P4. AMD might not say that in public, but that has always been the intention.
Is this a good thing or a bad thing? I don't see that it really matters. The replaced one TOTALLY MEANINGLESS number (clock speed) with another totally meaningless number (model number).
Personally I like the Opteron's numbering system best.
Socket 7?!?! Good &deity no! Talk about a TERRIBLE design!
If we kept socket 7 we would: a.) still be stuck at ~1GHz processors because the socket did not provide enough power or grounding pins for todays faster processors. b.) would have TERRIBLE memory throughput, the real-world performance of this socket was terrible even if when the theoretical numbers were ok.
Perhaps most importantly though, it wouldn't help anything. You would STILL need to buy new motherboards to support new chips. In fact, it would probably be a LOT worse because you wouldn't know just what processor your particular version of socket 7 board supported! So instead of having 4 sockets to worry about you would have THOUSANDS of different motherboards, all with a lists of dozens of processors that they do and do not support.
There's MUCH more to making a chip compatible with a motherboard than just the socket. In fact, the physical socket is a rather trivial part of it.
IBM seems ot be well ahead of AMD in moving towards a 90nm manufacturing process. The new Xserve G5's have already been announced with 90nm PPC970 chips and they should ship in 6-8 weeks, making it a close second to Intel's 90nm chips (which should actually be available in two to three weeks time). AMD is going to be a little bit late to the show (not surprising, they don't have the financial resources of Intel or IBM, so they're always ~6 months behind in bringing out a new process generation).
Of course, I wouldn't hold my breath for 3GHz G5's by Q3. 2.5GHz PPC 970 processors by mid-summer sure, no problem. But 3.0GHz? It's possible, but certainly not a sure thing. I fully expect the clock speeds of these two chips to remain pretty comperable for some time to come.
The marketing of the Athlon64 FX has become a bit confusing. It kind of made sense for the initial launch to combat a precieved weakness of the design compared to Intel's P4 though. With the P4 you get up to 6.4GB/s of memory bandwidth, while the first Athlon64 would only have 3.2GB/s of memory bandwidth. Now, it turns out that the extra bandwidth doesn't actually buy you much on most applications, but this was seen as a weakness, hence the Athlon64 FX. With Intel bringing out the P4EE to compete with the FX, now AMD might need to keep the chip, even if it isn't a worthwhile product (The P4EE isn't a worthwhile product either).
*Is it really a good idea to have the memory controller on the CPU?
Yes, yes it is a good idea. A VERY good idea in fact. Memory latency has only improved by about one order of magnitude in the past 15 years. Meanwhile everything else in the system has gone up by at least two orders of magnitude. Virtually everything that is being done in CPU design these days is to hide memory latency (larger caches, out-of-order executation, branch prediction, even SMT).
Integrating the memory controller reduces latency by 20-30%. At 2.0GHz this makes a BIG difference (this is the main reason why a 2.0GHz Athlon64 is faster than a 2.2GHz AthlonXP), at 4 or 5GHz the difference will be huge.
but it hasn't lowered mainboard costs
You can buy new Athlon64 motherboards for only $100, only 3 months after the chips release. It took ages for Athlon or P4 motherboards to reach that price point. What's perhaps even more impressive is the dual-processor boards that are only $200. In short, it HAS reduced motherboard costs, whether you've noticed or not. It also means that ALL Athlon64's support ECC, chipkill and a few other nifty reliability features, regardless of how badly VIA screws up their chipset design.
*Why in the world introduce an AthlonFX based on Socket 940, especially at the outrageous price, when you're moving to socket 939 imminently?
The Athlon64 FX was a bit of a last minute decision I believe. They found a marketing weakness and wanted the quickest and easiest solution they could find. The answer? Sell your server chip as an "enthusiast" chip. Intel did exactly the same thing for the same reason with the P4EE.
Also, it's actually VERY normal to switch sockets soon after releasing a new processor. Intel's upcoming Prescott will use Socket 478 for only about 6 months before switching to socket 775. The original P4 used Socket 423 for a very short time before switching to socket 478. The original Athlon used Slot A for a year or so before switching to Socket A. The PIII came out in Slot 1 form but then switched to Socket 370 about a year later. The Celeron followed the same path a couple years before.
I think it would have been more of a slam-dunk as a platform and a "brand" to release Athlon64 as all dual-channel, all Socket 939 (or some standard), and left Opteron as the high-end platform. Any other takers?
In retrospec that might seem like a good idea, hindsight is 20-20 after all. However the original split of ALL Athlon64 chips being socket 754 and ALL Opteron's being socket 940 seemed to make the most sense when AMD was desigining them. It wasn't until market conditions changed and a new perceived weakness was discovered that AMD felt they need a consumer chip with a 128-bit wide memory bus. By that time the chip was already late to market and designing a new socket would have added more delay to the equation.
There's also the question of budget chips. AMD hopes to move their entire product line to the Athlon64/Opteron platform by the end of 2004. That means they need a budget chip, and socket 939 with it's 128-bit wide memory bus is problematic for that. Hence the continued existance of Socket 754 and the AthlonXP for that platform.
But the PowerPC line has superior signal processing capabilities these days
According to who? And why? SSE2 provides plenty of instructions for signal processing, and SSE3/PNI will fill in a couple of the last remaining holes. AMD64 also doubles the number of registers for SSE2 as well as the general purpose registers.
Exactly what advantage does the PowerPC have for signal processing? They do have a nifty multiply-add instruction that is missing on x86/SSE2, but on the flip side their vector processing engine (Altivec) seems to be limited to single precision floating point math unless I missed something.
is much easier for compilers to cater to
Ohh, so THAT'S why there are WAY more x86 compilers out there than PPC ones.
And runs cooler and more efficient than the X86 offerings
The PPC 970, used in Apple's Powermacs is listed at a "typical" power of ~48W at 1.8GHz. Maximum power is probably a good 20-40% higher. In other words, power consumption is pretty much identical to AMD's Athlon64/Opteron line and pretty darn close to Intel's P4 line. It also has 52 million transistors, nearly the same as the 55 million transistors of the P4. The IBM Power4+ consumes a ton of power, somewhere up around 100W.
AFAIK, x86 units since the i686 have all used a RISC-like core that runs x86 ops by breaking them down into micro-ops and reconstituting them. It -works- but whay do that when the real thing is available?
Why do it? Becuase hardware is cheap, software is expensive. They also do it because it doesn't really make a difference, ALL chips start by decoding their instructions, executing them and then putting it all back together again. It's not just an x86 thing.
The main downsides to x86 are as follows:
- Too few general purpose registers and restrictions on what registers can be used for which instructions
- 32-bit pointers
- Stack based FPU
- Somewhat more complex instruction decoder
AMD has addressed the first two points with AMD64. They've extended the general purpose and FPU (SSE) registers from 8 to 16 and removed the restrictions on what registers can be used for which instruction. Not quite the 32 registers of some other architectures, but still pretty good.
The stack-based FPU is in the process of being replaced by SSE2. While SSE may have started out just as a vector SIMD engine, with SSE2 it's really a full fledged floating point unit to replace the old x87 unit.
The more complex decoder has also been largely addressed by Intel in their P4 design with their trace cache. This stores already decoded instructions, so about 90% of the time the decoder is not needed (and therefore can be made much simpler).
And the Pentium Pro core lasted them until 1.4GHz, but that didn't stop Intel from marketing it under the PII and later PIII brand names, not to mention the Xeons and Celerons.
It's likely that the Netburst core of the P4 will be here for a while yet. I don't know about the 11GHz number, but I suspect that it Intel will continue to use the core until at least 2006 or 2007. However I also suspect that we'll see at least a "Pentium 5" in that time frame, and a "Pentium 6" is not out of the question. It's also likely (in fact, almost certain) that Intel will put two Netburst cores on a single chip in 2005 or 2006, at least for their Xeon line, and that may extend the basic Netburst architecture for a while longer.
Records often sound good, but it's not because the quality of the medium is good. In fact, it's just the opposite. Many people like the distortion added by vynal records, and most older music was written with the intention that such distortion would be added in. When you listen to the music on a CD with no distortion, it just doesn't sound as good because the original analog was mixed with the intention of getting some distortion on playback.
Sure, encoding to digital will lose you some data, and CDs aren't ideal. The use of a linear codec instead of a logrithmic one seems to me to be it's biggest mistake (this causes some problems at low volumes, which isn't exactly a strong-point for records either).
Most people who think vinyl is a better medium than CDs are under the mistaken belief that a) CD's can not reproduce an analog signal of an exact frequency (they can, up to their 1/2 the sampling frequency of 44.1KHz, ie 0 to 22.05KHz), or b) that the best records and sound system CAN exactly reproduce the amplitude of the sound wave (the signal to noise ratio of even the very best records and sound systems is not as large as the dynamic range of a CD).
Of course, as mentioned above, a more exact reproduction of the original does not always equate a better "sounding" copy.
Open relays are defintiely yesterday's technology, they are rarely used by spammers anymore (mainly because there aren't nearly as many of them anymore and there are so many easier methods for spammers to use).
These days it's all done through open proxies or special spamming applications. These are setup using worms and proxies (SoBig, MiMail, etc.) and run on the systems of your average everyday joe-user with broadband connections. Sure, you could fine all these users (though you could only do so in a country by country basis), but do you really want to fine someone's grandmother just because she didn't install the latest patch to her Windows box (or her Mac, or whatever other OS she happens to be running that was compromised)? This would certainly not be a popular law with anyone except the most ovezealous geeks.
Err, not if you're me apparently. I just went through the Apple Online store and brought up the configuration of a bog-standard dual-processor G5. Estimated shipping time, 6-8 weeks. Tried again for the single-processor system with the same result, 6-8 weeks.
Err, have you bothered to check the datasheets? The P4 is hardly a cool running chip by any stretch. It may have a wider power consumption range than the AthlonXP, but when the chips were running flat out they sure didn't run any cooler.
The only reason why P4's used to run cooler than Athlons was because people would stick a 60mm x 60mm heatsink on their Athlon and an 80mm x 80mm heatsink on their P4. Both of these chips consume a lot of power, and both drain laptop batteries like crazy if you use the highest powered parts (Intel actually produces some P4 "mobile" chips with a TDP of 70W!, while AMD's brand new "mobile" Athlon64 chips consume over 60W at full throttle).
Now, the Pentium M... well that's another story altogether.
Err, have you ever installed a P4 or an Athlon64/Opteron processor? They DO use locking mechanisms to sandwich the processor between the motherboard and the heatsink!
The G5 at 90nm may have been announced by Apple, but it's not yet shipping from what I can tell. Given traditional Apple schedules, it will actually start shipping in one to two months time, or roughly the same time frame as Intel starts shipping their first 90nm processor (Prescott, not Tejas, that's a future chip not scheduled for 9+ months).
The XServe G5 is using new IBM PPC 970 processors made on a 90nm fab (as comapred to the 130nm of the desktop PowerMac G5). This should reduce the power consumption somewhat and bring it into the right range to cram it into a laptop. It'll probably still consume a reasonable amount of power at 2.0GHz and the standard voltage, but Apple could probably order up some 1.6 or maybe even 1.8GHz chips running at ~1.0V that would drop the power consumption down into the same basic range as the G4. Certainly getting power consumption down to about 30W TDP for a 1.6GHz part should be no trouble, and that's rather easy enough to cool (AMD and Intel both have chips that consume up to about 70W of power that they sell as "Desktop Replacement" mobile processors, and even the regular Pentium 4 M consumes 30-35W of power).
I suspect that we will see Powerbook G5's in the not-too-distant future, probably even sooner than this time next year. Maybe mid-summer? When's the next big Apple show?
Yes the XServe G5 and Opteron cost about the same, have similar features and for many applications will have similar performance. However, now people have a nice extra choice. For some situations, the G5 is a LOT faster than the Opteron. For others, the Opteron is a LOT faster than the G5. They are rather different architectures with different instruction sets.
There are a few nice points about the XServeG5. First off they now support ECC memory, a glaring omission in the standard G5 machines that made them totally unsuitable for servers of large clusters. Second, they are now using chips manufactured on a 90nm fab process. This should hopefully result in lower power consumption and heat dissipation. The 130nm G5 and 130nm Opteron had pretty similar power consumption numbers, and when the 90nm Opterons come out in mid-2004 they will likely have similar power consumption figures again, but for the time being the G5 has a slight edge here.
Pentium-M 1.7GHz has a TDP of 24.5W The Ultra Low Voltage Pentium-M at 900MHz has a TDP of 7.0W. The Transmeta Crusoe's power consumption of their lowest powered 1GHz Crusoe 5800 chip is 6.5W (or 8.5W for their cheapest 1GHz Crusoe 5800). These numbers are not defined 100% the same, so Transmeta might actually have a slighly larger advantage in real-world situations, but on the other hand it's a MUCH slower processor, so the 900MHz ULV Pentium-M will spend most of it's time in a lower power state as compared to the Crusoe. Neither chip consumes much of anything when idle.
In the end, the Crusoe doesn't end up looking very competitive except if the price is REALLY cheap. A few watts here or there for the ULV Pentium M are, for most people, MORE than made up for by the fact that it's roughly twice as fast.
Really a more accurate comparision is not with the Pentium M, but with the ULV Celeron processors, which also consume 7.0W at 800MHz. These chips are still a hell of a lot faster than the Crusoe (though maybe only 40 or 50% faster than a 1GHz Crusoe) while being in the same price range and same power consumption range.
Transmeta might sell a handful of laptops, webpads and thin clients, but probably less than a million a year in total.
ARM, on the other hand, vastly outsells all x86 chips combined (admittedly many of these are not for mobile devices, merely embedded systems). We're talking something on the order of 200-300 million chips (maybe more, last numbers I heard were from a few years back). Even if you only counted all of the PocketPC and Palm systems being sold with ARM chips you're probably looking at more ARM chips being sold every month than the total number of Crusoe chips Transmeta has ever sold.
According to Sandpile.org, the 90MHz Pentium consumes a grand total of 9.0W of power when going full-bore. Transmeta's claiming power consumption numbers of about 6 or 7W for their chips. The Transmeta chips will have a slight advantage in terms of dynamic power management (almost non-existant on the old Pentium classic), but really we're talking about a few watts here. Throughout the course of a year you're looking at a difference of about $2 to $3 in electrical costs. At that rate, the Transmeta chips have got to be REAL cheap to beat out spending 5 or 10 bucks on a Pentium system.
MiniITX'ers already have the VIA C3 chips that offer similar power consumption numbers and equal or better performance and a very low price tag.
Of course, this announcement by Transmeta is in direct response to VIA's announcement of physically smaller chips a few months back. VIA's Antaur processor is the exact same idea, take a small, low-cost and low powered x86 core and package it up real small. VIA's chip is 35mm x 35mm for the entire package. Transmeta comes in a slightly smaller 21mm x 21mm package.
In the end, the Transmeta has a slight power consumption advantage (clock for clock) and a slight size advantage, but really neither are very significant, and VIA may have a performance advantage (or, conversly, they can negate the power consumption advantage by using a lower-clocked processor) and traditionally has had quite a large price advantage. Both are fighting for a pretty small and low-profit market though.
Slashdot Mirror
Uhh, yeah, because Creative Labs in the pinnical of driver quality? Umm.. right!
I haven't used an Audigy, so I can't comment on that line, but every Creative card I have used, from the old SB Pro up to an SBLive! had TERRIBLE drivers! Creative is one of the WORST companies out there for driver quality! They do NOT play well with any other components in the system. The SBLive blatently violated the PCI spec, and despite this being a known problem, Creative still hadn't fixed their drivers a year later when I finally dumped the card!
Yeah, it's terrible how nVidia MADE 3Dfx screw up their entire distribution channels, how they made them buy out STB and try to become a card manufacturer. Absolute horrible how they made 3Dfx deliver all their products a year late (or more) and missing much-needed features. And it was especially bad how nVidia made 3Dfx release crappy drivers (or no drivers) for so long.
Face it, 3Dfx killed themselves, nVidia just moved in to pick up the slack. Even if the whole lawsuit between the companies had gone in 3Dfx's favor it's unlikely that they would have managed to survive long enough to see the results of it. A combination of bad decisions and products that were a day late and a dollar short (but still expensive) killed them, not nVidia.
Try to find a faster RISC chip with lower power consumption than an Athlon64 or P4? I'll give you a hint, you aren't likely to find one. DEC's old Alpha chips were faster than anything around in their day, but they also consumed more power (over 100W in some Alphas, and that was a few years back). IBM's Power4 is also up in the 100W+ range. Intel's VLIW Itanium consumes up to 107W.
Heck, even the IBM PowerPC 970, that lots of people point to as being a cool-running chip in their Powermac G5s is in the same ballpark as AMD's Athlon64/Opteron. It's listed as 48W typical power consumption at 1.8GHz. Add about 10% to get the typical power consumption at 2.0GHz and another 20-40% to get maximum power consumption and it's right up there with the 60-80W maximum power of the P4 or the Athlon64.
Honestly there really isn't a difference betwen RISC and CISC these days. CISC chips look like RISC chips on the inside and the instruction set of RISC chips has mostly become more complicated. All of the chips read instructions from memory and execute different operations on the inside, just like how it's done on x86 chips. The biggest remaining difference is that x86 chips have only 8 visible general purpose registers (16 GPRs in the case of AMD64 chips) while most RISC chips have 32 GPRs and VLIW chips like Itanium have a whole lot. Of course, once you factor in rename registers, even that difference isn't all that huge.
FWIW AMD lists the maximum power consumption of their Desktop Replacement Athlon64 at 81W, and the Mobile Athlon64 as 62W. Both are pretty high, though the Athlon64/Opteron core seems to have a pretty wide power range (ie typical power is quite a bit lower than the maximum power listed). I'm not sure whether this machine uses the Desktop Replacement Athlon64 or the Mobile Athlon64, but I believe that it's the latter (AMD only lists a 3200+ speed grade for the DTR chip, and the laptop comes with a 3000+ speed grade).
Certainly not the sort of chip that you can stuff into a super-small laptop, but about the same as it's main competitor. Intel's Mobile P4 processors have a Thermal Design Power (almost-sorta-kinda maximum power) of up to 70W.
In both cases these machines are really targetting the desktop repalcement crowd. They're designed for people who need to move their computer from one place to another every once in a while but rarely need to use them on the road.
Some of what you say has some validity, but it ignores the fact that sound is digitized at a much higher sampling rate and higher bit rate than a CD. Then you just throw some digital filters at it to get it down to the CD specifications.
The result is that you actually DO get an exact reproduction of the analog signal frequency (except maybe for the first and last half-second of the record) and very nearly an exact replication of the amplitude of the signal (at the low-end of the volume scale it's not quite as exact as it should be due to the use of a linear codec for a very logrithmic problem).
Think of this as being like scanning the image at 32-bit colour and very high resolution, but then cropping it down to 16-bit colour and a lower resolution. The resulting image almost always looks much better than if you scanned it initially at that same 16-bit colour and low resolution.
The "2800+" IS the speed. No, it's not the clock speed that the chip runs at, but that really doesn't make it any more or less relevant to the speed.
Intel decided to market their processors using one totally meaningless measure of performance, the clock speed.
AMD decided to market their processors using a different totally meaningless measure of performance, the model numbers.
The problem with the chips is not so much sleezy marketing as with are ridiculous focus on clock speed. That's like buying a car solely based on the displacement of the engine. Sure, a 3.0L engine might produce more power than a 2.5L engine, but that certainly isn't always the case, and it sure doesn't tell you much else about the car. But if the public bought cars based ONLY on the displacement of the engine like they buy PC's based ONLY on the clock speed of the processor, you better believe that companies would bring out "3.0+" engines with 2.5L of displacement.
In any case, AMD's marketing numbers have been extremely successful for the company. They almost instantly increased AMD's profit when they first brought them out years ago. Argue all you like, but the fact is that the computer buying public is, by and large, generally uniformed and buys systems based on two numbers alone, clock speed/model number of the processor and the price.
That's the official story, and AMD is sticking to it!
Of course, in reality the model numbers are in direct comparison to the P4. AMD might not say that in public, but that has always been the intention.
Is this a good thing or a bad thing? I don't see that it really matters. The replaced one TOTALLY MEANINGLESS number (clock speed) with another totally meaningless number (model number).
Personally I like the Opteron's numbering system best.
Socket 7?!?! Good &deity no! Talk about a TERRIBLE design!
If we kept socket 7 we would: a.) still be stuck at ~1GHz processors because the socket did not provide enough power or grounding pins for todays faster processors. b.) would have TERRIBLE memory throughput, the real-world performance of this socket was terrible even if when the theoretical numbers were ok.
Perhaps most importantly though, it wouldn't help anything. You would STILL need to buy new motherboards to support new chips. In fact, it would probably be a LOT worse because you wouldn't know just what processor your particular version of socket 7 board supported! So instead of having 4 sockets to worry about you would have THOUSANDS of different motherboards, all with a lists of dozens of processors that they do and do not support.
There's MUCH more to making a chip compatible with a motherboard than just the socket. In fact, the physical socket is a rather trivial part of it.
Been there, done that, but the PPro didn't sell.
Actually I'm not sure how much backwards compatibility there was in the PPro socket, but it WAS called "Socket 8".
IBM seems ot be well ahead of AMD in moving towards a 90nm manufacturing process. The new Xserve G5's have already been announced with 90nm PPC970 chips and they should ship in 6-8 weeks, making it a close second to Intel's 90nm chips (which should actually be available in two to three weeks time). AMD is going to be a little bit late to the show (not surprising, they don't have the financial resources of Intel or IBM, so they're always ~6 months behind in bringing out a new process generation).
Of course, I wouldn't hold my breath for 3GHz G5's by Q3. 2.5GHz PPC 970 processors by mid-summer sure, no problem. But 3.0GHz? It's possible, but certainly not a sure thing. I fully expect the clock speeds of these two chips to remain pretty comperable for some time to come.
The marketing of the Athlon64 FX has become a bit confusing. It kind of made sense for the initial launch to combat a precieved weakness of the design compared to Intel's P4 though. With the P4 you get up to 6.4GB/s of memory bandwidth, while the first Athlon64 would only have 3.2GB/s of memory bandwidth. Now, it turns out that the extra bandwidth doesn't actually buy you much on most applications, but this was seen as a weakness, hence the Athlon64 FX. With Intel bringing out the P4EE to compete with the FX, now AMD might need to keep the chip, even if it isn't a worthwhile product (The P4EE isn't a worthwhile product either).
*Is it really a good idea to have the memory controller on the CPU?
Yes, yes it is a good idea. A VERY good idea in fact. Memory latency has only improved by about one order of magnitude in the past 15 years. Meanwhile everything else in the system has gone up by at least two orders of magnitude. Virtually everything that is being done in CPU design these days is to hide memory latency (larger caches, out-of-order executation, branch prediction, even SMT).Integrating the memory controller reduces latency by 20-30%. At 2.0GHz this makes a BIG difference (this is the main reason why a 2.0GHz Athlon64 is faster than a 2.2GHz AthlonXP), at 4 or 5GHz the difference will be huge.
but it hasn't lowered mainboard costs
You can buy new Athlon64 motherboards for only $100, only 3 months after the chips release. It took ages for Athlon or P4 motherboards to reach that price point. What's perhaps even more impressive is the dual-processor boards that are only $200. In short, it HAS reduced motherboard costs, whether you've noticed or not. It also means that ALL Athlon64's support ECC, chipkill and a few other nifty reliability features, regardless of how badly VIA screws up their chipset design.
*Why in the world introduce an AthlonFX based on Socket 940, especially at the outrageous price, when you're moving to socket 939 imminently?
The Athlon64 FX was a bit of a last minute decision I believe. They found a marketing weakness and wanted the quickest and easiest solution they could find. The answer? Sell your server chip as an "enthusiast" chip. Intel did exactly the same thing for the same reason with the P4EE.
Also, it's actually VERY normal to switch sockets soon after releasing a new processor. Intel's upcoming Prescott will use Socket 478 for only about 6 months before switching to socket 775. The original P4 used Socket 423 for a very short time before switching to socket 478. The original Athlon used Slot A for a year or so before switching to Socket A. The PIII came out in Slot 1 form but then switched to Socket 370 about a year later. The Celeron followed the same path a couple years before.
I think it would have been more of a slam-dunk as a platform and a "brand" to release Athlon64 as all dual-channel, all Socket 939 (or some standard), and left Opteron as the high-end platform. Any other takers?
In retrospec that might seem like a good idea, hindsight is 20-20 after all. However the original split of ALL Athlon64 chips being socket 754 and ALL Opteron's being socket 940 seemed to make the most sense when AMD was desigining them. It wasn't until market conditions changed and a new perceived weakness was discovered that AMD felt they need a consumer chip with a 128-bit wide memory bus. By that time the chip was already late to market and designing a new socket would have added more delay to the equation.
There's also the question of budget chips. AMD hopes to move their entire product line to the Athlon64/Opteron platform by the end of 2004. That means they need a budget chip, and socket 939 with it's 128-bit wide memory bus is problematic for that. Hence the continued existance of Socket 754 and the AthlonXP for that platform.
But the PowerPC line has superior signal processing capabilities these days
According to who? And why? SSE2 provides plenty of instructions for signal processing, and SSE3/PNI will fill in a couple of the last remaining holes. AMD64 also doubles the number of registers for SSE2 as well as the general purpose registers.
Exactly what advantage does the PowerPC have for signal processing? They do have a nifty multiply-add instruction that is missing on x86/SSE2, but on the flip side their vector processing engine (Altivec) seems to be limited to single precision floating point math unless I missed something.
is much easier for compilers to cater to
Ohh, so THAT'S why there are WAY more x86 compilers out there than PPC ones.
And runs cooler and more efficient than the X86 offerings
The PPC 970, used in Apple's Powermacs is listed at a "typical" power of ~48W at 1.8GHz. Maximum power is probably a good 20-40% higher. In other words, power consumption is pretty much identical to AMD's Athlon64/Opteron line and pretty darn close to Intel's P4 line. It also has 52 million transistors, nearly the same as the 55 million transistors of the P4. The IBM Power4+ consumes a ton of power, somewhere up around 100W.
AFAIK, x86 units since the i686 have all used a RISC-like core that runs x86 ops by breaking them down into micro-ops and reconstituting them. It -works- but whay do that when the real thing is available?
Why do it? Becuase hardware is cheap, software is expensive. They also do it because it doesn't really make a difference, ALL chips start by decoding their instructions, executing them and then putting it all back together again. It's not just an x86 thing.
The main downsides to x86 are as follows:
- Too few general purpose registers and restrictions on what registers can be used for which instructions
- 32-bit pointers
- Stack based FPU
- Somewhat more complex instruction decoder
AMD has addressed the first two points with AMD64. They've extended the general purpose and FPU (SSE) registers from 8 to 16 and removed the restrictions on what registers can be used for which instruction. Not quite the 32 registers of some other architectures, but still pretty good.
The stack-based FPU is in the process of being replaced by SSE2. While SSE may have started out just as a vector SIMD engine, with SSE2 it's really a full fledged floating point unit to replace the old x87 unit.
The more complex decoder has also been largely addressed by Intel in their P4 design with their trace cache. This stores already decoded instructions, so about 90% of the time the decoder is not needed (and therefore can be made much simpler).
And the Pentium Pro core lasted them until 1.4GHz, but that didn't stop Intel from marketing it under the PII and later PIII brand names, not to mention the Xeons and Celerons.
It's likely that the Netburst core of the P4 will be here for a while yet. I don't know about the 11GHz number, but I suspect that it Intel will continue to use the core until at least 2006 or 2007. However I also suspect that we'll see at least a "Pentium 5" in that time frame, and a "Pentium 6" is not out of the question. It's also likely (in fact, almost certain) that Intel will put two Netburst cores on a single chip in 2005 or 2006, at least for their Xeon line, and that may extend the basic Netburst architecture for a while longer.
Records often sound good, but it's not because the quality of the medium is good. In fact, it's just the opposite. Many people like the distortion added by vynal records, and most older music was written with the intention that such distortion would be added in. When you listen to the music on a CD with no distortion, it just doesn't sound as good because the original analog was mixed with the intention of getting some distortion on playback.
Sure, encoding to digital will lose you some data, and CDs aren't ideal. The use of a linear codec instead of a logrithmic one seems to me to be it's biggest mistake (this causes some problems at low volumes, which isn't exactly a strong-point for records either).
Most people who think vinyl is a better medium than CDs are under the mistaken belief that a) CD's can not reproduce an analog signal of an exact frequency (they can, up to their 1/2 the sampling frequency of 44.1KHz, ie 0 to 22.05KHz), or b) that the best records and sound system CAN exactly reproduce the amplitude of the sound wave (the signal to noise ratio of even the very best records and sound systems is not as large as the dynamic range of a CD).
Of course, as mentioned above, a more exact reproduction of the original does not always equate a better "sounding" copy.
Open relays are defintiely yesterday's technology, they are rarely used by spammers anymore (mainly because there aren't nearly as many of them anymore and there are so many easier methods for spammers to use).
These days it's all done through open proxies or special spamming applications. These are setup using worms and proxies (SoBig, MiMail, etc.) and run on the systems of your average everyday joe-user with broadband connections. Sure, you could fine all these users (though you could only do so in a country by country basis), but do you really want to fine someone's grandmother just because she didn't install the latest patch to her Windows box (or her Mac, or whatever other OS she happens to be running that was compromised)? This would certainly not be a popular law with anyone except the most ovezealous geeks.
Err, not if you're me apparently. I just went through the Apple Online store and brought up the configuration of a bog-standard dual-processor G5. Estimated shipping time, 6-8 weeks. Tried again for the single-processor system with the same result, 6-8 weeks.
Err, have you bothered to check the datasheets? The P4 is hardly a cool running chip by any stretch. It may have a wider power consumption range than the AthlonXP, but when the chips were running flat out they sure didn't run any cooler.
The only reason why P4's used to run cooler than Athlons was because people would stick a 60mm x 60mm heatsink on their Athlon and an 80mm x 80mm heatsink on their P4. Both of these chips consume a lot of power, and both drain laptop batteries like crazy if you use the highest powered parts (Intel actually produces some P4 "mobile" chips with a TDP of 70W!, while AMD's brand new "mobile" Athlon64 chips consume over 60W at full throttle).
Now, the Pentium M... well that's another story altogether.
Err, have you ever installed a P4 or an Athlon64/Opteron processor? They DO use locking mechanisms to sandwich the processor between the motherboard and the heatsink!
The G5 at 90nm may have been announced by Apple, but it's not yet shipping from what I can tell. Given traditional Apple schedules, it will actually start shipping in one to two months time, or roughly the same time frame as Intel starts shipping their first 90nm processor (Prescott, not Tejas, that's a future chip not scheduled for 9+ months).
The XServe G5 is using new IBM PPC 970 processors made on a 90nm fab (as comapred to the 130nm of the desktop PowerMac G5). This should reduce the power consumption somewhat and bring it into the right range to cram it into a laptop. It'll probably still consume a reasonable amount of power at 2.0GHz and the standard voltage, but Apple could probably order up some 1.6 or maybe even 1.8GHz chips running at ~1.0V that would drop the power consumption down into the same basic range as the G4. Certainly getting power consumption down to about 30W TDP for a 1.6GHz part should be no trouble, and that's rather easy enough to cool (AMD and Intel both have chips that consume up to about 70W of power that they sell as "Desktop Replacement" mobile processors, and even the regular Pentium 4 M consumes 30-35W of power).
I suspect that we will see Powerbook G5's in the not-too-distant future, probably even sooner than this time next year. Maybe mid-summer? When's the next big Apple show?
Yes the XServe G5 and Opteron cost about the same, have similar features and for many applications will have similar performance. However, now people have a nice extra choice. For some situations, the G5 is a LOT faster than the Opteron. For others, the Opteron is a LOT faster than the G5. They are rather different architectures with different instruction sets.
There are a few nice points about the XServeG5. First off they now support ECC memory, a glaring omission in the standard G5 machines that made them totally unsuitable for servers of large clusters. Second, they are now using chips manufactured on a 90nm fab process. This should hopefully result in lower power consumption and heat dissipation. The 130nm G5 and 130nm Opteron had pretty similar power consumption numbers, and when the 90nm Opterons come out in mid-2004 they will likely have similar power consumption figures again, but for the time being the G5 has a slight edge here.
Pentium-M 1.7GHz has a TDP of 24.5W The Ultra Low Voltage Pentium-M at 900MHz has a TDP of 7.0W. The Transmeta Crusoe's power consumption of their lowest powered 1GHz Crusoe 5800 chip is 6.5W (or 8.5W for their cheapest 1GHz Crusoe 5800). These numbers are not defined 100% the same, so Transmeta might actually have a slighly larger advantage in real-world situations, but on the other hand it's a MUCH slower processor, so the 900MHz ULV Pentium-M will spend most of it's time in a lower power state as compared to the Crusoe. Neither chip consumes much of anything when idle.
In the end, the Crusoe doesn't end up looking very competitive except if the price is REALLY cheap. A few watts here or there for the ULV Pentium M are, for most people, MORE than made up for by the fact that it's roughly twice as fast.
Really a more accurate comparision is not with the Pentium M, but with the ULV Celeron processors, which also consume 7.0W at 800MHz. These chips are still a hell of a lot faster than the Crusoe (though maybe only 40 or 50% faster than a 1GHz Crusoe) while being in the same price range and same power consumption range.
Roughly 0%. Seriously.
Transmeta might sell a handful of laptops, webpads and thin clients, but probably less than a million a year in total.
ARM, on the other hand, vastly outsells all x86 chips combined (admittedly many of these are not for mobile devices, merely embedded systems). We're talking something on the order of 200-300 million chips (maybe more, last numbers I heard were from a few years back). Even if you only counted all of the PocketPC and Palm systems being sold with ARM chips you're probably looking at more ARM chips being sold every month than the total number of Crusoe chips Transmeta has ever sold.
According to Sandpile.org, the 90MHz Pentium consumes a grand total of 9.0W of power when going full-bore. Transmeta's claiming power consumption numbers of about 6 or 7W for their chips. The Transmeta chips will have a slight advantage in terms of dynamic power management (almost non-existant on the old Pentium classic), but really we're talking about a few watts here. Throughout the course of a year you're looking at a difference of about $2 to $3 in electrical costs. At that rate, the Transmeta chips have got to be REAL cheap to beat out spending 5 or 10 bucks on a Pentium system.
MiniITX'ers already have the VIA C3 chips that offer similar power consumption numbers and equal or better performance and a very low price tag.
Of course, this announcement by Transmeta is in direct response to VIA's announcement of physically smaller chips a few months back. VIA's Antaur processor is the exact same idea, take a small, low-cost and low powered x86 core and package it up real small. VIA's chip is 35mm x 35mm for the entire package. Transmeta comes in a slightly smaller 21mm x 21mm package.
In the end, the Transmeta has a slight power consumption advantage (clock for clock) and a slight size advantage, but really neither are very significant, and VIA may have a performance advantage (or, conversly, they can negate the power consumption advantage by using a lower-clocked processor) and traditionally has had quite a large price advantage. Both are fighting for a pretty small and low-profit market though.