Am I the only one to notice a weird correlation between the race for putting more blades on razors (three or four) and the race for putting more layers on next gen DVD formats (three or four).
I am not sure if even a dual 2.7GHz G5 can run H.264 encode on SD (720x480) material in real time using the H.264 encoder that comes with QT7 pro.
I was at the Apple store the other day and even dual 2GHz PowerMacs appeared to be struggling to keep up the frame rate when decoding 720p or 1080i H.264 streams. The complexity of encoding is at least 10X that of decoding, if not more.
I am not aware of a real-time H.264 codec that can encode SD video in real-time, even on a 3.73GHz P4, let alone HD. With the advent of the upcoming dual core CPUs, real-time H.264 encoding of SD content may become a reality soon, but real-time encoding of 720p material on a general purpose processor is probably still a few years away.
There are a few companies working on special HW (i.e., chips) for real-time H.264 encoding, though most of the first generation products are focusing on standard definition. Since video encoding is a parallelizable problem, I wouldn't be surprised to see specialized HW that is capable of real-time H.264 encoding 720p material in a couple of years.
Even a more general purpose device like the Sony/IBM Cell chip may find an application in this context. By that time, the density of flash cards or microdrives should have reached the sufficient capacity to be able to hold a reasonable amount of MPEG4/H.264 encoded video material. As with the current gen MPEG4 camcorders, I wouldn't be surprised if the encoded video quality of the initial batch of devices are not that great.
Even though hardware encoders are likely to be initially targeted at the small form factor high definition camcorder market, we are likely to see a batch of hardware HD MPEG4/H.264 encoder cards (similar to the current SD MPEG2/MPEG4 cards) as soon as such chips are commercially available.
Of course, the other requirement for HD video conferencing would be an increase of average broadband access rate to 6-10Mb/s range. Even with H.264 encoding, a somewhat static 720p stream can easily burn 2-3Mb/s of net bandwidth.
But the team didn't just take an existing core like the PowerPC 970FX and build an SoC around it. The core for Cell is new and appears to have been designed before the clock-frequency-is-dead era. The core was designed to reach certain power and die-size goals and is designed to be able to run at clock frequencies in the 4+GHz range. The engineering theam did simply some of the core design (for example, it's an in-order design and only a dual-issue superscalar) and used some dynamic logic in the design in certain critical timing areas.
The core complies with the PowerPC instruction-set architecture version 2.02 (and the 2.01 public version of teh specification). The core was designed with a particular balance of die size, clock speed and architectural efficiency that is different from that of PowerPC 970. This instantiation of the Power Architecture still has a relatively long pipeline, much like the Power 4 and PowerPC 970, but the Cell design does not have a very wide issue pipeline or out-of-order execution, nor does it have as many functional units.The Cell Power core has hardware fine grain multi-threading.
So it looks like the PS3 core is a lot simpler than even the simplified Power4 core in the PPC970. Looks like they decided that instruction level parallelism does not help with game code and went with a smaller dual issue design with reduced number of instruction units.
This is quite insteresting. Unlike general purpose processors, which are often optimized for a set of specific benchmarks, the processor for a game console is actually designed to optimize the performance for a specific set of applications, i.e., 3D games. The most demanding applications driving the performance of high end PCs today also happen top be 3D games. I wonder if we are going to see a transition to back to simplified cores with higher clock speeds soon. Given the current trend to integrate multiple cores on a single processor die, a multi-core design with a large number of simple, high speed processors would be an interesting design trend.
The multi-threading feature of the Cell core may be ported over from the Power5 design as a way to deal with memory latency at high clock speeds.
I think it would be pretty safe to assume that the PowerPC core in the Xbox360 chip is very similar, if not the same design.
Here is an IBM paper that shows, at least in the lab, they were able to run the cell processor above 4GHz.
It is not clear how the "declining" numbers for the broadcast stations were calculated in this study. There are several decent braodcast stations in my area, and all of them have Internet simulcast of their programs. I frequently listen to the Internet stream of these stations, especially in my office where I have very poor FM reception. Does the time I spent listening to the Internet stream count as lost broadcast listener time?
I also noticed that a lot of people from outside of the station's broadcast area call in or send e-mail to these stations every once in a while, frequently from a different state and sometimes from a different country. Some of the better on the air stations look like they are actually capitalizing on the capability to stream their content on the internet.
This may be the best possible reference case for the average IT guy trying to convince his/her boss that FireFox is a good solution for a corporate environment.
I am not keeping my hopes high on this. The response rate for the regular snail mail is about 1%, that is 99% of the junk mail people receive basically goes to trash. Still, corporate America is willing to spend more than $46 billion on direct mail marketing every year. Considering that postal junk mail is a lot more expensive to send than e-mail spam, a 0.1 or even 0.01% response rate for e-mail spam may be sufficient to sustain the current spam rate, if not continued growth.
Comparing the X2s against 130nm Athlon 64 is not fair. As expected, the new X2 Athlons do burn more power than a single core Athlon 64 built on the same core/process at the same frequency. The amazing thing is that the difference appears to be only around 20%, which is almost unbelievable.
You would expect to see less than 100% increase in the case of a dual core CPU due like the shared components which are not replicated:
- The X2 chips still have a single, 128-bit wide memory controller. Since the memory controller charges/discharges external bit lines going to DIMMs, they do burn quite a bit of power. This power consumption is not duplicated in the case of a dual core CPU.
- The X2 chips still have a single HyperTransport bus. The power consumption of this bus is the approximately the same between a dual core and single core CPU.
However, power scaling due to these shared components would probably not explain how a dual core chip can burn only 20% more power.
For both of the above cases, you could argue that one should expect to see higher utilization of the memory bus and the HyperTransport bus, so the exact power consumption contribution is not entirely clear.
One thing to note that, AMD Athlon 64 cores tend to burn much less power in idle state compared to Intel chips. This is probably due to choices AMD made both in architecture and process. So the fundemantal reason why AMD X2 chips have such minimal incremental power consumption over single core chips is that one of the cores is typically underutilized most of the time and therefore burns much less power.
It would be easy to fight this provision in the senate: Just attach a simple amendment that requires gun dealers to scan a customer's Real ID before making a sale!
First, the method used by LostCircuits is not very accurate to begin with. The Fluke 80i-410 probe they used has accuracy of +-5% and a measurement floor of 2.5A. A current probe with a lower measurement floor like this one would have been a better choice. There is at least another case (XbitLabs) where a similar measurement showed that Venice uses more power than Winchester at the same frequency. Unfortunately, XBitLabs test doesn't mention which current probe was used.
Even if we assume that the current measurements were accurate, it is almost impossible to come to conclusions about the Venice core being more efficient than the Winchester core based on observations from one sample each. Note that the observed current consumption between the Venice core and the Winchester core is within a few percent of each other in most of the tests run by LostCircuits. You may see more than that much difference between samples from different production runs of the same core.
The only thing that the Lostcircuits test proves is that AMD's 90nm cores are more power efficient that their 130nm cores...
If I go to HPs site and configure the lowest dual Xeon xw6200 with 512MB memory and a 160GB HD, the price is $1793. This happened to be the only piece of data in your original post, and you managed to get it wrong for the second time.
I really don't know how you got that. I went back to their site to make sure I didn't miss anything, and with the same dual-Xeon 2.8GHz xw6200 system, 512MB RAM, and 160GB drive space, I get $1,599. That's 99$ higher, so perhaps I did miss something the first time, but I can't figure out how you end up spending $200 more. You must have selected an extra option in there somewhere.
You have to add the dual-layer DVD Writer (a $169 option) to make it a somewhat fair comparison. Actually, I forgot to add the $49 IEEE1394 card option, which brings the total to $1842.
Again, completely wrong. For example, AMD's price for the Athlon X2 4400 (dual 2.2GHz core) is $581, whereas you can buy a single core Athlon 3200 (2.2GHz) for $185.
No, you just happen to have found a single exception, out of about a dozen that are priced exactly as I said.
Again, another comment from you with no data to back it up. You mention "a dozen cases" and don't bother to quantify a single one. And you are the one complaining that I am not giving any numbers.
It turns out that the Athlon 4400 example I had in my previous post was one of the milder cases in terms of the price premium for the dual-core. Take a look at the Opteron 148 vs. 175. Both run at 2.2GHz, but the dual core chip is over 3.5x the price of the single core CPU:
AMD Opteron 175 (2.2GHz) $999
AMD Opteron 148 (2.2GHz) $278
The similar trend continues for 2xx series:
AMD Opteron 275 (2.2GHz) $1299
AMD Opteron 248 (2.2GHz) $455
And for the 8xx series:
AMD Opteron 875 (2.2GHz) $2649
AMD Opteron 848 (2.2GHz) $873
Since you have proven that you are not capable of supporting any of your arguments with any kind of data, I will not argue with you any more.
If you bothered to read the results of your Froogle search, you would have noticed that most of the modern motherboards that are the hit list are extended ATX motherboards. Many of the smaller form factor boards happen to be PII or PIII motherboards from good old times when CPU power budgets were under 40W. It was educational to see people still selling slot1 Xeon motherboards...
Not at all... even a mini-ATX case could handle two very large heatsinks/fans without any space problems. A mid-tower ATX cas has enough room for quad-processors, if you're so inclined, though that's starting to push it.
This comment alone leads me to believe that you have either never put together a high end multiprocessor system together or never bothered to test one. Let's not even get into a discussion about how to dissipate the 320-360W of heat just from the CPUs. The footprints of 4 80+W processors would pretty much not leave any room for anything else on a standard ATX motherboard.
Again, I don't know how you came up with the $1500 number. If I go to HPs site and configure the lowest dual Xeon xw6200 with 512MB memory and a 160GB HD, the price is $1793. If HP did not happen to be running a free 2nd processor promo for low end Xeon configs, the price would be more than $2K.
Upgrading the same system to dual 3.6GHz Xeon (not even the most expensive CPU choice for this system) and upgrading the HD to 250GB brings the price up to $3903.
That's not really true. There are plenty of people running dual-CPU Athlons (not MP), Pentiums, etc. Even low-end Opterons (which most people buy for uniprocessor systems) can be used in dual-processor configurations with no problems as well. If you only want dual-CPU, the price increase is not significant. It's only above dual-CPU setups (quads) that things get complicated and expensive.
Let's look at some prices at NewEgg today:
Opteron 248 (2.2GHz) $445
Athlon 3500 (2.2GHz) $260
Intel Xeon Nocona (3.6GHz) $729
Intel P4 560J (3.6GHz) $429
At the same frequency, there seems to be a very significant premium just on the processor. In most cases, the motherboard will add another $100-$200 to the additional cost.
AMD must not have anyone knowledgable about semiconductor technology working for them in that case, because they are selling dual-core chips for significantly less than double the price of single-core chips.
Again, completely wrong. For example, AMD's price for the Athlon X2 4400 (dual 2.2GHz core) is $581, whereas you can buy a single core Athlon 3200 (2.2GHz) for $185.
In any case, you haven't shown any numbers to back-up your claim (and I don't believe you have any at all) so whatever the case, you certainly are talking out of your ass.
I don't know how many more "numbers" I can show. This complaint is quite unfair. The only number in your entire post was a vague price quote of an HP workstation, which, happened to be incorrect.
The bottomline is, AMD and Intel are heavily marketing double cores to PC users as this is the only viable way for system manufacturers to provide performance boost to their next generation systems without having to tackle difficult technical and cost issues.
I normally don't bother to respond to random uninformed comments, but I have to correct your unsubstantiated claims in this case.
Intel/AMD processors don't "jump" in speed from one CPU to the next, but if you compare AMD/Intel to Apple over the same time-frame, it seems that Apple isn't keeping pace. That's not necessarily a problem, but your implication that Apple is keeping up with AMD/Intel is not true. In addition, dual-cores is a big jump in performance, and AMD processors have been getting more energy-effecient all the while, which also leaves Apple behind.
Actually, Intel/AMD processors have exhibited historical jumps in clock speed. These jumps were either due to major architectural changes (PII->PIII, PIII->P4), or until recently, due to process changes. The process change from 130nm to 90nm has not been able to generate significnat clock speed improvement for neither Intel nor AMD. If you look at the clock speed improvement for the three major processors over the last two years (since June 2003) the picture is not very promising:
G5: CPU clock up 35%, FSB clock up 35%
P4: CPU clock up 19%, FSB clock up 33%
K8: CPU clock up 44% FSB clock up 25%
Note that if we started the two year period from April 2003 instead, Apple would have looked much better in terms of the percentage improvement, but that would include a major architecture change from G4->G5.
>
It is very hard to put two "hot" processors on an ATX motherboard in an ATX case.
No it isn't. Not even slightly.
There are not many commercial dual-processor products in ATX form factor. Actually, Intel had to create an entire new form factor for dual-Xeon systems. It may be possible to fit two low-power Opterons in an ATX motherboard, but obviously, this has not been very popular. Most ATX cases have problems fitting two large-size CPU coolers in this configuration.
Besides, comparing Apple computers to white-box computers is crazy. You should compare them to computers from HP or another company, who also make non-ATX systems. Power Macs are also more comparable to PC Workstations, rather than a low-end PC.
Obviously, there are dual-processor workstations made by other companies. As you pointed out, nearly all of these use non-ATX form factor, and they are even more expensive than PowerMacs in many cases. This is precisely my point.
PowerMac is a strange product family as is covers a wide market space starting from the high-end home PC all the way to professional workstation. For the high-end PC segment, Intel had no choice but to come up with a dual-core architecture to get manufacturers to help drive the mainstream high-end consumer base to SMP.
> This way, all of the rest of the system components (motherboard, chipset, case, cooling system) can stay the same.
The case can stay the same either way. The motherboards, chipsets, and cooling system only requires very minor modifications for dual-processor systems.
Your statement is completely bogus. For a dual-processor PC (as opposed to a dual-core), you would need a different chipset, different motherboard and a different cooling system today. On top of that, you need different processors (Xeon or Opteron) which are typicaly at much higher price points.
> Overall, it may actually be more cost effective for Apple to ship multiprocessor system
It might be, and it might not be. You might be talking out of your ass, you might not be talking out of your ass.
If you talk to anybody who is somewhat knowledgeable about semiconductor technology, they will tell you that the cost of manufacturing a 120mm2 die typically costs a lot more than twice the cost of manufacturing a 60mm2 die. Unlike Intel, IBM's PPC970 also has a very efficient and simple SMP architecture (uni-directional dedicated links to each CPU as opposed to a shared high-speed bus), which makes it possible for Apple to design dual-processor systems with reduced complexity and cost.
As for your last comment, I think it is pretty obvious who is talking out of his/her ass in this discussion...
Going briefly over the available documents on this, it appears that this technique consumes orders of magnitude more energy than it produces. This would preclude energy generation as one of the potential applications, which is usually regarded as the most promising potential application of cold fusion. Most of the other potential applications mentioned in the articles use this as a neutron generator, but there are other well known ways of achieving that...
This PowerMac revision is obviously not a major upgrade and Apple treated it as such. Apple's home page, is mostly full of Tiger stuff and the new PowerMac intro is just a small image on the lower left corner.
Lot of people are complaining about the "just 200MHz" speed bump for the high end model. 8% may not be that much of a speed bump, but neither Intel or AMD has been able to pull off dramatic clock frequency jumps lately. Clock speed stagnation seem to be a general problem in the processor design industry.
As for the dual cores, obviously AMD and Intel have much more incentive. The entire PC world is built around a standard form factors: ATX motherboards and ATX cases. Intel's efforts to move to a new form factor (BTX) has been quite unsuccessful so far. It is very hard to put two "hot" processors on an ATX motherboard in an ATX case. PC market is also driven by cut throat price pressure and low margins. There is a huge price difference between the prices of single processor motherboards and dual processor motherboards. Given the stagnation in the clock frequency, the only practical way for Intel and AMD to drive the mainstream PC to higher performance is the SMP model through dual-core chips. This way, all of the rest of the system components (motherboard, chipset, case, cooling system) can stay the same.
Apple does not have this constraint. Apple has been manufacturing mainstream multiprocessor desktops for manty years. Overall, it may actually be more cost effective for Apple to ship multiprocessor system. It may be a lot cheaper for IBM to manufacture two instances of a small die like the PPC970 FX (less than 60mm2) than a larger dual core die. As for Apple, having the source of the heat distributed accross two chips makes thermal management somewhat easier than dealing with one extremely hot dual core chip.
I am sure Apple will eventually move to dual core PPC970MP chips, potentially later this year, but this will most likely be in the context of being able to offer quad systems (two dual-core processors) for higher performance.
As for the choice of the base graphics card, the 9600 or 9650 is a perfectly reasonable choice. The primary driving force behind high end graphics cards in the PC world are 3D games. PowerMac G5 is obviously not the best 3D game platform. Most people buy PowerMacs to use in professional applications. Many pro applications do not require super-duper 3D performance. For those who are planning to do serious 3D work, the 6800 Ultra upgrade is the reasonable choice. There is no reason to burden all customers with an expensive (and potentially loud) graphics card.
There doesn't seem to be any readily available commercial multi-bank DC adapters out there. This is quite surprising since the solution is pretty simple.
The solution requires a switching power supply that generates a DC voltage that is somewhat higher (at least 3V) than the highest voltage required to be generated, and a bank of LM317 programmable voltage regulators.
In this configuration, each LM117 can provide up to 1.5A of current. If necessary, LM318 or LM150 devices that support higher current can be used.
LM117s package sell for about 50c each, so this would be a relatively low cost solution.
Actually, ATX power supplies have a -12V output. Combining this with +12 or +5V terminals, you can create 24V and 17V outputs at relatively low current values to supply devices which do not need a common ground (i.e., which will not be connected to any one of the other devices which is also being powered from the same ATX supply).
It is rather amazing that there appears to be a consensus among industry experts that there has not been any improvement in code quality over the past 30 years or so despite the development of a vast number of new tools and languages. It is true that the size and scope of the average application has grown by leaps and bounds. But most likely, the primary contributing factor to these kind of quality problems is the prevalent time-to-market pressuer in the software industry which is typically coupled with severe underestimation of time and resources required for projects.
Even if CS came up with a scientific solution to improve code quality, it would be an interesting exercise to see if the industry will be willing to absorb the costs associated with such a solution. Especially in an environment where end customers are well-trained to accept and deal with software quality issues.
Would executives of companies who "share" their customers' private data through negligence also qualify for the maximum three year jail sentence introduced by this law? In many cases, I would consider protection of private data more important than protection of copyrighted material.
So let me suggest *again* that the reason that these companies lose data is not because they *cannot* avoid it, but because they don't give two shits about it since there are exactly zero penalties for losing other people's personal data. I guarantee you that if the CEO had to sign a Sarbanes-Oxley style document each year certifying data integrity, you'd see these stories once in a blue moon. Why? Because when the higher-ups have some skin in the game, suddenly you start seeing attention paid and resources dedicated.
This is so true. As long as these companies are not liable for losing user data, they have not much incentive to invest any money into making their systems and/or data more secure.
This is one of the cases where opening up these cases for litigation by passing some kind of a liability law will help. True, if there are lawsuits, most of the payout will go the lawyers, not to the people whose personal data was compromised, but at least, it will provide a bit of incentives to invest more time and effort into protecting customer data.
Of course, a new breed of consulting companies are the most likely group to benefit from such legislation. They will probably come up with an insanely complicated (and therefore not very effective) methodology recommendation for protecting customer data. Worse, we may even see some kind of a standard (ISO9042?) that dictates how the effort to protect private data need to be documented...
Oh what marketing fluff. Headlines mention 2005, the first paragraph in the article says next year, and then the next paragraph says 2007. Which one should we believe?
Toshiba is actually the first to bring the perpendicular recording technology to market. We are likely to see the 40GB and 80GB Toshiba drives with perpendicular recording technology in iPods real soon (June?).
If they are going to do a 3D retake, they might as well make it a 3D-animated version so that we are not subjected to furhter abysmal acting as we had in Episodes I and II..
Slashdot Mirror
Am I the only one to notice a weird correlation between the race for putting more blades on razors (three or four) and the race for putting more layers on next gen DVD formats (three or four).
I am not sure if even a dual 2.7GHz G5 can run H.264 encode on SD (720x480) material in real time using the H.264 encoder that comes with QT7 pro. I was at the Apple store the other day and even dual 2GHz PowerMacs appeared to be struggling to keep up the frame rate when decoding 720p or 1080i H.264 streams. The complexity of encoding is at least 10X that of decoding, if not more.
I am not aware of a real-time H.264 codec that can encode SD video in real-time, even on a 3.73GHz P4, let alone HD. With the advent of the upcoming dual core CPUs, real-time H.264 encoding of SD content may become a reality soon, but real-time encoding of 720p material on a general purpose processor is probably still a few years away.
There are a few companies working on special HW (i.e., chips) for real-time H.264 encoding, though most of the first generation products are focusing on standard definition. Since video encoding is a parallelizable problem, I wouldn't be surprised to see specialized HW that is capable of real-time H.264 encoding 720p material in a couple of years. Even a more general purpose device like the Sony/IBM Cell chip may find an application in this context. By that time, the density of flash cards or microdrives should have reached the sufficient capacity to be able to hold a reasonable amount of MPEG4/H.264 encoded video material. As with the current gen MPEG4 camcorders, I wouldn't be surprised if the encoded video quality of the initial batch of devices are not that great.
Even though hardware encoders are likely to be initially targeted at the small form factor high definition camcorder market, we are likely to see a batch of hardware HD MPEG4/H.264 encoder cards (similar to the current SD MPEG2/MPEG4 cards) as soon as such chips are commercially available.
Of course, the other requirement for HD video conferencing would be an increase of average broadband access rate to 6-10Mb/s range. Even with H.264 encoding, a somewhat static 720p stream can easily burn 2-3Mb/s of net bandwidth.
This is quite insteresting. Unlike general purpose processors, which are often optimized for a set of specific benchmarks, the processor for a game console is actually designed to optimize the performance for a specific set of applications, i.e., 3D games. The most demanding applications driving the performance of high end PCs today also happen top be 3D games. I wonder if we are going to see a transition to back to simplified cores with higher clock speeds soon. Given the current trend to integrate multiple cores on a single processor die, a multi-core design with a large number of simple, high speed processors would be an interesting design trend.
The multi-threading feature of the Cell core may be ported over from the Power5 design as a way to deal with memory latency at high clock speeds.
I think it would be pretty safe to assume that the PowerPC core in the Xbox360 chip is very similar, if not the same design. Here is an IBM paper that shows, at least in the lab, they were able to run the cell processor above 4GHz.
I also noticed that a lot of people from outside of the station's broadcast area call in or send e-mail to these stations every once in a while, frequently from a different state and sometimes from a different country. Some of the better on the air stations look like they are actually capitalizing on the capability to stream their content on the internet.
This may be the best possible reference case for the average IT guy trying to convince his/her boss that FireFox is a good solution for a corporate environment.
I am not keeping my hopes high on this. The response rate for the regular snail mail is about 1%, that is 99% of the junk mail people receive basically goes to trash. Still, corporate America is willing to spend more than $46 billion on direct mail marketing every year. Considering that postal junk mail is a lot more expensive to send than e-mail spam, a 0.1 or even 0.01% response rate for e-mail spam may be sufficient to sustain the current spam rate, if not continued growth.
Doesn't the "Office" video for Firefox look a bit too much like one of the original PowerMac G5 ads where the user was literally blown away.
You would expect to see less than 100% increase in the case of a dual core CPU due like the shared components which are not replicated:
- The X2 chips still have a single, 128-bit wide memory controller. Since the memory controller charges/discharges external bit lines going to DIMMs, they do burn quite a bit of power. This power consumption is not duplicated in the case of a dual core CPU.
- The X2 chips still have a single HyperTransport bus. The power consumption of this bus is the approximately the same between a dual core and single core CPU.
However, power scaling due to these shared components would probably not explain how a dual core chip can burn only 20% more power. For both of the above cases, you could argue that one should expect to see higher utilization of the memory bus and the HyperTransport bus, so the exact power consumption contribution is not entirely clear.
One thing to note that, AMD Athlon 64 cores tend to burn much less power in idle state compared to Intel chips. This is probably due to choices AMD made both in architecture and process. So the fundemantal reason why AMD X2 chips have such minimal incremental power consumption over single core chips is that one of the cores is typically underutilized most of the time and therefore burns much less power.
It would be easy to fight this provision in the senate: Just attach a simple amendment that requires gun dealers to scan a customer's Real ID before making a sale!
First, the method used by LostCircuits is not very accurate to begin with. The Fluke 80i-410 probe they used has accuracy of +-5% and a measurement floor of 2.5A. A current probe with a lower measurement floor like this one would have been a better choice. There is at least another case (XbitLabs) where a similar measurement showed that Venice uses more power than Winchester at the same frequency. Unfortunately, XBitLabs test doesn't mention which current probe was used.
Even if we assume that the current measurements were accurate, it is almost impossible to come to conclusions about the Venice core being more efficient than the Winchester core based on observations from one sample each. Note that the observed current consumption between the Venice core and the Winchester core is within a few percent of each other in most of the tests run by LostCircuits. You may see more than that much difference between samples from different production runs of the same core. The only thing that the Lostcircuits test proves is that AMD's 90nm cores are more power efficient that their 130nm cores...
You have to add the dual-layer DVD Writer (a $169 option) to make it a somewhat fair comparison. Actually, I forgot to add the $49 IEEE1394 card option, which brings the total to $1842.
No, you just happen to have found a single exception, out of about a dozen that are priced exactly as I said.Again, another comment from you with no data to back it up. You mention "a dozen cases" and don't bother to quantify a single one. And you are the one complaining that I am not giving any numbers. It turns out that the Athlon 4400 example I had in my previous post was one of the milder cases in terms of the price premium for the dual-core. Take a look at the Opteron 148 vs. 175. Both run at 2.2GHz, but the dual core chip is over 3.5x the price of the single core CPU:
- AMD Opteron 175 (2.2GHz) $999
- AMD Opteron 148 (2.2GHz) $278
The similar trend continues for 2xx series:- AMD Opteron 275 (2.2GHz) $1299
- AMD Opteron 248 (2.2GHz) $455
And for the 8xx series:- AMD Opteron 875 (2.2GHz) $2649
- AMD Opteron 848 (2.2GHz) $873
Since you have proven that you are not capable of supporting any of your arguments with any kind of data, I will not argue with you any more.If you bothered to read the results of your Froogle search, you would have noticed that most of the modern motherboards that are the hit list are extended ATX motherboards. Many of the smaller form factor boards happen to be PII or PIII motherboards from good old times when CPU power budgets were under 40W. It was educational to see people still selling slot1 Xeon motherboards...
Not at all... even a mini-ATX case could handle two very large heatsinks/fans without any space problems. A mid-tower ATX cas has enough room for quad-processors, if you're so inclined, though that's starting to push it.
This comment alone leads me to believe that you have either never put together a high end multiprocessor system together or never bothered to test one. Let's not even get into a discussion about how to dissipate the 320-360W of heat just from the CPUs. The footprints of 4 80+W processors would pretty much not leave any room for anything else on a standard ATX motherboard.
Again, I don't know how you came up with the $1500 number. If I go to HPs site and configure the lowest dual Xeon xw6200 with 512MB memory and a 160GB HD, the price is $1793. If HP did not happen to be running a free 2nd processor promo for low end Xeon configs, the price would be more than $2K. Upgrading the same system to dual 3.6GHz Xeon (not even the most expensive CPU choice for this system) and upgrading the HD to 250GB brings the price up to $3903.
That's not really true. There are plenty of people running dual-CPU Athlons (not MP), Pentiums, etc. Even low-end Opterons (which most people buy for uniprocessor systems) can be used in dual-processor configurations with no problems as well. If you only want dual-CPU, the price increase is not significant. It's only above dual-CPU setups (quads) that things get complicated and expensive.
Let's look at some prices at NewEgg today:
- Intel Xeon Nocona (3.6GHz) $729
- Intel P4 560J (3.6GHz) $429
At the same frequency, there seems to be a very significant premium just on the processor. In most cases, the motherboard will add another $100-$200 to the additional cost.AMD must not have anyone knowledgable about semiconductor technology working for them in that case, because they are selling dual-core chips for significantly less than double the price of single-core chips. Again, completely wrong. For example, AMD's price for the Athlon X2 4400 (dual 2.2GHz core) is $581, whereas you can buy a single core Athlon 3200 (2.2GHz) for $185.
In any case, you haven't shown any numbers to back-up your claim (and I don't believe you have any at all) so whatever the case, you certainly are talking out of your ass. I don't know how many more "numbers" I can show. This complaint is quite unfair. The only number in your entire post was a vague price quote of an HP workstation, which, happened to be incorrect. The bottomline is, AMD and Intel are heavily marketing double cores to PC users as this is the only viable way for system manufacturers to provide performance boost to their next generation systems without having to tackle difficult technical and cost issues.
No, I don't.
I normally don't bother to respond to random uninformed comments, but I have to correct your unsubstantiated claims in this case.
Intel/AMD processors don't "jump" in speed from one CPU to the next, but if you compare AMD/Intel to Apple over the same time-frame, it seems that Apple isn't keeping pace. That's not necessarily a problem, but your implication that Apple is keeping up with AMD/Intel is not true. In addition, dual-cores is a big jump in performance, and AMD processors have been getting more energy-effecient all the while, which also leaves Apple behind.
Actually, Intel/AMD processors have exhibited historical jumps in clock speed. These jumps were either due to major architectural changes (PII->PIII, PIII->P4), or until recently, due to process changes. The process change from 130nm to 90nm has not been able to generate significnat clock speed improvement for neither Intel nor AMD. If you look at the clock speed improvement for the three major processors over the last two years (since June 2003) the picture is not very promising:- G5: CPU clock up 35%, FSB clock up 35%
- P4: CPU clock up 19%, FSB clock up 33%
- K8: CPU clock up 44% FSB clock up 25%
Note that if we started the two year period from April 2003 instead, Apple would have looked much better in terms of the percentage improvement, but that would include a major architecture change from G4->G5.> It is very hard to put two "hot" processors on an ATX motherboard in an ATX case.
No it isn't. Not even slightly.
There are not many commercial dual-processor products in ATX form factor. Actually, Intel had to create an entire new form factor for dual-Xeon systems. It may be possible to fit two low-power Opterons in an ATX motherboard, but obviously, this has not been very popular. Most ATX cases have problems fitting two large-size CPU coolers in this configuration.
Besides, comparing Apple computers to white-box computers is crazy. You should compare them to computers from HP or another company, who also make non-ATX systems. Power Macs are also more comparable to PC Workstations, rather than a low-end PC.
Obviously, there are dual-processor workstations made by other companies. As you pointed out, nearly all of these use non-ATX form factor, and they are even more expensive than PowerMacs in many cases. This is precisely my point.
PowerMac is a strange product family as is covers a wide market space starting from the high-end home PC all the way to professional workstation. For the high-end PC segment, Intel had no choice but to come up with a dual-core architecture to get manufacturers to help drive the mainstream high-end consumer base to SMP.
> This way, all of the rest of the system components (motherboard, chipset, case, cooling system) can stay the same.
The case can stay the same either way. The motherboards, chipsets, and cooling system only requires very minor modifications for dual-processor systems.
Your statement is completely bogus. For a dual-processor PC (as opposed to a dual-core), you would need a different chipset, different motherboard and a different cooling system today. On top of that, you need different processors (Xeon or Opteron) which are typicaly at much higher price points.
> Overall, it may actually be more cost effective for Apple to ship multiprocessor system
It might be, and it might not be. You might be talking out of your ass, you might not be talking out of your ass.
If you talk to anybody who is somewhat knowledgeable about semiconductor technology, they will tell you that the cost of manufacturing a 120mm2 die typically costs a lot more than twice the cost of manufacturing a 60mm2 die. Unlike Intel, IBM's PPC970 also has a very efficient and simple SMP architecture (uni-directional dedicated links to each CPU as opposed to a shared high-speed bus), which makes it possible for Apple to design dual-processor systems with reduced complexity and cost.
As for your last comment, I think it is pretty obvious who is talking out of his/her ass in this discussion...
Going briefly over the available documents on this, it appears that this technique consumes orders of magnitude more energy than it produces. This would preclude energy generation as one of the potential applications, which is usually regarded as the most promising potential application of cold fusion. Most of the other potential applications mentioned in the articles use this as a neutron generator, but there are other well known ways of achieving that...
Lot of people are complaining about the "just 200MHz" speed bump for the high end model. 8% may not be that much of a speed bump, but neither Intel or AMD has been able to pull off dramatic clock frequency jumps lately. Clock speed stagnation seem to be a general problem in the processor design industry.
As for the dual cores, obviously AMD and Intel have much more incentive. The entire PC world is built around a standard form factors: ATX motherboards and ATX cases. Intel's efforts to move to a new form factor (BTX) has been quite unsuccessful so far. It is very hard to put two "hot" processors on an ATX motherboard in an ATX case. PC market is also driven by cut throat price pressure and low margins. There is a huge price difference between the prices of single processor motherboards and dual processor motherboards. Given the stagnation in the clock frequency, the only practical way for Intel and AMD to drive the mainstream PC to higher performance is the SMP model through dual-core chips. This way, all of the rest of the system components (motherboard, chipset, case, cooling system) can stay the same.
Apple does not have this constraint. Apple has been manufacturing mainstream multiprocessor desktops for manty years. Overall, it may actually be more cost effective for Apple to ship multiprocessor system. It may be a lot cheaper for IBM to manufacture two instances of a small die like the PPC970 FX (less than 60mm2) than a larger dual core die. As for Apple, having the source of the heat distributed accross two chips makes thermal management somewhat easier than dealing with one extremely hot dual core chip.
I am sure Apple will eventually move to dual core PPC970MP chips, potentially later this year, but this will most likely be in the context of being able to offer quad systems (two dual-core processors) for higher performance.
As for the choice of the base graphics card, the 9600 or 9650 is a perfectly reasonable choice. The primary driving force behind high end graphics cards in the PC world are 3D games. PowerMac G5 is obviously not the best 3D game platform. Most people buy PowerMacs to use in professional applications. Many pro applications do not require super-duper 3D performance. For those who are planning to do serious 3D work, the 6800 Ultra upgrade is the reasonable choice. There is no reason to burden all customers with an expensive (and potentially loud) graphics card.
This must be a plot to return the French language back to its world wide popularity!
There doesn't seem to be any readily available commercial multi-bank DC adapters out there. This is quite surprising since the solution is pretty simple. The solution requires a switching power supply that generates a DC voltage that is somewhat higher (at least 3V) than the highest voltage required to be generated, and a bank of LM317 programmable voltage regulators. In this configuration, each LM117 can provide up to 1.5A of current. If necessary, LM318 or LM150 devices that support higher current can be used. LM117s package sell for about 50c each, so this would be a relatively low cost solution.
Actually, ATX power supplies have a -12V output. Combining this with +12 or +5V terminals, you can create 24V and 17V outputs at relatively low current values to supply devices which do not need a common ground (i.e., which will not be connected to any one of the other devices which is also being powered from the same ATX supply).
Even if CS came up with a scientific solution to improve code quality, it would be an interesting exercise to see if the industry will be willing to absorb the costs associated with such a solution. Especially in an environment where end customers are well-trained to accept and deal with software quality issues.
Would executives of companies who "share" their customers' private data through negligence also qualify for the maximum three year jail sentence introduced by this law? In many cases, I would consider protection of private data more important than protection of copyrighted material.
This is one of the cases where opening up these cases for litigation by passing some kind of a liability law will help. True, if there are lawsuits, most of the payout will go the lawyers, not to the people whose personal data was compromised, but at least, it will provide a bit of incentives to invest more time and effort into protecting customer data.
Of course, a new breed of consulting companies are the most likely group to benefit from such legislation. They will probably come up with an insanely complicated (and therefore not very effective) methodology recommendation for protecting customer data. Worse, we may even see some kind of a standard (ISO9042?) that dictates how the effort to protect private data need to be documented...
Monster branded Oxyride batteries that cost 4 times more than regular alkaline batteries: coming to a Circuit City near you soon!
Oh what marketing fluff. Headlines mention 2005, the first paragraph in the article says next year, and then the next paragraph says 2007. Which one should we believe? Toshiba is actually the first to bring the perpendicular recording technology to market. We are likely to see the 40GB and 80GB Toshiba drives with perpendicular recording technology in iPods real soon (June?).
If they are going to do a 3D retake, they might as well make it a 3D-animated version so that we are not subjected to furhter abysmal acting as we had in Episodes I and II..