Notice that the Via C3 does phenomenally better than the 1.7GHz P4-Celeron. Now consider how utterly weak the floating point unit (FPU) of the Via C3 processor is, its use of SDRAM, its lower bandwidth memory bus than the Celeron. I won't call the P4-Celeron a performance processor, but VIA's C3 should *NOT* be able to beat it in ANY performance benchmark.
What this says is that SySMark is really poorly coded, not "optimized" to favour Intel silicon. Incompetancy isn't evil. It's just...incompetant. This would explain why most serious benchmark runs seem to lack SysMark these days...
For a year and a half AMD managed to blow Intel out of the water in the performance arena, from the time they both reached 1.0GHz with the P3 and Athlon, to the time Intel started releasing Northwood core P4s with the extra 256KB L2 cache.
The longer pipeline of the P4 has nothing to do with performance. All it does is enable the processor to ramp in clock frequencies easily. A general rule of thumb is the longer the pipeline, the lower the IPC (instructions per clock/cycle) and the LOWER the actual number-crunching performance.
What is also interesting to note about this article is the hints it gives as to Microsoft's future plans for software security. The idea of having independant secured partitions within a computer is not new of course, but it's nice to note that MS is doing *something* about their rather poor security history.
Oh GOD, please pray that they don't integrate Outlook Express with the *secure OS* portion of Palladium
Amazingly enough, this one is able to analyze most of the knowledgebase around "Palladium" and boil it down to the more interesting core issues.
I would've appreciated a little more insight along the lines of what such a strategy as Palladium does to the role of the PC however. Generally speaking, PCs are multipurpose machines, which are *fully* programmable, and do pretty much whatever you tell them to. They manipulate data in any way *you* the *user* see fit. What Microsoft is attempting with "Palladium" is going to place restrictions about what a PC can do, and leave these restrictions up to the content producers. I won't comment on the stance of the content producer, but I will mention that this is a departure from what has been a central tenet of the computer: "it's yours". The trend seems to be shying away from "it's yours" to "you didn't buy it, you paid us to ALLOW you to use it - in a way we deem appropriate". Of course, "we" being the content producers. Microsoft really doesn't care what we dow with our music and movies - they just don't want the MPAA/RIAA/Legislators breathing down their neck.
Intel's Xscale architecture -should- have impressive performance. However, it seems that circumstances have conspired to keep it from showing its potential. In order to retain binary backwards compatibility, Microsoft has kept their software compiler rather basic, ensuring that it will work for multiple architectures. This lack of optimization also means that any architectural improvements Xscale has over Arm V4 or whatever it's called will not mean anything.
Hopefully the history of Xscale will work like netburst architecture's history, where about 1 year after inception, software that makes use of its architecture efficiently (like SSE2 with the P4) will start to appear.
I can identify with NPR and web businesses concerning the issue of deep linking. The business model of "free content with revenue generatioin via paid ads" is unable to handle deep linking from other websites. By having deep links bypass a website's splash page and index, the struggling website loses views of the paid ads. Since those paid ads are the only things that keep many websites in existence, I can see how they would take exception to deep linking.
On the other hand, maybe this situation is showing just how flawed and vulnerable this business model really is.
What is interesting is how the building block of the internet, IE, hypertext, does not lend itself to commercial advertising in this manner...Methinks someone needs to cook up some profitable web business models, and fast. Paid ads based on page views aren't working.
Considering how corrupt the current broadcast radio industry is, I'm surprised that online broadcasters are able to license any song to broadcast, whenever they want. With conventional radio, stations play what the labels pay them to play.
I can really see these new guidelines thinning out the more amateurish broadcasters, and leaving the more polished, better set-up ones intact. Personally, I wouldn't mind having some sort of radio subscription service for my favourite anime stations...
It may well be a situation where IBM would like to provide support, and is able to provide the manpower to making the drivers necessary to make the notebook line Linux compatible. However, one must realize that there is a difference between writing a set of drivers for a product, and actually being accountable for them. My bet is that IBM would rather use its resources in more efficient tasks than supporting such a small user base.
Browsers by default have a history folder that is only what, 15 days long? Websites rarely last longer than, what, 2 years?
The "internet" seems to be a transitory medium. Unlike paper, digital information is intangible, and can be easily wiped and replaced, or edited. A perfect example of this is the way that news sites often take their articles offline once they've been up for a week, or just look at webtracking software, which shows that links are dying faster than ever before.
If this trend continues (and given the current architecture of the 'net, I don't see it changing) then we might have a serious problem. I won't analyze it, but there is definitely something wrong when data is forever lost after existing for such a short period of time.
I did not say anything in my original post about marketshares, I only posted the technologies and products that Intel and Amd would be rolling out in 2002/2003, and likely, when these products would be launched.
Remember, AthlonXP thoroughbred core is 88mmsq i believe. small formfactor athlons are better suited to the mobile market.
However, as the voltage and heat properties of Hammer are not yet public knowledge, I do not know how suited Amd's new processor is to mobile computing.
I'm sorry, but I doubt that Amd will be able to capture 40% of the world's desktop market, and 60% of the world's notebook market.
With Dresden at full capacity producing AthlonXPs, Amd would have enough for about 30% of the world's desktop market. However, Dresden is not up to full capacity yet, and Amd is scaling down their K7 production, getting ready for Their AthlonPro64 (Clawhammer) and Opteron (Sledgehammer) CPUs.
Amd still has a problem with the heat production of their CPUs. I don't know about their new mobile AthlonXPs with Small Form Factor (SFF), but Amd laptops are still a rarity, atleast in the North American market. Hopefully that will change, but I don't see it.
Where did you get your info about 90nm process Athlons? To my knowledge, Amd just partnered up with UMC and some other companies to research 90/65nm fabrication technology, but I would say it's far too early to speculate as to when that will go into full production.
It is going to take UMC quite some time to work out the kinks and get their Athlon CPU lines in full swing. Don't expect too much from them until the second half of 2003.
I sincerely doubt that Intel's management is going to be forced out by irate stockholders.
And as for IA64. I hope that Intel continues researching and developing processors for this technology, as I really want X86 to die. Amd's X86-64 will hopefully provide say, 5-7 years more life for the x86 architecture, and hopefully by then the elegance and beauty that is EPIC will have matured enough to be realized fully in the mainstream market.
I do not know enough about the flash market to speculate on Amd or Intel's fortunes with it. However, I do hope that it returns to being the cash cow for both Intel and Amd because they need more revenue in order to develop and sell processors:)
the performance that the 800mhz clawhammer 256k l2 cache had...that was in pure 32bit mode. The hammer cpus can run in pure32, "compatibility" mode where both 64bit and 32bit registers are utilized, and in pure 64bit mode, doubling up some of the 32bit registers.
All programs need is a recompile to use the 64bit instructions. That's it.
Actually, the inclusion of 64bit space in the cpu does not do much to improve its performance. The performance boost of roughly 40% (so far in quake3) comes from the ondie memory controller and some other tweaks, like better branch prediction.
Games are NOT written to the lowest common denominator. Quake, Quake2, Quake3, Doom3, Max Payne, Giants, Black & White, etc etc. Some games are written to the lowest common, most aren't.
The inclusion of 64bit space was meant for servers, and future programs which could effectively utilize the registers. It also gives AMD bragging rights over Intel (hey, we have a 64bit desktop cpu, you don't, nya nya!)
just my 2C
According to all the reports I've read around the 'net, this seems to be the most likely situation concerning AMD and Intel in the coming year:
Q2 2002: Amd is set to announce the "thoroughbred" AthlonXPs , at 2200+ (1.8GHz) and possibly up to 2500+ (2.0GHz). This seems to be confirmed by some European/Asian online vendors now listing these speed grade Athlons for pre-order.
Sometime later during 2002: Amd is set to discontinue the morgan (duron) CPU, and replace it with low speed AthlonXPs, under the Duron moniker.
Q4 2002: Amd paper launches Clawhammer, probably named as Athlon Pro64, as well as Sledgehammer, under the name Opteron. Both CPUs will be made at the Dresden FAB, using 130nm with SOI process and the metal die slug, and should have impressive thermal performance compared to the AthlonXP cpu. The Clawhammer cpu will possibly start at 1.6GHz-2.0GHz for a PR rating in the range of 3000+ to 3400+. This cpu will have an ondie 64bit DDR333 bus and memory controller. It will also feature Intel's SSE2 SIMD extensions, and utilize a 700ish pin socket (not certain as to exact pin count). As for Sledgehammer, this CPU will have the aforementioned features of the Clawhammer, but use a 900ish pin socket, and have a dual-channel (128bit effective) DDR333 memory controller. Due to the packaging complexity and the dual channel memory controller, this CPU should command a higher price than the Sledgehammer.
Amd is also releasing a new chipset with Hammer. The Amd8000 series chipset will include an AGP3.0 (AGP8X) "tunnel" as well as a PCI "tunnel" and a conventional southbridge "tunnel". The reason for this naming convention is that the Hammer bus is a point to point protocol, and that the conventional "northbridge" does not exist in this platform, as it's main purpose of serving as a RAM controller, is now taken by the Hammer cpu.
Expect the Hammer CPUs to not become widely available until Q3/Q4 2003.
Now, as for Intel...
October 2002: Intel is expected to launch the last of the P4 "northwood" cpu series processor steppings, at 2.8GHz, utilizing a 533MHz QDR bus, much like the new 2.26GHz and 2.53GHz P4s.
In early 2003, Intel should paper launch the "Prescott" Pentium 4, with widespread availability of this cpu no earlier than late early Q4. This cpu will utilize a 667MHz bus (166MHz quadpumped), and will be produced in 90nm process, using the metal slug that the current P4 is equipped with. This cpu will have "Prescott new instructions", possibly named SSE3, but otherwise known as Symmetric Multithreading. Basically, the cpu behaves as if it were 2 independant processors, and can execute different threads and instructions simultaneously (this is a very poor explanation of SMT I know).
"Prescott" will be paired with a new motherboard chipset, utilizing Dual channel DDR333 for an effective 5.3GHz/s bandwidth. As well, this chipset should support Serial ATA as well as USB2.0, and (possibly, unconfirmed as of yet) AGP3.0, aka AGP8x support.
Whew, that's pretty much all the info I have on the near future of Amd and Intel.
Slashdot Mirror
What this says is that SySMark is really poorly coded, not "optimized" to favour Intel silicon. Incompetancy isn't evil. It's just...incompetant. This would explain why most serious benchmark runs seem to lack SysMark these days...
For a year and a half AMD managed to blow Intel out of the water in the performance arena, from the time they both reached 1.0GHz with the P3 and Athlon, to the time Intel started releasing Northwood core P4s with the extra 256KB L2 cache. The longer pipeline of the P4 has nothing to do with performance. All it does is enable the processor to ramp in clock frequencies easily. A general rule of thumb is the longer the pipeline, the lower the IPC (instructions per clock/cycle) and the LOWER the actual number-crunching performance.
AMD did the data-mining. Just like a poor fansite, Vans just posted the PR complete with their own pretty graphs made from the numbers AMD ran.
What is also interesting to note about this article is the hints it gives as to Microsoft's future plans for software security. The idea of having independant secured partitions within a computer is not new of course, but it's nice to note that MS is doing *something* about their rather poor security history. Oh GOD, please pray that they don't integrate Outlook Express with the *secure OS* portion of Palladium
Amazingly enough, this one is able to analyze most of the knowledgebase around "Palladium" and boil it down to the more interesting core issues. I would've appreciated a little more insight along the lines of what such a strategy as Palladium does to the role of the PC however. Generally speaking, PCs are multipurpose machines, which are *fully* programmable, and do pretty much whatever you tell them to. They manipulate data in any way *you* the *user* see fit. What Microsoft is attempting with "Palladium" is going to place restrictions about what a PC can do, and leave these restrictions up to the content producers. I won't comment on the stance of the content producer, but I will mention that this is a departure from what has been a central tenet of the computer: "it's yours". The trend seems to be shying away from "it's yours" to "you didn't buy it, you paid us to ALLOW you to use it - in a way we deem appropriate". Of course, "we" being the content producers. Microsoft really doesn't care what we dow with our music and movies - they just don't want the MPAA/RIAA/Legislators breathing down their neck.
what punishment? Microsoft can _not_ "punish" Intel. Who else can they use, that has the fabrification capacity even close to Intel's? No one.
One of the primary tenets of OpenBSD and NetBSD is security, correct? This is just another little bit of bytecode that improves security even more...
Intel's Xscale architecture -should- have impressive performance. However, it seems that circumstances have conspired to keep it from showing its potential. In order to retain binary backwards compatibility, Microsoft has kept their software compiler rather basic, ensuring that it will work for multiple architectures. This lack of optimization also means that any architectural improvements Xscale has over Arm V4 or whatever it's called will not mean anything. Hopefully the history of Xscale will work like netburst architecture's history, where about 1 year after inception, software that makes use of its architecture efficiently (like SSE2 with the P4) will start to appear.
This is true...personally, I consider NPR's case an anomaly in the bigger issue of deep linking.
I can identify with NPR and web businesses concerning the issue of deep linking. The business model of "free content with revenue generatioin via paid ads" is unable to handle deep linking from other websites. By having deep links bypass a website's splash page and index, the struggling website loses views of the paid ads. Since those paid ads are the only things that keep many websites in existence, I can see how they would take exception to deep linking.
On the other hand, maybe this situation is showing just how flawed and vulnerable this business model really is.
What is interesting is how the building block of the internet, IE, hypertext, does not lend itself to commercial advertising in this manner...Methinks someone needs to cook up some profitable web business models, and fast. Paid ads based on page views aren't working.
Considering how corrupt the current broadcast radio industry is, I'm surprised that online broadcasters are able to license any song to broadcast, whenever they want. With conventional radio, stations play what the labels pay them to play. I can really see these new guidelines thinning out the more amateurish broadcasters, and leaving the more polished, better set-up ones intact. Personally, I wouldn't mind having some sort of radio subscription service for my favourite anime stations...
It may well be a situation where IBM would like to provide support, and is able to provide the manpower to making the drivers necessary to make the notebook line Linux compatible. However, one must realize that there is a difference between writing a set of drivers for a product, and actually being accountable for them. My bet is that IBM would rather use its resources in more efficient tasks than supporting such a small user base.
Browsers by default have a history folder that is only what, 15 days long? Websites rarely last longer than, what, 2 years?
The "internet" seems to be a transitory medium. Unlike paper, digital information is intangible, and can be easily wiped and replaced, or edited. A perfect example of this is the way that news sites often take their articles offline once they've been up for a week, or just look at webtracking software, which shows that links are dying faster than ever before.
If this trend continues (and given the current architecture of the 'net, I don't see it changing) then we might have a serious problem. I won't analyze it, but there is definitely something wrong when data is forever lost after existing for such a short period of time.
I did not say anything in my original post about marketshares, I only posted the technologies and products that Intel and Amd would be rolling out in 2002/2003, and likely, when these products would be launched.
Remember, AthlonXP thoroughbred core is 88mmsq i believe. small formfactor athlons are better suited to the mobile market. However, as the voltage and heat properties of Hammer are not yet public knowledge, I do not know how suited Amd's new processor is to mobile computing.
I'm sorry, but I doubt that Amd will be able to capture 40% of the world's desktop market, and 60% of the world's notebook market. With Dresden at full capacity producing AthlonXPs, Amd would have enough for about 30% of the world's desktop market. However, Dresden is not up to full capacity yet, and Amd is scaling down their K7 production, getting ready for Their AthlonPro64 (Clawhammer) and Opteron (Sledgehammer) CPUs. Amd still has a problem with the heat production of their CPUs. I don't know about their new mobile AthlonXPs with Small Form Factor (SFF), but Amd laptops are still a rarity, atleast in the North American market. Hopefully that will change, but I don't see it. Where did you get your info about 90nm process Athlons? To my knowledge, Amd just partnered up with UMC and some other companies to research 90/65nm fabrication technology, but I would say it's far too early to speculate as to when that will go into full production. It is going to take UMC quite some time to work out the kinks and get their Athlon CPU lines in full swing. Don't expect too much from them until the second half of 2003. I sincerely doubt that Intel's management is going to be forced out by irate stockholders. And as for IA64. I hope that Intel continues researching and developing processors for this technology, as I really want X86 to die. Amd's X86-64 will hopefully provide say, 5-7 years more life for the x86 architecture, and hopefully by then the elegance and beauty that is EPIC will have matured enough to be realized fully in the mainstream market. I do not know enough about the flash market to speculate on Amd or Intel's fortunes with it. However, I do hope that it returns to being the cash cow for both Intel and Amd because they need more revenue in order to develop and sell processors :)
the performance that the 800mhz clawhammer 256k l2 cache had...that was in pure 32bit mode. The hammer cpus can run in pure32, "compatibility" mode where both 64bit and 32bit registers are utilized, and in pure 64bit mode, doubling up some of the 32bit registers. All programs need is a recompile to use the 64bit instructions. That's it. Actually, the inclusion of 64bit space in the cpu does not do much to improve its performance. The performance boost of roughly 40% (so far in quake3) comes from the ondie memory controller and some other tweaks, like better branch prediction. Games are NOT written to the lowest common denominator. Quake, Quake2, Quake3, Doom3, Max Payne, Giants, Black & White, etc etc. Some games are written to the lowest common, most aren't. The inclusion of 64bit space was meant for servers, and future programs which could effectively utilize the registers. It also gives AMD bragging rights over Intel (hey, we have a 64bit desktop cpu, you don't, nya nya!) just my 2C
comments, anyone?
According to all the reports I've read around the 'net, this seems to be the most likely situation concerning AMD and Intel in the coming year:
Q2 2002: Amd is set to announce the "thoroughbred" AthlonXPs , at 2200+ (1.8GHz) and possibly up to 2500+ (2.0GHz). This seems to be confirmed by some European/Asian online vendors now listing these speed grade Athlons for pre-order.
Sometime later during 2002: Amd is set to discontinue the morgan (duron) CPU, and replace it with low speed AthlonXPs, under the Duron moniker.
Q4 2002: Amd paper launches Clawhammer, probably named as Athlon Pro64, as well as Sledgehammer, under the name Opteron. Both CPUs will be made at the Dresden FAB, using 130nm with SOI process and the metal die slug, and should have impressive thermal performance compared to the AthlonXP cpu. The Clawhammer cpu will possibly start at 1.6GHz-2.0GHz for a PR rating in the range of 3000+ to 3400+. This cpu will have an ondie 64bit DDR333 bus and memory controller. It will also feature Intel's SSE2 SIMD extensions, and utilize a 700ish pin socket (not certain as to exact pin count). As for Sledgehammer, this CPU will have the aforementioned features of the Clawhammer, but use a 900ish pin socket, and have a dual-channel (128bit effective) DDR333 memory controller. Due to the packaging complexity and the dual channel memory controller, this CPU should command a higher price than the Sledgehammer.
Amd is also releasing a new chipset with Hammer. The Amd8000 series chipset will include an AGP3.0 (AGP8X) "tunnel" as well as a PCI "tunnel" and a conventional southbridge "tunnel". The reason for this naming convention is that the Hammer bus is a point to point protocol, and that the conventional "northbridge" does not exist in this platform, as it's main purpose of serving as a RAM controller, is now taken by the Hammer cpu.
Expect the Hammer CPUs to not become widely available until Q3/Q4 2003.
Now, as for Intel...
October 2002: Intel is expected to launch the last of the P4 "northwood" cpu series processor steppings, at 2.8GHz, utilizing a 533MHz QDR bus, much like the new 2.26GHz and 2.53GHz P4s.
In early 2003, Intel should paper launch the "Prescott" Pentium 4, with widespread availability of this cpu no earlier than late early Q4. This cpu will utilize a 667MHz bus (166MHz quadpumped), and will be produced in 90nm process, using the metal slug that the current P4 is equipped with. This cpu will have "Prescott new instructions", possibly named SSE3, but otherwise known as Symmetric Multithreading. Basically, the cpu behaves as if it were 2 independant processors, and can execute different threads and instructions simultaneously (this is a very poor explanation of SMT I know).
"Prescott" will be paired with a new motherboard chipset, utilizing Dual channel DDR333 for an effective 5.3GHz/s bandwidth. As well, this chipset should support Serial ATA as well as USB2.0, and (possibly, unconfirmed as of yet) AGP3.0, aka AGP8x support.
Whew, that's pretty much all the info I have on the near future of Amd and Intel.