• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

The future of Intel

DruSi3r

Gawd
Joined
Nov 4, 2002
Messages
937
Guys I was just wondering what your thoughts were about the future of intel. Specifically, in regards to market share versus amd. Recently, it seems as though Intel has hit some sort of road block. As clock speeds increase, more bandwidth is needed. Is this why intel is switching to a PR rating system ala amd? How is intel, the speed king, going to sell 1.4Ghz Dothans over 2.8Ghz P4s? Are they having trouble implimenting 64bit or just waiting for the OS architechure to catch up? Will there dual core save the day? Will it be too late? Your thoughts.
d
 
well, they might be losing in the enthusiast market, but it would require actual effort on Intel's part to loose majority of their market share to AMD.
 
I fully agree, Intel is huge and wont wont be dethroned easily. The only way i see it happening is if the US Gov steps in somehow. But I want to see some performance improvements. The 915 and 925 chipsets are dissappointing for the moment.
 
How can you blame it on the chipsets? i mean were not looking at a new core here or anything? a 3.2ghz 478 Prescott should be the same as a 3.2ghz 755 prescott - the main changes with the chipset are in thefeatures of the southbridge

As for ddr2, Wasnt there like 3% difference in sdr and ddr1?

As for intel, Yeah i think it will take a act of god to remove them from the king of chips
 
People seem to forget Intel exists outside the P4...you have the P4-M, Xscale, Lcos, network adapters, and on and on.
 
I didn't forget about Intel's existance outside the processor and chip arena. But as for their performance, DDR wasn't any improvment over SDRAM initially. It took time for the chipsets to mature to allow DDR to really shine.

The Netburst architecture hasn't been that great. Intel went with it mostly for marketing reasons. They could ramp the clock speed to insane levels and that sells chips. And now Intel has hit a wall. They can't ramp the clock speeds any higher without increasing the heat past Prescott's rediculous levels. So they've been working on improving the actual architecture of their chips to be more efficient per clock cycle. Using less power. Dothan has now become the logical choice for their desktops. BTW the Dothan is HEAVILY based on the P6 microarchitecture of the Pentium Pro/Pentium II/Pentium III chips.

Those chips designed for Mobile use have become very successful. I've got some Pentium M 1.4GHz Notebooks at work and they scream. For a 1.4 they are quite amazing. It only made sense to fallow suite with the rest of the market and go to model numbers instead of clock speeds. They would never sell Dothan's if they sat on the shelf next to Prescott's and all people saw was the clock frequencies.

Again it's about marketing. AMD and Apple have been fighting that misconception of clock speed is all that matters for some time. Now Intel has to do it to sell their Pentium M architecture.

But the Dothan is pretty impressive. a 2.2GHz Dothan is about as fast or faster than a 3.0GHz Pentium 4 Northwood. Not bad in my opinion.
 
I haven't had the pleasure of playing with the Pentium M's yet, unfortunately. Both my systems are running Northwood cores. And I have a feeling that Intel knows they've hit a road block but are just kinda pushing it ahead of them as they also make a sharp right hand turn and pave a completely new road for themselves. We're not gonna see huge performance increases right off the bat. I don't think we have with very many things as soon as the roll out. But as the technology matures some and actually gets implemented and becomes commercially available, it's gonna get alot better. We all know that the Prescott is very good at certain things and for others, it sucks hind tit. I think Intel knows that too and that's why they scrapped Tejas in favor of going toward dual-core. I believe that they wouldn't have scrapped it in the first place if they didn't already have an alternate well in hand. I think that the change over to LGA775 was just a beginning step on the way toward Intel's dual cored solutions. I for one am kinda excited to see how well they perform once they hit the streets. But, that's not gonna be for at least another 9 months to a year. I also think that when they release the 925XE chipset toward the end of this year, it's gonna be yet another step along the road to dual core.....

Just an odd question, while I'm at it, what FSB speeds are the Pentium M's running at? The reason I ask is b/c I've heard that the 925XE's are supposed to introduce the 1066MHz FSB.... Gonna be an interesting trip on the Intel side of the camp, that's all I can say.
 
I read somewhere intel announced they are getting in to the display market. Well not exactly they are making a new chip to compete with TIs DLP chip. Are they worried? Or just diversifying...
 
Dothan looks promising, but still isn't as effecient as an A64. A 2.2GHz Dothan is about as good as a 3.0C? Well an A64 running at 2.2GHz is about as good as 3.4C, and even 1 or 2% better. Wonder if they could somehow increace the IPC, and ramp it's effeciency instead?
 
DruSi3r said:
I fully agree, Intel is huge and wont wont be dethroned easily. The only way i see it happening is if the US Gov steps in somehow. But I want to see some performance improvements. The 915 and 925 chipsets are dissappointing for the moment.

it's not about dethroning or any of this black/white stuff - it IS about market share and i have to say AMD made some really clever moves with the introduction of Opteron - and man was that an entry into the server market! If i was to buy shares of any of the two it surely would be AMD - I'd bet my money on AMD gaining quite some market share in 2004, while, of course, intel is not going to be dethroned anytime soon.
 
Ummm... AMD didnt start the PR system, IBM (Cyrix) was doing it back in the 166MHz CPU days.

Im sure that someone did it before them too.

Kroz
 
Kroz said:
Ummm... AMD didnt start the PR system, IBM (Cyrix) was doing it back in the 166MHz CPU days.

Im sure that someone did it before them too.

Kroz

Actually no, it did start with Cyrix/IBM. Cyrix/IBM and AMD were all partners in the PR rating system. All thier processors fallowed it. It was for the same reasons too.

Cyrix's 6x86 PR200+ was actually 150MHz and kicked the Pentium's ass in everything but FPU/Integer performance. (Games) So AMD and Cyrix/IBM played the Megahertz don't matter game becuase they couldn't beat Intel at clock frequencies.
 
DruSi3r said:
Are they having trouble implimenting 64bit or just waiting for the OS architechure to catch up?
d

This is definitely not an issue...as Intel has had 64-bit CPU's out far longer than AMD has (itanium and itanium 2) If you had a use for them, and could afford them they are doing awesome in the business market where exteremly large amounts of bandwidth is used. IMO Intel took to HyperThreading as their big thing...they're riding that out until there is a need for 64 bit computing. Right now theres very few OS's that are running good 64 bit, and even less speed increase from them. The Athlon64's are fast yes....but that has absolutely NOTHING to do with being 64 bit. AMD just re-did 32-bit and did it very well. I think Intel has quite a few tricks up their sleeve and by the end of the year we'll see some pretty cool stuff from them.
 
M4d-K10wN said:
Dothan looks promising, but still isn't as effecient as an A64. A 2.2GHz Dothan is about as good as a 3.0C? Well an A64 running at 2.2GHz is about as good as 3.4C, and even 1 or 2% better. Wonder if they could somehow increace the IPC, and ramp it's effeciency instead?

Actually I believe that I read on some site(anandtech or tomshardware) that clock for clock, the dothans beat the athlon 64. voodoo's website said that the pentium m's performed around that of a pentium 4 1.9 times as fast, thus making the 2.2GHz Dothan about as fast as a 4GHZ p4, If voodoo's correct.
 
bountyhunter said:
Actually I believe that I read on some site(anandtech or tomshardware) that clock for clock, the dothans beat the athlon 64. voodoo's website said that the pentium m's performed around that of a pentium 4 1.9 times as fast, thus making the 2.2GHz Dothan about as fast as a 4GHZ p4, If voodoo's correct.



no, thats not true, i was hearing that 1.6ghz Banias' were just as fast as 2.4ghz P4-Ms, at the time of Banias' release, 2.4ghz was the fastest P4-M chip. At least if memory serves me right...
 
M4d-K10wN said:
Dothan looks promising, but still isn't as effecient as an A64. A 2.2GHz Dothan is about as good as a 3.0C? Well an A64 running at 2.2GHz is about as good as 3.4C, and even 1 or 2% better. Wonder if they could somehow increace the IPC, and ramp it's effeciency instead?
EXCUSE ME? My 3.0c @ 3.6 beats the crap out of the entire A64 line.
 
rayman2k2 said:
no, thats not true, i was hearing that 1.6ghz Banias' were just as fast as 2.4ghz P4-Ms, at the time of Banias' release, 2.4ghz was the fastest P4-M chip. At least if memory serves me right...

that's impossible, because I play farcry better with a p4m banias 1.6 then I did on my p4 2.5, and the 2.5 had more RAM at it's disposale.
 
bountyhunter said:
that's impossible, because I play farcry better with a p4m banias 1.6 then I did on my p4 2.5, and the 2.5 had more RAM at it's disposale.
thats still not a controlled scientific comparison of both processors
 
Yule Bryner said:
thats still not a controlled scientific comparison of both processors

so, it still means that the 1.6banias wipped the 2.5 northwood with twice the RAM
 
Yule Bryner said:
EXCUSE ME? My 3.0c @ 3.6 beats the crap out of the entire A64 line.

Sorry, That statement does not make sense without a definition of the CPU/user -specific workloads.

I only use GIMPS and NFSNET Software for DC leisure-time but even with that small selection of workloads it's clear that GIMPS shines with the P4s and NFSNET with AMD64s.
 
The main problem with Pentium M Dothans right now is the lack of desktop support with a modern dual channel chipset, along with still running on a 100 fsb. I am more than impressed with the performance of my Dothan 2.0 and it's performance, especially seeing that it is running in a laptop with no real tweak options in bios and is using the i855 chipset instead of something more modern like i865 or i875. There is no reason why they couldn't be run on either chipset though as they are electrically the same as P4, buswise. I think that if a mobo manufacturer would come out with a desktop mobbo that is set up with a socket compatible with P-M that there would be some big surprises in store for a lot of people. I've been folding now with my new laptop for around 3 weeks or so and it's produced in the neighborhood of 800 points/week average, which is better than what my mobile AXP2600 at 2.5 produces. Plus, it still runs cool even when folding; no small feat for a laptop.

BTW, does anyone know how Speedstep is supported? It came to me today that these P-M procs are downward unlocked on the multiplier in order for them to use Speedstep while on battery. Does it have to be supported by actual circuits in the chipset itself or is it just supported by settings in the bios? If it's just supported by bios settings, then it would be possible for the lower multipliers to be implemented using the 865 and 875 chipsets, which would then allow you to run a high fsb speed for more memory bandwidth if someone would make a desktop board for the processors.

P-M would make an ideal SFF proc due to it's very low TDP of around 21 watts with my 2.0. And I would say that it's IPC is at least on par with A64 and superior to AXP right now on the 100 fsb.
 
bountyhunter said:
Actually I believe that I read on some site(anandtech or tomshardware) that clock for clock, the dothans beat the athlon 64. voodoo's website said that the pentium m's performed around that of a pentium 4 1.9 times as fast, thus making the 2.2GHz Dothan about as fast as a 4GHZ p4, If voodoo's correct.
The architecture of the Pentium-M and the Athlon64FX is very similar. While is it possible, that seems like a stretch. The L2 cache of the Pentium-M would have to help is a lot for it to outperform a 4.4 GHz Pentium 4.

Tweak Monkey said:
This is definitely not an issue...as Intel has had 64-bit CPU's out far longer than AMD has (itanium and itanium 2) If you had a use for them, and could afford them they are doing awesome in the business market where exteremly large amounts of bandwidth is used. IMO Intel took to HyperThreading as their big thing...they're riding that out until there is a need for 64 bit computing. Right now theres very few OS's that are running good 64 bit, and even less speed increase from them. The Athlon64's are fast yes....but that has absolutely NOTHING to do with being 64 bit. AMD just re-did 32-bit and did it very well. I think Intel has quite a few tricks up their sleeve and by the end of the year we'll see some pretty cool stuff from them.
First of all, many people call the Itanium/2 the itanic because no one is really supporting the software. The lack of 32 bit extensions for the Itanium chip as well as its ability to run very well solely if the software is based on the epic architecture are stumbling blocks that hamper widespread usage.

Sir-Fragalot said:
I didn't forget about Intel's existance outside the processor and chip arena. But as for their performance, DDR wasn't any improvment over SDRAM initially. It took time for the chipsets to mature to allow DDR to really shine.

The Netburst architecture hasn't been that great. Intel went with it mostly for marketing reasons. They could ramp the clock speed to insane levels and that sells chips. And now Intel has hit a wall. They can't ramp the clock speeds any higher without increasing the heat past Prescott's rediculous levels. So they've been working on improving the actual architecture of their chips to be more efficient per clock cycle. Using less power. Dothan has now become the logical choice for their desktops. BTW the Dothan is HEAVILY based on the P6 microarchitecture of the Pentium Pro/Pentium II/Pentium III chips.

But the Dothan is pretty impressive. a 2.2GHz Dothan is about as fast or faster than a 3.0GHz Pentium 4 Northwood. Not bad in my opinion.
Well the reason why Intel didn’t want to keep its P6 architecture wasn’t only for marketing. They were actually too lazy to shrink the die to the 130 nm process. Somehow AMD was able to ramp up clock speeds, but Intel wasn’t. Their logic was 20% less efficient, 40% faster. Personally I feel that if the Pentium-M was based on the NetBurst architecture (just with a shorter pipeline), it would be better then it is now. The NetBurst’s sole goal is an attempt to regain IPC lost by the lengthened pipeline. Thus the rapid execution engine and trace cache were born. The trace cache is basically memory that sits between the execute and decode stages in the pipeline so when it flushes, you can get the data back a lot quicker. Those two factors and a shorter pipeline would have made the Pentium 4 the greatest processor for the time.

PhoenixAshes said:
I haven't had the pleasure of playing with the Pentium M's yet, unfortunately. Both my systems are running Northwood cores. And I have a feeling that Intel knows they've hit a road block but are just kinda pushing it ahead of them as they also make a sharp right hand turn and pave a completely new road for themselves. We're not gonna see huge performance increases right off the bat. I don't think we have with very many things as soon as the roll out. But as the technology matures some and actually gets implemented and becomes commercially available, it's gonna get alot better. We all know that the Prescott is very good at certain things and for others, it sucks hind tit. I think Intel knows that too and that's why they scrapped Tejas in favor of going toward dual-core. I believe that they wouldn't have scrapped it in the first place if they didn't already have an alternate well in hand. I think that the change over to LGA775 was just a beginning step on the way toward Intel's dual cored solutions. I for one am kinda excited to see how well they perform once they hit the streets. But, that's not gonna be for at least another 9 months to a year. I also think that when they release the 925XE chipset toward the end of this year, it's gonna be yet another step along the road to dual core.....

Just an odd question, while I'm at it, what FSB speeds are the Pentium M's running at? The reason I ask is b/c I've heard that the 925XE's are supposed to introduce the 1066MHz FSB.... Gonna be an interesting trip on the Intel side of the camp, that's all I can say.

The Pentium-M’s FSB is at 400 MHz, although intel is going to push it to 533 MHz for an increase performance increase. Its not that they’ve hit a roadblock, but more like the industry has hit a road block. The 90 nm process results in more leakage then before and thus the power consumption for the Prescott core shoots up. The fact that Intel used Strained Silicon didn’t help the leakage problem either. Its been said that the increase in the pipeline length is attributed to the excessive heat output from the transistor gate leaking. Intel’s tri-gate technology should fix that.
 
DruSi3r said:
I read somewhere intel announced they are getting in to the display market. Well not exactly they are making a new chip to compete with TIs DLP chip. Are they worried? Or just diversifying...

diversifying keeps you a float... :D
 
Yule Bryner said:
EXCUSE ME? My 3.0c @ 3.6 beats the crap out of the entire A64 line.
in what?? my p4 cant hang with my 64 in games but. when multi tasking the p4 really hands the 64 a beating. for gaming i use a athlon3200+64@2450mhz, i use the p4 for work.
 
I buy mostly AMD chips.
I buy mostly Intel stock. :p

This is because Intel has too much power! They will not lose significant market share unless they really try. And having an exclusive deal with Dell helps too. Intel is trusted and the average computer knows about Intel but does NOT know about AMD. I get "What's an AMD?" and I answer, so they respond with, "So it's a bootlegged Intel chip?".

Intel is a juggernaut with lots of money and a stable future.
 
M4d-K10wN said:
Even an AFX @ 3.6GHz?
I would like to see an Athlon FX @ 3.6GHz, but i'm pretty sure the heat would be a big problem.

As for Intel's future, the new chipsets aren't too impressive yet. They do have a lot of features that were re-done. As soon as Intel improves their CPU's and the preformance is noticeable, I think they'll stay at the top. I personally would jump to either company when my next upgrade is needed.
 
blinkdev said:
I would like to see an Athlon FX @ 3.6GHz, but i'm pretty sure the heat would be a big problem.
Around 175W with only a 0.25v bump isn't too bad, is it? :p

------------------------------------
DruSi3r said:
Guys I was just wondering what your thoughts were about the future of intel.
I don't think much will change in the next 6 months, or at least until WindowsXP 64-bit gets mainstream support. And even then it might not be much of a problem if Intel releases desktop EM64T chips at semi-reasonable prices (close to current NW/Prescott prices). If Intel does lose marketshare (I expect they will a little), it should rebound like it has in the past.

Ditching NetBurst wasn't a sign of weakness. Intel probably could have kept releasing chips through the 130-140W range, but that would pretty much be a dead end sooner than later. Moving up dual core CPUs by almost a year was a better strategy.

One thing to keep in mind is that the corporate market is much larger than the home and enthusiast markets. Service and support, along with acceptable speed, sell a lot more systems than just having a good CPU available. Witness that the Athlon XP got virtually no corporate marketshare. The A64 will probably be different, but the slow pace of building up the whole 64-bit infrastructure (read: 64-bit drivers and 64-bit software) will give Intel ample time to release a 64-bit desktop chip on their own schedule, and minimize marketshare decay.
 
pxc said:
Ditching NetBurst wasn't a sign of weakness. Intel probably could have kept releasing chips through the 130-140W range, but that would pretty much be a dead end sooner than later. Moving up dual core CPUs by almost a year was a better strategy.
Some guy from Intel said that we need a new approach to processor architecture. The leakage thats occuring w/prescott is serious. When you shrink a die, the temps are supposed to go down. They do have tri-gate which will fix the leakage really well, but they dont use it. So, if they did use the tri-gate transistor, temps would drop and they could continue ramping their clock speeds to unheard of speeds. I think if they were to incorperate parts of the NetBurst architecture into its future chips that would be really good. It would be interesting to see a dual core Pentium-M based chip w/ the Rapid execution engine, trace cache (had to have been tweaked cause it uses a moderate amount of power) and HT. I think if they were all used, AMD would be screwed and Intel would yet again take the lead. Also, overclockers.com thinks the Prescott is overvolted already. They cound a paper stating that the default vcore for prescott should be 1.2 volts.
 
blinkdev said:
I would like to see an Athlon FX @ 3.6GHz, but i'm pretty sure the heat would be a big problem.

As for Intel's future, the new chipsets aren't too impressive yet. They do have a lot of features that were re-done. As soon as Intel improves their CPU's and the preformance is noticeable, I think they'll stay at the top. I personally would jump to either company when my next upgrade is needed.
Your wish can be granted. Pm DaveX, and ask him to link to the thread at extremeforum. The FX-53 SledgeHammer (Opteron based, s940) was multiplier overclocked, with dual cascade cooling, running at -111˚C. I posted that, however, to invalidate Yule Bryner's argument. Furthermore, there are AMD64 chips that can beat your overclocked 3.0c,while staying at stock themselves. Such as A64 3800+, or s939 FX-53.
 
Back
Top