New Posts  All Forums:Forum Nav:

64-bit Pentium M

post #1 of 17
Thread Starter 
Just some rumors to read while we wait for Dothan.

The rush to 64-bit Pentium Ms?

Quote:
PERSISTENT RUMOURS from people close to Intel engineers are suggesting that the chip firm may have canned its work on Tejas, the chip that is supposed to follow Prescott.

These sources tell the INQUIRER, and as yet this is unconfirmed, that the project has been shelved before tape out, and layout resources are no longer working on it.

The speculation on the semiconductor street is that Intel is frantically attempting to take the Pentium M to the next shrink and move it as rapidly as possible into the pole position.

And the legendary skunk works in Oregon, the same sources suggest, have already designed a 64-bit wide P6/Pentium M core, perhaps dropping the legacy stuff and including 32-bit compatability.

Intel's deskbook CPU platform merger plan

Intel Accelerates Dual-Core Microprocessors Introduction?

Intel plots 4MB L2, 64-bit desktop CPU
post #2 of 17
Thread Starter 
Hmm, I guess it's nearly "official" now...

Intel to 'ditch' Pentium 4 core after Prescott

Quote:
Intel will today tear its Pentium 4 and Xeon roadmaps to shreds and announce the cancellation of its 'Tejas' and 'Jayhawk' processors and their successors.

Both chips represent what was to have been the next generation of 90nm Pentium 4 and Xeon, respectively, and were scheduled to arrive in the Q2 2005 timeframe. Chips derived from them include 'Cedarmill' on the desktop, and 'Potomac' and 'Tulsa' in the Xeon series. The future of these chips is now also in question.

According to Reuters, Tejas and Jayhawk are now history. The report doesn't say what Intel will offer instead, but it seems clear that the chip giant has decided to accelerate plans to roll-out desktop processors derived from its Pentium M architecture.
Intel to Shelve New Desktop, Server Chips (Reuters)
post #3 of 17
P6??? LOL, they should just call it good old P3 x2.
post #4 of 17
Thread Starter 
Intel Scraps Desktop, Server Chip Plan (offical word from Intel)

Quote:
Intel will focus on making chips that contain the cores of two microprocessors -- "like putting two cylinders in a car instead of having one big cylinder," Brookwood said.

Intel plans to introduce dual-core chips for desktop computers in 2005 and plans to start shipments of dual-core chips for notebook computers the same year.
post #5 of 17
hmm, there's rumors about AMD Opteron dual cores drop in replacement with current sockets on the horizon. Could that be.... pushing Intel to move their butts? Thank you AMD, I love the free market competition. I wonder where we'd be if not for AMD, maybe P3 1gig at double the price I bet.
post #6 of 17
Quote:
Originally Posted by exe
P6??? LOL, they should just call it good old P3 x2.
P6 refers to the core generation, not the brand name.
post #7 of 17
The Pentium 3 as well as the PentiumPro, Pentium2 and Pentium-M are all 6th generation chips based on the same core. They are all P6.
The P4 was actually a 7th generation(just not a very good one).
The AMD side uses K to denote generation. K7, K8, etc. The AMD64 is a K8.
post #8 of 17
Quote:
Originally Posted by exe
hmm, there's rumors about AMD Opteron dual cores drop in replacement with current sockets on the horizon. Could that be.... pushing Intel to move their butts? Thank you AMD, I love the free market competition. I wonder where we'd be if not for AMD, maybe P3 1gig at double the price I bet.
Intel's always planned on going dual cores. It was a question of when. AMD's dual core implementations are still a good ways out too.

I don't think Intel's change was so much AMD pushing is that they're hitting a wall in processor design and manufacturing. Prescott is a testament to this and the probelms obviously couldn't be resolved with Tejas. IBM and AMD are also encountering these problems.

http://www.overclockers.com/tips00579/

"Somewhere between 130-nm and 90-nm the whole system fell apart. Things stopped working and nobody seemed to notice. . . . Scaling is already dead but nobody noticed it had stopped breathing and its lips had turned blue." -- Bernie Meyerson, IBM CTO.

All 3 of them are in the same boat. CPUs are we know them are going bye-bye.

Super
post #9 of 17
Superguy, good points all. However AMD is not really in the same boat as Intel on this. They have their top of the line product and will continue to move it forward. They have been quicker to implement things such as SOI, strained silicon and copper interconnects and so while they too will have a problem scaling down 90nm, it is not as important since they are still at 130nm and have more head room there due to their adoption of these other technologies.
The winners are videographers, photographers and programmers which will benifit the most from the move to multiple cores. For gamers, its quite pointless.

Also, I don't think Intel was always planning to go multicore on the desktop. This is a move based on desperation. MHZ sells (which is why they jumped from the p3 to the p4). The idea that multiple cores on a single die came out more or less when they began to realize that that might have problems scaling as well as they had in the past.
With AMD the multi-core thing is pretty much a "me-too" reaction.
post #10 of 17
Thread Starter 
Quote:
Originally Posted by Superguy
http://www.overclockers.com/tips00579/

"Somewhere between 130-nm and 90-nm the whole system fell apart. Things stopped working and nobody seemed to notice. . . . Scaling is already dead but nobody noticed it had stopped breathing and its lips had turned blue." -- Bernie Meyerson, IBM CTO.
Thanks for the link. Good reading and a great quote there.
post #11 of 17
Quote:
Originally Posted by Bjorn
Superguy, good points all. However AMD is not really in the same boat as Intel on this. They have their top of the line product and will continue to move it forward. They have been quicker to implement things such as SOI, strained silicon and copper interconnects and so while they too will have a problem scaling down 90nm, it is not as important since they are still at 130nm and have more head room there due to their adoption of these other technologies.
Yes and no there. Going to copper really didn't help much in the P3. Aluminum was working fine. It needed to be done eventually, but it wasn't a "must do" right then. SOI seems to help and it's something that should be looked into. AMD can thank IBM for both of those techs.

Intel actually was before AMD on strained silicon.

Quote:
Originally Posted by Bjorn
The winners are videographers, photographers and programmers which will benifit the most from the move to multiple cores. For gamers, its quite pointless.
True, but I think that will change in the future. Hyperthreading has already shown that being able to process more than one thread is great. I love it for the smoothness in multitasking. I love the smoother feel.

The problem with HT is that it can't handle two large threads.

Intel's probably going to keep HT with it's multicores, so it has some serious potential. I wouldn't mind 4 threads running on one proc at a time.

Quote:
Originally Posted by Bjorn
Also, I don't think Intel was always planning to go multicore on the desktop. This is a move based on desperation. MHZ sells (which is why they jumped from the p3 to the p4). The idea that multiple cores on a single die came out more or less when they began to realize that that might have problems scaling as well as they had in the past.
With AMD the multi-core thing is pretty much a "me-too" reaction.
I believe Intel was planning it, but not until .65 micron processors came out. It's just now starting to get feasible from a cost point to do so. AMD had a better chance of doing it first, but that's due to their architecture. Hypertransport could glue the procs together much easier than Intel could at the time. So Intel put HT out as a way of boosting things in the meantime.

MHz did sell (don't forget AMD really hammered that (no pun intended ) when they beat Intel to 1GHz), but I think Ed brings up a good point that PCs are to the point where they're fast enough for most users so the focus needs to change.

Multicores were originally meant for Itanium at first. Personally I think since Itanium is a high margin/low production item that using it to be its guinea pig for multicores wasn't a bad idea. They can work out the mistakes there and then have it work well on their bread and butter processors.

Being a "Me too" isn't necessarily a bad thing. AMD has done it in the past and has sometimes done things better than Intel because of it.

Good thoughts. I enjoyed them.

Super
post #12 of 17
Quote:
Originally Posted by Superguy
True, but I think that will change in the future. Hyperthreading has already shown that being able to process more than one thread is great. I love it for the smoothness in multitasking. I love the smoother feel.

The problem with HT is that it can't handle two large threads.

Intel's probably going to keep HT with it's multicores, so it has some serious potential. I wouldn't mind 4 threads running on one proc at a time.
Well, If Intel is going with the Pentium M core, I doubt they will need hyperthreading. HT really only applies to really long pipelines like those present in the current Pentium-4 CPU. Although I don't have any concerte numbers, I believe that (judging from the Mhz and actual performance) P-M has a much shorter pipeline, like those found in Athlon CPU, and HT won't help a bit at all.

Multicore, now that would be sweet. SMP without the price
post #13 of 17
ur right steptix

the pm pipeline is longer than a p3 but shorter then a p4... its no where inefficient like teh p4 pipeline... p4's needed HT to make up for its inefficiency...

-------

the advancement with amd's clawhammer to newcastle and beyond is that theyre getting higher speed by reducing the cache size

existing 3200 3400 clawhammers have 1mb cache and once the newcastle predecessors come out they are halving the cache to 512kb to kick the clockspeed higher

-------

i think intel made a good movie in ditching the p4...
i remember seeing some review on techforums where they tested the new 3.4 presscott and under full load gaming it ran 76c with the case OFF
post #14 of 17
You're right that the pipeline is shorter than the P4's. As Rince said, it's somewhere in between.

I don't necessarily see Intel ditching HT. It's certainly a great concept and it didn't take that much chip real estate to implement.
post #15 of 17
i kinda foresaw this coming from reading anandtechs detailed review and breakdown of banias when it first came out...

http://notebookforums.com/showthread...2&page=2&pp=15
post #16 of 17
Sun has already done this, Their Sparc-IV chip can handle two threads at the same time by basically having two Sparc-III cores on one chip:

http://www.sun.com/processors/UltraSPARC-IV/

Sun's vision called Throughput Computing (which can scale to 4-way and beyond - I saw a note about 32 threads via eight cores):

http://www.sun.com/processors/throughput/

--TSK


Quote:
Originally Posted by Superguy
Intel's always planned on going dual cores. It was a question of when. AMD's dual core implementations are still a good ways out too.

I don't think Intel's change was so much AMD pushing is that they're hitting a wall in processor design and manufacturing. Prescott is a testament to this and the probelms obviously couldn't be resolved with Tejas. IBM and AMD are also encountering these problems.

All 3 of them are in the same boat. CPUs are we know them are going bye-bye.

Super
post #17 of 17
Yup.

Pentium Ms are future. DIE P4 DIE!

Poor bastards who are stuck with a P4 in their laptops
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Notebook Forums - General