How do M processors compare?

I

ikellensbro

[H]ard|Gawd
Joined
Feb 14, 2004
Messages
1,410
I'm thinking of possibly getting a laptop :D , & most of the new ones have the M chips so:
How do they compare in performance with older chips?
And what exactly do Mobile chips do?????????????????????? :confused:
 
Ignoring the must better A64 laptops, theyre the best mobile procs you can get. Run cool, use low power. Theyre based off the PIII, surprisingly enough, not the P4.
 
The Pentium M is sort of a hybrid cross between the P3 and P4, with it using an advanced data prefetch and a fsb electrically compatible to the P4 and using a short pipeline like the P3 procs. Like kronchev said, they are a very good choice for a mobile proc. IMO, they are a better choice than A64 due to their much lower power consumption and heat than A84, but both are worlds better in that regard than the P4 architecture for mobile use. I'm eagerly waiting for my new Dothan 2.0 laptop to come in and get some first hand experience with this fast a Pentium M.
 
Do these work in desktops like the xp-m's? Just out of curiosity

[sneaks back over to AMD forum]
 
No, unfortunately. The pinouts on them are different than a P4 proc. But can you imagine how nice one of these would be in a SFF box like a shuttle though? It would make it so much easier to keep that box running cool.
 
so far, no.........

If I'm not mistaken the Pentium M's are about equivelant to a pentium 4 clocked at 1.9 times the speed of the questionable Banias cpu, but I'm not sure how that works with Dothans, though I assume they are better. and the pentium M's clocked at the same speed as an athlon 64 will outpreform the a6f4.
 
I don't know the actual ratio. The Pentium-M line is faster than the Pentium 4. But the Dothan and the Banias are VERY similar. The Dothan has more cache memory and the Dothan has more transistors than even Prescott. This is partially due to cache memory and partially due to architectual differences.

The Dothan and Banias are superior clock for clock to an Intel. Arguably they are much much closer to AMD. However Intel doesn't have the ability to ramp up the clock speed on these CPU's the same way they do with Northwood and Prescott.

So we won't see a full transition to Dothan based chips until Intel gets everyone going on their whole model number scheme as opposed to the clock frequency.

Architectually speaking they combine the best things about the PIII and Pentium 4 lines. There are a few things taken from the P4 core. But it does more closely resemble a Pentium III's inner workings. Clock for clock however Dothan is only like 5% faster than Banias. This is mainly due to double the L2 cache I would think.
 
Benny Blanco said:
Do these work in desktops like the xp-m's? Just out of curiosity
Click to expand...

At least in Europe they do. However, boards for the Socket 479M are very expensive and most are laid out for HTPC uses. The Pentium-M ITX boards gotta be nice though. Oh well if I win the lottery...
 
ohhh i wish they were socket 478. Imagine how much you probably could oc them
 
There are no problems architecturally with having the mobo manufacturers adapt present boards to Pentium M AFAIK, besides wondering if there would be enough sales to make it profitable for them to do so. As far as ramping clock speed on them, I read an article where someone took an Asus notebook and changed some stuff on the mobo where they could change the fsb speed on it and they were able to overclock the Dothan 2.0 to 2.4 GHz. That looks good to me, with their clock speed being able to ramp up that much already with the first release silicon already showing that much headroom.

There are a few ITX mobo's out there that are made for Pentium M but the cost on them is outrageous (>$400) due to them not being a mass produced design. I think that we should start sending email to the major motherboard makers expressing our interest in motherboards using the Pentium M. Maybe someone like Abit or Shuttle would listen as the Pentium M seems to be tailor-made for a SFF machine like Digidice or the XPC boxes. The Pentium M would also seem to be an ideal choice for blade servers too, due to their IPC efficiency and low power requirements.
 
Mwarhead said:
I have a 1.6 ghz pentium m. Find that it is much faster then my Athlon XP 1700.
Click to expand...

And I find that to be no surprise at all :cool:

Pentium M 1.6's are usually compared to P4 2,2's...
 
Pentium M's kick ass. I could have sworn I saw a desktop version of them on newegg.com. Don't know if it would work in a socket 478 board though.
 
Sir-Fragalot said:
Pentium M's kick ass. I could have sworn I saw a desktop version of them on newegg.com. Don't know if it would work in a socket 478 board though.
Click to expand...

Sorry, that would really be great, and it is probably in intel's future, but right now there is no such thing. Pentium M is for Socket 479M and for that only. However, that isn't such a bad thing as nothing will keep Mobo Manufacturers from mass-producing Socket 479M desktop-boards sooner or later...
 
muddocktor said:
No, unfortunately. The pinouts on them are different than a P4 proc. But can you imagine how nice one of these would be in a SFF box like a shuttle though? It would make it so much easier to keep that box running cool.
Click to expand...

mc_P said:
ohhh i wish they were socket 478. Imagine how much you probably could oc them
Click to expand...


http://www.newegg.com/app/ViewProdu...frcode=0&propertycode=&propertycodevalue=5170
:confused:
Model: Intel Pentium M
Core: Dothan
Operating Frequency: 1.8GHz
FSB: 400MHz
Cache: L1/32K+32K; L2/ 2MB
Voltage: 0.956V-1.052V
Process: 0.09Micron
Socket: Socket 478***************
Multimedia Instruction: MMX, SSE, SSE2
Warranty: 3-year MFG
Packaging: Retail box (with Heatsink and Fan)
 
And you are both wrong. Right on page 8 of the linked .pdf:
The Intel® Pentium® M processor utilizes socketable Micro Flip-Chip Pin Grid Array (Micro-FCPGA) and surface mount Micro Flip-Chip Ball Grid Array (Micro-FCBGA) package technology. The Micro-FCPGA package plugs into a 479-hole, surface-mount, zero insertion force(ZIF) socket, which is referred to as the mPGA479M socket.
Click to expand...

It's only a one pin difference, physically, but the pinouts aren't even close.

And yes, they do describe the socket in another place as being 478 pins, but the physical layout of that socket does NOT match the physical layout of the Pentium 4's Socket 478. Take a look at page 37, Pin B2 exists on the P4's socket but not on the P-M.
 
janus said:
Sorry, that would really be great, and it is probably in intel's future, but right now there is no such thing. Pentium M is for Socket 479M and for that only. However, that isn't such a bad thing as nothing will keep Mobo Manufacturers from mass-producing Socket 479M desktop-boards sooner or later...
Click to expand...

i was told PERSONALLY by a marketing guy at intel that future Pentium M's are going to be compatiable with desktop boards :)
 
By the tiem we see mass production of desktop mobos they'll be running dual banias cores hopefully. And I'll take 3...
 
will desktop dothans ever be available for my socket? if so i might stick with my 2.4b longer than planed :)
 
Knowing Intels policy (for every new cpu a new socket/mobo revision) you wont be able to put a "new" cpu in todays mobos.
 
I guess our only real hope to run a Pentium M in a present desktop board would be if someone like Upgradeware or Powerleap would come out with a socket adapter, much like they've done for Tualatin and socket 423 P4 and I don't have the techical knowledge to know if that is even doable. We do know the bus is electrically compatible with Pentium M but just remapping the pinouts would be a nightmare, I imagine.
 
ufokillerz said:
i have 3x p4-mobile 1.6ghz and they work fine in my 478pin boards
Click to expand...
That works, but the Pentium M is different. It is electrically compatible on a signalling level, however the pin arrangement is different. Look at the datasheets, its like Intel took all of the pins and moved them over diagonally one row.

Yea, I can see it working with an adaptor, but Powerleap would have a helluvatime... I don't think there is a single pin that matches. I guess it would be on par with a 423 to 478 adaptor... at least in the case of the Pentium M to P4 socket there is the same number of pins... dunno about same number of pins per function...
 
You must log in or register to reply here.
Back
Top