TR Forums

I've recently seen some Pentium M 715a processors in laptops.

What's the difference between a 715a and a regular 715? They're both 1.5GHz, 2MB L2, 90nm, 400MHz FSB chips.

I believe that the 715A supports the XD bit, while the 715 doesn't.

If that's the case, why aren't they advertising "Centrino laptops with Pentium M 715A processors have integrated virus protection* (or something to that effect, and not the same as what AMD uses)"?

(The * is to say somewhere in a mess of footnotes that you should still have AV software, of course)

Now I'm curious about this myself. I looked around the Web a bit to see if I could find what the difference was, but I couldn't find anything differentiating the 715 from the 715A. The Pentium M 715 isn't on Intel's price list anymore (the slowest regular-voltage Pentium Ms still on there are the 725 and 730, both 1.6GHz but having 400 and 533MHz FSBs, respectively), so that's no help. I don't see how it could be a bus speed difference because 1500MHz doesn't divide into 133 or 66MHz for a good multiplier, and I can't imagine that they would tack the "A" on there just to indicate a newer stepping of the chip, since if they both have 2MB L2 cache they must be Dothan cores. I actually do sort of know an Intel employee who works on processor design (he claims to have managed the L1 cache design team on Prescott and last I saw him he said he was working on Nehalem)- he's the husband of my sister-in-law's best friend. I'll see if I can ask him, and even if he doesn't know, I imagine he could ask someone who does.

bhtooefr wrote:

If that's the case, why aren't they advertising "Centrino laptops with Pentium M 715A processors have integrated virus protection* (or something to that effect, and not the same as what AMD uses)"?

(The * is to say somewhere in a mess of footnotes that you should still have AV software, of course)

They're backing off from saying statements like 'integrated virus protection' with regards to XD / NX bit because it's relatively easy for a virus to bypass the NX / XD protections. It may stop some older viruses, but it won't be too long before the virus writers work out how to mod their shellcode to bypass NX.

I just got an email reply from Alan, my sister-in-law's best friend's husband who is an engineer at Intel. Sadly, he said he didn't have a clue what the difference was, and didn't even know whom to ask about it, because that's a marketing thing and they're far separated from the marketing people. He did say he's still working on Nehalem, though he didn't provide any other details than that.

Evan wrote:

I just got an email reply from Alan, my sister-in-law's best friend's husband who is an engineer at Intel. Sadly, he said he didn't have a clue what the difference was, and didn't even know whom to ask about it, because that's a marketing thing and they're far separated from the marketing people. He did say he's still working on Nehalem, though he didn't provide any other details than that.

So that means that Nehalem wasn't canned? I was under the impression that it was considering that Intel is moving away from Netburst to a new architecture designed on the same principles of the Pentium-M.

I guess this means that we can expect to see 10GHz in 5 years or so.

The plot thickens...

Looking on Intel's Processor Spec Finder, there are three sSpecs for a 1.5GHz Pentium M:

SL7GL, which is a 90nm part with 2MB L2, and in a PGA package. The Spec Finder also says (when you click the sSpec) that "This is an Intel Pentium M 715."
SL6F9, which is a 130nm part with 1MB L2, and in a PGA package.
SL6F6, which is an SL6F9, but in a BGA (read: permanently installed on mobo) package.

I'm currently pulling up the Spec Update on the Dothan...

It also lists an SL7GK for the 715 number. Said SL7GK is the same stepping as the SL7GL - so, it's probably the BGA version.

No mention of the 715a, though.

Edit: After more googling, I've found that the 725, 735, 745, 720, 730, 740, 750, and 760 are also available in A variants...

This is kind of interesting. Intel has a page on their website which lists all of their processors:

Intel's processor list

but they don't list any of the a variants.

Intel's price list wrote:

Intel® Pentium® M processor May '05 (05/01) June '05 (06/12) % Decrease
Mobile (uFCBGA / uFCPGA) Price Price
770 (2M L2 cache 2.13 GHz 533 MHz FSB 90nm) $637 $637 -
765 (2M L2 cache 2.10 GHz 400 MHz FSB 90nm) $637 $637 -
760 (2M L2 cache 2A GHz 533 MHz FSB 90nm) $423 $423 -
755 (2M L2 cache 2 GHz 400 MHz FSB 90nm) $423 $423 -
750 (2M L2 cache 1.86 GHz 533 MHz FSB 90nm) $294 $294 -
745 (2M L2 cache 1.80 GHz 400 MHz FSB 90nm) $294 $294 -
740 (2M L2 cache 1.73 GHz 533 MHz FSB 90nm) $241 $241 -
735 (2M L2 cache 1.70A GHz 400 MHz FSB 90nm) $241 $241 -
730 (2M L2 cache 1.60B GHz 533 MHz FSB 90nm) $209 $209 -
725 (2M L2 cache 1.60A GHz 400 MHz FSB 90nm) $209 $209 -

They list the A in the speeds, though (and B on some 533FSB versions)... but only on the clock speeds that are also Banias speeds - which makes sense - it fits in with Intel's old procedure - the Celeron 300 was a Covington, the 300A was a Mendocino. The Celeron 1.0 was a Coppermine, the 1.0A was a Tualatin, and the 1.0B is a Shelton (the bastard child of the Celeron M).

They've discontinued the 715, FWIW - I guess the Celeron M 370 was competing with it...

Also, I don't think that A is the code for XD - I thought that J was the code for XD:

Intel's price list wrote:

Intel® Celeron® M processor May '05 (05/01) June '05 (06/12) % Decrease
Mobile (uFCBGA / uFCPGA) Price Price
370 (1M L2 cache 1.50 GHz 400 MHz FSB 90nm) $134 $134 -
360/360J (1M L2 cache 1.40 GHz 400 MHz FSB 90nm) $107 $107 -
350/350J (1M L2 cache 1.30 GHz 400 MHz FSB 90nm) $86 $86 -

Shining Arcanine wrote:

So that means that Nehalem wasn't canned? I was under the impression that it was considering that Intel is moving away from Netburst to a new architecture designed on the same principles of the Pentium-M.

If I'm remembering correctly, I saw an article a few months ago somewhere (The Inquirer, perhaps?) saying that Nehalem, despite what most everyone thought, wasn't cancelled after all, but its scope changed from being the codename for a particular CPU (which was going to be a totally new architecture, not based on P6 or P7/NetBurst at all) to being the codename for a family of new CPUs based on that brand new architecture. I wrote Alan back asking if he could tell me anything at all about it, but he hasn't replied yet. I barely know the guy so he might not write back again.

Shining Arcanine wrote:

I guess this means that we can expect to see 10GHz in 5 years or so.

Actually I doubt that frequencies are going to go too much higher than they are now because of heat dissipation issues. Processors are moving toward multi-core stuff since technical problems like this make it more and more difficult to squeeze more single-thread performance out of processors. Intel's long-term plans involve having lots, I mean tens of cores, on one CPU, and software is going to have to change to be very, very multi-threaded.

Evan wrote:

Shining Arcanine wrote:

So that means that Nehalem wasn't canned? I was under the impression that it was considering that Intel is moving away from Netburst to a new architecture designed on the same principles of the Pentium-M.

If I'm remembering correctly, I saw an article a few months ago somewhere (The Inquirer, perhaps?) saying that Nehalem, despite what most everyone thought, wasn't cancelled after all, but its scope changed from being the codename for a particular CPU (which was going to be a totally new architecture, not based on P6 or P7/NetBurst at all) to being the codename for a family of new CPUs based on that brand new architecture. I wrote Alan back asking if he could tell me anything at all about it, but he hasn't replied yet. I barely know the guy so he might not write back again.

Shining Arcanine wrote:

I guess this means that we can expect to see 10GHz in 5 years or so.

Actually I doubt that frequencies are going to go too much higher than they are now because of heat dissipation issues. Processors are moving toward multi-core stuff since technical problems like this make it more and more difficult to squeeze more single-thread performance out of processors. Intel's long-term plans involve having lots, I mean tens of cores, on one CPU, and software is going to have to change to be very, very multi-threaded.

Nehalem was the codename for the first processor from a new architecture that would supercede the Netburst family and break the 10GHz barrier. I had thought that the processor had been canned along with Netburst. Not being canned could mean that they're still working on getting to 10GHz although Intel might have torn up its plans and then reused the codename for something else.

Alan finally wrote back, but unfortunately he didn't say much. He said that since he's not sure what he's allowed to divulge and what is secret, he doesn't want to risk spilling anything he shouldn't, but he did say that the direction of Nehalem has changed from what it was before, and that maybe in six months or so he'd be able to reveal more since he should know by then what he can and can't say about it. So, we'll have to be patient.