lhgpoobaa

Illustrious
Dec 31, 2007
14,462
1
40,780
unless something significant changes i assume so.

might be slightly less likely that standard cpu's work in mp possibly. amd might want to crack down on that to push up sales of their MP cpu's.

Today on Toms: Trisexual hamsters, anal applications of peanut butter and Marrage councilling!
 

jclw

Distinguished
Dec 31, 2007
1,255
0
19,290
I'm guessing that AMD will follow intel's lead and "break" the SMP capability in the standard/desktop/XP/whateveryouwanttocallthem 0.13 micron cores in order to boost sales of the server chips, which after all, is where the money is.

- JW
 

Matisaro

Splendid
Mar 23, 2001
6,737
0
25,780
I'm guessing that AMD will follow intel's lead and "break" the SMP capability in the standard/desktop/XP/whateveryouwanttocallthem 0.13 micron cores in order to boost sales of the server chips, which after all, is where the money is.

Why all of a sudden, the athlon mp has been around for a while and the xp was never broken?

So why do you think amd will suddenly up and break axp mp functionality? Especially since the hammer line is coming soon and the point will be moot, and all breaking smp will serve to do is piss off amd fans.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 

jclw

Distinguished
Dec 31, 2007
1,255
0
19,290
Maybe I misread the original question. I think AMD will "break" the SMP on the XP cores, but of course leave it enabled on the MP cores.

If we look back in history we'll find that SMP has always been "broken" during major core changes, mainly die shrinks.

Intel "broke" the celeron SMP moving from the 0.25 to 0.18 cores.

Intel "broke" the PIII SMP (on 256k desktop chips) moving from the 0.18 to 0.13 cores.

That's why I think AMD will "break" the SMP on the XP cores when it moves to 0.13 micron.

AMD cannot hope to expand while only selling $200 chips to us, while intel is selling $600 (Xeon DP) to $3500 (Xeon MP) chips to large companies as fast as they can make them. If AMD wants to keep up the competition they need to break into that high end market to support R&D down the road. While AMD has some great 0.18 fabs, to my knowledge they only have one 0.13 fab and I believe it uses considerably smaller wafers then the intel fabs. That's probably why we havn't seen any 0.13 AMD chips yet. I think if they have to piss off the guys that buy $200 chips to sell more $600+ chips they should (and will) do it, because it'll be better for us (the small guys) in the future.

BTW:
Xeon DP - What most of us call "Xeons" - support two way SMP
Xeon MP - support four and more way SMP

- JW<P ID="edit"><FONT SIZE=-1><EM>Edited by JCLW on 05/10/02 02:43 PM.</EM></FONT></P>
 

mbetea

Distinguished
Aug 16, 2001
1,662
0
19,780
i agree. i think they will make sure nonMP thoroughbreds ARE infact smp disabled. why would they start then going through the latest batches of xp chips, cutting the L5 bridge?

[insert philosophical statement here]
 

Matisaro

Splendid
Mar 23, 2001
6,737
0
25,780
Maybe I misread the original question. I think AMD will "break" the SMP on the XP cores, but of course leave it enabled on the MP cores.

If we look back in history we'll find that SMP has always been "broken" during major core changes, mainly die shrinks.

And what I am saying is that amd has not given any indications that they intend to break smp, in fact they have known axps are routinely used in smp systems and have done nothing to change that.

Intel likes to gouge the consumer, and perhaps when amd is as big as them they will do the same, but for the time being amd generally takes steps to ensure that they offer a good value, be it via overclocking or smp.



While AMD has some great 0.18 fabs, to my knowledge they only have one 0.13 fab and I believe it uses considerably smaller wafers then the intel fabs. That's probably why we havn't seen any 0.13 AMD chips yet.

Umm, amd has 1(2 if you count austin which will be closed soon) .18 fab making processors, and that same fab is .13 making processors, and its called dresden, and its a megafab and is larger than any intel fab (iirc), it also has a very good yield and process control which is legendary(.13 class gate lengths on a .18 process!!!).

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 

Matisaro

Splendid
Mar 23, 2001
6,737
0
25,780
agree. i think they will make sure nonMP thoroughbreds ARE infact smp disabled. why would they start then going through the latest batches of xp chips, cutting the L5 bridge?

See heres the thing, cutting a bridge is not "disabling" smp, anyone with the will can reenable the smp, so amd may do this(and probably will) but they will not break smp on the chip itself like intel has.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 

mbetea

Distinguished
Aug 16, 2001
1,662
0
19,780
i think it would be considered disabling it, BUT that is assuming amd thinks noone will do a hardware mod. but i guess it's pointless now seeing hammer MP boards will have a different pin count/socket all together. sorry i also agree with you in that it is not at the level intel went with with their chips.

[insert philosophical statement here]<P ID="edit"><FONT SIZE=-1><EM>Edited by mbetea on 05/11/02 03:26 PM.</EM></FONT></P>
 

jclw

Distinguished
Dec 31, 2007
1,255
0
19,290
Remember AMD tried to get MPX motherboard manufacturers to disable SMP for XP chips in BIOS when they were first released. There were many "excuses" released later... "We wanted to make them stable with MP chips first", etc.

AMD may have a nice 0.13 fab, but there aren't many 0.13 AMD chips floating around yet.

- JW