Intel imperfect AMD64 implementation

TheRod

Distinguished
Aug 2, 2002
2,031
0
19,780
<A HREF="http://www.siliconinvestor.com/stocktalk/msg.gsp?msgid=20289129" target="_new">http://www.siliconinvestor.com/stocktalk/msg.gsp?msgid=20289129</A>

Not a good news! Intel still have work to do!

--
It's tricky to use words like <b><font color=green>AMD</font color=green></b> or <b><font color=blue>Intel</font color=blue></b> in a signature some users could think your are biased.
 
LOL! Good link.. looks like a rush job and Intel has some learning curve ahead in the art of cloning 😛 Doesnt look too serious though, should be relatively easy to work around this bug (if it isnt fixed in the stepping intel will release), and 36 physical address bits isnt quite a huge limitation either. 64 GB per cpu ought to be enough for 99% of xeon installs, and where it isnt enough, Intel wants you to buy Itanium anyhow.

What i'm unsure about however, is how this carries over to large SMP systems. For 1 or 2 way systems, a 64GB ram limit sounds reasonable, for 4-8-16 way though, it is unacceptable, but I assume the limit applies per node/chipset.

= The views stated herein are my personal views, and not necessarily the views of my wife. =
 
and 36 physical address bits isnt quite a huge limitation either.
If this problem persist coult it hurt software compatibility? I doubt it and Intel will surely correct this for their official AMD64 launch! Humm.. Scuze me for their EM64T "paper" launch! :smile:

--
It's tricky to use words like <b><font color=green>AMD</font color=green></b> or <b><font color=blue>Intel</font color=blue></b> in a signature some users could think your are biased.
 
>If this problem persist coult it hurt software
>compatibility?

Not on anything else as the OS. The OS should recognize Nocona and work around the bug (if it isnt fixed by then), but no application would ever suffer. Well, not from this one at least :)

= The views stated herein are my personal views, and not necessarily the views of my wife. =
 
I'm not sure.. I'm not sure what "caring" means for a company. But apparently, intel found the 64 bit issue important enough to rush a cpu to market which isnt ready yet, release technical documentation that is seriously flawed (copy/pasted from AMD and not applicable to their own design.. ouch.). Mind you, all this is for a *server* cpu, not a desktop part.. If intel "didnt care" for 64 bits x86, I think it would have waited until everything was ready and ironed out. We all know what tends to happen the few times intel rushes something to market...

= The views stated herein are my personal views, and not necessarily the views of my wife. =
 
No no I tend to agree with trooper I really don't think they care.

Xeon

<font color=red>Post created with being a dickhead in mind.</font color=red>
<font color=white>For all emotional and slanderous statements contact THG for all law suits.</font color=white>
 
LOL, A64 can only address 40-bits.

Prescott 3.0E 1MB L2 HT
1GB PC 3200 Dual channel(PAT)
Asus P4P800 Bios 1016
PNY Geforce 6800 GT 256MB DDR3
57,397 Aquamarks
 
and what can EM64T address? some rumors going aorund that it wil only adress 36bits, but see no one can know becuase the noconas seem either under NDA for everyone to test x86-64 lol, or no one cares to review it becuase it doesnt perform, who knows, but the silence isnt a good sing is it?
 
No, it can adress 48 bits (virtual memory) in its current form, which is only 16 million terrabyte. If you know a way to exceed that, let me know. NOt all the harddisks in the world would exceed that. And A64/Opteron can address 1024 Gigabyte of RAM per cpu. If you know of a way plugging in that much memory, let me know as well.

Oh, and if you find that funny, then how funny is Nocona's limitation of 36 bits ? A brand new 64 bit server chip that can only address 6% of the RAM a desktop K8 can. Hilarious no ?

= The views stated herein are my personal views, and not necessarily the views of my wife. =