1 |
On Wednesday 05 March 2008, Chris Brennan wrote: |
2 |
> Here is the output of my cpuinfo, I'll point out where it will mean |
3 |
> 64bit. For the sake of this demonstration, I have only pasted one (1) |
4 |
> CPU. |
5 |
> |
6 |
> model name : AMD Athlon(tm) 64 X2 Dual Core Processor 4200+ <-- Here |
7 |
> will be your first indication. 64 bit processors, weather they are |
8 |
> SPARC, AMD or intel, must identify themselves clearly in the model |
9 |
> name field. |
10 |
|
11 |
In my case "must" equates to "should. but doesn't" |
12 |
The box on my desk identifies the cpu as Pentium4. But it's running 64 |
13 |
bit SLES, so something got omitted on the chip |
14 |
|
15 |
> flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca |
16 |
> cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext |
17 |
> fxsr_opt rdtscp lm 3dnowext 3dnow rep_good pni cx16 lahf_lm |
18 |
> cmp_legacy svm extapic cr8_legacy <-- you can look for flags that may |
19 |
> give away 64bit, such as lahf_lm, lm and nx. |
20 |
|
21 |
Is there a list somewhere of which flags are definitely 64 bit |
22 |
capabilities only, and which might be found on 32 bit chips? |
23 |
|
24 |
It would not surprise me if a manufacturer announced they had |
25 |
implemented ssse3 on 32 bit silicon for example |
26 |
|
27 |
> address sizes : 40 bits physical, 48 bits virtual |
28 |
|
29 |
How about this? Current amd64 implements 48 bit addressing, how wide is |
30 |
it for 32 bit? OTOH my Core2 Duo notebook does not have this field at |
31 |
all |
32 |
|
33 |
> For reference, this is a good read talking about 64 Bit processors |
34 |
> and how to identify them. http://en.wikipedia.org/wiki/X86-64 |
35 |
|
36 |
I see it's vastly more complex than I ever imagined. Thanks for the link |
37 |
though, learned a lot! |
38 |
|
39 |
-- |
40 |
Alan McKinnon |
41 |
alan dot mckinnon at gmail dot com |
42 |
|
43 |
-- |
44 |
gentoo-user@l.g.o mailing list |