uname -m for an Intel machine running FreeBSD

[Valeri Galtsev]

> On Jan 26, 2020, at 10:01 PM, Ihor Antonov <ihor at antonovs.family> wrote:
> 
> On 2020-01-26 21:49, Chris Hill wrote:
>> On Sun, 26 Jan 2020, Clay Daniels wrote:
>> 
>>> I'm working on building an efitool program written and used in various
>>> linux flavors that I would like to see work on FreeBSD.
>> 
>> [ huge snip ]
>> 
>>> I think I can manage  the code, except I don't have an Intel machine to
>>> know what "name" uname -m returns on a Intel box running FreeBSD.
>>> 
>>> Can someone with an Intel machine tell me what uname -m returns?
>> 
>> $ uname -m
>> amd64
>> 
>> This is on a:
>> CPU: Intel(R) Core(TM)2 Duo CPU     E8500  @ 3.16GHz (3158.81-MHz K8-class CPU)
> 
> Same on my Intel(R) Core(TM) i7-8850H CPU @ 2.60GHz
> 
> Linux: x86_64
> FreeBSD: amd64  - both baremetal and as KVM guest
> 

This correct.

Man name
…
     -m      Write the type of the current hardware platform…

This essentially denotes 64 bit CPU with with inscruction set haveng i386 instruction set as its subset. It is not by any means hardware (CPU) make, model etc identifier.

I like the name FreeBSD uses: amd64. Some Linuxes do use the same name, can not remember distros off hand. RedHat and clones and Debian ad clones use x86_64.

(I am not historian, so someone may correct me where my memory fails me). This architecture began by AMD releasing its Opteron CPU (the project was called “hammer”). This was 64 bit CPU with instruction set containg as subset i386 instructions. Simultaneously AMD made excellemt move about RAM organization: in multi-CPU machine each CPU had memory controller, and its chunk of memory. Thus two different CPU will not collide on access of segments of memory directly attached to them. For over decade (if not two decades) CPU speed grew up much faster that RAM speed. And this solution was known even before AMD implemented it in hammer (Opteron) project. Then, for access of the trest of RAM given CPU has to go through another CPU wo which necessary segment of prysical memory is attached directly. This is do through very fast “hypertransport channels” between CPUS which VIA developed for AMD (necessary electronics is insode each of CPUs.

At that moment Intel had different approach to 64 bit, and about that time they just developed their Itanium architecture. The instruction set was different (nothing bad about that), and in multi-processor boxes CPU was accessing RAM tghrugh memory bus. Which means, all CPU were colliding on slow process: accessing RAM. After some time Itanium faded away, and Intel gradually moved to amd64 architecture. There are still some differences (which I can only mention in general): AMD CPU finishes instruction in 4 ticks of the clock (read: its GHz specification), whereas Intel CPU takes 6 ticks. Intel CPU has more “hardware” inside, so it effectively can utilize one core for two threads. From the prospective of sysadmin who had machines sometimes in “not perfect” server room (like temperature may get high), AMD Opteron CPUs are more robust. Someone put it this way: you can boil water on it, and it still will run. I’ve had the following (about 10 years ago): they had to do maintenance in the server room, and for 2 hours the temperature was 96 F (36 C). Some of intel boxes - not all of them - hot sick. None of Opteron boxes did: they just kept running.

So, historically, I agree with FreeBSD who call it amd64 architecture. Even though we have mire Intel thank AMD boxes these days.

Valeri

>> -- 
>> Chris Hill               chris at monochrome.org
>> _______________________________________________
>> freebsd-questions at freebsd.org mailing list
>> https://lists.freebsd.org/mailman/listinfo/freebsd-questions
>> To unsubscribe, send any mail to "freebsd-questions-unsubscribe at freebsd.org"
> _______________________________________________
> freebsd-questions at freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-questions
> To unsubscribe, send any mail to "freebsd-questions-unsubscribe at freebsd.org"

++++++++++++++++++++++++++++++++++++++++
Valeri Galtsev
Sr System Administrator
Department of Astronomy and Astrophysics
Kavli Institute for Cosmological Physics
University of Chicago
Phone: 773-702-4247
++++++++++++++++++++++++++++++++++++++++