1 |
在 2006-06-25日的 19:50 +0200,Etaoin Shrdlu写道: |
2 |
> On Sunday 25 June 2006 19:02, 张韡武 wrote: |
3 |
> |
4 |
> > Thank you for the answer. Just another smaller question: from pruely a |
5 |
> > user's view (who simply wish to use it, rather then from developer |
6 |
> > view who might gain knowledge and fun doing it), is it worthy to try |
7 |
> > it (64bit sparc)? |
8 |
> > |
9 |
> > The original idea is the system as a print server is too slow (capable |
10 |
> > of doing fast dithering for the inkjet printer, but not fast enough to |
11 |
> > do real-time even-tone dithering to server the printer, usually cause |
12 |
> > printer to wait for server), a rough gain of 20% to 30% CPU power, by |
13 |
> > my guessing of observing top(1), should be enough for the printer. |
14 |
> > Dithering is always a CPU intensive job and we do a lot of printing. |
15 |
> > |
16 |
> > I observed kernel is taking only a very small percentage of CPU usage, |
17 |
> > if I try hours to get 64-bit kernel running but only obtain 5% speed |
18 |
> > gain, that may not be worth. Or should I try Solaris 10? Should |
19 |
> > Solaris provide better performance in this very case? (how much |
20 |
> > performance gain might happen? 10%? 20%?) I don't need a very close |
21 |
> > estimation, just if someone can give me a very rough estimation is |
22 |
> > very helpful, for I am not techincally able to do the decision:) |
23 |
> > |
24 |
> > If switch to Solaris for performance (no offensive, I am a 5 year's |
25 |
> > hardcore Linux user), generally speaking, is Solaris very compatible |
26 |
> > with PCI cards like NEC USB controller, Promise IDE controller and |
27 |
> > ethernet cards? |
28 |
> |
29 |
> Well, I'm not a sparc dev so probably I cannot give you the most accurate |
30 |
> answer you could get; moreover, I cannot speak about solaris since I |
31 |
> don't know it. Anyway... |
32 |
> From what I've read, the main reason gentoo stays 64 bit kernel/32 bit |
33 |
> userland is that switching to a full 64 bit system (perhaps with |
34 |
> multilib) will not be a considerable improvement, but rather just a |
35 |
> waste of memory. This is different from what happens in the x86 world, |
36 |
> where being 64-bit means getting a lot of benefits (like, for example, |
37 |
> many more cpu registers). In the sparc world, for an userland app "being |
38 |
> 64-bit" means, basically, "wasting more memory" (due to larger |
39 |
> executables binaries). In short, I think that (from the pure user's view |
40 |
> you were talking about), it's not worth trying. |
41 |
> However, I believe that a (highly experimental) userland 64-bit profile |
42 |
> exists (but of course don't expect that to work flawlessly or to be |
43 |
> usable in a production environment). |
44 |
> Read this thread on the forums, which explains all: |
45 |
> |
46 |
> http://forums.gentoo.org/viewtopic-t-442194.html |
47 |
> |
48 |
> btw, AFAIK solaris uses a similar 64bit-kernel/32bit-userland approach |
49 |
> (or, at least, *most* of the userland is 32bit). |
50 |
> |
51 |
> Hope this helps (and please somebody correct me if what I said is wrong). |
52 |
|
53 |
Blame me for being stupid, but I simply cannot understand. Truly in x86 |
54 |
world 64-bit is often marketed as much faster then 32-bit applications, |
55 |
if this is not ture in sparc system, the question is why people bother |
56 |
to build 64-bit stuff, being slower and wastes more memory at all? For |
57 |
sure I am not an arch guru, but there must be a good reason behind it. |
58 |
|
59 |
I am one of stupid graduated CS student the modern universities |
60 |
produced, I can understand perfectly well that 32-bit binary |
61 |
applications perform equally on 32-bit or 64-bit platform, but I was |
62 |
tought that is because the binary is 32-bit: if the source is compiled |
63 |
targeted to be 64-bit binary, usually it is faster especially for CPU |
64 |
intensive tasks, the compiler optimize the source and use 64-bit when |
65 |
possible (perhaps especially for float point). In my case I can |
66 |
re-compile gutenprint or cups to generate 64-bit binary. Especially the |
67 |
job like dithering for the printer. In my understanding dithering is |
68 |
purely computation of how to mix colors, few I/O and only some memory, |
69 |
that looks really like something can be optimized by using 64-bit. |
70 |
|
71 |
For the suggestions I can easily take all of them, just now I am |
72 |
curious:) If 64-bit is truly considered an enhancement for CPU, there |
73 |
must be some use of it, and if purely CPU intensive tasks (dithering) |
74 |
cannot benifit from enhancement for CPU, whatelse can take the advantage |
75 |
of 64-bit? (Okay, now I am not in purely user's view, I wish to dig into |
76 |
it a little bit because that makes me really curious). |
77 |
|
78 |
-- |
79 |
gentoo-sparc@g.o mailing list |