1 |
On Mar 15, 2013 7:31 AM, "Mateusz Kowalczyk" <fuuzetsu@×××××××××××.uk> |
2 |
wrote: |
3 |
> |
4 |
> On 14/03/13 23:52, Dale wrote: |
5 |
> > Mateusz Kowalczyk wrote: |
6 |
> >> On 14/03/13 22:41, Dale wrote: |
7 |
> >>> Grant Edwards wrote: |
8 |
> >>>> On 2013-03-14, Dale <rdalek1967@×××××.com> wrote: |
9 |
> >>>> |
10 |
> >>>>> I was wondering. Has anyone ever seen where a test as been done to |
11 |
> >>>>> compare the speed of Gentoo with other distros? Maybe Gentoo |
12 |
compared |
13 |
> >>>>> to Redhat, Mandrake, Ubuntu and such? |
14 |
> >>>> I just did a test, and they're all the same. |
15 |
> >>>> |
16 |
> >>>> CDs/DVDS of various distros dropped from a height of 1m all hit the |
17 |
> >>>> floor simultaneously [there are random variations due to aerodynamic |
18 |
> >>>> instability of the disk shape, but it's the same for all distros]. If |
19 |
> >>>> launched horizontally with spin to provide attitude stability (thrown |
20 |
> >>>> like a frisbee), they all fly the same. |
21 |
> >>>> |
22 |
> >>>> The point being, you're going to have to define "speed". |
23 |
> >>>> |
24 |
> >>>> Does speed refer to |
25 |
> >>>> |
26 |
> >>>> Installation time? |
27 |
> >>>> |
28 |
> >>>> Boot time? |
29 |
> >>>> |
30 |
> >>>> Linpack? |
31 |
> >>>> |
32 |
> >>>> Dhrystone? |
33 |
> >>>> |
34 |
> >>>> Whetstone? |
35 |
> >>>> |
36 |
> >>>> Time for me to figure out how to fix a configuration problem? |
37 |
> >>>> |
38 |
> >>>> Time to do to an update on a machine that's been unplugged for a |
39 |
year? |
40 |
> >>>> |
41 |
> >>>> Time to to produce a packaged version of some random C program that |
42 |
> >>>> comes with a Makefile that uses autotools? |
43 |
> >>>> |
44 |
> >>>> Time for a reported bug to get fixed? |
45 |
> >>>> |
46 |
> >>> |
47 |
> >>> OK. It appears not very many can figure out what I asked for. So, |
48 |
let |
49 |
> >>> me spell it out for those who are challenged. LOL ;-) Read some |
50 |
> >>> humor into that OK. |
51 |
> >>> |
52 |
> >>> Install a OS. Run tests on a set of programs and record the time it |
53 |
> >>> takes to complete a certain task. More tasks the better. |
54 |
> >>> |
55 |
> >>> Then install another OS on the same hardware. Run tests on a set of |
56 |
> >>> programs and record the time it takes to complete a certain task. |
57 |
More |
58 |
> >>> tasks the better. |
59 |
> >>> |
60 |
> >>> The object of this is, does Gentoo with the customization it allows |
61 |
run |
62 |
> >>> faster than some binary install that does NOT allow those controls? |
63 |
In |
64 |
> >>> other words, can a Gentoo based install perform more efficiently than |
65 |
a |
66 |
> >>> binary based install like Redhat, Ubuntu or some other distro? |
67 |
> >>> |
68 |
> >>> I am NOT concerned about compile times or the install itself. |
69 |
> >>> |
70 |
> >>> Does that put the dots closer together for the challenged ones? ROFL |
71 |
> >>> |
72 |
> >>> Dale |
73 |
> >>> |
74 |
> >>> :-) :-) |
75 |
> >>> |
76 |
> >> The point of the challenged ones was that while we can take |
77 |
measurements |
78 |
> >> like these, it's rather meaningless to do so. The result will be |
79 |
> >> different for every single person out there depending on their |
80 |
> >> configuration, USE, CFLAGS and who knows what else. |
81 |
> >> |
82 |
> >> I can compile a package with support for 3 different DEs, few WMs, oss |
83 |
> >> and alsa and about a billion things I will never use. Does this make |
84 |
for |
85 |
> >> a more or less of a meaningful test than doing the same test with no |
86 |
> >> flags what so ever? There is no correct answer as it varies per user |
87 |
> >> basis. The most meaningful measurements that we can probably take would |
88 |
> >> be between different USE flags configurations. Maybe we can say that |
89 |
> >> package ‘foo’ with certain USE and CFLAGS runs in less average time |
90 |
than |
91 |
> >> the same package on a distro Bar. |
92 |
> >> |
93 |
> >> In my opinion, it would be far more meaningful to measure the effect of |
94 |
> >> different USE flags on the same package, *in relative time* on the same |
95 |
> >> system. This would give us more idea about the impact of each flag as |
96 |
> >> opposed to a very limited view of ‘package foo with certain specific |
97 |
USE |
98 |
> >> flags runs 10ms faster than the same package on the same hardware on a |
99 |
> >> binary distribution’. If you still want such measurements and you want |
100 |
> >> them to be somewhat meaningful to you, it is you who will have to take |
101 |
> >> them. Unless there are some gross inconsistencies in run times on |
102 |
> >> different distributions, we have no use for such measurement. |
103 |
> >> |
104 |
> >> Everyone understood what you asked for. It's _you_ that misunderstood |
105 |
> >> their explanation for why it's meaningless to ask such a question in |
106 |
the |
107 |
> >> first place. |
108 |
> >> |
109 |
> > |
110 |
> > I didn't miss anything. I get what some are saying. The reason for my |
111 |
> > question is this. Gentoo allows a person to customize the OS to the |
112 |
> > specific hardware it is being run on. Redhat and other binary distros |
113 |
> > don't allow this, unless you compile your own packages which is no |
114 |
> > longer really a binary install. |
115 |
> > |
116 |
> > So, if I install Redhat on my machine, would it be less efficient than |
117 |
> > my Gentoo install which is customized for my hardware? Has someone else |
118 |
> > tested this and made it public? |
119 |
> > |
120 |
> > If people can't get this, never mind. |
121 |
> > |
122 |
> > Dale |
123 |
> > |
124 |
> > :-) :-) |
125 |
> > |
126 |
> |
127 |
> I don't think that it's plausible to take such measurement. We could set |
128 |
> every USE flag possible for the package we are benchmarking to try and |
129 |
> replicate the support for everything that the binary package is likely |
130 |
> to have. We also have to do this for all its dependencies (and their |
131 |
> dependencies and so on) to have nothing that could potentially influence |
132 |
> the measurement. Assuming that portage complies with this (it won't), we |
133 |
> compile the package with optimizations for our hardware. The result? We |
134 |
> probably have the same result on Gentoo and the other distro. |
135 |
> |
136 |
> Why? The reason is simple: binary distributions provide packages |
137 |
> compiled with optimizations turned on for specific architectures. Unless |
138 |
> you are doing some unheard of optimizations for your obscure model of |
139 |
> the CPU, I don't imagine you'd get much advantage at all. If the |
140 |
> maintainer of the binary package compiles it with optimizations for an |
141 |
> i7 and you do the same, why should the performance be different? |
142 |
> |
143 |
> If you install RedHat on your machine, it probably will be less |
144 |
> efficient than Gentoo but for a different reason. It won't be |
145 |
> (noticeably) more efficient because you've got your CFLAGS set in a |
146 |
> particular way but it because you only install the packages you actually |
147 |
> want and remove support for things you don't need. Why bloat a package |
148 |
> and waste cycles having it try to poll a service you might never |
149 |
> actually use? Gentoo lets you get rid (or never install in the first |
150 |
> place) of this kind of… bloat while you don't have much choice on a |
151 |
> binary package short of compiling everything by hand at which point you |
152 |
> should be using Gentoo to do it for you. |
153 |
> |
154 |
> RedHat maintainers aren't stupid (you can probably tell I've never used |
155 |
> RH) – they will release packages optimized for architectures they will |
156 |
> run on. Overall you might get very slight performance boost because of |
157 |
> some CFLAG you enable but you might as well have worse performance |
158 |
> because you don't know as much about optimizations as the RH maintainers |
159 |
> and developers. Bah, you can even find examples on Gentoo wiki where |
160 |
> compiling certain packages with certain flags actually makes them slower |
161 |
> and not faster where usually the opposite is the case. |
162 |
> |
163 |
> -- |
164 |
> Mateusz K. |
165 |
> |
166 |
|
167 |
AFAIK, binary distros have only two kinds of targets for Intel/AMD |
168 |
processors: x86 or amd64. |
169 |
|
170 |
I think most binary distros set their -march or -mtune to nothing newer |
171 |
than a Pentium Pro (x86 case) or Pentium 4 (amd64 case). This means, newer |
172 |
generation instruction sets are not used. |
173 |
|
174 |
Of course, the amount (%) of code that can be optimized is very small |
175 |
compared to the rest of the code. On most programs, there will be no |
176 |
benefit of mtune/march to a specific processor. |
177 |
|
178 |
However, if you happen to frequently use a program whose major processing |
179 |
time happens to be optimizable with the newer instructions, you hit a |
180 |
jackpot. |
181 |
|
182 |
TL;DR : Just like everybody say, Gentoo might be faster, but depends |
183 |
totally on the jobs one do on it. On majority of cases, the faster speed is |
184 |
totally inconsequential. |
185 |
|
186 |
Rgds, |
187 |
-- |