Most powerful machine you've put gentoo on?

[krinn]

emerge -ep system != emerge -ep @system

[Ormaaj]

emerge -ep system != emerge -ep @system

The set is exactly the same.

[joaopft]

You guys are comparing apples with oranges.
emerge -ep system will build all the dependencies of the packages in system, and that depends on the USE flags set in
/etc/make.conf

For instance, (on a core i7, 3.2GHz, desktop profile with kde and other software):

Code: Select all

emerge -ep --nodeps system |  genlop -p

...

Estimated update time: 36 minutes.

while

Code: Select all

emerge -ep system |  genlop -p

...

Estimated update time: 2 hours, 11 minutes.

On the first example, we are building only the system set of packages (no deps) and on the second example we also build the dependencies created by active USE flags. As we can see, emerge -ep system is not a good benchmark. I would suggest, as a minimalistic bench that tests CPU Integer and floating point ops, Memory r/w and de-alloc and multithreading:

Code: Select all

emerge bashmark
bashmark

I will post some benchs later. If you use cpufreq, don't forget to set the 'performance' governor.

[joaopft]

Assorted bashmark results (using code display because this requires a fixed width font):

Code: Select all

#######################################################
:  T   E   S   T        :    :S C O R E :  : R A T I O:
:-----------------------------------------------------:
:Cpu, Integer           :    :      1032:  :       -4%:
:Cpu, Floating point    :    :       849:  :      +12%:
:                       :    :          :  :          :
:Memory r/w (cached)    :    :      1567:  :      +30%:
:Memory de-/alloc       :    :       349:  :      -47%:
:                       :    :          :  :          :
:Multithreading         :    :       861:  :      -65%:
#######################################################
:           S  Y  S  T  E  M     I  N  F  O           :
-------------------------------------------------------
2x Intel(R) Pentium(R) 4 2995.767MHz, L2 512KB
Linux 2.6.28-gentoo-r5
GCC 4.3.2
100KB binary size


#######################################################
:  T   E   S   T        :    :S C O R E :  : R A T I O:
:-----------------------------------------------------:
:Cpu, Integer           :    :      1909:  :      +78%:
:Cpu, Floating point    :    :      1760:  :     +132%:
:                       :    :          :  :          :
:Memory r/w (cached)    :    :     14047:  :    +1068%:
:Memory de-/alloc       :    :      1407:  :     +115%:
:                       :    :          :  :          :
:Multithreading         :    :      4889:  :      +96%:
#######################################################
:           S  Y  S  T  E  M     I  N  F  O           :
-------------------------------------------------------
1x AMD Athlon(tm) 64 3000+ 2493.787MHz, L2 512KB
Linux 2.6.15-gentoo-r5
GCC 3.4.4
116KB binary size


#######################################################
:  T   E   S   T        :    :S C O R E :  : R A T I O:
:-----------------------------------------------------:
:Cpu, Integer           :    :      7376:  :     +587%:
:Cpu, Floating point    :    :      6138:  :     +709%:
:                       :    :          :  :          :
:Memory r/w (cached)    :    :     16176:  :    +1245%:
:Memory de-/alloc       :    :      3191:  :     +388%:
:                       :    :          :  :          :
:Multithreading         :    :      9469:  :     +280%:
#######################################################
:           S  Y  S  T  E  M     I  N  F  O           :
-------------------------------------------------------
8x Intel(R) Core(TM) i7 920 @ 3199.482MHz, L2 8192KB
Linux 2.6.34-gentoo-r11 
GCC 4.4.3
96KB binary size


#######################################################
:  T   E   S   T        :    :S C O R E :  : R A T I O:
:-----------------------------------------------------:
:Cpu, Integer           :    :      1685:  :      +57%:
:Cpu, Floating point    :    :       896:  :      +18%:
:                       :    :          :  :          :
:Memory r/w (cached)    :    :      2104:  :      +75%:
:Memory de-/alloc       :    :       321:  :      -51%:
:                       :    :          :  :          :
:Multithreading         :    :      1435:  :      -42%:
#######################################################
:           S  Y  S  T  E  M     I  N  F  O           :
-------------------------------------------------------
2x Intel(R) Atom(TM) N280 @ 1666.638MHz, L2 512KB
Linux 2.6.34-gentoo-r6 
GCC 4.4.3
100KB binary size


#######################################################
:      R E F E R E N C E   S Y S T E M   I N F O      :
-------------------------------------------------------
Reference system was Geno's pc with:
Athlon XP 1800+ 1575.631MHz, 256KB
Linux 2.6.11-ck1
GCC 3.4.3-20050110 (compiled with standard cflags)
glibc 2.3.4 (with nptl)
128KB binary size
Scores gathered on March, 30th. 2005 with bashmark 0.6

[Chiitoo]

Gave bashmark a quick test as well.
For no particular reason, tried teh --detailed option as well, it is the second result which is more or less obvious.

Code: Select all

#######################################################
:  T   E   S   T        :    :S C O R E :  : R A T I O:
:-----------------------------------------------------:
:Cpu, Integer           :    :      1218:  :      +14%:
:Cpu, Floating point    :    :      1086:  :      +43%:
:                       :    :          :  :          :
:Memory r/w (cached)    :    :      2279:  :      +89%:
:Memory de-/alloc       :    :      1013:  :      +55%:
:                       :    :          :  :          :
:Multithreading         :    :      2550:  :       +2%:
#######################################################
:  T   E   S   T        :    :S C O R E :  : R A T I O:
:-----------------------------------------------------:
: 8bit integer          :    :      1453:  :      +33%:
:16bit integer          :    :      1454:  :      +15%:
:32bit integer          :    :      1455:  :       +2%:
:64bit integer          :    :       508:  :       -2%:
:32bit fp               :    :      1455:  :      +66%:
:64bit fp               :    :      1456:  :      +67%:
:128bit(80real) fp      :    :       348:  :      -34%:
:                       :    :          :  :          :
:128B Mem r/w (cached)  :    :      2427:  :      +77%:
:256B Mem r/w (cached)  :    :      2124:  :     +105%:
: 32Byte Mem de-/alloc  :    :      1217:  :      +55%:
: 64Byte Mem de-/alloc  :    :      1217:  :      +61%:
:128Byte Mem de-/alloc  :    :       606:  :      +44%:
:                       :    :          :  :          :
:MThreading u-/lock     :    :      2477:  :       -1%:
#######################################################
:           S  Y  S  T  E  M     I  N  F  O           :
-------------------------------------------------------
6x AMD Phenom(tm) II X6 1090T 3616.167MHz, L2 512KB
Linux 2.6.34-gentoo-r11 
GCC 4.4.4
100KB binary size
#######################################################
:      R E F E R E N C E   S Y S T E M   I N F O      :
-------------------------------------------------------
Reference system was Geno's pc with:
Athlon XP 1800+ 1575.631MHz, 256KB
Linux 2.6.11-ck1
GCC 3.4.3-20050110 (compiled with standard cflags)
glibc 2.3.4 (with nptl)
128KB binary size
Scores gathered on March, 30th. 2005 with bashmark 0.6

Funky.
Don't know about powerful but this is the most powerful one I've yet to use Gentoo (kinda first one too, been going for weeks now never used Linux that long before).

Got Asus ENGTX275 doing them graphics,
Creative X-Fi Xtreme Gamer for sounds,
and other even less special things hehe...

[jathlon]

Hey Chiitoo. Something's not quite right with your system (or maybe just the way the benchmark is running on your system). Same processor (running at spec GHz) here and I'm flat out stomping your system.

Code: Select all

 ~ # bashmark -d
#######################################################
:  T   E   S   T        :    :S C O R E :  : R A T I O:
:-----------------------------------------------------:
: 8bit integer          :    :     11425:  :     +948%:
:16bit integer          :    :     10249:  :     +714%:
:32bit integer          :    :      8766:  :     +516%:
:64bit integer          :    :      5116:  :     +882%:
:32bit fp               :    :      8645:  :     +889%:
:64bit fp               :    :      8037:  :     +820%:
:128bit(80real) fp      :    :      3757:  :     +609%:
:                       :    :          :  :          :
:128B Mem r/w (cached)  :    :     32961:  :    +2302%:
:256B Mem r/w (cached)  :    :     12525:  :    +1111%:
: 32Byte Mem de-/alloc  :    :      2821:  :     +259%:
: 64Byte Mem de-/alloc  :    :      2842:  :     +275%:
:128Byte Mem de-/alloc  :    :      1939:  :     +361%:
:                       :    :          :  :          :
:MThreading u-/lock     :    :      3702:  :      +49%:
#######################################################
:           S  Y  S  T  E  M     I  N  F  O           :
-------------------------------------------------------
6x AMD Phenom(tm) II X6 1090T 3200.000MHz, L2 512KB
Linux 2.6.36-gentoo 
GCC 4.4.5
96KB binary size
#######################################################
:      R E F E R E N C E   S Y S T E M   I N F O      :
-------------------------------------------------------
Reference system was Geno's pc with:
Athlon XP 1800+ 1575.631MHz, 256KB
Linux 2.6.11-ck1
GCC 3.4.3-20050110 (compiled with standard cflags)
glibc 2.3.4 (with nptl)
128KB binary size
Scores gathered on March, 30th. 2005 with bashmark 0.6

m/b: ASUS M4A87TD EVO
ram: Kingston KVR1333D3N9K2/8G (two kits for 16G)
sdd: Intel X25-V (where all the system binaries live)

So.. not really an extreme system. Farthest thing out there might be considered the ram. /var/log is mount on tmpfs so doing the genlop thing isn't really an option for me.

I think this system might keep me happy for a while.

joe

[Chiitoo]

Hey jathlon!


Yeah, I was actually expecting such results from someone else, or even hoping for it.
I've had a lot of weird problems indicating just that, I was indeed hoping for results for the same banchmark with the same CPU because when I first got this hardware (CPU, MOBO, MEM was replaced), I would get weird performance issues with the OS I used back then (winxP), most noticeable when several games were running at the same time.

It would result to even just the windows menus opening and moving sluggishly, this problem I did not have with my previous setup that had a dual-core AMD.
I then tried windoze7 and while I didn't exactly experience the same kind of things as under xP, I would however get bluescreen crashes a lot. Still, benchmarks like 3Dmark would produce more or less "correct" results.

I then decided to pick up Linux again and this time for realz!
I've had one lock-up while working with KDE, one time the computer had restarted while I was AFK and one time.. I can't remember what happened but I think I've had only 3 crashes with Gentoo in total.

I even contacted AMD about it back when I was running winxP and they would give me a pretty generic reply which did not help.
I can only suspect it's either the MOBO or the PSU at this time, I have my old functional PSU to try though it is only 500W I think while this one is 850W but it keeps making weird buzzing noises at times. o.o

Everything is new so you can imagine my frustration, save for the HDs and the sound-card which did not produce any problems before.
Memory I can't be sure of either but memtest did not result to any errors after running for 11 hours or so I think.
It's rather cheap'ish 1333MHz though, can't even remember the manufacturer right now...


Thanks for the results, I really appreciate it and it really does more prove to myself that there indeed is something wrong.
I have been too excited with Gentoo lately so I frankly did not care too much about the problem.

One more thing I thought of mentioning now is that the Gigabyte board I have was advertised to be ready for the CPU but I had to update the BIOS to get support for it and it was another task I would have not wanted to deal with since all my old CPUs are AM2 or AM2+ so they don't fit into this AM3 socket. Hoewver I managed to update the BIOS from a USB stick though it took a while to get to the Flash-utility due to the lock-upses at POST...


Also, in this thread, Phenom II X6, there might soon be more benchmarking related to this processor if you're interested. ^^

Anyways, thank you, really.
This helps!

[Chiitoo]

One thing I noticed is that bashmark seems to actually utilize the CPU very little, is this normal? :S

The most it would show up using was 4% I think, that in the system monitor applet of KDE.
TOP shows a bit more as it does for all not really sure how it differs from the other monitor, does it show it for single cores or what, haven't really looked into the command yet.

[hielvc]

bogomips=2*cpu_freq or cpu_Hz=bogo/2

Code: Select all

uname -a
Linux amd_1055 2.6.36-gentoo #4 SMP Sat Oct 23 09:42:41 PDT 2010 x86_64 AMD Phenom(tm) II X6 1055T Processor AuthenticAMD GNU/Linux

grep "bogo" /proc/cpuinfo
bogomips	: 6440.05
bogomips	: 6439.90
bogomips	: 6439.90
bogomips	: 6439.88
bogomips	: 6439.92
bogomips	: 6439.92

Cpu 3220 MHz. Standard AMD fan no voltage tweaking. FSB 230
~amd64
CFLAGS="-march=amdfam10 -O2 -fomit-frame-pointer -pipe"
MAKEOPTS="-j6"

What we all can relate to is compile times. Here is my kernel, gcc, glibc, and openoffice genlop times.

Code: Select all

amd_1055 linux # make clean; time make -j6

real	1m23.091s
user	6m51.253s
sys	0m36.009s

hielvc@amd_1055 ~ $ genlop -t gcc
Mon Oct 25 12:38:47 2010 >>> sys-devel/gcc-4.4.5
       merge time: 10 minutes and 47 seconds.

hielvc@amd_1055 ~ $ genlop -t glibc
Mon Oct 25 12:03:28 2010 >>> sys-libs/glibc-2.12.1-r1
       merge time: 8 minutes and 21 seconds.

hielvc@amd_1055 ~ $ genlop -t openoffice
Sun Oct 24 23:25:58 2010 >>> app-office/openoffice-3.2.1
       merge time: 39 minutes and 41 seconds.

Oh and

Code: Select all

emerge -eqp --nodeps @system |  genlop -p
Estimated update time: 45 minutes.

[Chiitoo]

Well this is pretty interesting.
Testing right now with my old PSU which is underpowered (it doesn't even have "native" PCI-e power lines and my GTX takes 2!) and I'm getting pretty much the EXACT same results for every test I run. I even tried running them while I have like 4 different games running! 4 via Wine and 1 built for Linux.
One that has 3 clients open (windows app) has some serious memory leaking going on at times even which tends to make the computer slow at times.

There is something different, though.
The whole GUI (KDE + the games) are not going into a sluggish-mode same way they went on xP and Linux with the new, more powerful PSU... so I think my doubts had some truth in them, but the CPU tests are mind-boggling, truly.
I'm going to suspect the "AM3 ready!!!123" motherboard is the next one to blame. >.>

[Muso]

Chiitoo, those PCI-e power cables that your GTX wants are needed for cooling... all you're doing by not using the proper power supply is damaging your graphics card.

[aidanjt]

Chiitoo, those PCI-e power cables that your GTX wants are needed for cooling... all you're doing by not using the proper power supply is damaging your graphics card.

And takes the strain off the PCIe bus.

[krinn]

emerge -ep system != emerge -ep @system

The set is exactly the same.

Not for all portage versions.

Code: Select all

emerge -pv @system
!!! '@system' is not a valid package atom.
!!! Please check ebuild(5) for full details.

[redwold]

I never install gentoo myself . My sister give her laptop(intel core 2 due processor) to me with gentoo installed and i m using this(love this)

[Chiitoo]

Chiitoo, those PCI-e power cables that your GTX wants are needed for cooling... all you're doing by not using the proper power supply is damaging your graphics card.

Of course I'm not running it without power going there.
If it's anything like my old card (8800), it would make a painfully loud noise if started without doing so. I don't think I ever tried / forgot the connections on this one so I don't know for sure, but I'd assume it's the same.

I have 4 different 4-pin molex connectors via 2 converters going into it so while it's not powerful enough for sure, I think it's doing fine (it seems as if the system is actually working better than with the proper PSU).
Thanks for pointing it out, though. ^^


Either way, going to post any further findings in the Phenom II X6 thread as it is more relevant.

[bhos13]

Code: Select all

Memory: 24729748k/25690108k available (4494k kernel code, 534788k absent, 425572k reserved, 3449k data, 532k init)
Brought up 24 CPUs
Linux xXxXxXxX 2.6.34-gentoo-r12 #2 SMP Fri Dec 10 05:55:24 CST 2010 x86_64 Intel(R) Xeon(R) CPU X5660 @ 2.80GHz GenuineIntel GNU/Linux

Total of 24 processors activated (134347.86 BogoMIPS)

Samba compiles in under a minute with -j26, GCC in under 6

[Spastik]

Code: Select all

Memory: 24729748k/25690108k available (4494k kernel code, 534788k absent, 425572k reserved, 3449k data, 532k init)
Brought up 24 CPUs
Linux xXxXxXxX 2.6.34-gentoo-r12 #2 SMP Fri Dec 10 05:55:24 CST 2010 x86_64 Intel(R) Xeon(R) CPU X5660 @ 2.80GHz GenuineIntel GNU/Linux

Total of 24 processors activated (134347.86 BogoMIPS)

Samba compiles in under a minute with -j26, GCC in under 6 :D

Two hexa-core (6, w/e you call it) processors with HT?

[72_6f_6c_61_6e_64]

I installed Gentoo on two Netbooks (Intel Atom N450/N280 1,66 GHz, 1GB Ram).

Now I am installing Gentoo on a Notebook (Intel Core 2 Duo 2,2GHz, 2 GB Ram) to use DistCC for the Netbook.

LG Roland

[pgu]

Code: Select all

grep bogo /proc/cpuinfo 
bogomips        : 6588.03
bogomips        : 6588.43
uname -p
Intel(R) Core(TM)2 CPU 6600 @ 2.40GHz

[72_6f_6c_61_6e_64]

Code: Select all

 grep bogo /proc/cpuinfo 
bogomips        : 3333.30
bogomips        : 3332.82
uname -p
Intel(R) Atom(TM) CPU N450 @ 1.66GHz

[barul]

The most powerful machine I have is my desktop. I've got a Gentoo Linux - Windows 7 dual boot.

CPU : AMD Phenom II X6 1055T (2.8Ghz)
RAM : 4GB DDR3 1600Mhz
GPU : Gigabyte GTX460 1GB
HDD : WD 500GB 7200tpm

Also on my 2 laptops
1st :
CPU : Intel C2D T6400 (2Ghz)
RAM : 4GB DDR3 1033Mhz
GPU : NVIDIA 9600M GS
HDD : WD 500GB 7200tpm

2nd :
CPU : Intel Pentium Dual Core T4500 (2.3Ghz)
RAM : 4GB DDR3 1033Mhz
GPU : NVIDIA G310M
HDD : 500GB 5400tpm

[Chiitoo]

After nearly exactly a year from [post=6461893]my first tests[/post], while upgrading the Kernel, I finally decided to take a more thorough look-see to the Processor type and features section of menuconfig and Lo and behold, the results sort of hath a slightly different kind of an air around them now:

Code: Select all

 ~ $ bashmark
#######################################################
:  T   E   S   T        :    :S C O R E :  : R A T I O:
:-----------------------------------------------------:
:Cpu, Integer           :    :      9133:  :     +751%:
:Cpu, Floating point    :    :      6799:  :     +796%:
:                       :    :          :  :          :
:Memory r/w (cached)    :    :     17345:  :    +1342%:
:Memory de-/alloc       :    :      5425:  :     +730%:
:                       :    :          :  :          :
:Multithreading         :    :      4182:  :      +68%:

Code: Select all

 ~ $ bashmark -d
#######################################################
:  T   E   S   T        :    :S C O R E :  : R A T I O:
:-----------------------------------------------------:
: 8bit integer          :    :     12803:  :    +1075%:
:16bit integer          :    :     11159:  :     +786%:
:32bit integer          :    :      9196:  :     +546%:
:64bit integer          :    :      4776:  :     +817%:
:32bit fp               :    :      9051:  :     +936%:
:64bit fp               :    :      8193:  :     +837%:
:128bit(80real) fp      :    :      3903:  :     +636%:
:                       :    :          :  :          :
:128B Mem r/w (cached)  :    :     38577:  :    +2712%:
:256B Mem r/w (cached)  :    :     12894:  :    +1147%:
: 32Byte Mem de-/alloc  :    :      6546:  :     +734%:
: 64Byte Mem de-/alloc  :    :      6605:  :     +771%:
:128Byte Mem de-/alloc  :    :      4063:  :     +865%:
:                       :    :          :  :          :
:MThreading u-/lock     :    :      4133:  :      +66%:

I didn't go about to test which option it was exactly that made such a drastic change concerning this particular test, though I think it is a combination of sorts, as I was able to get them lower by about 50% or so, can't remember, but I couldn't get them THAT low again. But as I mentioned, I didn't spend enough time to find out exactly. I think Tickless System had some effect at least. May bee.

I was also pleased to notice quite a change in some games, specifically some applications via Wine. The frame-rate and overall performance feels a lot better. This I would assume might have something to do with the Timer frequency setting, which was at 100 HZ, and has probably always been for me. After some reading, I tried 1000, and noted definite changes.

Might do some more testing now as it seems Portage is offering me a new Kernel once more again.

[john.newman]

this one right here

ASUS P6T
i7 950
24 GB 1600 (yes i do use a lot of that)
64 SSD raid1
2 TB raid10
GTX 570

compile times .. emerge -e world is a piss in the wind.

soon though .. it will all be giving way for this thing http://www.techpowerup.com/153916/EVGA- ... tured.html can't wait

[Mr. M]

Not sure if it counts but I have a gentoo-prefix installation on my workstation at work

- Intel Xeon X5660 (6 cores, 2.8GHz)
- 48GB RAM
- Nvidia Tesla C2050 (because the CPU is too slow for some things )

[john.newman]

^ that's YOUR personal workstation?? that's like one of our production servers...

wow... can i work for you guys