BOINC benchmark: post results

[AMD-C]

Starting BOINC client version 6.2.28 for windows_intelx86
log flags: task, file_xfer, sched_ops
Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
Data directory: C:\Documents and Settings\All Users\Application Data\BOINC
Processor: 4 GenuineIntel Intel(R) Xeon(R) CPU           E5504  @ 2.00GHz [x86 Family 6 Model 26 Stepping 5]
Processor features: fpu tsc pae nx sse sse2 mmx
OS: Microsoft Windows XP: Professional x86 Editon, Service Pack 3, (05.01.2600.00)
Running CPU benchmarks
Suspending computation - running CPU benchmarks
Benchmark results:
   Number of CPUs: 4
   2138 floating point MIPS (Whetstone) per CPU
   4734 integer MIPS (Dhrystone) per CPU

[ReKcOlNu]

Well unless noted in the members post I'm assuming if you are running a 64bit OS then Boinc is 64-bit also?

[MrDucky]

well i hope so for the people otherwise it sucks lol xD

[ReKcOlNu]

well i hope so for the people otherwise it sucks lol xD

I just figured both x86 and x64 versions were in the Boinc installer and it would auto install the 64-bit version if it detected a 64-bit OS?

EDIT OK got a PM from PaganII and cleared that up as he's running boinc x86 on 64-bit OS,.... so take the Bionc "bit" in the charts with a grain of salt.

[slaveondope]

Thanks! Wheres your X6?

As requested. All stock(undervolted) on Win7x64 with BOINC 6.10.18 64bit
The 955 was on 6.10.18 64bit too.

[ReKcOlNu]

As requested. All stock(undervolted) on Win7x64 with BOINC 6.10.18 64bit
The 955 was on 6.10.18 64bit too.

Thanks!

Sort by Integer


Sort by Floating Point

[mAJORD]

I've noticed, for some reason the x64 version of boinc gives a much Lower score for FPU on AMD CPU's.. (about 25% higher on x86) Anyone know why?

[kiwi]

16,5k integer? Holy crap... looks like Linux does beat Win in this bench. Curious to see the difference in actual crunching times

yep, just run on ubuntu x64 and getting 4k/21K @4Ghz E8400. I will try running a few nights under ubuntu to see the actual output. Win7 was getting 4.3/8.8

[bluestang]

Work PC as listed in sig...

10/23/2010 12:43:40 PM		Running CPU benchmarks
10/23/2010 12:43:40 PM		Suspending computation - running CPU benchmarks
10/23/2010 12:44:12 PM		Benchmark results:
10/23/2010 12:44:12 PM		   Number of CPUs: 2
10/23/2010 12:44:12 PM		   4216 floating point MIPS (Whetstone) per CPU
10/23/2010 12:44:12 PM		   8454 integer MIPS (Dhrystone) per CPU
10/23/2010 12:44:13 PM		Resuming computation

[Nikolasz]

Thats mine BOINC benchmark: post results

quad 9550 :
2884 whetstone per cpu
8456 dhrystone per cpu

and cuda gpu
432 GFLOPS peak




Is not much i know, i dont like ocing. It kills smooth ride like weard os behavior, os crushes, and exploding electronics

[haschioz]

29/10/2010 7:36:20 PM|| Number of CPUs: 2
29/10/2010 7:36:20 PM|| 1848 floating point MIPS (Whetstone) per CPU
29/10/2010 7:36:20 PM|| 3490 integer MIPS (Dhrystone) per CPU
t4200

[kaok]

Q6600 G0 3.0GHz, Ubuntu 10.04 x64.

2785 floating point MIPS (Whetstone) per CPU
14642 integer MIPS (Dhrystone) per CPU

Dhrystone MIPS in linux is somewhy much better. Not sure though, how much that makes overall difference.

[sRHunt3r]

Here's my new one with Win7 64 and boinic 6.10.58 64.

[PoppaGeek]

What am I not understanding here. A one cpu @3.5gz gets 8824 and another @2.3 gets 10342. What do these numbers prove? I am lost here.



Top one Opteron 1354 @2.4gz bottom Opteron 8356 at 2.3GZ

[Nikolasz]

i call that overclock data loss.

your not spesific anyway.
Maybe some overvolt will help, or some other too tight timings


if your using one pc and oc it up and down...
and enable/dissable cores it will free up cache, so then rest of cores got more cache..

[bearcatrp]

Dual quad E5420's running ubuntu
2940 whetstone
15000 dhrystone

Dual quad E5640's running win7 pro
2668 whetstone
6923 dhrystone

Not sure why my harpertowns are beating my westmeres.

[PoppaGeek]

Going by the charts above my 2.4gz opterons are beating a Athlonx2 @4gz, a Phenom2x4 @4gz and a i7 920 @3.8gz. I get very similar numbers on the 8356 @2.3. Of course they do not show this in PPD so what do these numbers mean?

[xVeinx]

The Dhrystone and Whetstone scores show something close to the max possible performance using simple calculations that fit entirely into cache (ie, the instructions can fit into the L1, so there is no instruction fetching). Since the work units are composed of much more complex code and don't fit entirely into L1 cache, architecture differences come into play far more. Actual performance (as measured in PPD) will vary based on the particular work units being processed. Each work unit type does something different in code, and some may be more optimized than others (generating more PPD). Linux is going to run most/all work units faster than on Windows, typically with a 10-20% speedup from what I can tell. This isn't captured by the PPD output, and I have no idea as to why one nets less PPD although having increased performance in linux.


I'm sure others here can explain/fill in details far better than I.

[D_A]

Basically, Windows benchmarks are as good or better optimised than Linux, while the units take longer to actually process in Windows. The net result is that Windows will net more points per unit while Linux completes more units.

[PoppaGeek]

Just ran the benchmarks on a 1354 2.4gz running Windows Vista 64 and got 2018 fp and 6223 integer. Seems closer to reality but still leaves me thinking these numbers don't mean alot. The Linux are 64 as well.

[[Ion]]

Main laptop (x64):


Dedicated cruncher (x86):

[kjeldoran]

i7 860 @ 3255mhz
2974 Floating
10464 integer

Windows Server 2008 R2 64 bit

Athlon x4 640 @ 3.4ghz
2559 Floating
10451 Integer

Ubuntu 10.04 64bit

Sossaman 2ghz
1893 Floating
3868 Integer

Windows XP Pro 32 bit

[bearcatrp]

Dual quad E5420's running ubuntu
2940 whetstone
15000 dhrystone

Dual quad E5640's running win7 pro
2668 whetstone
6923 dhrystone

Not sure why my harpertowns are beating my westmeres.

Deleted

[kjeldoran]

Just realized I am using boinc 32 bit. Would boinc 64 bit have much affect on my westmere box? I know the wu's are 32 bit but have read the 64 bit version is a bit faster.

Well, I thought the 64bit gives 10% more points than the 32bit version. I am not sure if effects the benchmark though.

[Nikolasz]

I think moust important is math processor, floating unit, fpu. Some cpus got more fpu in it..
I remember days when cpus didnt have fpu in inside cpu, and without fpu jubs takes 800% longer to do.

As you may see fpu units got more or less same score, depend on frequency of fpu.
Then again there are good fpu units and wery good fpu units with extra cache for fpu.
And there are graphic cards that do cpu jub nowdays

At programable fpus are abit slow, there are logic chips that may do same jub for less cycles.
Well idunno abaut much of electronics.

Os load thats important too, if your using all kinds of crap instaled on your pc it slows down your boinc.
Why is that?
Coz it cpu calculates your windows, WITH Boinc

I simply saying this, os must be clean and install only boinc and do only crunch. Then you get best boinc scores


Idunno if this link may help you alot, i wrote it for game/s, hope you can use it
http://www.bz2maps.com/phpBB/viewtopic.php?t=1597