Rosetta Benchmark

[[XC] moddolicous]

Sempron 2800 @ 2390mhz (8x298) (mem running 1/2 divider )
2135 double precision MIPS (Whetstone) per CPU
3992 integer MIPS (Dhrystone) per CPU
Thats without the tweaked boinc thing. I'll do that now and report back.

[[XC] moddolicous]

Alright, installed the tweaked boinc client, and here are the results now:
3303 double precision MIPS (Whetstone) per CPU
10010 integer MIPS (Dhrystone) per CPU
Hopefully for my brothers birthday I'll talk my parents in getting some real ram.

[L0$t Pr0PhEt]

Alright, installed the tweaked boinc client, and here are the results now:
3303 double precision MIPS (Whetstone) per CPU
10010 integer MIPS (Dhrystone) per CPU
Hopefully for my brothers birthday I'll talk my parents in getting some real ram.

omg your sempron gets more than my a64 with same clockspeed, im gonna get me a sempron Is that a s754 and what timings are you running?

[[XC] moddolicous]

yea, s754, and 2x256 @ 2.5-3-3-7 @ 149fsb (half divider). If I had some 1:1 ram or some 2-2-2-5 ram @ ~230 (any suggestions) I'm sure they will be higher. This week I'm trying to find a stable ground for my A-XP 2000. Right now its memtesting @ 2171mhz (16x136), and for some reason 18x133 didnt boot (no screen). Maybe more volts are needed.

[L0$t Pr0PhEt]

yea, s754, and 2x256 @ 2.5-3-3-7 @ 149fsb (half divider). If I had some 1:1 ram or some 2-2-2-5 ram @ ~230 (any suggestions) I'm sure they will be higher. This week I'm trying to find a stable ground for my A-XP 2000. Right now its memtesting @ 2171mhz (16x136), and for some reason 18x133 didnt boot (no screen). Maybe more volts are needed.

Very strange, my a64 is at 2400MHz with 300fsb 3:2 divider, 200fsb @ 2,5-2-2-5 and my score is a bit less than yours. Do the s754 semprons oc quite well or not?

[Movieman]

This is what I just got with the "optimised" driver: (HT is on)
1/10/2006 12:02:06 AM|| Number of CPUs: 4
1/10/2006 12:02:06 AM|| 1857 double precision MIPS (Whetstone) per CPU
1/10/2006 12:02:06 AM|| 2052 integer MIPS (Dhrystone) per CPU
1/10/2006 12:02:06 AM||Finished CPU benchmarks
Now either I got the slowest PC short of granny's old P1-166 or I did something wrong!
This is on a DX 3600/2mb/800,2 gig DDR2-400, Raptors in raid
(Image of me getting the sledgehammer out and making lots of little PC's out of one big one!)

[lv_dicedealer]

Tested the Crunch3r Optimized clients on my 3GHz DX rig (15*200)

Here are the average results of 6 benchmarks....
5.2.11 Client : 1462.83 Double Precision MIPS / 2097.33 Integer MIPS
5.2.13 Client : 2001.33 Double Precision MIPS / 1720.83 Integer MIPS

Now the question is..... which is more important DP MIPS or Int MIPS?

[L0$t Pr0PhEt]

Tested the Crunch3r Optimized clients on my 3GHz DX rig (15*200)

Here are the average results of 6 benchmarks....
5.2.11 Client : 1462.83 Double Precision MIPS / 2097.33 Integer MIPS
5.2.13 Client : 2001.33 Double Precision MIPS / 1720.83 Integer MIPS

Now the question is..... which is more important DP MIPS or Int MIPS?

They are equally important, your score is calculated like this:

(Double precision + integer) / 480 = your score per hour

[ex2cib]

They are equally important, your score is calculated like this:

Double precision + integer / 480 = your score per hour

somehow, i dont think thats right, if that were the case, i would be doing around 850 a day with my rig, which i dont see happening. 500 maybe, 800, no

[Frederik]

This is what I just got with the "optimised" driver: (HT is on)
1/10/2006 12:02:06 AM|| Number of CPUs: 4
1/10/2006 12:02:06 AM|| 1857 double precision MIPS (Whetstone) per CPU
1/10/2006 12:02:06 AM|| 2052 integer MIPS (Dhrystone) per CPU
1/10/2006 12:02:06 AM||Finished CPU benchmarks
Now either I got the slowest PC short of granny's old P1-166 or I did something wrong!
This is on a DX 3600/2mb/800,2 gig DDR2-400, Raptors in raid
(Image of me getting the sledgehammer out and making lots of little PC's out of one big one!)

You got 4 threads, that means: 3909 as CPU score x 4 (see the red per cpu, I suppose HT is on) equals 15636 as a total score. So each day you can get 15636/24 = 32,575 x 24 = 781,8 points out of that machine, which is nice. I assume you work on the pc so I estimate you can get around 600-700 ppd. Not bad.

@ex2cib:The score is calculated this way. 850 would probably be the score you get when you do nothing else on the machine, and even when it was a dedicated machine you're max score would probably be only 800 or something because of other programs running...

[L0$t Pr0PhEt]

somehow, i dont think thats right, if that were the case, i would be doing around 850 a day with my rig, which i dont see happening. 500 maybe, 800, no

If I calculate mine it's almost 100% correct and I didn't invent that formula, it comes from somewhere...

[Vapor]

Is it (DP + Integer)/480 or DP + Integer/480?

EDIT: NVM, I'm dumb, shoulda just run the numbers out.

And also, that number is slightly off with dual cores due to bandwidth sharing.

If I don't do anything I get 94% of 2x it (due to being dual core).

[Movieman]

You got 4 threads, that means: 3909 as CPU score x 4 (see the red per cpu, I suppose HT is on) equals 15636 as a total score. So each day you can get 15636/24 = 32,575 x 24 = 781,8 points out of that machine, which is nice. I assume you work on the pc so I estimate you can get around 600-700 ppd. Not bad.

@ex2cib:The score is calculated this way. 850 would probably be the score you get when you do nothing else on the machine, and even when it was a dedicated machine you're max score would probably be only 800 or something because of other programs running...

OK, so I'm dumb! I took a look at the numbers compared to other systems and was VERY disapointed! I'd heard that the P4's took a hit as compared to the AMD cpu's and when I saw the guy with the sempron's scores forgot to see that I was running 4 threads compared to his one.
Ten hours into this I'm showing 190 points posted. We'll see what it does over a 2-3 day timeframe.
Thanks for your imput..( Puts down the sledgehammer)

[WesM63]

P4 1.7ghz 400mhz FSB (stock)

Running CPU benchmarks
Benchmark results:
Number of CPUs: 1
1458 double precision MIPS (Whetstone) per CPU
2471 integer MIPS (Dhrystone) per CPU
Finished CPU benchmarks

Weak! thats even with the optimized files.

[cx323]

3568 double precision MIPS (Whetstone) per CPU
11019 integer MIPS (Dhrystone) per CPU

[[XC] moddolicous]

Is that with the rig in your sig??

[cx323]

Is that with the rig in your sig??

yea

[WesM63]

Athlon XP 1800+ (stock)

Running CPU benchmarks
Benchmark results:
Number of CPUs: 1
1606 double precision MIPS (Whetstone) per CPU
4336 integer MIPS (Dhrystone) per CPU
Finished CPU benchmarks

W/O optimized client 1303/2229
Psh.. gimme an AMD any day of the week.

[gpcola]

P4 1.3 Willamette @ 1.49 running 5.2.11 optimised client

1205 double precision MIPS (Whetstone) per CPU
3130 integer MIPS (Dhrystone) per CPU

[WesM63]

Brothers A64 3400+ (Single channel S754, w/e stock timings are for OCZ EB)

Benchmark results:
Number of CPUs: 1
2450 double precision MIPS (Whetstone) per CPU
7856 integer MIPS (Dhrystone) per CPU
Finished CPU benchmarks

Using the optimized client afaik. (testing something)

[Rodzilla]

Brothers A64 3400+ (Single channel S754, w/e stock timings are for OCZ EB)

Benchmark results:
Number of CPUs: 1
2450 double precision MIPS (Whetstone) per CPU
7856 integer MIPS (Dhrystone) per CPU
Finished CPU benchmarks

Using the optimized client afaik. (testing something)

That's not too bad! One of the better A64 benches I've seen with no memory tweaks

[Paladin]

Did some testing to get my DX rigs at peak points/day. Tests were on a PC-DL, 2.4M0@16x200, 1gb@2-2-2-5, 74gb Raptor. Scores averaged between multiple benches.
Note: scores were often wildly different from run to run, I suggest running it til you get the highest repeated and let it run with that result for actual use (at least til it recalculates automatically in ~week).

Linux
stock: 892 / 2228
boinc_5.2.13_i686-pc-linux-gnu.sh: 882 / 2215

(Linux client only reported 1 CPU, Windows reported 4)

Windows Server 2003
stock: 2079 / 1420
Boinc_5.2.13_P4_Athlon64_SSE2.zip: 2080 / 2170


Boinc_5.2.13_P4_Athlon64_SSE2.zip vs Boinc-5.3.6.tx11-sse2.zip
These tests were on a different DX (PC-DL, 1.6D1@15x207, 512mb@2-3-2-6, WinXPpro).

stock: 1334 / 1045
Boinc_5.2.13_P4_Athlon64_SSE2.zip: 1950 / 1960
Boinc-5.3.6.tx11-sse2.zip: 1785 / 1967

[2fink]

Intel(R) Pentium(R) 4 CPU 3.00GHz @ 3.3ghz
Number of CPUs 2
Memory 1007.48 MB
Measured floating point speed 2116.07 million ops/sec
Measured integer speed 2051.92 million ops/sec

[ex2cib]

is that optimized 2fink? looks low to me

[2fink]

is that optimized 2fink? looks low to me

yes, it is optimized. will reinstall the optimation later and rerun the bench. my p4 dont has inet at the moment so i cant test it!