tested it yet? how's it clock?Quote:
Originally Posted by STEvil
Printable View
tested it yet? how's it clock?Quote:
Originally Posted by STEvil
Win2Kpro SP4
XP1700 & NF7-S (from PS2pcGamer)
Swiftech H2O system & 2x256MB BH-5 (from Jeff)
XP1700 @ 12x200 2-2-2-11 165 sec
XP1700 @ 12x200 2.5-4-4-11 166 sec
XP1700 @ 13x185 2-2-2-11 169 sec
All results were duplicated to insure consistancy.
Do timings make a difference? Yes, but very little.
Does FSB make a difference? Yes, but not much.
Seems to be like 3D rendering (e.g. Lightwave) where raw Mhz is king.
shmaa - 2600mhz @ 1.58v so far (used DMM to test voltage from missing capacitor spot near cpu, shows at ~1.51v in windows).
Does 2t make a difference? Like 2-2-2-5 2t and 2-2-2-5 1t?
probably not much, i'll give it a go when I remember to ;)
Wonder if mutiprocessor/uniprocessor driver (ricky tweak) affects it?
mine wont get past the "Press Enter" part.
Dual 2.8GHz Xeon HT enabled 1 instance - 269s
Dual 2.8GHz Xeon HT enabled 4 instances - 108.94s
Dual 2.8GHz Xeon HT enabled 8 instances - 85.45s
So is it better to run more than 4 instances?
you have to go to \D2OL\res\data\bin and copy dockwin & gridwin. I had that problem also.Quote:
Originally Posted by Magnj
I have it there. I have to make a copy of them though?
yea, copy them and paste them into the folder.
u have the benchmark in the folder with dockwin and gridwin already? No copies are needed if thats the case...Quote:
Originally Posted by Magnj
Wow... you sure you calculated the times right? :slobber:Quote:
Originally Posted by Helheim
Pretty sure, I used your formula.. Took an average and then divided that by the number of instances running. I'll try again tonight as well.
Then that's an interesting result!Quote:
Originally Posted by Helheim
running 2T on winchester instead of 1T was a 2 second penalty.
Does anyone else have trouble running 8 instances of DBench? On two different machines now two instances just stop running(ie. window closes) during the test. I've tried 4 times now and the same thing happens. One machine has 512Mb and the other has 1024Mb. :shrug:
ive tabulated most of the results just need to add ram timings and order them. should i do it by processor type or by time?
btw the attachment is the table so far done in excel then zipped coz it wont attach .xls
Ok so it's a bit flawed because 4 of the processes finish way before the others and if they were still running like it would be running regular d2ol they wouldn't stop and allow the others to finish..
http://www.xtremesystems.org/forums/...id=27341&stc=1
(521+447+510+490+883+848+801+889)/64=84.20s each
This is interesting:
Dual Xeon 2.8Ghz
Win2003 1 instance - 261s
Win2003 4 instances - 435,435,436,437 = 108.94s
Win2000 SP4 1 instance - 249s
Win2000 SP4 4 instance - 404,404,403,403 = 100.88s
Both were fresh installs with nothing else installed.
Jeff... i get the same thing... and if I can get them to run all the way through 4 of them finish well before the other 4... so it throws the numbers way off...
Well i finally got round to testing it
I got 147 on my 754 2800+ AMD 64 @2475 Mhz @ 275 x 9 7, 3, 3, 3
seems fairly good for a 745 based AMD 64
Anyone else get very mixed results with there dual xeon machine?
Iv been tryiong a few things like changing the priority of the dbench and am finding very mixed results, varying from 84 seconds to 95+ seconds.
the 84 seconds had 2 in the 280 second places and 2 in the 400+ seconds. Thats at realtime priority. So obviously the first 2 processes are being done at 80-90% of one cpu's power ( ht or not).
Also i found running at 3.5Ghz gave me slower timeings than 3.4Ghz, i would assume its something to do with it being near the limit of its overclock maybe throwing errors into it.
3.4Ghz is 100% stable for me while 3.5 is very close to being stable and i belive when it isnt run on an old generic powersupply will clock to 3.5 fully stable.
Just got myself a 83.5 Second stable with 267+263+416+390
obviously this is infact worse than Jeff's 84 seconds due to the fact on the later ending ones it does not take into account the ones that would still be going.
I did a fresh install of Win2000 and it was 261s (@3Ghz) running one instance, installed all the updates and everything and now it's 285s even after I disable all the unnecessary services, with them running it's 301s. :(
Well im baffled, really
2 runs one after another, NOTHING changed.
first one averages to 95, second to 80.
Im not convinced by running multiple copys of it.
Id love to know how repeatable it is for you other dual xeon guys
Mines been dead on 99% of the time...Quote:
Originally Posted by Smurfy