OPB clearly states pagefile should be 512Mb, maxmem=632.Quote:
Originally Posted by cadaveca
Printable View
OPB clearly states pagefile should be 512Mb, maxmem=632.Quote:
Originally Posted by cadaveca
Mimic as in identical or close to the same performance?Quote:
Originally Posted by G H Z
What did opb say? pagefile?:DQuote:
Originally Posted by mrlobber
I haven't applied that yet and already 0.8secs gain :D
Where can i do that :D
right click my computer, select properties, advanced tab, under performance button is pagefile settings.
yes, but do you understand why?Quote:
Originally Posted by mrlobber
No :D I have some ideas but until I don't get the tweak working, I wouldn't like to comment on them.Quote:
Originally Posted by cadaveca
Btw, I should correct myself, OPB said maxmem=600, 632Mb is for the files to be archived :)
ok i will try that tomorow :DQuote:
Originally Posted by cadaveca
I agree with this 100%Quote:
Originally Posted by Gautam
It will make a difference if your pagefile is on ramdisk, combined with copywaza. If you run superpi from ramdisk, the time it takes to start each iterartion should be less than if the data was sourced from HDD. This effect would be compounded by higher clockspeeds...furhter added to form starting Suerpi, and using copywaza to force the PI data out, and then windows puts it back. This is why you must wait a moment before benching...for windows to "reset" itself.
you can also get some extra time by imaging entire windows install, defragged first, of course, then placing it onto fully formatted drive, and benching from that.
Do some comparisons...I've done everything you mentioned and I'm summing it up into one sentence for you. Take my experience for what its worth.
I will, and I totally accept that you are right. What I don't accept is how this tweak could work. Just trying to look at every angle, I'm not saying you don't know what you're talking about. There's no emotion in this for me at all, just a bit of curiosity. I mean really, that's a real complicated bit of mumbo-jumbo...WTF would someone do that?
Same...I wish I could try it but I can't right now. It seems to me like it has some effect, but its just a variation on copy-waza, one of many variations.
yeah, but if you can get it right, you might end up with nothing but superpi in pagefile, windows vmem cache, and ram. I can understand how this may prove beneficial, and it also could expalin why some iterations are faster than others...it would depend on where windows sourced the data from. There's no waya 32MB run is all gonna come from cpu cache...and it really wouldn't run the same on AMD, for sure.
I do apologize that all these make several friends confused. However, if Elmor can try to pick up my "broken English" and reserch in depth, I don't see why some inteligent overclockers who just try to be picky on my writen English on my own thought whih I did NOT give any permission to anyone posting here and get this huge deabate?
Regardless if this is just a variation or one of many variations of copy waza, you have no RIGHT to critize my work like that. Gautam, you have no idea what did you do. Moreover, by showing that "not so brilliant 3.6g 32m run" , I sincerely suggest you go deep to get some improvement then post later.
Again what massman post here is not even express the whole CDT-IV from my own, that's merely just the quick explanation from the MSN conversation between I and one of my OCX members , Kain, and I thought I could forget that very soon , so I post immediately at coolaler's forum. Also, I have no idea why this post can be post here without having my permission?
Let 's not talking about any tweak or OS, let me invite you for taking this challenge Gautam, would you please match this run? not a competition or so, I am just talking about the raw power of memory subtimings;)
That 's all for now, I appreciate Fugger call me lastnight and being very nice to me. So long everyone.
Thanks for those who always support me.:)
I won't be back here and post until my name is cleared up here in public.:)
http://www.ocxtreme.org/opb/notweak36gspi1m.png
w00t :eek: :eek: :eek:
a) I don't have any hardware on hand.
b) As I've clearly said, I want to see a third party's times, not yours.
OPB, 49500pp at 3600 and with <600MHz DDR2 just isn't possible with Conroe/Kentsfield....heck, it's BARELY possible with DDR3 clocked sky-high.
Not only that, but here's a 48444pp you tried to pass off as the 36GP WR before:
http://xs220.xs.to/xs220/07426/36gpi1mfastest.jpg
That ^^^ is frankly impossible...
This tweak may be worth a few seconds on 32M, which is great...thank you for sharing. However, let's call a spade a spade and acknowledge what scores are impossible at the claimed clocks.
I didn't want to post this but some of the last posts, and the way somebody expressed himself here, I'm gonna get cheeky.
No screenshots and sh1te, wanna clear things up ?
Instructions for a reproducable tweak that shove of those ms for the superpi, not the best product score one, let's take the 8.91s result this time.
5550MHz C2D scoring 8.91s on 1M.
Scores from very good SuperPi benchers at the same frequency are like...hmm
9.11s @ 5548MHz... , 9.09 @ 5552MHz...
Now...a 5914MHz C2D of another well-known bencher ( a very respected one ;) ) scored 8.563s ... a 5728MHz C2D from another well-known bencher scored 8.938s
Give us a tweak that shoves up 200ms on 1M ( by itself ), because all the known tweaks have been used :)
so if you don't have anything on hand, then you better be quiet thenQuote:
Originally Posted by Gautam
as I have clearly said , I want you take the challenge, so I don't care if you have hardware or not, just yes or no you can take that.;)
vapor, there is NONE of your business, so don't even jump up into this , and BTW, don't even show me your lously calculation about that again, I don't think that can be any efficient for upcoming 45nm cpu onto any pi calculation anymore..:)
why your English is so bad I can't even understand what you're talking about then? None of your business either, even fugger gave me respect when he call me, what do you dare to post like that?;)Quote:
Originally Posted by BenchZowner
if you guys keep this attitude, , I am sorry that , I will ask fugger just delete my account, I won't be back to here and say anymore. Sad, this place is still like something discrimination forum huh? very good.hahaha
I don't think that it's too hard to understand what I've posted, looks like "I don't like this question, I'm gonna ignore it" to me ;)
As for the "none of your business part", well if you're going to react like that, why bother posting in here ?
I guess it all comes down to who you'd like to listen & talk to.
As for the last two sentences of your post:
1) Fugger told me that, said so, did so... with all due respect to Charles, this doesn't mean anything.
People respected Bush, does that indicate or even prove that he wasn't/isn't involved into nasty stuff ? Nope.
2) What do I dare to post like that ?
Who am I talking to ? You're not God, are you ?
friendly as always,
BZ
Quote:
Originally Posted by BenchZowner
oh I am not god compare with you then, so why bother to listen to you then, I see and read enough for your post just try to stir things worse;)
Penryns already have a known 'good' performance product of roughly 45000. Barcelonas have one at around 70k. K8s have one at like 75k IIRC. Conroe/Kentsfield/DDR2 of just over 50k.Quote:
Originally Posted by Onepagebook
Each processor has a difference performance level, and the performance product is a VERY easy way to describe that.
Anybody cheating is everyone's business. It's unacceptable. The scores you've gotten are just not possible...
No discrimination here...your scores are just not possible. You're not a victim, you're the one who is trying to pass off 'WR' scores that are not possible.
This CDT tweak is good for a few seconds on 32M, it does not 'clear your name' as it still isn't coming close to the performance your scores claim to have.
This is the best I've ever done at 3600MHz Kevin.
http://www-personal.umich.edu/~gautamb/3607_13828.PNG
If you're willing to divulge whatever it is that you're doing leading to such fast times, I, and surely everyone else here are all ears.
edited