[Splave]

Originally Posted by karmakazi

Well your XP isnt properly tweaked then

I can do a clock for clock comparison with reverendmaynard's best time on hwbot. He has 8m25.xxx s @ 4505Mhz in Windows 7

I will use XPSP3 or 2k3 maybe both if I have time

In other news ill beat it by at least 3 seconds same clocks/ram etc

And im not just making numbers up. Use performance product (PP) calculation to estimate your pi efficiency (cpu clock * pi time in seconds). You can then extrapolate/interpolate times at different clock speeds. The lower the PP the better your efficiency.

Im at work so I can only use examples from hwbot profiles. For instance in Reverend's run, 505.680 seconds * 4505 cpu = 2278088.4 PP

Versus my run 421.3 seconds * 5009 cpu = 2154768.7 PP

If you want to see my pi time at 4505 simply substitute... x * 4505 = 2154768.7. X=478.306 seconds. So I should theoretically be able to beat that score by a half minute or so. Now if I match my ram clocks to his that gap would of course close a little, but XP is still faster.

Therefore drnips 2073600 pwns your sox off?

[karmakazi]

Originally Posted by Chuchnit

[puts on hat]

Gents, lets keep it clean and to the facts. If you have proof for the pudding then post it up. No need to start getting personal. Consider this a warning for no further attacks.

[/takes off hat]

Karmakazi, we are all about sharing information around here. If you got something you want to share, then share away

Maybe we need to start a thread to do some real performance evaluation between the two OS's? Of course you have to compare with proper tweaks applied; not fresh full install to fresh full install.

Sorry, you have my apologies. I like this forum alot, and if PURE hadnt formed I was pretty close to asking Kal if I could join hwbot team etc...

I am all for the sharing of information myself, but to ask someone to give away literally hundreds of hours of painstaking and boring testing... Superpi isnt a big secret like some ppl make it out to be. All the needed tweaks are out there, you just have to look, learn and most important test them. I learned more about superpi from testing and running 32M 150 times (for all the testing I did it was at least that many) than I learned from reading all the forums I did (english, french, russian, dutch, german, italian, chinese, japanese, etc).

I would be willing to help with any OS testing you guys want to do. For testing OS I would recommend only using LSC, copywaza and maxmem tweaks. Many XP reg tweaks dont work/arent effective on vista and 7.

[karmakazi]

Originally Posted by Splave

Therefore drnips 2073600 pwns your sox off?

What run? Im just browsing his hwbot profile and his 5070mhz @ 7m13.xxx = 2216319 PP

EDIT: I found it, if your talking about his 3200mhz LCC submission? Lower clocks is easier to achieve lower PP.

[Splave]

ahh I see that does make sense thanks for clearing that up karma

[DrNip]

I will clear this all up guys when I get home. I will run the same settings and same 5 tweaks on my Windows 7 HDD. My bet is on my XP install.

[3oh6]

i am going to chime in here with my thoughts as i would fall under the 'pro's' category still using XP. i too believe XP is faster in SPi 32M, with that said, i haven't spent a significant time tweaking W7...actually couldn't say i ran 32M on W7 ever. i have never seen anything to contradict this either.

for you guys getting lower times in W7, can you run your best most tweaked run at 4GHz - 1000MHz 7-7-6 or 7-8-7 even - 4GHz uncore? possibly on a Classified? just for a reference time as there is a ton of data at 4GHz. the average is around 8m 44s for a well tweaked XP on Classified. a screen with CPU Tweaker for memory sub-timings would be great for comparison purposes. different motherboards are going to give different results, RTL/B2B CL/etc... all get set differently by different boards...but from another board would be fine as well.

any other W7 screens at 5GHz or higher would be great as well if you have them so i can compare to my slew of XP work. again Classy would be great but on any MB would do. unfortunately i am not at home right now so i can't partake in any direct comparisons at this time, just able to compare to screenshots i happen to have access to.

[Chuchnit]

Originally Posted by karmakazi

Sorry, you have my apologies. I like this forum alot, and if PURE hadnt formed I was pretty close to asking Kal if I could join hwbot team etc...

I am all for the sharing of information myself, but to ask someone to give away literally hundreds of hours of painstaking and boring testing... Superpi isnt a big secret like some ppl make it out to be. All the needed tweaks are out there, you just have to look, learn and most important test them. I learned more about superpi from testing and running 32M 150 times (for all the testing I did it was at least that many) than I learned from reading all the forums I did (english, french, russian, dutch, german, italian, chinese, japanese, etc).

I would be willing to help with any OS testing you guys want to do. For testing OS I would recommend only using LSC, copywaza and maxmem tweaks. Many XP reg tweaks dont work/arent effective on vista and 7.

karma, I wasn't asking you to hand them out bud. I was just throwing the offer out there. Everybody has there view on tweaks and have every right to keep them to themselves if they wish. Just wanna make that clear.

[karmakazi]

I will freely share and trade tweaks with any cool peeps, just hit me up on MSN

[dinos22]

lol amusing discussion here
i'm really enjoying the "experts" at work LOL


btw karma it's nice to know that you would have left us for another team as well had you not been invited to PURE....what did we do to you :(

[DrNip]

OK here we go. Some what of a comparison. Really my Windows 7 has a big disadvantage here as I am using a slimmed and trimmed nlite version of XP that has all the BS out of it where as I am using a fully loaded Windows 7. Also I was not able to use two of my tweaks on my Windows 7. One of them being Copy Waza as I don't have a partition on this drive and can't set it up the way I want to. That needed to be done on the install. The other tweak I'm not able to disclose right now. Whether that tweak actually has any real world performance, I can't say 100%. I have seen people rave about it but need to do some tests after the competition. Also this Windows 7 has been installed for months now where as my XP is a fresh install as of last week. I used the other three remaining tweaks on this run but whether they have the same effect as they did on XP I dunno as I haven't fully tested them. I think my score prolly would have been at least 3-5sec faster with the other 2 tweaks. There might also be tweaks that I dismissed on my XP install that didn't give me any gains that might of on my Windows 7. Along side my testing of my tweak that I can't reveal I will also do a standard XP and standard Winwos 7 test. Both will be full blown, bloated up installs with no tweaks what so ever to see what the real difference is between the two. I still have my money on XP in the end being faster. Enough of the talk. Here is my disadvatage Windows 7 run with 3 of the 5 tweaks I used on my XP run.