27s in SuperPi 1M Test
P4 3,4E @ 4634; P4C800-E; 2x512MB OCZ EB 3500 @ 272.6 @ 3-2-2-8
http://www.forumdeluxx.de/gallery/da...4927Screen.JPG
All I've done is the droop and vdimm mod. 1024MB 272.6mhz is nice on i875P, right?
Vcore 1.6V, Vdimm 3.4V
Printable View
27s in SuperPi 1M Test
P4 3,4E @ 4634; P4C800-E; 2x512MB OCZ EB 3500 @ 272.6 @ 3-2-2-8
http://www.forumdeluxx.de/gallery/da...4927Screen.JPG
All I've done is the droop and vdimm mod. 1024MB 272.6mhz is nice on i875P, right?
Vcore 1.6V, Vdimm 3.4V
Not bad :D ... but - dead link!!!Quote:
Originally Posted by xsky
fixed it ;)
very impressive :toast:
Dual channel? Wow. That rocks, my German fellow :thumbsup:
thanks a lot :)
yes it's DC with PAT and PAM enabled
it was cooled by stock LS (if I didn't mention before)
Quote:
Originally Posted by xsky
You did not mention it ;)
"stock" LS you gotta love this place =) Myself i get 35 secs in a moderately OCed barebon thats supersilent.
now warming up for 32M Test with good old Kingston BH-5.
starting with 1.45V Vcore and 3.4V Vdimm
http://www.forumdeluxx.de/gallery/da.../944924_46.JPG
Hello Im new here, and this my one
P4 2.8c @ 3.36 (240*14, 1:1)
Abit ic7
twinmos 400mhz 256mb *2 2.5-3-3-8
http://s03.imagehost.org/0070/SuperPI-40s.jpg
Welcome to the madness Andrew :)
Nice start. Are you sure you cant push the cpu any further and eventually running with a divider?
TNX,Quote:
Originally Posted by bias_hjorth
240mhz x 14 (CPU:Mem 1:1) its high level of my P4 2.8c D1 :\ bad cpu for OC
I cannot boot my comp at 240mhz with ratio 5:4 ;( (i think the problem with CPU and not with memory )
Nett, my friend! ;)Quote:
Originally Posted by xsky
Get it on... :D
Thx!
Here my first sub 24min bench:
P4E 3.4 DO @ 4511.6MHz
P4C800E-Deluxe
2*256MB Kingston HyperX 3200 @ 265.4MHz 2-2-2-5
http://maitre.physik.uni-kl.de/~mkrauss/23.58.JPG
Nice, how you do 1mb ?Quote:
Originally Posted by xsky
my result should be some posts before
27s
http://maitre.physik.uni-kl.de/~mkrauss/27Screen.JPG
this is done with 2*512MB OCZ EB 3500
niceQuote:
Originally Posted by xsky
I activated the pat on my P4P800SE and FSB250 stable now because the dram latency is at 101 , before it was 215
a little comparison with the pat disabled and dram latency at 111...
http://membres.lycos.fr/cpulloverclo...atdisabled.gif
http://membres.lycos.fr/cpulloverclo...patenabled.gif
P4 3.4E S478, wcooled, 1.45v in full, but u have to look at the Pat enabled at this fsb with the P4P800-SE :D
http://membres.lycos.fr/cpulloverclock79/31s.gif
http://membres.lycos.fr/cpulloverclock79/5m47s.gif
Hmm, cpu, no differences to spot there besides the pat thingy ;) :D
/edit. first post, I mean :D
yep, pat disabled + dram latency at 111 VS pat enabled + dram latency at 101 = same perf in superpi 104... elsewhere I dont knowQuote:
Originally Posted by SAE
Aaahhhh... now I understood. You wanted to show that the dram latency makes up for missing PAT!?Quote:
Originally Posted by cpulloverclock
Kewl :D
yes exactly :DQuote:
Originally Posted by SAE
Dunno how legit this patch is that updates FISTTP instructions for prescott:
http://www.vr-zone.com.sg/Shamino/AS8/patch.JPG
www.vr-zone.com.sg/Shamino/AS8/pipatch.zip
But it gave me a 23 second result:
http://www.vr-zone.com.sg/Shamino/AS8/23sec.JPG
Think it shaves off 2secs for the scotties
I test it at 4077MHz
wooooow. at 4ghz my score changed from 32s to 29s. please god let it be legit :)
as i understand it it's no cheat as the program does still the same calculation but optimized for p4e?
impressiveQuote:
Originally Posted by xsky
Makes it pretty much impossible to compare with original superpi results - which is not a good thing IMO. Where's the A64 patch?? :DQuote:
Dunno how legit this patch is that updates FISTTP instructions for prescott:
Yeah, but with this out there its impossible to tell which version its runnning so just run the fastest version? :p:
20-21s, the new world record??
Damn where is the A64 patch :(
After the patch its good for comparing different Prescotts but thats about it. Silly prescott-only bench.
I mean superpi has been the same for several years now and become a popular 'standard bench'. It was fairly easy to see the performance level of any given hardware (AMD or intel) just by running the SPi 1M test. Now someone (working for intel??) got a great idea and edited the code and made it 'prescott-optimized'. He should have edited the name of the benchmark too.
and I think that's not just optimized for the 1M, 8M and 32m too?
I test
It's not funny if there is nothing for the A64
yeh it would be more fun if there was an a64 patch but somehow it's ok as it is a prescott feature. maybe simular things are possible in real useful tasks and prescott becomes superior to the rest.
i don't know enough about these things but maybe someone can tell me...
Its not funny that it wont run on any other CPU than Prescott (just tested W/ my P4EE and it crashes instantly).Quote:
It's not funny if there is nothing for the A64
I dont see how this edited-superpi has anything to do with any kind of realworld performance now.
:banana::banana::banana::banana:ty thing that in weekend gonna cool with co2 and ln2 myh precious FX-53 and with out that patch. :( It ain't so funny anymore
5m16s at 4097MHz!!!!!!!!!!!!!!!!!!
bad trip
hehe =)Quote:
5m16s at 4097MHz!!!!!!!!!!!!!!!!!!
bad trip
I wonder if Aguri has been useing this patch for his latest DDR2 Superpi benchmarks? If not then it seems that he is the only person on the planet getting decent performance out of DDR2 mobos :D
edit: or maybe there is a 'DDR2 patch' available too??
yes he has because he gets 31s at 4148MHz like me but with 230/306Quote:
Originally Posted by macci
http://members.jcom.home.ne.jp/aguri...50-E0_test.htm
Yep, its definitely way off in terms of Real world performance
Have no idea at all who wrote the patch
Quote:
Originally Posted by LardArse
Hmm checked ABOUT in the patch
-={ CodeFusion Runtime }=-
Generator: CodeFusion Wizard
Super PI was reverse engineered by
Igor Levicki
Here's an improved one of mine, nothing special, but i was happy to get it.
http://www.sevenlabs.net/sucka/pictures/superpi5.jpg
Sorry, just keep trying for more...
http://sevenlabs.net/sucka/pictures/superpi6.jpg
Here's mine AKA Vflux/Henesseynalert
Sucka, whats up!!bro!!! :)
Air Cooling 3.2E 5:4 Cas2.5-3-3-7 BL8 PAM Auto/Standard
http://images.dr3vil.com/uploads/Prescottsuperpi2.JPG
LS cooled 3700+ @ 1.66v:
http://server5.uploadit.org/files/nexus2525-1M30s.JPG
Messy WinXP, stuff run in background. Only had this 3700+ for less than a day. I think it can do better.
it's realy "ScrenShot" :DQuote:
Originally Posted by Tek
hmm CO2 didn't improve my 1M time at all, but 8M 3 second faster than with prometia machII r404. Now meh best time in 8M is 4min 49s
SPi 8M (non-patched) w/ 2.8A (D0):
http://www.akiba-pc.com/2.8A/424.gif
Here's the 8M Ranking at the moment and as far as I can tell none of the top 7 results is done w/ the patch.
http://www.akiba-pc.com/2.8A/spi8mrank.gif
Neither is mine at place #10 ;)
5468Mhz and 4:36
Congrats on the 4:24 time!
:banana:
Still way behind even with ln2 coz couldn't get 13 multier get working.
best 8m time with 12x270 4min 46s
Macci, Is that SuperPI on your download site unpatched?
TIA
Henessey
In Finland we NEVER use cheat patch, that patch sucks so hard...
Fair Play :toast:
I just tried the one from macci's site and the one with patch. I got the same score..........What's the deal!!!I'd like to run mine clean. No pressy optimisation ofcourse.... :toast:Quote:
Originally Posted by Verdict
the one at akiba-pc.com is unpatched and was uploaded there long before the patch was even available. see if you find a difference between the versions in 2m or 4m test? 1m may in some cases give the same result I guess.
Why can't people just bench by the rules instead of reprogramming every-f***ing peice of software to score a lil better :mad: :confused:
Bro, I got the same results from 1M-4M with the SuperPI i downloaded from your site and the one i runned my 32sec record. I guess i've been running the un-patched version all along then..Since there's no way i can tell which one is patched or what not....I just had to make sure I trully earned my 32sec respectively. :toast:Quote:
Originally Posted by macci
Respect,
Henessey
Yup Your more than likely running the unpatched version - like most of us. :)
Haven't seen pre-patched version upped anywhere yet so basically your running the original version if you haven't downloaded the patch and patched the superpi.exe file.
Nope, I didnt even know there was a patch ti'll I read all through page 1 from this thread. All i know is that i've had this copy of SuperPI back then in april....So i guess my 32sec is legit!!! :banana:Quote:
Originally Posted by macci
Thx for clarifying.......
Respect,
Henessey
Someone can give me link for this Patch, and explain me what it do ?
Thx
according to your sig you dont have a prescott and you shouldnt anyway, its cheating. but the link is earlier in the thread.
Ok .. TNXQuote:
Originally Posted by aoc007
My base for 8M Superpi (Unpatched) I'll slowly break it soon. All done aircooling.
http://images.dr3vil.com/uploads/Prescott8msuperpi.JPG
Henessey :toast:
my new record :D with my single stage phase change.
WTG Bro, nice :D :toast: :DQuote:
Originally Posted by Unseen
Finally!
http://people.freenet.de/funkflix/35secpi.jpg
But now im out of the nforce2-World! S939 is waiting for me :cool:
Results next week then.
Impresive results Unseen....Quote:
Originally Posted by Unseen
Keep going lower... :toast:
You was sure, on my NW i cant run the patchQuote:
according to your sig you dont have a prescott and you shouldnt anyway, its cheating. but the link is earlier in the thread.
Its only for presscott.
http://www.takorautaa.net/img/tuloks..._superpi8m.jpg
Too bad I forgot to use the i875-tweaker :( Could have shaved off a second perhaps. Not that i'd care for that one second anymore, since with a prescottpatch I can shave off 20 seconds :rolleyes:
http://www.takorautaa.net/img/tuloks...scottpatch.jpg
This is not a legit score ;)
whats a Prescott patch?
page 44Quote:
Originally Posted by jjcom
Quote:
Originally Posted by LardArse
Congrats Hilppa :toast: :toast:
Old pap needs to figure out what went wrong with myh 13 multiplier :mad:
Some tests between the S478 and the S775 with a superpi (unpatched) :
P4C800 + 3.4E = 31s at 4040MHz full timings and tweaks
P4P800SE + 3.4E = 31s at 4077MHz full timings and tweaks
AA8 + 3.6E = 31s at 4152MHz DDR2 in asynch mode
P5P800 + 3.6E = 31s at 4160-4165MHz full timings and tweaks
The S775 has very bad perfs in DDR + DDR2...
best frequency in Water cooling for a superpi 1M with my 3.6E : 4286MHz and always 31s :rolleyes:
Why was the Prescott patch released anyway - A very stupid idea.
http://users.pandora.be/gamer/Gskill/31.JPG
personal best.
Whats the batch code of this 2.8a?Quote:
Originally Posted by Hilppa
very nice, LN2?
I got 33 seconds with my LS...check my specs and tell me what I can change to make this better!
P4 3.4c@4.27 251x17 -27 degrees celsius per BIOS
Corsair XMS 2.4-4-4-8 will not boot any lower timings running at 1:1
Vapochill LS
ASUS P4C800-E Deluxe BIOS 18
CPU voltage is 1.95
DDR voltage 2.65
AGP/PCI locked at 66.6/33.3
33 seconds SuperPI @ 4.27 P4 phase changed
http://img.photobucket.com/albums/v3...ro/superpi.jpg
3500+ @ 2.5ghz 1.6v
Corsair 3200XLPro 250HTT 2.5-3-3-7 2.85v
Msi Neo2
Nothing like you guys best its my best :D
My best time with my 3.4E S478 D0 water cooled at 1.52v
the highest frequency with superpi is 4238MHz, not bad
My P4P800SE is better than my P4C800dlx with worse tweaks
4077MHz for 31s with my P4P800SE (dram latency 01)
4084MHz for 31s with my P4C800dlx (dram latency 111)
!!!!
SUPERPI UNPATCHED
http://membres.lycos.fr/cpulloverclock79/30s.gif
29s at 4320MHz
SuperPI 1M at 4258Mhz for the best frequency
http://membres.lycos.fr/cpulloverclock79/5m38.gif
http://membres.lycos.fr/cpulloverclock79/5m38..gif
could someone tell me, what the actual SuperPI-records for Intel and AMD are?
i think 23s for bothQuote:
Originally Posted by p4z1f1st
lol, ich hab doch extra hier gefragt, um DIR eine antwort geben zu können @ P3D ;)Quote:
Originally Posted by blander
this setup: http://www.try-this.dk/Projekt%20Iso...lering17,2.jpg
with liquid @ -4,2ºC from my waterchiller
gives me this (never mind the vcore, it flucates :rolleyes: ):
http://www.try-this.dk/33sek,1.jpg
EDIT: pictures of isolation here: http://www.tweak.dk/forum/thread.php...id=1&styleid=1
CPU Prescott 2.8 D0 @315*14
Mobo: ASUS P4C800E-D rev2.0 droop modded
Ram: Corsair 3200LL BH5 256×2 Vdimm 3.54v with DDR booster
VGA Card: Hercules GF2 Ultra
Cooling: Mach2
Pi 1M=29s :toast:
http://digi.pchome.net/2004/10/29/nVIDIA_1099033467.jpg
better than all the japanese :DQuote:
Originally Posted by cpulloverclock
http://xoomer.virgilio.it/kikoguru/img/3042.png
http://www.biwa.ne.jp/~yok/P4-3G-GP/P4-3G-GP-TABLE.htm
Ciao
getting a bit better.
:banana: Ah yeh finally cracked 30 seconds. But, used my Mushkin BH5.
:toast:
A little AIR-COOLED Winchester action....
30s @ 2810mHz
And same winchester with boxed cooler 30s and 2729Mhz.
Quite amaising prosessors these are. :D
nice...Quote:
Originally Posted by Mauri Antero
does tras=0 help your sPi?
tras 0 improves mem speed a bit, so is also helps with superpi
8M superPi runs in 5min 35s
I might give that a try, previous testing on a S754 chip showed no improvement in scores from Tras=5 down to 0, but I never tested for improved stability ;)Quote:
Originally Posted by Mauri Antero
2 seconds better result with 8M 5min 33s with boosted air cooling (cold morning and windows open)
25 sec at last, on win2k too ;)
http://home.exetel.com.au/codered/fx53/superpi25sec.jpg
It wouldve been easier to just install win2003 than squeeze that last 20 MHz out of this chip :D
Looks like the MSI is doing a good job :toast:Quote:
Originally Posted by CodeRed
Thats some real highspeed action there CodeRed :toast:
Nice timings :D