DFI 875P-T @ 6Ghz

Printable View

Show 100 post(s) from this thread on one page

11-03-2004, 04:48 AM
Xassius

Quote:

Originally Posted by Mini

Tell me why the new patch is a cheat???

Isnt it just a patch that utilizes the new instructions and stuff that the prescott has???

Thats like saying that when ATI optimizes the drivers for X800XT core then they cheat?

Or that its cheat when you use a 64Bit OS and 64bit superpi???

Its just an advantage the CPU has... I think its fine...

it may not be strictly a cheat...but certainly doesn't promote a universality (is there a word?) for comparing against other results...as you cannot tell whether it is patched or not.
11-03-2004, 05:04 AM
Mini

Quote:

Originally Posted by Xassius

it may not be strictly a cheat...but certainly doesn't promote a universality (is there a word?) for comparing against other results...as you cannot tell whether it is patched or not.

True... But still... I mean its a few feaures the CPU has...

Then 3dmark05 is a cheat??? All that are not pro users cant choose if they want to bench with PS3.0 og 2.0b???

Thats just silly... You compare everything that the CPU can...

Not just some silly Mhz... Thats silly???

½ of WR are made on win2000 with tweaks and all sorts of stuff??? Isnt that cheat as well???

Thats just pure silly IMO...
11-03-2004, 05:07 AM
Xassius

that's why you post both screenies to avoid any argument ;)

EDIT: Before you know it...there will be DOOM 3 Athlon 64 Edition and DOOM 3 Prescott Edition for tweaks/optimisations :p:
11-03-2004, 05:42 AM
Mini

That would be great...

You cant compare like that...

Thats like comparing SCSI 15k HDD with a 5400 PATA Harddisk :D

Except its a cheat if the SCSI runs at 15k RPM...

You should show the FULL potential of every peace of hardware...

Limiting a piece of hardware like that is just sad...
11-03-2004, 05:43 AM
Kunaak

to me, this "patch" is somewhere closer to being like the wireframe hack, or PS-null cheats on 3dmark.
it gives a unfair advantage that other people with prescotts won't have.
personally- this makes me doubt a whole lot of 5 ghz+ overclocks now.
makes me like seeing the A64's and thier screenshots even more, cause I know they are directly comparable to what we all should get if we had the same settings.

if 1 computer is 4 ghz, and getting 34 seconds, and another is 4 ghz and getting 31 seconds, somethings weird there, and thats more then "optimising" to me.

using a64 tweaker to get 1/10th of a second faster per tweak on the ram is one thing, but using a whole different version of pifast so you can get a 5 second faster score, is another thing.

just like how theres "invalid" versions of Pi-fast out there.
it now seems super pi, has it's own invalid version.
I personally find it hard to believe a prescott would have that much of a advantage over a northwood, even with SS3 and such...
11-03-2004, 05:46 AM
Xassius

Quote:

Originally Posted by Kunaak

to me, this "patch" is somewhere closer to being like the wireframe hack, or PS-null cheats on 3dmark.
it gives a unfair advantage that other people with prescotts won't have.
personally- this makes me doubt a whole lot of 5 ghz+ overclocks now.
makes me like seeing the A64's and thier screenshots even more, cause I know they are directly comparable to what we all should get if we had the same settings.

if 1 computer is 4 ghz, and getting 34 seconds, and another is 4 ghz and getting 31 seconds, somethings weird there, and thats more then "optimising" to me.

using a64 tweaker to get 1/10th of a second faster per tweak on the ram is one thing, but using a whole different version of pifast so you can get a 5 second faster score, is another thing.

just like how theres "invalid" versions of Pi-fast out there.
it now seems super pi, has it's own invalid version.
I personally find it hard to believe a prescott would have that much of a advantage over a northwood, even with SS3 and such...

Yeah...2 seconds gain in 1M superpi is already quite hard for your non-prescott cpu. Prescott cpu's now can just apply this patch...

hmm...

First..where did this patch come from? Who wrote it? Any release notes?
11-03-2004, 07:28 AM
chilly1

Quote:

Originally Posted by pc ice

what's it benchable @

5.6 /5.7 was benchable but is more a limit of the agp buss right now... Fugger was working on it last I knew.. PC ice the R14 will be in tomorrow....
11-03-2004, 07:34 AM
Ailleur

This whole cheating thing is ridiculous, it calculates Pi, patched or non patched.
Take full advantage of every one of the cpu's advantage.
I guess if a 64bit version of superpi was released it would be invalid? (If we assume superpi would gain from a 64bit port).
If amd doesnt use sse3, or intel doesnt use 64bit (well, you know) its not the users problem. The user should try to get the most it can out of his hardware.
11-03-2004, 07:52 AM
kcnyc

Quote:

Originally Posted by dpa

/me like patch :D

http://www.home.no/dpa/Ftp/Bilder/med%20patch%203.9.JPG

With Patch:

http://www.home.no/dpa/Ftp/Bilder/med%20patch.JPG

Without Patch:

http://www.home.no/dpa/Ftp/Bilder/uten%20patch.JPG

Where do we get the patch?
11-03-2004, 08:06 AM
Scyphe

Quote:

Originally Posted by Ailleur

This whole cheating thing is ridiculous, it calculates Pi, patched or non patched.
Take full advantage of every one of the cpu's advantage.
I guess if a 64bit version of superpi was released it would be invalid? (If we assume superpi would gain from a 64bit port).
If amd doesnt use sse3, or intel doesnt use 64bit (well, you know) its not the users problem. The user should try to get the most it can out of his hardware.

Yes. But this is benchmarking, to be able to compare two systems. If the program doesn't say it's patched or not, how do you know if the benches are compatible? Without the patch, the numbers are comparable to other CPU's without the same "optimizations", ie. they will have a common ground to be compared on.

Imagine having two highspeed trains racing. One train is running on plain old standard railroad, the other is running on special tracks designed only for speed. That would make the race unfair and the results invalid. If they run on the same tracks, that would be a totally different matter, then it would be up to the train and crew to beef up the engine.

Lol, sorry for the metaphore...
11-03-2004, 08:14 AM
Ailleur

Im not interested in knowing which cpu is fastest under certain circumstances, or in this case, constrains. I want to know, which cpu, when i stick it on the motherboard and turn on the pc, will be the fastest. If only that cpu has access to those "special rails" more power to it.
11-03-2004, 10:38 AM
WesM63

Quote:

Originally Posted by Ailleur

This whole cheating thing is ridiculous, it calculates Pi, patched or non patched.
Take full advantage of every one of the cpu's advantage.
I guess if a 64bit version of superpi was released it would be invalid? (If we assume superpi would gain from a 64bit port).
If amd doesnt use sse3, or intel doesnt use 64bit (well, you know) its not the users problem. The user should try to get the most it can out of his hardware.

they're will be a 64bit superpi, its not a cheat, it is where the world is going. Even using 64bit XP with 32bit Superpi gives you a 1sec gain over 2000. (both fresh installs tested by me)
11-03-2004, 11:17 AM
Scyphe

Quote:

Originally Posted by Ailleur

Im not interested in knowing which cpu is fastest under certain circumstances, or in this case, constrains. I want to know, which cpu, when i stick it on the motherboard and turn on the pc, will be the fastest. If only that cpu has access to those "special rails" more power to it.

Well, that's ok. But the comparisons in this thread is about circumstances and common platforms to compete on, otherwise it'll be very difficult to get straight comparisons. The aim here is to show us what a P4 @ 6+Ghz can do with the same program and settings that other benches have been made with. If someone suddenly use an optimized version of a program, the comparison will be skewed and useless to others using a non-optimized version.

Of course, if everyone used only the patched version, it wouldn't be a problem, but then comparisons between CPU's without those extra instructions would be skewed (northwood's, amd's). And it's possible to do 2 benches as someone posted earlier.
11-03-2004, 11:21 AM
Think

Where is this patched version. I would like to try a comparitive view as well with my P42.8C coupled with the DFI Lanparty 875bPro verB1.
11-03-2004, 11:31 AM
dpa

Here is the patch:

http://www.home.no/dpa/Ftp/Bilder/pipatch.zip
11-03-2004, 11:57 AM
death metal

Congratulations fugger and chilly1.....really awesome breakthrough in the P4 arena....woot!!!!
11-03-2004, 02:27 PM
cpulloverclock

http://www012.upp.so-net.ne.jp/ohij2/p4.htm

M/B: ASUS P5P800 (48MG) BIOS: 1002
CPU: Pentium4 560 3.6GHz SL7J9(D0)
CPU Cooling: クリPro3+ARC Bed Rev3.3+LN2
Memory: TwinMos PC2700 BH-6 256MB
VGA: ミレ PCI
HDD: IBM HDS722540VLAT20&740DX 20G
Power Supply: Zippy-460w(3.3V&5V改)
Vcore: 1.68V(実測)
Vddr: 3.62V (実測) Vio:3.68V
OS: Windows Server 2003
Memory Timing: CL2.5-3-3-8-8-E-A-A(DDR266 3:2)

http://www012.upp.so-net.ne.jp/ohij2/348.gif
11-03-2004, 02:28 PM
Bennah

Whos that by?
11-03-2004, 02:28 PM
cpulloverclock

Quote:

Originally Posted by Bennah

Whos that by?

Hideo
11-03-2004, 02:50 PM
P4rthen0n

Japs are getting closer :p

It's out of this world, they ALWAYS get the oc WR nomather the kind of cpu they're dealing with :)
11-03-2004, 06:09 PM
holst1202

6..It's a really great work. :toast:

And I think maybe you can try to disable the
"CPC Addr/control" in asvanced chipset features.
It will have benefits when you runs in 5:4 OR 3:2.
11-03-2004, 06:15 PM
corvus_corax

awesome this LGA.....6.3 Ghz with 1.616V wow
11-03-2004, 06:29 PM
computerpro3

and I was just getting hyped up about my impending fx-55/vapochill LS rig... :rolleyes:

Awesome Job Fugger!!! :toast:
11-03-2004, 07:09 PM
Kameleon

On the patch front, there are Linux versions of superpi that will calculate 1M in 10 seconds, should they be allowed too? If the patch allows the Prescotts to do things that they can do in normal programs then it's debateable whether they're valid scores, if it's a superpi-only optimisation then it's very dodgy to me.

On the clock front, that's a damn nice job FUGGER :D
11-03-2004, 09:42 PM
Hilppa

Quote:

Originally Posted by FUGGER

Memesama's 23 seconds I am guessing is a patched, I have 2K to load and see if it helps.

It can't be. I got 22 seconds with an identical setup at 5,3Ghz (patched).
24s still at 5,5Ghz for me though, need to work on it.

Show 100 post(s) from this thread on one page