32m Low Clock Challenge

Printable View

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

03-06-2007, 07:12 PM
dinos22

B1 vs B2

~20 seconds

B2 E6600
32M SuperPI 3.6GHz CPU + 550MHz RAM 4-4-4-8 timings G.Skill HZ RAM
http://img248.imageshack.us/img248/3...4m04656wd9.jpg
[/B]

B1 E6600
3.6GHz 32M SuperPI 550MHz on RAM 4-4-4-8
http://img242.imageshack.us/img242/7...4m24875py1.jpg
03-06-2007, 07:16 PM
Gautam

^^That's more reasonable.
03-06-2007, 07:25 PM
s7e9h3n

There's a difference here though....This cpu that I'm running is one of the first ES x6800's to come off the lot. Although Dino's tests show his cpu's aren't that far apart, I have feeling that it's not the same case for the x6800's. If I recall correctly, Kingpin posted a 32m run with his B1 @ a super high clockspeed as well as memspeed only the time was amazingly ugly for those clocks. His run must have been AT LEAST 5% off. Even the jap Oc'r Omega admitted that a RETAIL B2 x6800 Conroe has a 1-2 HUNDRED Mhz advantage over an ES B1 x6800 @ 5ghz+.....so, for example, a 5ghz B2 retail XE= a 5.2ghz ES B1 XE...

And Dinos, although you're comparing b1 vs. b2, both your cpu's are ES's. Therefore the B2 isn't necessarily an indication of how the retail B2's are.....
03-06-2007, 07:42 PM
s7e9h3n

Quote:

Originally Posted by Gautam

Some differences that I notice between you and eva-

You're at CAS 4, he's running CAS 5. Your FSB is 450, his is 400. Now one would think that both of those would be working towards your advantage, but 680i is such an odd chipset, its a chance both are hurting. :S Perhaps the 3:4 divider in particular as an additional latency hit, or perhaps you're on a looser strap or something similar as you would be on i965 by running a higher FSB.

Yes, B1's are slower, but IIRC its more like 10-15 seconds or so and not 30. I only tried 32M once on a B1, and that was months ago. I don't recall the time being too far off the norm. But then again, perhaps the effect is more pronounced on 680i than on the Intel chipsets. :shrug:

OK, well then what about the benches I ran earlier using the SAME FSB as eva....only I was running 4-3-3-4 @ 600 vs. his 5-3-3-3 @ 600? His time was almost 40 seconds faster than mine. Lets, for a second, just simply ignore the fact that 4-3-3-4 is faster than 5-3-3-3 under ANY conditions and compare only the cpu's. Eva's is a retail B2 which - according to your theory - has a 15 second MAX advantage over the ES B1 I'm using. So doing a little math 40-15 = 25 seconds <--- that's how much faster Eva's run would be compared to mine IF it was a level playing field. Please, then, someone tell me the 25 SECONDS worth of tweaks for 32m that I'm either NOT doing or don't know about........:confused:
03-06-2007, 07:56 PM
Gautam

It is very strange, I won't deny it. :shrug: And I don't know anything about 680i at all or how to run it.

And my B1 ES is a 6700, and my 6800 a retail B2. So who knows, perhaps that could be skewing things.
03-06-2007, 08:00 PM
s7e9h3n

Quote:

Originally Posted by Gautam

It is very strange, I won't deny it. :shrug: And I don't know anything about 680i at all or how to run it.

And my B1 ES is a 6700, and my 6800 a retail B2. So who knows, perhaps that could be skewing things.

Tell ya what...Send me your b2 and I'll test it on the same rig....I guess I'll have to make this sacrifice for the good of the OCing community :D Bleh...Intel....:p:
03-06-2007, 09:37 PM
fhpchris

Quote:

Originally Posted by s7e9h3n

Tell ya what...Send me your b2 and I'll test it on the same rig....I guess I'll have to make this sacrifice for the good of the OCing community :D Bleh...Intel....:p:

You could try running something like 12x multi @ 300 FSB 1:2 unlinked :)

I might be able to pencil some voltage in and get 600mhz to pass 32m so we could compare to my quad...

Quote:

*Note: Hipro5 latencies are not applicable to 99.99% of the population. You have to either possess some sort of voodoo majic or have hardware manufactured @ Mt. Olympus to run such timings

Some of us just call them fatbodies :)
03-06-2007, 09:47 PM
s7e9h3n

Quote:

Originally Posted by fhpchris

You could try running something like 12x multi @ 300 FSB 1:2 unlinked :)

I might be able to pencil some voltage in and get 600mhz to pass 32m so we could compare to my quad...

Some of us just call them fatbodies :)

Why would I run with the 12x multi? To see if the FSB has an impact on the 680i boards?

Your quad is a b1 :p: ? IIRC, all of them are b3's....right?

Yes, I know....but sure as hell, my fatties can't even come close to what Hipro runs his at......:D
03-06-2007, 09:48 PM
fhpchris

Quote:

Originally Posted by s7e9h3n

Your quad is a b1 :p: ? IIRC, all of them are b3's....right?

Yes, I know....but sure as hell, my fatties can't even come close to what Hipro runs his at......:D

My Quad is a B3 ES :) Some were 1s IIRC..

This is From GHZ..
http://img.photobucket.com/albums/v31/canny/4ghz.jpg

What do you 1m @ that cpu speed? I know he is running 1T, but we have 170+ mhz ram on him with tighter timings :)
03-07-2007, 01:30 AM
k|ngp|n

Quote:

Originally Posted by s7e9h3n

There's a difference here though....This cpu that I'm running is one of the first ES x6800's to come off the lot. Although Dino's tests show his cpu's aren't that far apart, I have feeling that it's not the same case for the x6800's. If I recall correctly, Kingpin posted a 32m run with his B1 @ a super high clockspeed as well as memspeed only the time was amazingly ugly for those clocks. His run must have been AT LEAST 5% off. Even the jap Oc'r Omega admitted that a RETAIL B2 x6800 Conroe has a 1-2 HUNDRED Mhz advantage over an ES B1 x6800 @ 5ghz+.....so, for example, a 5ghz B2 retail XE= a 5.2ghz ES B1 XE...

And Dinos, although you're comparing b1 vs. b2, both your cpu's are ES's. Therefore the B2 isn't necessarily an indication of how the retail B2's are.....

B1 es is slow as hell in 32m, negligible difference in 1m. 27-28 sec difference in 32m betwen b2 es and b1 es. The b2 retail chip is just about identical to b2es and didnt really see any difference there at all in 32m. My b1 es was one of the very first as well.
So b1es<b2es=b2retail :D
03-07-2007, 01:42 AM
s7e9h3n

Quote:

Originally Posted by fhpchris

My Quad is a B3 ES :) Some were 1s IIRC..

This is From GHZ..
http://img.photobucket.com/albums/v31/canny/4ghz.jpg

What do you 1m @ that cpu speed? I know he is running 1T, but we have 170+ mhz ram on him with tighter timings :)

Hmmm...I've run it in the past, but not with this pair of mems...I'll try later. But FWIW, I can run a 1m @ 3600 at a bit under 14 sec. Don't remember exactly.....

Thnx Vince....but one thing that would make it easier on both me and anybody else reading this thread - when we talk about the differences clock for clock between different steppings, it would be nice if we could state advantage and say @ what speed. A 30 second difference on a sub 10-min 32m is a much larger percentage than a 30 second difference on a 20-min run :)
03-07-2007, 08:44 AM
k|ngp|n

Quote:

Originally Posted by s7e9h3n

Hmmm...I've run it in the past, but not with this pair of mems...I'll try later. But FWIW, I can run a 1m @ 3600 at a bit under 14 sec. Don't remember exactly.....

Thnx Vince....but one thing that would make it easier on both me and anybody else reading this thread - when we talk about the differences clock for clock between different steppings, it would be nice if we could state advantage and say @ what speed. A 30 second difference on a sub 10-min 32m is a much larger percentage than a 30 second difference on a 20-min run :)

Your right, didnt think of that. Well for me when I say 27-28 diff between the the two steppings, I am talking about aprox 3300-3400mhz cpu speed using 2:3 div config with 4-4-4-5.
03-07-2007, 03:55 PM
Johnny Bravo

So far so good....

http://www.itsjohnnybravo.com/Overcl...2M%2012-39.JPG

next stop is some software tweaks ;)
03-08-2007, 07:32 AM
before

hey Johnny; I'll reply your PM ASAP ;)

So let me share this little run... first try with one of those Wintec kits I've received this morning.

AMPX PC2-8000C5 @2.61V DMM with a fresh install of Win XP and only LSC=1 + real time priority as tweaks... nothing more! All the other stuff was running, explorer.exe included.

I must say Wintec roxxes!! I definitely love this brand! TBH I didn't really expect performance like that; bloody hell!... D9GKX I guess. :rolleyes:

http://img100.imageshack.us/img100/9...m09672sfx6.png

So; I'll try to improve it under 2k3 and few tips.
03-08-2007, 08:08 AM
alpha0ne

Thats worth looking at before :toast: :clap: :banana:
03-08-2007, 08:23 AM
Rol-Co

like i said in the other topic ,the wintec has gkx mem irev210 has taken of the heatspreaders of one set. :)

can you try cl5-5-5 ?
03-08-2007, 08:44 AM
Gautam

Quote:

Originally Posted by before

hey Johnny; I'll reply your PM ASAP ;)

So let me share this little run... first try with one of those Wintec kits I've received this morning.

AMPX PC2-8000C5 @2.61V DMM with a fresh install of Win XP and only LSC=1 + real time priority as tweaks... nothing more! All the other stuff was running, explorer.exe included.

I must say Wintec roxxes!! I definitely love this brand! TBH I didn't really expect performance like that; bloody hell!... D9GKX I guess. :rolleyes:

Looks great! I see things are gonna be heating up quick. :D
03-08-2007, 08:48 AM
before

@Rol-Co: not yet :) what about yours?

@Gautam: hehe :D I've just started the copy-wazza a minute ago...
03-08-2007, 08:52 AM
Rol-Co

cant test them , they are still not here :( a few day's

edit:
and daniel tested only at dual 32m 4-4-4-5-4-35-3-10-5-12 2.45v p5b deluxe.
03-08-2007, 09:27 AM
kemo

Quote:

Originally Posted by Rol-Co

like i said in the other topic ,the wintec has gkx mem irev210 has taken of the heatspreaders of one set. :)

can you try cl5-5-5 ?

I can also confirm that they are GKX but don't know anything about their quality and the PCB
03-08-2007, 10:02 AM
before

ahh too bad; I need to install 2k3 again... actual looks like corrupted. :(

I'll receive 25l of LN² tomorrow along with R1150 and a X6800 L624A... I need a perfect windows. :D
03-08-2007, 01:08 PM
s7e9h3n

Quote:

Originally Posted by fhpchris

What do you 1m @ that cpu speed? I know he is running 1T, but we have 170+ mhz ram on him with tighter timings :)

Here...ran last night:

http://img232.imageshack.us/img232/7713/10x4001mmg6.jpg
03-08-2007, 01:17 PM
Zeus

What's wrong with your setup man?
With memspeed that high and timings that tight you should have around 12.500 seconds or better.

I did better with my RAM at only 444MHz and 4-4-4-12 timings.

It's time for a better CPU man, that B1 seem to suck, even in 1M.
03-08-2007, 01:27 PM
s7e9h3n

Quote:

Originally Posted by Zeus

What's wrong with your setup man?
With memspeed that high and timings that tight you should have around 12.500 seconds or better.

I did better with my RAM at only 444MHz and 4-4-4-12 timings.

It's time for a better CPU man, that B1 seem to suck, even in 1M.

Yep....it's a shame...this very x6800 used to be considered the TOP x6800 in all the land....now it's just an over-the-hill, handicapped cpu with a case of mild retardation (In other words -> it's slow) :D
03-08-2007, 01:33 PM
kiwi

Quick test on 650i, p5n-e

http://img57.imageshack.us/img57/2604/52327522ok3.jpg

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