Asus P5W way faster than Bad Axe

At least for me it has been.

I keep saying it and I can't figure out why but the Asus P5W DH is WAY WAY
faster than the Bad Axe 304. Unless I have the Bad Axe setup wrong, with
exact same hardware and settings. The Asus gives the Bad Axe a whipping
in Super Pi 32M. However looking around the forum, my Bad Axe 32m times
are in line with other members. While on the other hand, my Asus times are
similar to Kyosen-san's Asus results at similar clocks.

Both at 3.5ghz with 1.4v and exact same ram and ram timings at 2.2v.
Nearly 1 min 45s difference at 3.5ghz.

Can't quite figure it out.... :eh:

Asus:
http://www.jpop.com/gocchin/asus32m.png

Bad Axe:
http://www.jpop.com/gocchin/badaxe32m.png

What's the strap on that BadAxe run, gocchin? 1333fsb or 1066fsb?

I thought I noticed a small difference myself, but now Im gonna have to do a side by side comparison.....That is a HUGE difference....I didnt even think to check and see if it was this dramatic....Ill post back later and see if my results are similar. But a quick question....what's your strap on the BadAxe? That can make a world of a difference...check some of the Xtreme bandwith threads to read about the quirks....Not that you wouldnt think to check this, im just trying to help any way i can

Hmmmm I had it forced at 1333fsb... I better go read. :)

... rather interesting
same is refered here by some of our forum members
but as far as I remember noone has done an exact side by side comparison

and ..hmmm this is huge , isn`t it

will you try a 1066 fsb run ?
think it will be much closer
though I think the Asus will still be faster

Quote:

---

Originally Posted by mine

... rather interesting
same is refered here by some of our forum members
but as far as I remember noone has done an exact side by side comparison

..hmmm this is quite huge , isn`t it

---

Though its only a glimpse right now of one benchmark, the difference is huge relatively speaking.

Quote:

---

Originally Posted by gocchin

Hmmmm I had it forced at 1333fsb... I better go read. :)

---

That's exacty why :fact: It seems that forcing 1333fsb strap on BadAxe does "something" to this board that allows for supposedly HUGE memory OCs. Now, there are two options:

1. either 1333fsb strap looses some MCH settings to the point that memory performance is just crap but allows for some unreal frequecies of up to DDR-1100 or DDR2-1200 8-3-3-3.0 or 8-2-2-3.0 or
2. none of the monitoring software can detect the real speed and/or latencies the RAM is actually running at

Your finding seems to prove that at forced 1333fsb, BadAxe seems to have some horrible memory performace. Thanks for posting that :)

lol maybe that extra 4 Mhz on Asus might made the difference :P j/k, but possible

also is it possible for this kyeosen to run other test like 3d mark, pc mark ... and such??

Quote:

---

Originally Posted by theteamaqua

lol maybe that extra 4 Mhz on Asus might made the difference :P j/k, but possible

also is it possible for this kyeosen to run other test like 3d mark, pc mark ... and such??

---

Those are my results above I was just mentioning that Kyosen had similar results. I already posted my prelimnary 3D results in the 3D section :).

Quote:

---

Originally Posted by bachus_anonym

That's exacty why :fact: It seems that forcing 1333fsb strap on BadAxe does "something" to this board that allows for supposedly HUGE memory OCs. Now, there are either two options:

1. either 1333fsb strap looses some MCH settings to the point that memory performance is just crap but allows for some unreal frequecies of up to DDR-1100 or DDR2-1200 8-3-3-3.0 or 8-2-2-3.0 or
2. none of the monitoring software can detect the real speed and/or latencies the RAM is actually running at

Your finding seems to prove that at forced 1333fsb, BadAxe seems to have some horrible memory performace. Thanks for posting that :)

---

Anytime bachus... we're all here to learn :). Well I'm going to run 32m again this time back at 1067 and see what the time is. :)

Quote:

---

Originally Posted by bachus_anonym

That's exacty why :fact: It seems that forcing 1333fsb strap on BadAxe does "something" to this board that allows for supposedly HUGE memory OCs. Now, there are either two options:

1. either 1333fsb strap looses some MCH settings to the point that memory performance is just crap but allows for some unreal frequecies of up to DDR-1100 or DDR2-1200 8-3-3-3.0 or 8-2-2-3.0 or
2. none of the monitoring software can detect the real speed or latencies the RAM is actually running at

Your finding seems to prove that at forced 1333fsb, BadAxe seems to have some horrible memory performace. Thanks for posting that :)

---

This certainly does seem to be the missing variable. Though I think we do need to try a side by side comparison of some of these boards (I have the BadAxe and the Asus, but I wouldnt mind putting some numbers together) Anyways, I await to hear if Gocchin can pull the BadAxe numbers closer or if his Asus is smokin it ;)

gocchin, try badaxe at 1067fsb and 350fsb 4:5 ;)

this was done at 2.2v vdimm 4-4-3-15 with 2x256MB Centon PC4300 Advance LOL with 1067fsb strap.. not sure how much a penalty 2x256mb would be heh

http://fileshosts.com/intel/Intel975..._single_tn.png

for conroe what's the diff between 2x256mb vs 2x512mb vs 2x1GB for 32m ?

also timings for tRAS differ it seems for your runs ?

I'm running 1066 with the same settings now and after a few loops its already faster... give me 10 more minutes to post a result. :)

Interesting... with everything the same except using 1066 I was UNABLE to complete a 32M run. I had to go back in the bios and raise VMCH a notch to run it again.

while waiting .... what a nice girl you are goccin and such a good oc er :-)

:toast:

Quote:

---

Originally Posted by mine

while waiting .... what a nice girl you are goccin and such a good oc er :-)

:toast:

---

lol if gocchin is a girl really...im in love with her! :banana:

i||uSi0n^

Quote:

---

Originally Posted by mine

while waiting .... what a nice girl you are goccin and such a good oc er :-)

:toast:

---

For all what I have seen so far it depends of the memory, options and software.

Really for you Asus is better in any way.

I wonder if she is a girl.:rolleyes:

hee... I was first :banana: :banana: :banana:

So gocchin, by ur experience, is the Asus better than the Badaxe?

A girl that likes OCing... i want one! j/k :p:

Quote:

---

Originally Posted by Metroid

For all what I have seen so far it depends of the memory, options and software.

Really for you Asus is better in any way.

---

Well every factor is the same in the above screen shots except for the mobos.

Anyway back to the testing.

1067 is indeed way faster than 1333 bootstrap but still lagging behind
the asus by quite a bit:

http://www.jpop.com/gocchin/badaxe1067.png

maybe should check memset?
afraid some mem timing different:confused:

i have been looking at both these Motherboards, to be honest i think i have made my mind up on the Asus.

As for the chip i am looking at a 6700, and hopefully the G Skill 6400 2gb set.

together with the X1900 in crossfire...i cant wait!!

now if andré with the P5 de luxe and the 965 chipset could chime in

we would have a nice comparison 975 vs. 965 @ 32 M

and we could see whats up with the vastly improved new intel 965 memory controler ............

gocchin is your asus p5w dh volt modded ?

Thats bad, hope intel can fix this 1333 strap error with a Bios update. Good to know anyway that Bad Axe doesnt have terrible performance overall, thanx for the research guys!

OT: Gocchin in that avatar is some kind of Japanese pop star, my guess is that Gocchin@xtremesystems is one of her biggest fans :)

Quote:

---

Originally Posted by mine

while waiting .... what a nice girl you are goccin and such a good oc er :-)

:toast:

---

This is for you guys in Europe: http://www.jpop.com/overclock

Look, listen and learn.:D

Quote:

---

Originally Posted by mine

while waiting .... what a nice girl you are goccin and such a good oc er :-)

:toast:

---

lol :rofl: , you are so wrong.

Quote:

---

I had to go back in the bios and raise VMCH a notch to run it again.

---

what voltage are you using for mch ? :)

Did you run some other benchmarks?

Quote:

---

Originally Posted by dogsx2

This is for you guys in Europe: http://www.jpop.com/overclock

Look, listen and learn.:D

---

:slobber: :slobber: :slobber: :D :toast:

Quote:

---

Originally Posted by mine

now if andré with the P5 de luxe and the 965 chipset could chime in

we would have a nice comparison 975 vs. 965 @ 32 M

and we could see whats up with the vastly improved new intel 965 memory controler ............

---

I am sure P965 chip memory control is not better than 975x.

Quote:

---

Originally Posted by Dynasty

lol :rofl: , you are so wrong.

Quote:

---

while waiting .... what a nice girl you are goccin and such a good oc er :-)

:toast:

---

:)

---

Dynasty :-)

You didn `t see the joke in it ...... my young friend


:banana: :banana: :banana: :banana:


now look at hicookie what a nice girl and a nice oc er... :-)



@ andre

finally the word I waited for .. and in complete coincidence
with what I thought as well ...

thanks champ

whats bootstrap

Guys

remember PAT...all that was is a way of forcing an overclock on the chipset...IE running the chipset at a different strap than the fsb range usually will allow.

So, running 1066 strap will drop the max fsb the CPU will reach but internal latency on the chipset will be a LOT LOWER...If you remember the IC7 days we all used to run 667 or 533 strap at this boosted memory performance over the 800 strap but you needed good ram and motherboard tweaks to get the fsb high.

Now badaxe has 800,1066 and 1333 strap, to get the high fsb's you need the 1333 but you losen the internal latency and so lose the memory performance. The ONLY way to claw some back is 1:1 and ultra tight timings at high FSB's...Now the alternative is run the 1066 or the 800 if you have a good board. i quoted in another thread the DFI was faster as it was stuck on the 1066 strap but FSB is limited to 370 ish tops.

So what does this mean, it means you are looking for a CPU with a high multi, IE 13 or 14 and run 1066 strap to gain all this lovely low chipset latency and gain a ton of memory speed. Im sure some one will sort out a way to get the MCH working at over 370 and we probably need to pump more fsb voltage also but for now run a high multi and 1066strap to keep the speed up with conroe.

This is yet another reason to "dump 'n forget" current BadAxe... From what people were showing so far, ASUS P5W DH is already hitting up to 440FSB (no vmods) and 1066fsb bootstrap, plus kyosen hit 500FSB (with $hit load of mods). The sooner we forget about freakish BadAxe, the better :D

I still think and are with what macci and hipro5 suggested in other threads in XB - RAM is not running at what CPU-Z reports in all those cases we've seen so far :) PAT, no PAT, I don't buy it :p: Slap me If I'm proved wrong, though...

Tony, any info if the R600 chipset based boards will have this "problem"? I think a lot of users (like me) will have E6400-E6600 CPUs and if the board tops out at 375 our OC will be relatively low. And as we are seeing, the 1333 strap is showing lack of performance, so i think that i will have to spend more money on the CPU...

Quote:

---

Originally Posted by bachus_anonym

This is yet another reason to "dump 'n forget" current BadAxe... From what people were showing so far, ASUS P5W DH is already hitting up to 440FSB (no vmods) and 1066fsb bootstrap, plus kyosen hit 500FSB (with $hit load of mods). The sooner we forget about freakish BadAxe, the better :D

I still think and are with what macci and hipro5 suggested in other threads in XB - RAM is not running at what CPU-Z reports in all those cases we've seen so far :) PAT, no PAT, I don't buy it :p: Slap me If I'm proved wrong, though...

---

Internal chipset latency is key here, use Everest etc to measure and you will see. PAT was just a BS way of saying "hey overclock this and its faster"..LOL!! do the same on badaxe and it will be faster...trust me.

I have a feeling CPU-Z may be wrong, 4:5 may actually be 1:1 someone needs to scope it and see.

Onto the RD600, I have the reference here, its an awesome board and does have a way of setting the 1333 strap, at the moment they are still tuning the board for high fsb overclocking so its unfair to comment. I can tell you its not like the Intel...IE you need 1333 strap to get the high fsb's but this does not affect the memory controller...so once they have it tuned up we will see some awesome results from it, I will say its already VERY good!!!

Its already hitting 450fsb so don't worry about it being limited to 370 ish as that was a 1066 strap limitation

Totally independant memory overclocking in 1fsb incrememnts to ddr1200, set the cpu to 349fsb and the ram to 436fsb if you like, yup its that flexable LOL

1T to ddr800ish, Low latency ram settings in bios and some nice new ram tweaks will have you begging to get one...I know I want a final board as my workstation ;)

so some 3dmark and see if theres such a difference there 2

Tony...would this be akin to the MAL and RP settings from A64? The effects are VERY similar...

@ Tony - thank you once again for your very clear technical explanation :toast:

When can we expect RD600 hit retail?

Sounds like a killer board

Quote:

---

Originally Posted by Tony

(...)Totally independant memory overclocking in 1fsb incrememnts to ddr1200, set the cpu to 349fsb and the ram to 436fsb if you like, yup its that flexable LOL(...)

---

I'm in :D

ok now i get it..

is there no way to get above 37x tony?.. im stuck on 379fsb.

Quote:

---

Originally Posted by bachus_anonym

I'm in :D

---

I have been cleared to post up screenies etc, im just waiting on a better CPU as the A stepping I have is a little slow etc, i have a friend with a B5 so I may use his for a day or so and show you some cool pics etc ;)

Pedro, i try and explain it like i see it, no point in the bull 5h1te you sometimes see as this confuses people...thank you for posting thanks though ;)

Intel chipsets have always needed overclocking to be fast, once you do this you will find they are as good as it gets but you may have to hardware mod to get the best from them. The new Nvidia and ATI chipsets do 1T though, this will kill off the need for Intel to have the 975 as we have seen from the news.

Quote:

---

Originally Posted by Waus-mod

ok now i get it..

is there no way to get above 37x tony?.. im stuck on 379fsb.

---

Yes, there is... At this moment, get ASUS P5W DH and you'll get 400+ @ 1066fsb bootstrap :fact: :D Or wait for RD600, from what Tony says...

Quote:

---

Originally Posted by Waus-mod

ok now i get it..

is there no way to get above 37x tony?.. im stuck on 379fsb.

---

MCH voltage, VFSB and a lot of cooling. Asus may have found a tweak that forces the MCH to a lower strap and sets the CPU to 1333 to allow the higher fsb's maybe..hard to say as i don't have the board to test yet.

Tony thanks for ur explanation :)

I hope the board will hit the retail market as soon as possible, with ur explanaiton you have made me to want one so badly:p:

Quote:

---

Originally Posted by Tony

I have been cleared to post up screenies etc, im just waiting on a better CPU as the A stepping I have is a little slow etc, i have a friend with a B5 so I may use his for a day or so and show you some cool pics etc :)

---

Lame excuse :lol: On with screenies, Tony... maybe in another thread so we don't hijack this one, though.

Quote:

---

Originally Posted by bachus_anonym

Lame excuse :lol: On with screenies, Tony... maybe in another thread so we don't hijack this one, though.

---

Sounds good to me, lets get back to PAT on 965 LOL as thats all it is ;)

Tony is right:

400fsb 3-3-3-8 1067strap = memtest error
445fsb 3-3-3-8 1333strap = fine

3.6ghz 1067 (1:1) 4-4-3-8 gave me a better 32m than 3.616ghz (4:5) 3-3-3-8

At least my board/cpu can do 3.6ghz dual prime (1067 strap)

hey Asus P5W DH's 400 fsb is good enough for me. E6600 : 9 x 400Mhz = 3.6Ghz. Although i might get RD600 or nforce 590 later for better performance on dual gfx

Quote:

---

Originally Posted by Tony

Guys

remember PAT...all that was is a way of forcing an overclock on the chipset...IE running the chipset at a different strap than the fsb range usually will allow.

So, running 1066 strap will drop the max fsb the CPU will reach but internal latency on the chipset will be a LOT LOWER...If you remember the IC7 days we all used to run 667 or 533 strap at this boosted memory performance over the 800 strap but you needed good ram and motherboard tweaks to get the fsb high.

Now badaxe has 800,1066 and 1333 strap, to get the high fsb's you need the 1333 but you losen the internal latency and so lose the memory performance. The ONLY way to claw some back is 1:1 and ultra tight timings at high FSB's...Now the alternative is run the 1066 or the 800 if you have a good board. i quoted in another thread the DFI was faster as it was stuck on the 1066 strap but FSB is limited to 370 ish tops.

So what does this mean, it means you are looking for a CPU with a high multi, IE 13 or 14 and run 1066 strap to gain all this lovely low chipset latency and gain a ton of memory speed. Im sure some one will sort out a way to get the MCH working at over 370 and we probably need to pump more fsb voltage also but for now run a high multi and 1066strap to keep the speed up with conroe.

---

thanks for the explanation

Hi gocchin and all,

bachus_anonym> kyosen hit 500FSB (with $hit load of mods).
--
No, no...at least, yet...upto 475 as shown in another thread, so far:)

Anyway, P5W-DH(Performance Mode=Turbo, HyperPath3=Disabled) is
slightly faster than D975XBX(1067 strap), at least 1:1 mode.

8M result on D975XBX:
http://222.151.149.56/c-board/file/M...G-1067-533.png

8M results on P5W-DH:
http://222.151.149.56/c-board/file/H..._boot-350T.png

interesting does this translate into 3dmarks as well. care to compare ?

is hyperpath3 = HT ?

okay found it

Quote:

---

We should also note that the P5WDG2-WS' BIOS incorporates an Asus proprietary feature dubbed HyperPath 3 that's designed to further reduce memory access latency, over and above what's typically offered with the 975X Express chipset. Here's what Asus has to saw about HyperPath 3:

Hyper Path 3
Data transfers within the system are made possible by various mechanisms that synchronize the sending and receiving of commends among components. Unfortunately, these mechanisms cause data transfer delays known as latency time. This unique feature significantly shortens latency time during data transfers, allowing users to experience the real power of their computers.

---

Quote:

---

Originally Posted by kyosen

Hi gocchin and all,

bachus_anonym> kyosen hit 500FSB (with $hit load of mods).
--
No, no...at least, yet...upto 475 as shown in another thread, so far:)

Anyway, P5W-DH(Performance Mode=Turbo, HyperPath3=Disabled) is
slightly faster than D975XBX(1067 strap), at least 1:1 mode.

---

Hi Kyosen-san,

When playing with the HyperPath3 did disabling it allow you to clock the fsb higher? :)

ok. i have to ask.

what's this "strap" you guys are talking about?

is this an es sample of the Chipset that has unlocked features?

or is this just inherent in the conroe design?

wow you guys have never played with hyperpath ?

basically what it does is adjust your memory timings on the fly.

turn it on and go into windows,, make sure your ram is tighter than the SPD ratings. or atleast equal to. (hhmm come to think of it,, i dont think that matters)

open cpu-z. go to the memory timings tab run Pi... or Prime.. watch the memory tab in cpu-z it will automatically loosen your ram timings.

sucks for bandwidth. but allows the ram to relax a little.
when the test is done it will return back to normal. (it did for me atleast)

did anyone ever figure out what the performance mode is ? Auto/Standard/Turbo i know that setting has been around for a number of years but what does it change ?

Ｈｉ gocchin,

> When playing with the HyperPath3 did disabling
> it allow you to clock the fsb higher?
--
HyperPath3=Enabled
-> can't boot at higher (mere ~380) FSB

HyperPath3=Auto/Disabled
-> can boot at higher FSB, but no difference for FSB limit, in my case.

so I always simply disabled HyperPath3...

Quote:

---

Originally Posted by Supertim0r

Tony is right:

400fsb 3-3-3-8 1067strap = memtest error
445fsb 3-3-3-8 1333strap = fine

3.6ghz 1067 (1:1) 4-4-3-8 gave me a better 32m than 3.616ghz (4:5) 3-3-3-8

At least my board/cpu can do 3.6ghz dual prime (1067 strap)

---

I say always that to the people 1/1 is the better way to get performance.

The ram will flow better than 4/5, but it demands less FSB.

I'm back on 1067fsb strap and with E6600 only having 9x multiplier had to do 9x389fsb 1:1 with ram at 389mhz 4-4-4-5 at 2.2v

seems with my 2x512MB PC5400ULv1.5 ram i can do 389fsb easy with 1067fsb strap straight boot from bios - stock unmodded badaxe rev304

http://fileshosts.com/intel/Intel975..._15m06s_tn.png

I loose 438-389 = 49mhz off memory clocks compared to your 10x350fsb 4:5 438mhz 4-4-4-5 and loose +20s off 32M with 2x512MB vs 2x1GB heh

Getting the hang of this board and cpu now, 1067fsb strap with 9x400fsb with 2x512MB PC5400ULv1.5 @400mhz 4-4-4-5 at 2.2v straight boot from bios with +50% burnin.

http://fileshosts.com/intel/Intel975...4min42s_tn.png

This is what it takes on Badaxe with 1304 bios for 2x 512MB modules on 1067fsb strap 1:1 cpu:mem to break 14min 28s single 32M. Booted straight from bios at 9x400fsb and used clockgen to raise fsb to 404.7fsb for 3646Mhz with ram @404mhz 4-4-3-5 at 2.2v

http://fileshosts.com/intel/Intel975...4min27s_tn.png


Crazy 1067fsb strap with 9x411fsb = 3703mhz with 2x512MB PC5400ULv1.5 ram @411mhz 4-4-3-5 at 2.2v totally unmodded badaxe rev304.

http://fileshosts.com/intel/Intel975...4min12s_tn.png

@9x415fsb 4-4-3-5 at 2.2v!

http://fileshosts.com/intel/Intel975...4min05s_tn.png

@9x418fsb 4-4-3-5 at 2.2v breaks 14min for 1st time :D

http://fileshosts.com/intel/Intel975...3min57s_tn.png

I wonder if how high fsb you can clock on 1067fsb is memory dependent ? i.e.

on 1067fsb with 2.2v vdimm 1304 bios

2x256MB Centon PC4300 Advance Micron D9 fatties = max 382fsb

2x512MB OCZ PC5400 Platinum XTC = max 378fsb

2x512MB Corsair PC5400ULv1.5 Micron D9 fatties = so far completed 32M upto 418fsb and doing 32M run @421fsb!

Interesting results, eva2000! It looks like it takes the right guy and the right mems to get this BadAxe working above 400Mhz and 1066fsb strap :up:

BTW, are you still having doubts that 1333fsb bootstrap criples memory performance in one way or another - more likely down-shifting ratios. Looks to me like over 1m difference between the both - 1066fsb vs 1333fsb :fact: You have TRP=3 at 1066fsb run but that does not make a lot of diffference, few secs at the most...

i never said 1333fsb gave better performance than 1067fsb strap just that it allows higher fsb max heh

I've hit the limit of 1067fsb strap for 2x512MB PC5400ULv1.5 at 4-4-3-5 at 2.2v. The max is 421fsb. I guess you can subtract 20+ seconds if i had 2x1GB modules hehe.

http://fileshosts.com/intel/Intel975...3min53s_tn.png

2x1GB will not give you 20s premium in SPi32M over 2x512MB. You have 2x1GB modules, don't you? Check it out and you'll see for yourself ;)
But your 421MHz is rockin', man! :toast:

for conroe, gocchin, opb and victor all say 2x1GB vs 2x512MB there's >20 sec diff... bachus you tried conroe or you only going by s939, am2 experience ?

no 2x1GB modules here right now :(

Yeah, I read what they said... I don't have BadAxe so I can't say anything about 2x1GB advantage over 2x512MB on this board, but I somehow doubt it it's that big :p: It's just me, I like to verify things myself :lol: Especially, after everyone is believing in their huge memory OCs on 1333fsb and BX I just got a bit sceptical about all this...

EDIT: When you compare your 15:06 with 2x512MB vs gocchin's 14:57 with 2x1GB. He has 50MHz advantage on RAM and some disadvantage on FSB but once you get to the same clocks, you might even better his score with your 2x512MB. We need to consider differences with OS config and the way SPi was run, but it's a good start...

Anyway, nope. I'll be joining you guys soon on that new platform :)

Wonder if the 7950 GX2 will work in this board.

Quote:

---

Originally Posted by malficar

Wonder if the 7950 GX2 will work in this board.

---

It works .

http://soj.mesdiscussions.net/forum2...0&new=0&nojs=0

http://soj.mesdiscussions.net/forum2...&nojs=0#t35278

Quote:

---

Originally Posted by Stelios

It works .

http://soj.mesdiscussions.net/forum2...0&new=0&nojs=0

http://soj.mesdiscussions.net/forum2...&nojs=0#t35278

---

Thanks, just ordered it from the EGG:

MB ASUS P5W DH DELUXE/WIFI-AP GREEN
CPU INTEL|PD 930 3.0G 775 2x2M R
CASE ANTEC|P180B BLACK RT
MEM 1Gx2|MUSHKIN DII800 996523

That should do for now. I forgot to ask, is there a specific BIOS that is best?

malficar you may have gotten the last one of this shipment, its out of stock again

Quote:

---

Originally Posted by Tony

Totally independant memory overclocking in 1fsb incrememnts to ddr1200, set the cpu to 349fsb and the ram to 436fsb if you like, yup its that flexable LOL

---

Are you serious? That would be one sweet feature if it works.

OK, stupid question - what the hell is the bootstrap? I assume on the conroe the default is 1067 right? So if I use the Bad Axe for a 30% OC, the FSB would be 1387, but the strap would still be 1067 right?

I guess I'm a bit confused on what setting is changing here...

Thanks for helping a noobie.

the D975XBX lets your change this default bootstrap. So you can run your processor at 1333 FSB and from there you can again have +/-50 % change.

The strap, also called the northbridge strap, determines the base FSB frequency of the CPU when you boot up. Let's say you have Conroe which has default FSB frequency of 266.66MHz, or 1067MHz quad-pumped. If you change the strap to 800MHz in the BIOS, the base FSB frequency will become 200MHz. With the 1333MHz strap, it's 333MHz, etc.

Straps are sometimes useful for overclockers because you can start at a higher base frequency and work your way up from there. But the strap you choose can affect things like latency, etc. So a higher strap is not always better.

Why would 2x1gb give better performance than 2x512?

Quote:

---

Originally Posted by ZX7891

Why would 2x1gb give better performance than 2x512?

---

Bank interleaving. ;)

Quote:

---

Originally Posted by sierra_bound

Bank interleaving. ;)

---

you can do bank interleave with 2x512

Quote:

---

Originally Posted by ZX7891

you can do bank interleave with 2x512

---

Yes, I know. You can do more with 2GB. ;)

I'm just going by personal experience with DDR2. There are some 2X1GB kits that are as fast, if not faster than 2X512MB. Granted, some of that has to do with the IC's used. But at least in the DDR2 world, there's no real advantage to having 2X512 over 2X1024.

Hey all so whats the result in " Asus P5W way faster than Bad Axe" i hope we could see more benchies in 3rd and real application.. =) ..just my one peso.. =)

Thanks for sharing.. =) ..mm maybe i'll wait for ati chipset and nvidia to come into play.. :))

Quote:

---

Originally Posted by bachus_anonym

Yeah, I read what they said... I don't have BadAxe so I can't say anything about 2x1GB advantage over 2x512MB on this board, but I somehow doubt it it's that big :p: It's just me, I like to verify things myself :lol: Especially, after everyone is believing in their huge memory OCs on 1333fsb and BX I just got a bit sceptical about all this...

EDIT: When you compare your 15:06 with 2x512MB vs gocchin's 14:57 with 2x1GB. He has 50MHz advantage on RAM and some disadvantage on FSB but once you get to the same clocks, you might even better his score with your 2x512MB. We need to consider differences with OS config and the way SPi was run, but it's a good start...

Anyway, nope. I'll be joining you guys soon on that new platform :)

---

managed to push my E6600 + Badaxe rev304 a bit harder but still 2x 512MB PC5400ULv1.5 at 2.2v heh

http://fileshosts.com/intel/Intel975...4min37s_tn.png

I've tested kentsfield on ASUS/Intel

300*11=3.3G
ram:fsb=3:5 , 500mhz 4-4-3-8


1. Intel:boot from 800fsb(in order to get 3:5 ram), 47% host(cannot boot higher), clkgen to 300.68fsb.
pi_32m=15min30sec
http://www.xtremesystems.org/forums/...1&d=1152025444


2. Asus: with hyperpath3=on
pi_32m=15min10sec
http://www.xtremesystems.org/forums/...1&d=1152025457


so....Intel, 20sec fall behind in pi_32m , 0.01sec slower in pi_1m.
I think maybe some unchangable ram_timings do have effect on that.
cpu's clk to clk is same due to almost the same pi_1m benching but huge difference in pi_32m , usually that's because of ram.

:p:

So let me get this straight. The 0502 BIOS will work with any Conroe, yes?

Quote:

---

Originally Posted by VictorWang

I've tested kentsfield on ASUS/Intel

300*11=3.3G
ram:fsb=3:5 , 500mhz 4-4-3-8


1. Intel:boot from 800fsb(in order to get 3:5 ram), 47% host(cannot boot higher), clkgen to 300.68fsb.
pi_32m=15min30sec
http://www.xtremesystems.org/forums/...1&d=1152025444


2. Asus: with hyperpath3=on
pi_32m=15min10sec
http://www.xtremesystems.org/forums/...1&d=1152025457


so....Intel, 20sec fall behind in pi_32m , 0.01sec slower in pi_1m.
I think maybe some unchangable ram_timings do have effect on that.
cpu's clk to clk is same due to almost the same pi_1m benching but huge difference in pi_32m , usually that's because of ram.

:p:

---

Good comparison Victor. I also don't notice much difference in 1M, but in 32M the difference is quite large... If you're going for 32M records the bad axe is NOT the board to use. :D

Quote:

---

Originally Posted by gocchin

Good comparison Victor. I also don't notice much difference in 1M, but in 32M the difference is quite large... If you're going for 32M records the bad axe is NOT the board to use. :D

---

really huh

Went for my new personal best 32M time. I think this is one of the fastest times under 3.9ghz and 4.0ghz at 1:1 ??

http://fileshosts.com/intel/Intel975...9s609ms_tn.png

hwbot HOF http://www.hwbot.org//hallOfFame.do?...pplicationId=7 :D

Very nice my friend :).... I would try 1:1 at that speed on my asus but it died due to conedensation. :(

oh dam... insulating these intel boards seems to be hard... but IIRC you didn't use any insulation for your DI runs ?

Quote:

---

Originally Posted by gocchin

Very nice my friend :).... I would try 1:1 at that speed on my asus but it died due to conedensation. :(

---

want the MSI 975X ? :D

Quote:

---

Originally Posted by gocchin

Very nice my friend :).... I would try 1:1 at that speed on my asus but it died due to conedensation. :(

---

Sorry to hear that gocchin.

Quote:

---

Originally Posted by eva2000

managed to push my E6600 + Badaxe rev304 a bit harder but still 2x 512MB PC5400ULv1.5 at 2.2v heh

http://fileshosts.com/intel/Intel975...e_14min37s.png

---

Told ya :p: :up: Congrats on your 3.81GHz time :) Nice onem but you definetely need more volts ;)

Quote:

---

Originally Posted by gocchin

Very nice my friend :).... I would try 1:1 at that speed on my asus but it died due to conedensation. :(

---

There goes the fun :( Sorry, bud... Those caps are crazy close to the socket on P5W DH... Where did the drops go? Got a backup one? :)

The Bad Axe has horrible placement for the PCI-E and PCI slots. I'm running X1900 in Crossfire and I had to take off one of the Zalman fans and put the stock cooler back on for them both to fit. I'm looking at the Asus now since it appears to have better placement with the PCI-E slots. Also, really only 1 PCI slot is available in the Bad Axe after both x1900 are put in.