Printable View
One other thing I noticed about using the 1333fsb is when booting the mobo powers off then on by iteself. I didnt have this problem before I started using the 1333fsb option... This only happens on a fresh boot not a reboot.. Does anyone else have this prob?
32m 1333+ i dont remember the %
OC only from bios
http://simon.webideal.ca/Bench/32m%203.96.JPG
same thing happened here ....Quote:
Originally Posted by njkid32
Another example here - your CPU is 450Mhz faster than gocchin's, you've lower latencies on RAM and yet only barely catching up with his P5W DH result...Quote:
Originally Posted by Supertim0r
BTW, in case some of you guys missed an interesting post from macci, regarding BadAxe 1066fsb vs 1333fsb:Quote:
Originally Posted by macci
Its not a problem really - the board just needs to do that in order to switch to 1333 mode.Quote:
One other thing I noticed about using the 1333fsb is when booting the mobo powers off then on by iteself. I didnt have this problem before I started using the 1333fsb option... This only happens on a fresh boot not a reboot.. Does anyone else have this prob?
If there is any logic in the BadAxe BIOS it should give 1:1 when 1333FSB and 667RAM modes are selected (as they both mean 333MHz in reality). This is how it works with 1066 + 533. However if you select such a combination (1333+667) you'll see CPU-Z reading it as 4:5 RAM divider (FSB 333, RAM 417). Then again Everest doesnt show the clock info when 1333+667 is selected.
it could be possible that CPU-Z is reading the clock incorrectly w/ FSB1333 strap and 1333:667 is really 1:1 and not 4:5. This would mean that selecting 1333strap + 533RAM from the BIOS (which reads 1:1 in CPU-Z) would actually enable 5:4 RAM divider - this would put the clocks to 333FSB:266RAM
Of course someone needs to scope the clocks to verify this - perhaps hipro5 already did this?
Does the 800MHz mode ("2:3") work for you when you select 1333strap? The reason I'm asking is that I'm able to run the memory up to around 560-570 (CL4-4-4) using 1066strap and 2:3 divider (378FSB:567RAM) but it wont boot up at 1333strap:800MHz (333:500).
The performance w/ 1333strap is rather low (Everest mem bench):
400FSB "1:1" CL4-4-4 = READ:6737MB/s, COPY:5653MB/s, LATENCY:63.1ns
400FSB:500RAM "4:5" = READ:7151MB/s, COPY:6322MB/s, LATENCY:61.4ns
NB "X" @ 400 1:1 CL4 = READ:9242MB/s, COPY:6754MB/s, LATENCY:52.8ns
interesting macci, but it would not explain why my OCZ PC5400 Plat XTC which max at 423-426mhz wont boot at 1333fsb 4:5 at 346fsb, but my corsair pc5400ulv1.5 do boot and run because if it was 5:4 my OCZ ram would of worked fine ????
I have 3 bad axe. One of them won't run 400fsb 3-3-3-8 but my best run 445fsb 3-3-3-8 stable...weird
weird was this with same bios on all 3 badaxes ?
interesting findings so far, wonder if this does translate into 3dmarks as well. care to compare ?
:D time for me to post some results of BadAxe
my badaxe, cap/vdimm/vcore/tcore moded. bios=1304.
set 1067fsb, 50% host=400fsb w/o any problem and stable like a rock.
1067fsb, host=50%, ram:fsb=4:5 4-3-2-4, multi=11x, pi_32m=12min15sec
http://vic.expreview.com/attachment/1151504368.jpg
set 1333fsb, with a E6700, the mobo achieved 495fsb.
http://vic.expreview.com/attachment/...+ddr2_1237.jpg
then, some pi_1m results for you to compare:
1)1067fsb, host=50%, ram:fsb=1:1 3-2-2-4, multi=10x, pi_1m=12.781sec
http://www.xtremesystems.org/forums/...1&d=1151919492
2)compare to 1333fsb, host=30%(bios 1181) clkgen to 445fsb, ram:fsb=1:1 3-3-3-8, multi=9x, pi_1m=12.844sec
http://www.xtremesystems.org/forums/...1&d=1151920153
3)compare to ASUS P5WDH
364fsb, 11x, ram:fsb=1:1 3-3-3-8, pi_1m=12.828sec
http://www.xtremesystems.org/forums/...1&d=1151920597
Quote:
Originally Posted by eva2000
Intel wins in CPU score of 3Dmark06, so far I've tested:
Quad 2.66G @ 320_10, cpu scroe=4999 with badaxe
Quad 2.66G @ 310_11, cpu scroe=4991 with p5w dh
and... for overclk ability/stability.
Intel wins both, compare with asus/msi/dfi
1.Highest fsb which can boot directly from bios.
2.same HIGHEST benchable clk with ASUS.
3.no vdroop, need 0.05v less to run dual_sp2004 with same cpu_clk compare to ASUS.
but isn`t it so Victor that a
cap/vdimm/vcore/tcore modded "bad axe"
seems to be only quite as fast as a unmodded asus P5W DH ? ....... :-)
and hmmmm....am i right , that your bad axe has selected handsoldered alu_solid_caps that replace most of the caps on the mobo ?
Hi Victor. Here is an unmodded totally stock Asus P5W DH Deluxe.Quote:
Originally Posted by VictorWang
4250mhz 10x425 Ram running 1:1 4-4-4-5
Even with 150mhz less and looser timings it's only slightly slower.
http://www.jpop.com/gocchin/1221.png
I saw ur thread yesterday and the result is damn good.:clap:Quote:
Originally Posted by gocchin
don't know if it's the higher fsb and higher ram_clk helps.
conroe is not sensitive to ram_timings 'cos of it has tooooo large L2_Cache :D
victor and gocchin what's the gains in 32m from going 2x1GB vs 2x512MB DDR2 at same timings/speed for ram ???
Victor I'm sure you will be under 12 minutes very soon! :)
eva, I think Kyosen-san had posted some times using SuperPi 8M with 2x512 and 2x1gb and the difference in 8M was around 5 seconds faster for the 2x1gb
I've tested , 512m*2 vs 1G*2 at same settings pi_32mQuote:
Originally Posted by eva2000
1gb*2 leads 20sec way ahead :eek: :D
so that translates to >20s diff at 32M ??Quote:
Originally Posted by gocchin
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!
Quote:
Originally Posted by VictorWang
talking about conroe mobos
found an interesting comparison :
_________________
coolalers 32 M run some time ago on a MSI 975 platinum
hmmm...300 MHz less and look @ the time...not bad at all :
http://img192.imageshack.us/img192/1523/coolaler6uv.jpg
Quote:
Originally Posted by mine
That looks to be just as fast as the Asus if not faster... I think we are all coming to the same conclusion here... the bad axe is slow....
gocchin try a 9x421fsb 1:1 4-4-3-5 run on asus i want to compare with badaxe stock unmodded with my custom 1304 1067strap bios
http://fileshosts.com/intel/Intel975...3min53s_tn.png
hehe
with clkgen, about 70% effiency in pi calculating clk to clk set from bios.
so, expecting gocchin and eva show us a 400*9 set from bios pi_32m
with ram 4:5 running at 500mhz 4-3-2-8.
that will be more easy for the comparision. :D:D
My corsair C3's arrive tomorrow... I have a much better chance at 500mhz 4-3-2-8 with them. :)Quote:
Originally Posted by VictorWang
don't think i'd get my 5400UL to those speeds at 2.2v vdimm lolQuote:
Originally Posted by VictorWang
ill see what my memory can really do tomorow :D
could someone remember where are the vdimm reading points ?