By making the mod, you are taking BSEL2 from L to H. That would move you from the first row int Table 17 to the last row in Table 17. It's safe to say that the last row in Table 17 is "RESERVED" for 333 MHz FSB.
Seems pretty straightforward to me.
Printable View
By making the mod, you are taking BSEL2 from L to H. That would move you from the first row int Table 17 to the last row in Table 17. It's safe to say that the last row in Table 17 is "RESERVED" for 333 MHz FSB.
Seems pretty straightforward to me.
kunaak,
have you been able to measure any increase in temps by simply jumping to 1333. either cpu or chipset?
i know its prob. not going to happen but it would be interesting to see what the cpu and/or chipset does to itself in terms of output wattage/heat when doing this mod. i know in appearance its nothing more than changing the FSB to 333. but in the internals of the cpu and motherboard,, whats happening.
is the cpu jumping up its default voltage?
is the chipset changing the latency? is it also increasing the voltage its asking for??
Quote:
Originally Posted by virtualrain
no its not safe to say that unless you can find what that registered is reserved for.
i mean obviously its happening, but its a risk,, a big one, and thanks to Big K for possibly sacrificing one of his chips to test it.
makes me want to tear apart my rig and try it.
so, its not possible an E4XXX (200MHz) become a 333MHz, because 200MHz are L-H-L and 333MHz are H-L-LQuote:
Originally Posted by virtualrain
we have to mod more pins (BSEL1 and 2) on E4XXX in order to work, like:
http://img127.imageshack.us/img127/6796/intel01ack3.jpg
make a jumper between G30 <-> J30 and isolate H30 ;)
or simply isolate H30 to have 266MHz, an E6600 with half cache.
/// EDIT
reading Intel's Xeon 5100 Series Datasheet I've found that on page 19:
http://img71.imageshack.us/img71/3898/intel07ib1.jpg
So, the reserved is for 333.33MHz, and it is H-L-L ;) . The pins are the same, G30.
how did it boot just like that on the quad core? didnt it need more voltage at 1333?:eek:
on my CPU no.Quote:
Originally Posted by GrimReaper
mine will do 3.4 ghz at stock voltage, and 3.6 ghz with 1.38 volts.
thats the tops I was able to easily keep stable in my P5B Deluxe.
same CPU tested on RD600 and 680i, and in all cases, the CPU doesnt need any extra voltage to reach 3 ghz.
my CPU's just a good one, but I also said many times...
make sure your CPU can do the stock multiplier x 333 at stock voltage, otherwise this mod would be useless to you.
heres how to make a E4300 boot at 9 x 266.
http://www.ocforums.com/showpost.php...1&postcount=40
shame,i was getting all exited . have you tried benching? to compare and see if theres any loss or gain ?Quote:
Originally Posted by over@locker886
nice work kunaak .
:eek: I need to grow some balls before I can try this on my 6600 :p:Quote:
Originally Posted by Kunaak
Quote:
Originally Posted by MarlboroMan
So far, the Pin-out for my Pentium 4 630 is absoluteley the same for 266mhz FSB. Does that mean i just have to isolate H30 to get the FSB 1066 working on my Processor?
I dont see though this oO...
Quote:
Originally Posted by Kunaak
somebody has tried with a e6700 ?
333 * 10 = 3330 mhz instead of 2667 ....
Wait a second. I think I'm getting a little rusty, but I recall when I implemented the dothan mod, I had to make a switch from high resistance to low, and did so by connecting to vss (ground point). I've never worked the other way, but always remembered high resistance being flagged by pulling a pin back in pga days or something like white out with the newer lga's.
Looking at the spec, 266 is LLL and 333 is HLL. Was it determined that connecting to vcc (power) will flag it to high resistance?!?! Shouldn't we be whiting out the land for bsel2 to make it high? Sorry if I'm sounding a little stupid. I'm a tad bit rusty
Hey Kunaak,
I never thought of shorting a VCC pin to the BSEL2 pin, back in August I attempted the same thing by insulating the BSEL2 pin on an E6400 to force the 1333 strap but unfortunately my DFI at the time didn't support the 1333 strap so I got no POST :p:
Well done! Awesome to see that it works :) Now to order and E4300 and have a play around.
Here's the thread I created on the subject:
http://www.xtremesystems.org/forums/...d.php?t=111510
Theoretically yes, but I believe connecting the VCC pin means that current will flow into the bsel pad on the CPU, instead of from the bsel pad into the lga array, thereby insulating it. The bsel pad must be lower resistance than the corresponding pin then and what we're doing is making the pad high resistance.Quote:
Originally Posted by ziddey
I think :rolleyes: :confused:
Hmm, interesting. Here's a more complete list I've compiled:
BSEL2 BSEL1 BSEL0 FSB
L L L 266
L L H 133
L H H 166
L H L 200
H H L
H H H
H L H
H L L 333
Not sure if the intermediate bsel flags would still work since it's still largely bios dependent as to whether it can interpret it properly or not
no, this was my mistake.. You can try isolating it, but i don't think it will work.Quote:
Originally Posted by cryptocat
You have to make a jumper between H30 (BSEL 1) and L30 (VSS - ground)
make sure you are doing right, download and check Pentiums datasheet on intel's website ;)
Erh... why L30? thats like 3 Pins away from the BSEL ones....
And by jumper you mean, connecting those 2 pins with silverpaint?
I cannot seem to find this 1901 bios for the P5W-DH anywhere.Quote:
Originally Posted by Kunaak
I'll give it a go with my 6700.
Anyone?
Nice find and good work Kunaak
if you don't know what are you doing, don't do itQuote:
Originally Posted by cryptocat
my advise for ya is not to do it or learn more b4 doing it ;)
Does anybody know a decent place to get an explanation of BSEL and all this other stuff you guys are talking about?
Go to intels website and look at the white sheets for the processors.Quote:
Originally Posted by MotF Bane
Motf bane, google up xeon pinmodding. Most of what you'd find would be about old stuff back when they were pga but there should be some good explanations. I wrote a decent article three years ago but the server went mia unfortunately. Google still has a cache of the writeup although the pictures (useful) are gone. You can find it at http://209.85.165.104/search?q=cache...lnk&cd=3&gl=us
It takes a long time to load for some reason, but I actually had to reference the cache today myself since I got a little rusty. Hope it helps.
Basically, bsel is a set of flags defined on the cpu that the motherboard reads to know what fsb to set as default. Input is measured in terms of high or low resistance. Thus, in the traditional sense, to achieve a high resistance from something that's originally low, you can use white out or nail polish or tape to cover up a land. In the pga days, it's be nail polish to insulate or just yanking a pin.
To make a high resistance into one of low resistance, one can either link it to a bsel of low resistance, or link it to a ground (vss).
It looks like from this technique as posted by op, it's been determined that you can link up a bsel that's originally flagged as low to a vcc (power) to bring it to high. I'm still thinking of the phenomena behind this myself, but it looks like it'd essentially be causing a relative resistance as seen by the motherboard to be high and thus be the same as isolation of the land/pin.
Hope this helps :)
how about testing this on one of the mATX boards where the pci-e lock breaks past 300.
would LOVE!!!!!! to see this work!
The problem is that the board itself has to have the bios to support the bsel flags. Otherwise, it'll either still boot at 266 or worse, not post at all.
I tried e6400 with p5b dlx, it says bios 1004 supports it, but no post.Quote:
Originally Posted by ziddey