There it is! BH-5 is back again!

[GazC]

It's a pity you cannot do direct purchasing from Twinmos.

[kryptobs2000]

dh5 > bh5 > ch5???

[Rabbi_NZ]

hmmm been looking for some good CH/BH/DH? too

[Shroomalistic]

so are these moduals actually produced by winbond or just winbond machines?

[enok]

great... let's see some DDR action! shall we?

god praise TwinMOS! i'm really starting to love them!

and show us some more tests Alonso! let's see if this is/isn't an improved BH-5 chip. (taking UTT scores as example.... )

[Bluebeard]

it´s the well known BH-5. Geting hotter than CH-5 in high voltage

@ kryotops

like we learned in elementary scool

a-b-c-d-....

[pik-ard v1.1]

so Alonso, you're saying that the DH-5 is WORSE then CH-5? considering BH-5 was always considered better then CH-5...

[enok]

IF (lot of if's here ) u consider A-B-C-D as A < B < C < D. for as lower process making, lower power consuming and lower temps... teorically better oc.
UTT is in fact what CH-5 should have been from the start?! considering the excelent results it has done, especially for 2x512 DC?
i think it isn't too bold to conclude that UTT is probably making as good results as initial BH-5!?
so ch should have been > to bh all along (if just winbond didn't have to make those damn budget "shortcuts"...), and probably dh will be even better. (right OPB? )

[Bluebeard]

so Alonso, you're saying that the DH-5 is WORSE then CH-5? considering BH-5 was always considered better then CH-5...

DH-5 is not built yet...

NEW CH-5 is not running as hot as OLD, new released, BH-5 in high-voltage and reaching almost same results...


@ enok

?

[enok]

like we learned in elementary scool

a-b-c-d-....

a < b < c < d?

[kryptobs2000]

I understand now, but if some people still don't, here's how it is, although it's been explained already, just summing it all up.

bh5 = 15nm (I thought it was always 13, but shows what I know)
ch5 = 13nm
dh5 = 11nm

Now, technically ch5 SHOULD be better than bh5, but it wasn't, they cut costs and used cheaper materials and the 13nm process wasn't perfected (my guess) so therefore ch5 < bh5. So dh5 has the potential to be better, lets just hope it is.

[kristos]

Production Date 07/2005

How does one recognize the production week/year?

S/N: 507M55M12111149

is this it? or is it something else?

[enok]

Now, technically ch5 SHOULD be better than bh5, but it wasn't, they cut costs and used cheaper materials and the 13nm process wasn't perfected (my guess) so therefore ch5 < bh5. So dh5 has the potential to be better, lets just hope it is.

that's why i said UTT might be what CH-5 should have been, if they didn't cut the budget then.

[Bluebeard]

S/N: 507M55M12111149

right! year 2005 week 07

[Rabbi_NZ]

both amd and intel had issues when first going to 90nm.... a new process will usually cause headaches to begin with. jus shows that the CH process wasnt quite ready during most of the CH days, and now that winbond are mastering it the UTT/CH being produced is VERY GOOD

I'm still hoping for DH>CH>BH

[THunDA]

** subscribing to the wonderfull thread **

[DaveX]

If the 0.11u process isn't perfected yet, then DH-5 will probably not be able to take a lot of volts, kind of like the old CH-5, meaning bad oc.

[Revv23]

seems to me that if DH-5 was going to be much better they would call it DH-4 am i right?

from my point of view it seems like they are using the die shrink solely to cut costs, which doesnt mean better speeds to me, as smaller process cannot handle as much voltage, and they may cut even more corners to save even more money like they did with CH...

i also would not expect them to try and make a -4 anything, why would they need/want to, when the DDR market will never go over DDR400, frankly, i think that we will not see better DDR then our BH/TCCD/UTT ever, as DDR2 is on its way in, with no market pressure to get DDR going faster, what im saying is why would engineers make something to go DDR500, when they could make a cheap DDR400, and cash in on the very last budget DDR people in a market where all the "big" ram manufacturers have alreadyt gone DDR2...

sorry its confusing what im trying to say...i may be wrong but thats what im thinking...

[kryptobs2000]

because not everyone uses an intel, those with amd won't see ddr2 for another year at least, and ddr2 isn't so hot right now from what I've seen anyways, especially even more so since amd's like low latencies.

[pik-ard v1.1]

i wouldnt go and say that....



by FUGGER.... read up: http://www.xtremesystems.org/forums/...2&page=1&pp=25

[Skip]

thats his centon memory, pretty sweet stuff.

[ZL1]

a bit confused

utt is ch5 ?
and new ch5 (utt) better than bh5 ???


Thanks
Dan

[kryptobs2000]

yeah, basically, it's bh5 only 13nm (vs 15nm)

[Enz0]

BH-5 can run at Trfc of 12.
But CH-5 or UTT can't.
Alonso, please set Trfc to 12 and check if your RAM can handle that.

[TekXoID]

Hey Enz0, we have pretty similar ram and FSB, I was wondering what A64 Tweaker settings
you use as I have been pretty unsucessful getting anything out of the tweak-tool myself.