Is bh-5 still alive? got 2x256mb running at ddr560 2-2-2-5

[Onepagebook]

mobo: MSK k8n neo2 1.3b6 beta bios
CPU : 3500+ 90nm s939
Cooling : vapo LS
OCZ DDR booster 3.7v

[Torin]

Nice. It seems the problems with burning out A64 mem controller from high VDIMM are long gone.

[naelq]

now that's what i call speed!! it's the almighty 2-2-2-x
which voltage!?

/// edit ///
WTF??! 3.7v !!! :banana4:

[STEvil]

i seem to remember macci running at 4.0v pushing for 310mhz+ with some bh-5.

[Arnar]

Real nice!!

Can you post some bh-5 vs gskill results?

then each at its max freq.
Im about to have 3x BH-5 kits (2x512mb and 1x1g kits) and was wondering if I should get some gskill..

thanks in advance

[Onepagebook]

Real nice!!

Can you post some bh-5 vs gskill results?

then each at its max freq.
Im about to have 3x BH-5 kits (2x512mb and 1x1g kits) and was wondering if I should get some gskill..

thanks in advance

I think that is totally different story, You can't just do it for 3.7v every day
but you can have samsung TCCD chip play 2.6v ddr550 every day.

[Arnar]

Hehe.. but cant you use 3.5v on BH-5 for every day?

It would though be nice to see 250 BH-5 (2225) vs gskill at 300

And have multiplier set so the cpu is at the sem freq

[Onepagebook]

Arnar, I can be very sure that
DDR500 2-2-2-5 is much much worse than TCCD running at ddr600 2.5-3-3-x(in terms of normal badwidth efficiency)
And the voltage for BH-5 running ddr500 is around 3.1~3.25v while TCCD only needs about 2.65v~2.75v to run DDR600(256mb x2 dual channel please)

[Arnar]

Ok, I was just looking for something to make up my mind wether to get Gskill ram.. since I recently bought a whole lot of BH-5 (3kits)

Voltages on the RAM doesnt matter ( I also have OCZ DDR Booster ).. just that they could live 24/7 on them ( or just couple of hours a day ) so I would be very thankfull if you could do that

[Onepagebook]

ya I did that earlier already, please check out here:
http://www.xtremesystems.org/forums/...ad.php?t=43710

you can see those results from TCCD doing over ddr600(at least).
compare with this bh5 done for ddr560 2-2-2-5, especially TCCD on ddr611(fsb 305)
has got better result 8543mb/s than bh5 though.

[Arnar]

ok thanks.. and would the results be the same for 3dmark01 also?

[Gogeta]

Very nice Kev
I love seeing those 8xxxMB/s Sandra screens.

[hipro5]

Very nice........What about unbuffered Sandra with TCCD and BH-5 at their maxed frequency?......

[Geforce4ti4200]

"TCCD only needs about 2.65v~2.75v to run DDR600(256mb x2 dual channel please)"

but at 3-4-4-8 timings which would cancel out. dont ever use worse than 2.5-3-3-7 timings, the super high latency will kill performance, esp with a64s

[drunkenmaster]

as pointed out buffered sandra is not a huge deal, it tells us something but unbuffered and the ddr500 2,2,2,5 vs ddr600 2.5,3,3,7 won't be that far apart i wouldn't think. Also wanna see 3dmark results to show general performance.

another thing is, jsut cos the tccd will run high on lower voltage, doesn't make that everyday use and bh-5 not. simply put there are a shedload of people that have run up to 3.5v for long periods of time and not had a stick die. its about running voltage that those memory chips can handle, tccd likes less voltage, that doesn't make runnning at 2.75v inherantly better. tbh i'd ru n3.7v without question all day long without caring. just not got a chip/board combo that wants to run high enough fsb to warrant that at the moment,


also, ath 64 deaths, other than that basically initial thread what, almost a year ago where only a few people actually had deaths and almost entirely all of them were on chaintech(early boards) or modded vdimm shuttle boards where almost all of them died from fluctuations, basically unsafe mod. Since then i ain't heard anything about dying ath 64's. Sure a few die here or there, no product in any history doesn't have some faulty components. My first 3200+ back last, sept? august? i ran at 3.3v for like 7 months straight before it was sold, to my m8, with the board, and its been at the same since then.

[macci]

DDR500 2-2-2-5 is much much worse than TCCD running at ddr600 2.5-3-3-x

Lets see a 3DMark2001 Lobby Hi test then?
12x250 1:1 2-2-2-5 vs. 10x300 1:1 2.5-3-3-x ?
edit: lets add 272x11 1:1 2-2-2-5 there too since you obviously got RAM capable of it
Can you do one? Thx

[Elisha]

no problem running BH5/6 at high voltage at all.

[althes]

That is a very nice score. Cant wait for those benches

[Onepagebook]

Lets see a 3DMark2001 Lobby Hi test then?
12x250 1:1 2-2-2-5 vs. 10x300 1:1 2.5-3-3-x ?
edit: lets add 272x11 1:1 2-2-2-5 there too since you obviously got RAM capable of it
Can you do one? Thx

Sorry Macci, I am talking about normal bandwidth. I certainly understand what your point is.

[Onepagebook]

PLEASE test it.

ya Iwill definitely do that when I got time.
need a spare HD to install win2k

[TEDY]

why win2k ?

[FlyingHamster]

why win2k ?

because for performance, generally:

amd + win2k

intel + winxp

[Onepagebook]

As the advice from Macci,(run 3d mark 2001se, game 3 high details only);please see here
Mobo: MSI k8n neo2 Bios: 1.3b6
RAM: 1.Mushkin PC 3500 Level II BH-5, 2x 256mb DC kit
2.GSkill TCCD DDR550 2.5-3-3-7, 2x256mb DC kit
OS: Windows 2000 SP4
CPU AMD s939 3500+ 90nm
Power: OCZ power stream 520W
HD: Maxtor 6E040L0 IDE

Let's see those two set of RAM:



1. BH-5 273x11 2-2-2-5:



2.TCCD 300x10; 2.5-3-3-6:



It shows that results are really close to each other.
Therefore, I think there is no need to do BH-5 at 12x250, 2-2-2-5 right?

Thank you everyone.

[TEDY]

sell me these mushkin

[blinky]

what volts on the BH5?