P4C800-E >260mhz 2-2-2-6

**eva2000** · 07-22-2004, 11:32 PM

I'm having loads of fun with this OCZ DDR Booster on my Asus P4C800-E rev2.0 with 1016 bios. But I've seem to have hit a ceiling now and can't get over 260mhz 2-2-2-6 1:1.

System:
2.8C SL6Z5 with Corsair Hydrocool does max 275-283fsb
Asus P4C800-E rev2.0 1016 bios stock NB heatsink with stock TIM
256MB Gecube X800Pro @ stock
80GB WD 800JB
550W Antec TrueControl
WinXP Pro SP1

2 pairs of Winbond BH-5 tested.

1. 2x 256MB Kingston HyperX PC3200 KHX3200K2/512 BH-5
2. 2x 256MB Corsair XMS3500C2v1.1 BH-5

----------------------------------------------------------------------------
vAGP = 1.6v unless otherwise stated below
PAM = AUTO
DDR Idle Timer = AUTO
DDR Refresh Rate = AUTO
PM = Standard
Spread spectrum = Disabled
AGP Aperture = 64MB
AGP/PCI = 66.66/33.33
USB Legacy = Disabled

1. 2x 256MB Kingston HyperX PC3200 KHX3200K2/512 BH-5
- did 240-245mhz 2-2-2-6 at 3.18v on AI7
- on P4C800-E with OCZ DDR Booster:
245mhz 2-2-2-6 at 3.0v
250mhz 2-2-2-6 at 3.2v
255mhz 2-2-2-6 at 3.3v
257mhz 2-2-2-6 at 3.4v

single channel 1x 256MB KHX3200K2/512 BH-5 stick does:

255mhz 2-2-2-6 at 3.1v
260mhz 2-2-2-6 at 3.3v
265mhz 2-2-2-6 at 3.4v
270mhz 2-2-2-6 at 3.5v

2. 2x 256MB Corsair XMS3500C2v1.1 BH-5
- did 245-252mhz 2-2-2-6 at 3.18v on AI7
- on P4C800-E with OCZ DDR Booster:
240mhz 2-2-2-6 at 2.9v
245mhz 2-2-2-6 at 3.0v
250mhz 2-2-2-6 at 3.1v
255mhz 2-2-2-6 at 3.2v
257mhz 2-2-2-6 at 3.3v
259mhz 2-2-2-6 at 3.4v with 1.7v vAGP ok - with 1.6v vAGP same freeze in memtest

But either pair in dual channel at 260mhz 2-2-2-6 with 3.4-3.6v vdimm and 1.6-1.8 vAGP will freeze in memtest86 in 1st or 2nd pass of Test #5 loop.

Sooooo, could it be I need better NB heatsink cooling ? Or some other tweaks ? Or just the limitation of my particular sticks of BH-5 ? Or limitation of the memory controller in dual channel mode ?

Any help is appreciated

**Minnyboy** · 07-23-2004, 12:04 AM

I think it's a limitation of your sticks in dual channel. Mine does probably 10mhz more in single channel as well.

What's your best with PAT enabled??

Better cooling for your N/B is a must imo.

Stick a fan at your ram & that should help cool you northbridge as well.

Replace the stock TIM with some AS5 perhaps??

I'm more inclined to thinking that it's your ram's limits in dual channel.

Uh, oh. Me thinks eva2000 is coming for my score like a rocket

Gotta get 14,987 to beat me now eva2000. That'd beat mine by a BIG 1 pt.

Gotta get me one of these little turbochargers for future boards ASAP.

Cheers,
Minnyboy

Cheers,
Minnyboy

**eva2000** · 07-23-2004, 12:07 AM

yeah i have 2x 40x25mm fans on top of the ram already.

PAT is always enabled on P4C800-E or do you mean PAM - Performance Acceleration Mode in chipset menu ?

**DrJay** · 07-23-2004, 12:19 AM

Never mind the DDR booster, that's a great OC on the CPU! Almost 4 gig with 'only' a Corsair hydrocool system! Doesn't it use 1/4" ID tubes?

**eva2000** · 07-23-2004, 12:22 AM

yeah Hydrocool uses externel 1/4" ID and internal 3/4" ID tubing.

I have 3x 2.8C SL6Z5 exactly the same 3343A746 /March 18, 2004 pack dates all 3 do 275-280fsb on air as well as on water

**althes** · 07-23-2004, 02:21 AM

very nice

**Minnyboy** · 07-23-2004, 02:36 AM

Originally posted by eva2000
yeah i have 2x 40x25mm fans on top of the ram already.

PAT is always enabled on P4C800-E or do you mean PAM - Performance Acceleration Mode in chipset menu ?

Yeah, PAM was what I was talking bout. Been a while since I had the P4C800-E...

PAT, GAT, PAM, MAM.... Same $h!t...

Just got meself a 2.8c as well. Free trade-up. The perks of working in a PC shop

.

Just need the luck with my 3rd X800Pro now...
Was thinking about leaving you way behind in 3DMark03 with a new card & all but alas, it was not to be

Cheers,
Minnyboy

**scottk** · 07-23-2004, 07:57 AM

What dimm slots are you using? For DC, I get better max FSB using A1 and B1 (vs. A2/B2).

Here are some things that helped my P4C800...

Reseat the NB with better TIM and put a fan on it. The NB gets really hot looping memtest T5. I've also tried Thermalright NB-1 and Swiftech MCX159 coolers, and settled with a Swiftech MCW20 on my NB. The three different aftermarket coolers yielded the same top end, I settled with the MCW20 because it's quiet. I recall another post where oppainter commented he didn't get noticeable gains until he used more extreme cooling on his NB.

If you have a USB mouse or keyboard, move them to the PS/2 connection. This is in addition to having Legacy USB disabled. This raised my "freeze limit" on T5.

Direct a fan over the PLL chip and mosfets. There is also a mosfet under the agp slot that gets REAL hot. Consider sinking them.

Raise vagp.

Note: On my board, 1.5 and 1.6 vagp setting in the bios measures ~1.6 at the mosfet leg. The 1.6, 1.7, and 1.8 settings work as expected. This is similar to the 2.55v and 2.65 vdimm settings... they both measure same at the mosfet leg.

**skate2snow** · 07-23-2004, 08:06 AM

Originally posted by Minnyboy
I think it's a limitation of your sticks in dual channel. Mine does probably 10mhz more in single channel as well.

What's your best with PAT enabled??

Better cooling for your N/B is a must imo.

Stick a fan at your ram & that should help cool you northbridge as well.

Replace the stock TIM with some AS5 perhaps??

I'm more inclined to thinking that it's your ram's limits in dual channel.

Uh, oh. Me thinks eva2000 is coming for my score like a rocket

Gotta get 14,987 to beat me now eva2000. That'd beat mine by a BIG 1 pt.

Gotta get me one of these little turbochargers for future boards ASAP.

Cheers,
Minnyboy

Cheers,
Minnyboy

The reason you have more in single channel is becoze the NB is not stressed as much.

**eva2000** · 07-23-2004, 08:31 AM

thanks guys, need to get an MCX159 then

I'm using Logitech Elite keyboard with USB to PS/2 adaptor + Logitech Trackman Trackball mouse with USB to PS/2 adaptor both connected to Aten CS-84A 4port kvm switch. So guess it's PS/2.

I noticed in windows @ 259mhz 2-2-2-6, than Everest Home and Aida reported my system bus frequency at 86mhz which is the AGP bus ???? 259mhz/3 = 86mhz ???

Or should i use clockgen or setfsb (which is better) and move up in windows ???

**skate2snow** · 07-23-2004, 08:35 AM

Both are doing the same thing, its just a different program. Both use about the same commands...

**eva2000** · 07-23-2004, 08:45 AM

Thanks... just plugged in a Promise 32bit/66mhz controller wonder if that helps ?

**scottk** · 07-23-2004, 08:48 AM

Originally posted by eva2000
Thanks... just plugged in a Promise 32bit/66mhz controller wonder if that helps ?

Unless you need the controller for additional ide devices, it shouldn't be necessary.

If you're worried about the the agp/pci lock, load up Clockgen; it reads values as expected and works fine on my p4c800e-d r2 board.

**eva2000** · 07-23-2004, 08:50 AM

thanks scottk, i am more comfortable with clockgen hehe

**thephenom** · 07-23-2004, 09:56 AM

ya, the asus stock NB sink sucks, i just glued a 40mm fan on it to help it a bit.

**Spec3** · 07-23-2004, 10:00 AM

I TEC cooled the NB of my p4c800 and it didn't get me 1 single MH higher...

**intakeem** · 07-23-2004, 10:41 AM

i have a p4c800-e dlx with vdimm@3.55 (antec 550w vmod) and i have try some ram, muskin pc3500, corsair pc3200, a-data, all with bh5 chips. it's impossible to pass 260mhz, when with my old p4c800 rev 1.03, i can go over 260mhz with less volts...i thinks that the ultimate p4c800 REV 2.0 printed on pcb it's not goog like a first p4c800 rev 1.03-->2.0

**OC_Newbee** · 07-23-2004, 10:49 AM

Originally posted by eva2000
I'm having loads of fun with this OCZ DDR Booster on my Asus P4C800-E rev2.0 with 1016 bios. But I've seem to have hit a ceiling now and can't get over 260mhz 2-2-2-6 1:1.

System:
2.8C SL6Z5 with Corsair Hydrocool does max 275-283fsb
Asus P4C800-E rev2.0 1016 bios stock NB heatsink with stock TIM
256MB Gecube X800Pro @ stock
80GB WD 800JB
550W Antec TrueControl
WinXP Pro SP1

2 pairs of Winbond BH-5 tested.

1. 2x 256MB Kingston HyperX PC3200 KHX3200K2/512 BH-5
2. 2x 256MB Corsair XMS3500C2v1.1 BH-5

----------------------------------------------------------------------------
vAGP = 1.6v unless otherwise stated below
PAM = AUTO
DDR Idle Timer = AUTO
DDR Refresh Rate = AUTO
PM = Standard
Spread spectrum = Disabled
AGP Aperture = 64MB
AGP/PCI = 66.66/33.33
USB Legacy = Disabled

1. 2x 256MB Kingston HyperX PC3200 KHX3200K2/512 BH-5
- did 240-245mhz 2-2-2-6 at 3.18v on AI7
- on P4C800-E with OCZ DDR Booster:
245mhz 2-2-2-6 at 3.0v
250mhz 2-2-2-6 at 3.2v
255mhz 2-2-2-6 at 3.3v
257mhz 2-2-2-6 at 3.4v

single channel 1x 256MB KHX3200K2/512 BH-5 stick does:

255mhz 2-2-2-6 at 3.1v
260mhz 2-2-2-6 at 3.3v
265mhz 2-2-2-6 at 3.4v
270mhz 2-2-2-6 at 3.5v

2. 2x 256MB Corsair XMS3500C2v1.1 BH-5
- did 245-252mhz 2-2-2-6 at 3.18v on AI7
- on P4C800-E with OCZ DDR Booster:
240mhz 2-2-2-6 at 2.9v
245mhz 2-2-2-6 at 3.0v
250mhz 2-2-2-6 at 3.1v
255mhz 2-2-2-6 at 3.2v
257mhz 2-2-2-6 at 3.3v
259mhz 2-2-2-6 at 3.4v with 1.7v vAGP ok - with 1.6v vAGP same freeze in memtest

But either pair in dual channel at 260mhz 2-2-2-6 with 3.4-3.6v vdimm and 1.6-1.8 vAGP will freeze in memtest86 in 1st or 2nd pass of Test #5 loop.

Sooooo, could it be I need better NB heatsink cooling ? Or some other tweaks ? Or just the limitation of my particular sticks of BH-5 ? Or limitation of the memory controller in dual channel mode ?

Any help is appreciated

Eva2000, very nice result. I find that my bh-5 stick of XMS go good with the P4C800e board. It scale higher than the epox and IC7 that I have.

Yeah maybe a better cooling to the NB and maybe raise the Vagp to 1.8v might help stability a little more. My pair of 256mb XMS only need 3.48 vdimm and 1.8 Vagp to be fully stable at test # 5 loop overnight at 260fsb 1:1 2-2-2-5 with everything enable(except for spread spectrum).

How is your unbuffer score at 260fsb 1:1 2-2-2-6?

Mine at 260fsb 1:1 2-2-2-5 2x 256mb XMS PC3500 @ 3.48v
vAGP = 1.8v
PAM = Enable
DDR Idle Timer = Infinite
DDR Refresh Rate = 64uSec
Spread spectrum = Disabled
AGP Aperture = 256MB
AGP/PCI = 66.66/33.33
USB Legacy = Disabled

Sandra unbuffered: 3600+
Sandra buffered: 6300+

**eva2000** · 07-23-2004, 10:58 AM

Originally posted by intakeem
i have a p4c800-e dlx with vdimm@3.55 (antec 550w vmod) and i have try some ram, muskin pc3500, corsair pc3200, a-data, all with bh5 chips. it's impossible to pass 260mhz, when with my old p4c800 rev 1.03, i can go over 260mhz with less volts...i thinks that the ultimate p4c800 REV 2.0 printed on pcb it's not goog like a first p4c800 rev 1.03-->2.0

interesting might try my P4C800-E rev1.02 instead then

as to my unbuffered

2.8C @ 259fsb
1:1
2x 256MB XMS3500C2v1.1 BH-5 @259mhz 2-2-2-6 at 3.4v

vAGP = 1.7v
PM = Standard
PAM = Auto
DDR Idle Timer = Infinite
DDR Refresh Rate = 64uSec
Spread spectrum = Disabled
AGP Aperture = 64MB
AGP/PCI = 66.66/33.33
USB Legacy = Disabled

Buffered = 6407 / 6401
Unbuffered = 3639 / 3710

see this is what everest home reports for chipset real clock = 86mhz

**mikead_99** · 07-23-2004, 11:11 AM

FWIW, I've had 265MHz 1:1 2-2-2-5 out of my Rev. 2 board, with vdimm@3.55 (antec 550w vmod). I'm actually hoping the DDRBooster will prove my volt-limited theory, not the 260 fsb theory. Looks not so good for me from what you guys are saying.

**eva2000** · 07-23-2004, 01:29 PM

mikead_99 could be each p4c800-e is slightly different ?

anyway, I did a few changes swapped Gecube X800Pro out for Albatron MX480EL card and put 80GB WD 800JB Hdd on Promise 32bit/66mhz controller instead of board's primary IDE channel and can manage to scrape in a few benchies @ 265mhz 2-2-2-6 at 3.6v

2.8C @ 265fsb
Asus P4C800-E rev2.0 1016 bios
Albatron MX480EL @ stock 275/266
80GB WD 800JB on Promise ATA133 32bit/66mhz controller
550W Antec TrueControl PSU

1:1
2x 256MB XMS3500C2v1.1 BH-5 @265mhz 2-2-2-6 at 3.6v

vAGP = 1.7v for clockgen benched superpi/aida32 and 1.8v for sandra bandwidth benchies
PM = Standard
PAM = Auto
DDR Idle Timer = Infinite
DDR Refresh Rate = 64uSec
Spread spectrum = Disabled
AGP Aperture = 64MB
AGP/PCI = 66.66/33.33
USB Legacy = Disabled

Sandra 2004 STD SP2
Buffered = 6555 / 6550
Unbuffered = 3782 / 3793

Aida32 v3.93
Read = 6468 MB/s
Write = 2026 MB/s
Ram SPD settings

SuperPi
1M = 36s
2M = 1min 25s

notes:
- this ain't 100% stable but working on it

- vAGP of 1.6 was fine if i booted into windows at 233fsb and used clockgen to reach 265mhz 2-2-2-6 and managed run off aida and superpi benchies but trying to boot into windows at 265mhz 2-2-2-6 froze at winxp welcome screen
- vAGP of 1.7v allowed me to boot into windows at 265mhz 2-2-2-6 but mouse and keyboard didn't respond
- vAGP of 1.8v allowed my mouse and keyboard to work fine and run sandra buff/unbuffered benchmarks

**masterofpuppets** · 07-23-2004, 01:34 PM

Bah, I've got the Revision 1 p4c800-e myself. Is the Rev.2 any better?

**eva2000** · 07-23-2004, 01:42 PM

i have both revisions and don't notice any difference but then again they're both stock with no volt mods so with 2.85v i can't really compare high 2-2-2-6 1:1 ocs like now with the OCZ DDR Booster

**neyoung** · 07-23-2004, 09:35 PM

I'm haveing the same problem. My mush bh5 and hyperx that I have both bsod when I get to 260 fsb. I have a tiger 1 nb cooler. I know that people get higher fsb on these boards so there must be a trick. Anyone care to share