DFI Lanparty 875P-T LGA775 Benchmark Adventures

[eva2000]

I was lucky enough to grab a P4 3.4E 550 cpu on Christmas Eve to go with my 2nd hand DFI Lanparty 875P-T motherboard that I bought off Oppainter at xtremesystems.org forums Had a few initial problems in that the system didn't boot up but loosening the G4 Storm waterblock's mounting screws a bit on the DFI Lanparty 875P-T allowed it to boot.

System:

P4 3.4E 550
SL7J8
Q433A423
Nov 26, 2004
Malaysia

DFI Lanparty 875P-T with 11/30 bios
2x 256MB KHX3200K2/512 BH-5
64MB Albatron MX480EL (for testing)
120GB WD 1200JB
Pioneer AO5 DVD Burner
580W Topower 686P6 Adjustable rail PSU

CPU cooled by:

Cathar's G4 Storm waterblock with revised stepped base addition for LGA775 cpus
Swiftec MCP650 Pump
Thermochill 120.3 with 3x 120x38mm Panaflo 103cfm fans on 3.5" Rheobus 13W per channel
Dangerden fill port
Clearflex 1/2" ID tubing

Stock benchmarks:

Winbond Hardware doctor


Sandra 2005








Everest Home
Read: 5,808 MB/s
Write: 2,056 MB/s
Latency: 80.3 ns

Super Pi unpatched
1M = 38s
2M = 1min 26s
8M = 6min 47s

Super Pi patched
1M = 34s
2M = 1min 19s
8M = 6min 18s

StressPrime 2004
16+ mins

Overclocking:

I still haven't found my way around the bios options and not sure how well this memory performs on the 875P-T yet but I was able to boot and take this screenie at stock vcore - not stable



Not bad temperature wise which was my main concern. I just have to test all my memory on this setup and see what works. I haven't installed my OCZ DDR Booster yet but I know it works with this motherboard so can pump more vdimm into my memory

So having a 875P chipset, does it also have the same affect in that 2x 512MB will result in more bandwidth than 2x 256MB memory modules ? I'll find out soon

Note: Seems the Winbond Hardware doctor's reported 12v rail is off off a bit, it shows 12.24-12.38v but from multimeter it says 11.98v idle and 11.94v load. So i'v raised it to 12.20v idle and see if that helps



.

[eva2000]

Reserved for overclocked benchmarks

Full Pics here

[$0m#0n#]

Nice. What r ur temp?
BTW I dont like screens @stock frequency

[eva2000]

added some pics to 2nd post

This is where i am now priming



i'm completing my full set of benchmarks @ 4097mhz and man is this 875P-T one screaming fast board.. think of Asus P4C800-E and Abit AI7 mixed into one, with fast stable cpu/system performance and the memory/GAT tweaks of AI7.

i.e. i'm able to pull off 3811MB/s memtest86+ v1.40 bandwidth @ 17x 241mhz 3-2-2-6 with 2x 512MB OCZ PC3700EB single sticks at 3.2v vdimm

[eva2000]

Time for some overclocked benchmarks This 875P-T is very aggressive on memory and really puts out some nice memory bandwidth!

3.4E 550 SL7J8 @ 4097mhz at 1.5-1.53v vcore load
DFI Lanparty 875P-T 11/30 bios
2x 512MB OCZ PC3700EB Singles @ 241mhz 3-2-2-6 at 3.2v vdimm
80GB WD 800JB
580W Topower 6P686 Adjustable rail PSU

Sandra 2005
CPU = 11,997 / 4,909/8594
CPU Multimedia = 28,922 / 38,838
Buffered = 6,080 / 6,066
Unbuffered = 3,692 / 3,741

Everest Home
Read = 7,231 MB/s
Write = 2,632 MB/s
Latency = 66.4 ns

SuperPI
1M = 31s
2M = 1min 10s
4M = 2min 34s
8M = 5min 35s
16M = 12min 23s
32M = 27min 05s

SuperPI Patched
1M = 28s
2M = 1min 05s
4M = 2min 22s
8M = 5min 11s
16M = 11min 29s
32M = 25min 10s

Bloody fast for the low mem speed


.

[IamAnoobieCheez]

That is bloody impressive.

[eva2000]

thanks mate

Not stable but here's more benchies

3.4E 550 SL7J8 @ 4182mhz at 1.568-1.6v vcore load
DFI Lanparty 875P-T 11/30 bios
2x 512MB OCZ PC3700EB Singles @ 246mhz 3-2-2-6 at 3.2v vdimm
80GB WD 800JB
580W Topower 6P686 Adjustable rail PSU

Sandra 2005
CPU = 12,156 / 5,017/8,714
CPU Multimedia = 29,518 / 39,625
Buffered = 6,190 / 6,195
Unbuffered = 3,769 / 3,824

Everest Home
Read = 7,312
Write = 2,426
Latency = 65 ns

SuperPI
1M = 30s
2M = 1min 09s
4M = 2min 31s

SuperPI Patched
1M = 28s
2M = 1min 03s
4M = 2min 19s

[eva2000]

dropping to 14x multiplier for some bandwidth benchies @ 251mhz 3-2-2-8 at 3.2v

Everest Home
Read = 7,343 MB/s
Write = 2,767 MB/s
Latency = 68.5 ns

Sandra 2005
Buffered = 6,239 / 6,239
Unbuffered = 3,686 / 3,747 (guess the lower 14x multipler reduced this considering i get higher @ 17x X 246mhz 3-2-2-6 here )

[xsky]

i fried two boosters on this board within seconds after first boot...

[eva2000]

i fried two boosters on this board within seconds after first boot...

yeah i fried mine too at 3.2-3.3v too in same exact area that Fugger fried his

[xsky]

yeah i fried mine too at 3.2-3.3v too in same exact area that Fugger fried his

welcome to the club

exactly that happened two times to me. my vdimm was set to 2.6V in bios and the booster died with 2.8V on display!!!!
it wasn't warm at all

[eva2000]

was it fried in the same trace/area as above pic ?

[xsky]

yes. absolutely the same area as on the pic BOTH times

[eva2000]

xsky, what psu and psu rail settings you had with the combo ?

[xsky]

first time: antec 550W True Power
second time: ocz powerstream 520W set to exactly 3.3/5/12V

[eva2000]

well that seems to rule out the psu being the problem

[Tony]

just bridge the burn with a wire and the booster will come back to life...i did the same as you guys...LOL

I also wacked the 3.3V line up on the PSU and had 3.5Vdimm at the 3.2v setting which was nice...

[eva2000]

oh you mean 3.3v rail upped without booster can give 3.5v ??? or only with booster in place ? got pics of the bridged trace ? i don't have the skill to do such a thing anyway... have to rma it once the retailer reopens heh

where do you measure vdimm without booster ? winbond hardware doctor on cd-rom displays vdimm but it beeps like crazy even at 2.8v vdimm, the newer winbond hardware doctor on web site at DFI removed vdimm display option

[Tony]

Tme you learn't how to solder there eva...i just bridged over the burn section to 2 solder points with a wire and the booster came back to life...real easy repair.Rear window defogger paint may do the same for you...apply it thick so it takes the current but the wire is a much better job.

No booster and 3.6v+ on 3.3v line upped the vdimm.

[Tony]

for those with burn't boosters from the 875P-T this should get the booster back up although it seems the booster is not 100% compatible with this board.

I would solder 3 wires to the 3 points from the 1 common point to spread the load. use wire that can carry 10A at 5V

[mike.elmes]

for those with burn't boosters from the 875P-T this should get the booster back up although it seems the booster is not 100% compatible with this board.

I would solder 3 wires to the 3 points from the 1 common point to spread the load. use wire that can carry 10A at 5V

My booster didn't burn like the above pic but doesn't work anymore. Is there a work around to replace the adjustable pot or are the new units different. I'm not sure the pot is fried, but it worked for one day and just stopped being adjustable.

[eva2000]

i can't even a tin a wire yet lol.. but i have 2 other boosters here so that's no prob for my other boards.. guess 875P-T doesn't like the booster... >3.3v psu rail for more vdimm sounds nice... all i need is to be able to see what vdimm i'm setting heh

[Tony]

My booster didn't burn like the above pic but doesn't work anymore. Is there a work around to replace the adjustable pot or are the new units different. I'm not sure the pot is fried, but it worked for one day and just stopped being adjustable.

Just remove the pot and solder a 10k linear 20 turn or so cermet pot on...all 3 legs

dead easy and much more finely adjustable.

AndyT has a thread in the overclocking forum showing his modded unit.

[xsky]

hi!
can anyone show a pic of vdimm read point?
@bigtoe did you remember the "special" bios setting on the board already?

[STEvil]

eva - make sure your solder is rosen core, otherwise you will never get it to tin