2.8C PX865PE PRO2 + Geil Golden Dragon PC3700

[eva2000]

CPU: DDR ratio = 1:1

2x 256MB GEIL PC3700 Golden Dragon Dual Channel @245mhz 2.5-4-4-7

=================================================
2.8C | SL6WJ | Malaysia | L308A678 | 4/4/03
AX478 + 80x38mm Delta 80cfm fan connected to TT fan speed controller
128MB Gigacube 9600PRO VIVO @default with ATI cat 3.5 drivers and control panel
120GB WD 8mb 1200JB connected to IDE1
520W Topower Gold Dual Fan PSU
WinXP PRO SP1 with DX9.0

=================================================
Albatron PX865PE PRO2 1.02 beta bios

vcore: 1.6v
vdimm: 2.85v
vAGP: 1.6v
AGP/PCI: Fixed 66/33
Enhanced Performance: Enabled
Overclocking Stability DDRAM: Standard

PCI Latency: 64
SATA: disabled
Promise Raid: disabled
=================================================

CPU: DDR ratio = 1:1

2x 256MB GEIL PC3700 Golden Dragon Dual Channel @245mhz 2-4-4-7

Room Temp: 21-23C

CPUZ & MBM 5.330 (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/cpuz.jpg)

Sandra Max buffered memory (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/mem-buff.jpg) = 5505 / 5568
Sandra Max unbuffered memory (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/mem-unbuff.jpg) = 2637 / 2728

PCMark2002 (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/pcmark2002.jpg) = 8511 / 9536 / 1097

Aida Read (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/aida-read.jpg) = 5138 MB/s
Aida Write (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/aida-write.jpg) = 1435 MB/s

3dmark2001SE (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/3dmark2001.jpg) = 12,488
3dmark2001SE @501mhz core/ 339mhz memory (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/3dmark2001.jpg) = 14,402
3Dmark2003 (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/3dmark2003.jpg) = 3499

Sciencemark (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/sciencemark.jpg) = 4518.83 MB/s

SuperPi 1m (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/superpi-1m.jpg) = 42s
SuperPi 16m (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/superpi-16m.jpg) = 17min 16s

Prime95 v23.4 Bench (http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/245-245-2.5447-ehon/prime95bench.jpg) = 59.069s

=================================================

[Hawk]

Nice eva2000,

Only thing is the memory scores in Sandra look a little low. Maybe due to the 2-4-4-7 timings? ?

[Chuck232]

Well wasn't it you Hawk that showed me that RAM timings had a huge impact of the memory scores?

IIRC you went from 2.5-4-4-7 to 2.5-3-3-7 and you got a pretty large boost.

[eva2000]

Originally posted by Hawk
Nice eva2000,

Only thing is the memory scores in Sandra look a little low. Maybe dur to thge 2-4-4-7 timings? ? there low due to

1. 865PE board and not 875P board
2. geil doesn't like cas 2 as it's rated for 2.5-3-3-7

probably Geil would do better at max supported voltage of 3.1v vdimm

[eva2000]

Originally posted by Chuck232
Well wasn't it you Hawk that showed me that RAM timings had a huge impact of the memory scores?

IIRC you went from 2.5-4-4-7 to 2.5-3-3-7 and you got a pretty large boost. that was at 233mhz 2.5-3-3-7

http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/233-233-2337-ehon/mem-buff.jpg

http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/233-233-2337-ehon/mem-unbuff.jpg

[Hawk]

Yes, but he is running those chips at CAS2, not 2.5 as they are spec'ed for. We are all seeing interesing things with the way the differnent boards handle the chip timings. This is a Sprindale board, mine is a Canterwood.





Originally posted by Chuck232
Well wasn't it you Hawk that showed me that RAM timings had a huge impact of the memory scores?

IIRC you went from 2.5-4-4-7 to 2.5-3-3-7 and you got a pretty large boost.

[eva2000]

Originally posted by Hawk
Yes, but he is running those chips at CAS2, not 2.5 as they are spec'ed for. We are all seeing interesing things with the way the differnent boards handle the chip timings. This is a Sprindale board, mine is a Canterwood. whoops typo :o

should be 245mhz 2.5-4-4-7 hehe

[Hawk]

Very interesting, your 245 Unbuffered is lower then 233 but the buffered is higher. . . . Eventhough different timings, there is enough a FSB difference to make up for it. . . .








Originally posted by eva2000
that was at 233mhz 2.5-3-3-7

http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/233-233-2337-ehon/mem-buff.jpg

http://www.fileshosts.com/pentium4/D1/albatron/benchmarks2/2.8C/geil/pc3700gd/233-233-2337-ehon/mem-unbuff.jpg

[eva2000]

Originally posted by Hawk
Very interesting, your 245 Unbuffered is lower then 233 but the buffered is higher. . . . Eventhough different timings, there is enough a FSB difference to make up for it. . . . yup, maybe sort of lends credibility to what ramguy said regarding xms3700 3-4-4-8 timings, the increased raw speed makes up for the looser timings ?

[Hawk]

eva2000, i'm not sure. . . :confused: I've seen a few weird memory scores lately that don't make sense. . . .

[eva2000]

well i'm hoping it's the PX865PE PRO2 1.02 beta bios playing up.. wish Albatron would release a bios update with more vcore and vdimm to 3.3v :D

[Hawk]

I would bet it is the BIOS, because if those are the same chips you tested in the IC7 from your other post, they are working way better in the IC7. . . .

[eva2000]

Originally posted by Hawk
I would bet it is the BIOS, because if those are the same chips you tested in the IC7 from your other post, they are working way better in the IC7. . . . no.. no.. nooooooo :)

IC7 only tested OCZ PC3700 Gold dual channel
PX865PE PRO2 only tested GEIL PC3700 Golden Dragon dual channel

haven't yet switched the GEIL memory to IC7 yet

[Hawk]

My bad. . . Read that too fast. . . .

[eva2000]

Originally posted by Hawk
My bad. . . Read that too fast. . . . you're forgiven :D

[Chuck232]

OCZ did say that the P4 dual channel platform benefited more from higher RAM speeds than tighter timings. That's why they "optimized" their OCZ PC3700 Gold to run at high clock speeds, but not at low lantency settings. So even at lower speeds, you may not be able to use stricter timings.