memtest86 is not unreliable. it's meant to test purely on ram.

you can pass memtest86 but fail to get into windows because other factor counts... like cpu stability and other components as well

tccd won't do 200htt @ 1T, sounds weird... is there a possible dat it was caused by mobo?

i wished i can have those 2*256 tccd module, juz wanna push d cpu + ram to d max...

i started to doubt d reliability of memtest coz my venice run fine @ 2.7 with my bh-5 (133 divider, error free), but w/ my fc, no matter how high d vdimm (1:1, error free) still it failed in super pi 32M...

gonna dig out d windows based memtest ledi. update u guys if i find sumthg new...

thx for all d input guys, esp raymond n master soulfly...

This post has been edited by ah_khoo: Sep 23 2005, 08:47 AM

i dunno, but mine can't do 2-2-2-5 @ 200mhz...

is okay using default spd, now 4 stick go for 250mhz, wif the most relax setting

Yes, there are some TCCD that cannot do 2-2-2-x 200mhz @1T, because TCCD is actually binned to run at 500mhz

same case with TCC5, which is actually binned at 466mhz, but results showing that more/most TCC5 can do 2-2-2-x 200mhz compare to TCCD

i'm doing a simple test...
and found out tat, tccd perform well @ 2-3-3-X
and may go 220 for tat speed...

now running 250...
still in testing progress...

my fc onli like 2.5-4-4-x, even 2.5-4-3-x will give thousands of errors, regardless 1t or 2t for htt > 275, never try lower thou...

i've tried d windows based memtest juznow, 2.5-4-4-8 @ 300HTT (2T) seems fine, but i need to feed 2.96V for d ram, any lower will give some errors.

so d method i've found to burn in ram is:

1. run dos based memtest til' nomore error in all test.

2. run sandra burn in wizard (mem. bandwidth) around 100-200 pass

3. test stability w/ windows based memtest.

how bout u guys?

This post has been edited by ah_khoo: Sep 23 2005, 08:09 PM

1000 round of test5 for memtest...
then only 3 days full test for memory

I run memtest with test #5 until it's fully stable without errors for hours.

Then I start playing with vcore, vchipset and vagp if it's not stable in windows. The method works quite well.

Actually TCCD is the most flexible chip ever because it can support low latencies at 200Mhz 1T and super high MHz at looser latencies.

It is quite rare that you will find a TCCD that can't do 2-2-2 at 200Mhz. If you can't do it, it's probably more of a mobo / CPU problem rather than the RAM itself.

Most TCCDs, be it DDR600/DDR550/DDR400, should be SPD programmed at 200MHz 2-2-2.

But I can't confirm this with TCC5 as they are slightly different than TCCD.

Also, try swapping the modules between the slots because they are quite picky.

TCC5 is actually the same chip of TCCD ... which is the Rev.F

It's just binned at lower speed to reduce manufacturing cost.... but technically, they're the same thing.

now I'm currently doing 300mhz 2.5-4-3-7-1T with only 2.7Vdimm after some fine tuning on the extra memory settings on DFI... but not totally perfect, got 6 errors for 200loops.

EDIT : with Idle Cycle Clock set to 64 Cycles, currently passing 300loops without any error

This post has been edited by soulfly: Sep 25 2005, 01:51 AM

Care to share ur genie bios settings

system:
AMD Sempron 64 2800+ "Rev. E6" LBBLE 0527EPAW
DFI LP UT nF3 250Gb (Bios 5/4)
1x 512MB OCZ EL PC-3200 Platinum Rev.2 1.1 (TCC5)

Genie BIOS Settings:

FSB Bus Frequency - 300
AGP Bus Frequency - 66mhz
Clock Spread Spectrum - Disabled
LDT Downstream Width - Auto
LDT Upstream Width - Auto
LDT/FSB Frequency - Auto
CPU/FSB Frequency Ratio - Auto


CPU VID Control - Auto
CPU VID Special Control - Auto
Chip Set Voltage Control - 1.60v
AGP Voltage Control - 1.50v
DRAM Voltage Control - 2.70v

DRAM Configuration Settings:

DRAM Frequency Set - 200 (DRAM/FSB:1/10)
Command Per Clock (CPC) - Enable
CAS Latency Control (Tcl) - 2.5
RAS# to CAS# delay (Trcd) - 04 Bus Clocks
Min RAS# active time (Tras) - 07 Bus Clocks
Row precharge time (Trp) - 03 Bus Clocks
Row Cycle time (Trc) - 07 Bus Clocks
Row refresh cyc time (Trfc) - 14 Bus Clocks
Row to Row delay (Trrd) - 02 Bus Clocks
Write recovery time (Twr) - 02 Bus Clocks
Write to Read delay (Twtr) - 02 Bus Clocks
Read to Write delay (Trwt) - 02 Bus Clocks
Refresh Period (Tref) - 4708 Cycles
Write CAS Latency (Twcl) - 1
DRAM Bank Interleave - Enabled



DQS Skew Control - Increase Skew
DQS Skew Value - 64
DRAM Drive Strength - Level 1
Max Async Latency - 07.0 Nano Seconds
Read Preamble Time - 05.0 Nano Seconds
IdleCycle Limit - 64 Cycles
Dynamic Counter - Disable
R/W Queue Bypass - 16x
Bypass Max - 07 x
32 Byte Granularity - Disable(8 Bursts)



some screenies .....


This post has been edited by soulfly: Sep 25 2005, 12:58 PM

an example of TCCD RAM which cant do 2-2-2-X @ DDR 400

Thanks for the info din.
Yup, a number of Gskill modules will not do 2-2-2 at DDR400 but I've had better luck with TCCD from different manufacturers.

now trying to run with divider... but unluckily the setting that i used for 1:1 is not applicable

actually i'm trying to run at 9/10 (300mhz ram, 338mhz htt) since i know my processor is capable of running at 2.7ghz stable. but seems like the ram need different memory setting. 2.8vdimm also no luck... maybe need 2.9vdimm.

EDITED: result : success

I played around with :
Max Async Latency
Read Preamble Time
IdleCycle Limit

These values should be affected/connected to divider and htt. Then I bump the voltage to 2.8V..... now can get more than 3 hrs memtest stable... and hopefully it's really stable for SuperPi 32M and prime

will post the screenie and setting afterwards after some more testing.

This post has been edited by soulfly: Sep 26 2005, 08:23 PM

bro din, but my problem is, set to 260, the ram will freeze my pc, @ 2.5-4-4-8.
so, possible to rma it?

RMA is only possible if your RAM does not meet its default SPD rating.

but problems is it freeze my pc when running 250mhz...

try use memtest86 to check whether it's your ram problem or your setting problem.