E6300 stability issue?

[tempoct]

Hi,
My setup

Asus P5B Deluxe
E6300 on stock cooler
Corsair XMS2 6400C4 2GB

I have my setup stable (orthos blend stable for 8 hrs) on my E6300 @ 3.15GHz last night (FSB = 450, memory @ DDR2-900). However, same setup failed orthos after 20 secs tonight, weird?

Jumperfree
AI Tuning.......................Manual
CPU External Frequency..450
MEMORY...........................DDR2-900
PCI-E Frequency.............100
PCI Clock.......................33.33
Spread Spectrum.............Disabled
Memory Voltage...............2.10
CPU VCore Voltage.......... 1.35
FSB Termination Voltage...1.40
Northbridge VCore...........1.55
Southbridge VCore...........Auto
ICH Chipset Voltage.........Auto

CPU configure
Modify Ratio Support........x7
C1E Support....................Disabled
Max CPUID Value Limit.....Disabled
Vanderpool Technology.....Disabled
CPU TM function...............Disabled
Execute Disable Bit...........Disabled
PECI...............................Disabled

Chipset
Memory Remap Feature....Disabled
CAS latency (CL)...............5
RAS to CAS delay (tRCD)...5
RAS precharge (tRP)..........5
Activate to Precharge............15
Wrire Recovery time ........6
DRAM TRFC.....................42
DRAM Act to Act Delay TRRD....10
Rank Write to Read Delay.....10
Read to Precharge Delay......10
Write to Precharge Delay......10
Static Read Control..........Auto


Tonight, I try to stress test my RAM. Set it to DDR2-1000 @ 2.25v 5-5-5-15. Run Orthos 1.5 hrs Large FFT no problem but it failed the blended test less than a minute! What does this mean?

Jumperfree
AI Tuning.......................Manual
CPU External Frequency..400
MEMORY.........................DDR2-1000
PCI-E Frequency.............100
PCI Clock.......................33.33
Spread Spectrum.............Disabled
Memory Voltage...............2.25*
CPU VCore Voltage.......... 1.35
FSB Termination Voltage...1.40
Northbridge VCore...........1.55
Southbridge VCore...........Auto
ICH Chipset Voltage.........Auto

CPU configure
Modify Ratio Support........x6 (so, force 2.4GHz to test RAM)

What now??

 

[myocardia]

Orthos small FFT and large FFT are stressing primarily the processor (small FFT is solely the processor). Blend is ~50% processor, ~50% RAM. That means that your RAM isn't going to be able to run @DDR900, at least at the voltage and timings you're trying to run it at. Try running it at 5-6-6-18, and raising your ICH voltage by .1v.

 

[tempoct]

I thought Large FFT is for RAM and small FFT is for CPU??
I'll try to loose timing a bit...
Also, I found Orthos is not so reliable as the stability tester, I tried dual Prime95 and one of the instance failed in few mins for my setting. Now I'm lowering multiplier and increase Vdimm also ease the timing...
Let's see how it turns out..
C2D seems

Originally posted by: myocardia
Orthos small FFT and large FFT are stressing primarily the processor (small FFT is solely the processor). Blend is ~50% processor, ~50% RAM. That means that your RAM isn't going to be able to run @DDR900, at least at the voltage and timings you're trying to run it at. Try running it at 5-6-6-18, and raising your ICH voltage by .1v.

 

[myocardia]

Originally posted by: tempoct

I thought Large FFT is for RAM and small FFT is for CPU??
I'll try to loose timing a bit...
Also, I found Orthos is not so reliable as the stability tester, I tried dual Prime95 and one of the instance failed in few mins for my setting. Now I'm lowering multiplier and increase Vdimm also ease the timing...
Let's see how it turns out..
C2D seems

Well, like I was saying, small FFT's is processor and cache only. Large FFT's is processor and a little RAM, but not much RAM. Blend is roughly 50/50.

 

[tempoct]

ok,
So, maybe you can suggest me how to find out the max of my RAM then.
Go back to good old Memtest?