Ok, can anyone help me diagnose this problem with an overclock?

[RyanGreener]

Okay, I know i've made way too many threads asking for help, but the situation changes a bit and I think I'm pretty much at the "peak" of the overclock. Here is my current setup and settings that would most likely matter:

AMD Athlon 64 X2 5600+ Windsor 90nm (2.8 GHZ stock)
4x1GB DDR2 800 MHz
Seasonic 520W PSU
MSI 785GT-E63 ATX Motherboard

Memory is error free according to memtest86, CPU and motherboard and PSU are proven to work fine. Here are the overclock settings:

2.8 ---> 3.29 GHZ
CPU: 1.495V
x14 Multiplier
235 MHz FSB
100 MHz PCIe (Locked)
x4 HT Multiplier (~900 MHz HT)
FSB RAM Ratio of 667 MHz, 1.66 (RAM speeds are around 740 MHz according to BIOS)
Stock CPU-NB, NB, and RAM Voltages.
Stock RAM Timings.
Passed 30 Minute OCCT Linpack Test (Max 55 Deg C)
Passed 2 Hour 40 Minute OCCT Standard Test (which is good enough to me)
Windows 7 64 Bit

Now, when I change the settings to anything higher than 235 FSB, I get certain BSODs that I've been trying to figure out. One was "BAD POOL CALLER" and the other was "MEMORY MANAGEMENT".

The FSB limit is definitely not at 240, because I have done 270x12, and 240x13 before. I keep reading many OCing guides and nothing seems conclusive.

I tried raising CPU-NB voltage, NB voltage which also raised the HT Voltage, and nothing works. I only raised them though only within the "safe limits" according to the motherboard (white for safe, yellow for performance, then red for unsafe).

Should I further change the HT multiplier, or memory divider? I want to bump this processor up as much as I can. I don't want to raise the voltage anymore because 1.5V is pretty high (although this seems to be the safe limit for 65nm, and this processor is 90nm, I still want to keep it a BIT safe).

RAM timings don't seem to be able to be made higher, except CAS Latency but that only allows it to go up to 7, which says "DH_Only". I tried it and was forced to reset my CMOS because my computer wouldn't boot up

The NB temperature doesn't seem to get high. System temperature seems to be NB Temperature according to some MSI Mobo forums, and when I touch it I barely feel any heat at all, even when I raised the voltage, although the voltage was a slight bump (less than 1V).

Any help would be appreciated....

 

[bryanW1995]

you've either hit an fsb limit or a hole. if it's a hole then you'll be able to regain stability at a higher fsb, if it's a limit then you're screwed.

drop multi down around x6 or x7, then try to find the fsb limits. also, read the athlon x2 overclocking guide here (if you can find it), if not then google it. it's been a LONG time since I oc'd one of those little guys or I would offer more help.

 

[RyanGreener]

you've either hit an fsb limit or a hole. if it's a hole then you'll be able to regain stability at a higher fsb, if it's a limit then you're screwed.

drop multi down around x6 or x7, then try to find the fsb limits. also, read the athlon x2 overclocking guide here (if you can find it), if not then google it. it's been a LONG time since I oc'd one of those little guys or I would offer more help.

I had it at a x13 multiplier + 240 MHz before, so I definitely know that 240 MHz is not the limit. I haven't tested for higher limits though. I'll try to find it some other time, but for now, 240 MHz is the general highest I've gone so far. Do you know what those BSODs could imply? I'll search for a guide on OCing this specific chip I guess.

 

[Bartman39]

4x1GB DDR2 800 MHz

Sometimes this can really hinder or over time become unstable...? I would memtest each stick to make sure they are all still good... Very random crashes for the most part I`ve found (this is with a machine that has been OC`d for some time) the memory is the weakest link... Possibly a MB going bad as well...?

 

[RyanGreener]

Sometimes this can really hinder or over time become unstable...? I would memtest each stick to make sure they are all still good... Very random crashes for the most part I`ve found (this is with a machine that has been OC`d for some time) the memory is the weakest link... Possibly a MB going bad as well...?

Yeah, I already ran the memtest and it said everything was error free. The motherboard is new as well (maybe under 2 weeks old). The memory has never been "overclocked" or "overvolted" because I used a memory divider, etc.

So, should I try to "relax" the RAM timings or something? Here are the timings according to CPU-Z:

-------JEDEC #1, JEDEC #2, JEDEC #3
CAS Latency: 4.0, 5.0, 6.0
RAS TO CAS: 4, 5, 6
RAS Precharge: 4, 5, 6
tRAS: 12, 15, 18
tRC: 16, 20, 24

 

[RyanGreener]

Hmm, the RAM timing thing seems to be a "fail". I can't raise it higher than the settings I already posted. The only one I could raise was CAS Latency, but it said "7 (DH_Only)" and when I did it, my computer wouldn't boot. I reset the CMOS battery and now I'm back to square 1. I'm testing my FSB limit and so far, I'm at 270 and climbing. I really don't think the FSB limit is a problem here.

 

[dguy6789]

You are running an Athlon 64 X2 and you're pretty darn close to the reasonable limit of expectations for overclocking one of those. I'd say your CPU just can't be pushed any further.

 

[RyanGreener]

You are running an Athlon 64 X2 and you're pretty darn close to the reasonable limit of expectations for overclocking one of those. I'd say your CPU just can't be pushed any further.

Yeah, this appears to be the case. I just got more interested because I saw in an AMD OC database, some people were hitting 3.5 GHZ. I either have to raise the voltage more, or just quit now

 

[RavenSEAL]

Your CPU voltage is @ 1.495?...isn't that waayy too high?

 

[RyanGreener]

Your CPU voltage is @ 1.495?...isn't that waayy too high?

According to the AT sticky, for 45nm the safe limit is 1.4V, and for 65nm it is 1.5V. My processor is a 90nm, so it's safe to assume that it's higher than 1.5V. On other forums, most people push theres REALLY high (1.7V). But assuming a linear relationship, I'm guessing that it could take 1.6V safely...

 

[RavenSEAL]

According to the AT sticky, for 45nm the safe limit is 1.4V, and for 65nm it is 1.5V. My processor is a 90nm, so it's safe to assume that it's higher than 1.5V. On other forums, most people push theres REALLY high (1.7V). But assuming a linear relationship, I'm guessing that it could take 1.6V safely...

Oh gotcha. Didn't know 90nm could be taken higher than 45nm as far as voltage.

 

[RyanGreener]

Oh gotcha. Didn't know 90nm could be taken higher than 45nm as far as voltage.

Yeah, I don't know much about it all but I'm guessing the smaller the process, the more risk of electromigration, which is why voltages have to be lower on those chips.