Help overclocking an F3 5600+:

[BeepBeep2]

I've got it at 3000 Mhz OCCT 2 1/2 hr stable. (1.37v)

I upped the vcore to 1.5, kept multipler @ 14, set FSB to 230, and HT link to 4x resulting in roughly 3230 Mhz and it failed OCCT CPU within 10 minutes. (Temps hovering 56c max)

Max suicide was 1.5v @ 3280, did superPi 1.5 in 26.1 seconds. (Failed OCCT in 14 seconds.)

My RAM is fine, it still doesn't want to do 3200 half-stable with CAS 5. (DDR2-920)

System: 1 1/2 years old, the beast:
AMD 5600+ Windsor F3 @ 3Ghz Xigmatek HDT SD-964
2GB Super*Talent DDR2-800 (860) 4-4-3-8-17 TRC (two RAM banks screwed on my mobo...tried to RMA to XFX but I had to send it back and they said they would "contact me to let me know if it's approved"...I have no other PC, at least for now.)
(XFX) Foxconn 590 SLi
Seagate 7200.10 250GB
Maxtor 120GB 5400RPM, last of old XP install
OCZ 700w (FSP platform)

I'm thinking of reseating the heatsink, temps seem a little high.

I'd like to keep the voltage @ 1.45 or less, just need help getting it stable at higher frequencies.

[uwackme]

The XFX has a flaw in the design, just like the old DFI NF4 boards.... it uses a Riktek 9715 voltage regulator for the DDR2, which outputs huge spikes under heavy load.

This limits the memory OC and holds you back.

I'm able to get 3.5Ghz and 270Mhz on the memory bus with OCZ 1066 SLI 2x1G but the ram can do alot better and is held back by this crappy voltage regulator.

Switching to a Gigabyte 790GX board next week. Looking forward to no more 9715, between NF4 and this board, it's been 3 years lg with limitation. Was even worse on the BH5 in the DFI NF4. Someone with a storage scope had taken images on the NF4 to prove just how bad that regulator really is... its on here somewhere.

[cbjaust]

You're running 230 x 14 for ~3220MHz on the CPU and ~3220/7 for 460MHz (DDR800 divider) for the memory, yeah?

If that is so I would suggest that it could be the RAM that is holding you back.

Have you had the HT frequency higher than 230MHz?

Try 267 x 12, DDR667 divider, HT->K8 multi at 4x and see if you can push it up from there (this will give you 3200 CPU, 400 RAM and 1066 HT)

Also make sure you have vDIMM set correctly for the manufacturer's recomendation and try 4-4-4-12 2T 18 (TRC) for the RAM timings. What other RAM timing option do you have available in the BIOS?

Cheers

[BeepBeep2]

You're running 230 x 14 for ~3220MHz on the CPU and ~3220/7 for 460MHz (DDR800 divider) for the memory, yeah?

If that is so I would suggest that it could be the RAM that is holding you back.

Have you had the HT frequency higher than 230MHz?

Try 267 x 12, DDR667 divider, HT->K8 multi at 4x and see if you can push it up from there (this will give you 3200 CPU, 400 RAM and 1066 HT)

Also make sure you have vDIMM set correctly for the manufacturer's recomendation and try 4-4-4-12 2T 18 (TRC) for the RAM timings. What other RAM timing option do you have available in the BIOS?

Cheers

I set the ram to the DDR667 divider, CAS5...no go. It's not the ram.
(I'm not able to check the IC's on the RAM, heatsinks like nearly superglued to the PCB) There is half a chance they are Micron D9's, half chance Elipda.

The ram is awsome, does DDR2-800 4-4-3-8 @ 2.075v. vdimm standard is 2.2v

I'll try that multi/fsb combo and see.

[BeepBeep2]

My highest benches/screens:

[knopflerbruce]

Depending on the stepping, this may be right. If it's a CCBVF or CCB6F chip, then you got a bad one;) The CCB8F's are much better.

[BeepBeep2]

CCB8F, 0711, FPAW

which is why I'm frustated.

I've run memory @ 4-4-3-6 @ 860, and 13 tRC. 5-4-4-14 21 tRC @ 1000. It aint the memory.

[gOtVoltage]

CCB8F, 0711, FPAW

which is why I'm frustated.

I've run memory @ 4-4-3-6 @ 860, and 13 tRC. 5-4-4-14 21 tRC @ 1000. It aint the memory.


CPU Vcore may be too low,,
Also try Nbchipset @ 1.31v will help with keeping CPU stable,,Trust me.

MY X2's under 120 thermalRight . They both liked Volts..provoded you use a good Aircooler or WBsetup.


My 0711 TPMW 5600+

http://i168.photobucket.com/albums/u192/darrellsierra/ASUS3.jpg

MY 0720 XPMW 6000+

http://i168.photobucket.com/albums/u192/darrellsierra/3.jpg

[BeepBeep2]

Xigamatek SD-964 so yes, high-end air. (about 5c off the TRUE)

Ihmo, 1.55v was the spi run @ 3.28 CPUZ isn't reading volts right. How long can you run the HT Link that high? I heard to stay under 1200.

Anyway, looks as if I got a horrible batch or something.

[gOtVoltage]

Xigamatek SD-964 so yes, high-end air. (about 5c off the TRUE)

Ihmo, 1.55v was the spi run @ 3.28 CPUZ isn't reading volts right. How long can you run the HT Link that high? I heard to stay under 1200.

Anyway, looks as if I got a horrible batch or something.

It all depends what HT the Mobo likes really. I ran Both CPU's on a M2N 32 SLI deluxe ( HT speed for mobo was rated for 2000ht) I went up to 3000ht (1500ht)with no problems at all..

It does take some tweaking to get the HT up but in all the 5600+ and 6000+ are made to run . So if the Mobo can do 1500ht and the Chipset stayscool,, I JUST give what the Mobo wants freqency wise(HT). If it seems buggy then I dont run it. its that simple.

. With the X2 when running Higher HT you should notice A faster boot and overall responsiveness in everything else:D

May help ya tweak your Ram/ HT and FSB/HTT to optimal for your CPU multi..

AM2 Overclock Multi Chart.

http://i168.photobucket.com/albums/u192/darrellsierra/AM2multisCPU.jpg

[BeepBeep2]

Hi guys, thx for the help. 12x258.5 is at 3101 Mhz and 1.37v loaded 100% as I speak for 41 minutes. (Load temp is 56c)

If it makes it 2 hours I'm gonna leave it and call it stable. I might go a bit higher to try 3.2, as load temps at 2900Mhz were also 54c. Load doesnt seem to go up much but it's idling 41c right now. (No CnQ)

I'm also going to try the ram at 1034 5-5-5-15 and see if it's also stable. If not I'll revert to the 775 it's at now.

Oh, I think heat might be an issue...I'm gonna reseat the heatsink as the "Heat-Pipe Direct Touch" heatsinks are touchy.

[cbjaust]

Hi guys, thx for the help. 12x258.5 is at 3101 Mhz and 1.37v loaded 100% as I speak for 41 minutes. (Load temp is 56c)

If it makes it 2 hours I'm gonna leave it and call it stable. I might go a bit higher to try 3.2, as load temps at 2900Mhz were also 54c. Load doesnt seem to go up much but it's idling 41c right now. (No CnQ)

I'm also going to try the ram at 1034 5-5-5-15 and see if it's also stable. If not I'll revert to the 775 it's at now.

Oh, I think heat might be an issue...I'm gonna reseat the heatsink as the "Heat-Pipe Direct Touch" heatsinks are touchy.

So if your're running at 12 x 258.5 and your RAM is 775 (DDR) this means that you are on the "DDR667" RAM setting in the BIOS, yeah?

If you want you could try an intermediate RAM frequency which would be 886 (DDR), or 443MHz ('real') by putting the RAM setting back to "DDR800" in the BIOS and using any multiplier from 12.5 up to your max of 14 if half mutis are available:

1. 12.5 x 248
2. 13.0 x 239 recommended
3. 13.5 x 230
4. 14.0 x 222 recommended

Up to you as to what your HTT frequency end up as using either the x4 or x5 setting.

Also check the mem bandwidth with everest with each setting. My experience with my 6000+ shows that the half multipliers increase the latencies by a good 50% and drops the bandwidth in the order of 30%. See here (http://www.csd.dficlub.org/forum/showthread.php?p=61654). So check for yourself, however I would recommend you stick to the integer multipliers

Cheers

[BeepBeep2]

Alright guys, a month later (sorry for resurrecting this ancient thread) but

I'm stable at 3000.01 Mhz with 12x275 and 1.45v.

I had to keep the side panel off for stability testing because temps were hitting 60c with it on.