PHENOM II AM3/DDR3 Ram Overclocking List

[LightSpeed]

Guys, need some help please. persuaded a friend to go pII instead of the E8400 he was getting, now need information

check out the following RAM modules and let me know which is most likely to do 1600mhz 9-9-9-x stable (or better) on a M4A78-T with a Phenom II X3 720

this, this, this or this

many thanks, and sorry if i posted in the wrong thread, il gladly repost somewhere else then and mods can delete this

edit: wrote the wrong mobo model

[xPliziT]

hey light speed

lol check out the modules i got

patriot spec 2000MHz 9 9 9 24

running them at 7 7 7 24 1600 MHz 2.0 Volt

ths should solve it for you

[chew*]

Guys, need some help please. persuaded a friend to go pII instead of the E8400 he was getting, now need information

check out the following RAM modules and let me know which is most likely to do 1600mhz 9-9-9-x stable (or better) on a M3A78-T with a Phenom II X3 720

this, this, this or this

many thanks, and sorry if i posted in the wrong thread, il gladly repost somewhere else then and mods can delete this

I'm partial to the golds I always get lucky with them, however theres really now way to know unless someone has that exact setup that can give a definite yes or no.

[Tony]

I think the voltage you can apply depends more on PCB used......the higher rated voltage D9's most likely use a heavier duty PCB to handle the voltage rating.

Its nothing to to do with the PCB, its the IC's themselves that are actually rated down as low as 1.3V in some cases.
We are increasingly using more LV IC's on kits now and they all seem to max out around 1.65V. I actually think this is the byproduct of Intel demanding 1.65V max on i7..(although we all know you can run higher)

Animal series is cas9 micron 2000 re badged as 1333 or 1600 to give you guys something cheap that clocks. The 1600Plat new AMD special kit 7-7-7- rated is based off similar IC's.

The main issues are the limit the IMC seems to have around 1800MHZ...1333 6-6-6-24 is as fast as 7-7-7-24 1600, what you would think would be a fast combo to run often is not...its like the cpu has tRD (performance level) and it loosens by auto as you clock up the HTT or in some cases change the CPU multi.

You can look for this as you clock up HTT, or change multi and test with everest. Sometimes clocking down the multi gets the ram working faster...very weird.

Anyway...try different timings with different speeds, don't limit the tests to 6-6-6 or 8-8-8- etc, try everything and see what you find.

[G.Foyle]

I have had better results in spi16M at 1666 7-6-5-16 1T than at 6-6-5-16 1T
Also NB is tougher to keep stable at CL6, I would think less stable setting is more stressing and faster. I don't know why would CL7 be faster.

[shaolin95]

The only thing I could think of is memory errors....other than that there shouldnt be any reason for a higher cas to be better than a lower cas if all other settings remain the same.

[chew*]

The only thing I could think of is memory errors....other than that there shouldnt be any reason for a higher cas to be better than a lower cas if all other settings remain the same.

i've been using 5-5-5 timings in a alot of things on the bot and time and time again I've beat those at 4-4-4 , fact is depending on what clockspeeds your at and what NB speeds seem to make a diff........

It almost like when your at High clocks memory all of a sudden becomes a bottleneck and the low 4-4-4 latency no longer matters.... with ddr II even though testing on air at lower clocks yielded better results....

Really think that @ certain memory speeds its almost like the IMC changes latency.......

[shaolin95]

. Sometimes clocking down the multi gets the ram working faster...very weird.

Funny that you mention this because at 3.7Ghz my ram was giving me about 11,370 according to memtest but when I dropped to 3.6Ghz I actually got 11760!

[G.Foyle]

I also noticed spi16M is noticeably faster after a failed boot, i.e. when I dial 10 GHz in bios, it can't boot, restart, press f1 after a failed boot attempt, dial your clocks back, hit f10 & enter, and it's faster than if I did a regular reboot. It's also faster on first boot after changing timings in BIOS.

[BrowncoatGR]

OCZ3G1600LV2G 1600mhz 7-7-7-24-36-2T 1.75v Phenom 2 X3 720 @ 3.3Ghz NB 2.2Ghz

[Oese]

Oese | 800Mhz | 1.77v | 7-6-6-18 1T | x4 955 @ 3.7g | 2x2GB Crucial Value DDR3-1066 | D9JNL | DFI LP DK 790FXB-M3H5 | 790FX |

[shaolin95]

Its nothing to to do with the PCB, its the IC's themselves that are actually rated down as low as 1.3V in some cases.
We are increasingly using more LV IC's on kits now and they all seem to max out around 1.65V. I actually think this is the byproduct of Intel demanding 1.65V max on i7..(although we all know you can run higher)

Animal series is cas9 micron 2000 re badged as 1333 or 1600 to give you guys something cheap that clocks. The 1600Plat new AMD special kit 7-7-7- rated is based off similar IC's.

The main issues are the limit the IMC seems to have around 1800MHZ...1333 6-6-6-24 is as fast as 7-7-7-24 1600, what you would think would be a fast combo to run often is not...its like the cpu has tRD (performance level) and it loosens by auto as you clock up the HTT or in some cases change the CPU multi.

You can look for this as you clock up HTT, or change multi and test with everest. Sometimes clocking down the multi gets the ram working faster...very weird.

Anyway...try different timings with different speeds, don't limit the tests to 6-6-6 or 8-8-8- etc, try everything and see what you find.

So Tony, I did some tests with my OCZ Ram and according to the Memtest bandwith, I get the same bandwidth with either 1600 7-7-7-21, 1333 6-6-6-18 or 1066 5-5-5-15.
So which is a better choice in your opinion or is it irrelevant? In other words, does my CPU has to work harder thus more limiting to OCing when running 1600 or is it harder to run tighter timings even at 1066?
Thanks

[chew*]

So Tony, I did some tests with my OCZ Ram and according to the Memtest bandwith, I get the same bandwidth with either 1600 7-7-7-21, 1333 6-6-6-18 or 1066 5-5-5-15.
So which is a better choice in your opinion or is it irrelevant? In other words, does my CPU has to work harder thus more limiting to OCing when running 1600 or is it harder to run tighter timings even at 1066?
Thanks

6-6-6 is going to be faster, jury is still out on 5-5-5 don't have enough comparisons yet....that and my ram needs super octane to run 5-5-5

[shaolin95]

Nice...I need to run some tests myself of course...I just couldnt do it today except for testing stability and I can do either using 1.65v which is very nice. :-)
I will report back later with my results but I guess it would be a matter of finding the right balance.
Do you think everest latency test has any importance as in having lower latency becoming a priority?

[chew*]

results tuned for everest are good for everest, thats about all

[demonkevy666]

question...
is the Am3 temperature senors any better ?

[namegt]

955BE / ASUS M4A79T Deluxe 32 PI

Patriot Viper DDR3 PC3-16000 2X2G
955BE@3584MHz, NB@2912MHz / 1792MHz 7-7-7-21 1T 2.0v 32MB PI




BIOS SET





1840MHz 8-8-8-20 1T 2.0v

[Oese]

any idea which chips?

[namegt]

any idea which chips?

I do not know what chip ...
I want to know.

[G.Foyle]

Finally got CL6 stable:



xoqolatl | 833MHz | 6-6-5-15 1T | 2,30V | 2x1GB Corsair Dominator DDR3-1800 | D9??? | ASUS M4A79T Deluxe | 790FX | Screen

I had to blur the time - it's a surprise to one guy I am having a 16M low clock challenge with.
I can do 32M at that speed and timings but I have misplaced the screenshot somewhere.

Hey chew, what was the tip to keep my IMC happy again?

[Oese]

nice but HIGH voltage xD

[G.Foyle]

The can take it all day long, I was benching them like this for four days and everything's OK, no degradation, no problems.

[shaolin95]

Its impressive but I dont expect them to take it for too long though.

[Oese]

i wonder what voltage level is safe all the time. heat dissipation is lower, but structure size and durability too i think? someone info on this?

[G.Foyle]

Saaya recommended max 2.1V for Micron D9 DDR3. I don't really care about them since they were abused by the previous owner and were supposed to be dead or severely degraded. Seems they run fine after all