PHENOM II AM3/DDR3 Ram Overclocking List

[chew*]

is 1840 the highest anyone has hit on a AM3 rig?

Its the highest unstable screenshot someone has taken, get back to us when it runs at usable 100% benchable settings, thus the request originally for 32mPi by the OP.

[cdawall]

Its the highest unstable screenshot someone has taken, get back to us when it runs at usable 100% benchable settings, thus the request originally for 32mPi by the OP.

i'm going to try loosening it up to CL7-7-7-21 2.4v and see if it can run through 32m. do you have any insite on if anyone has topped even unstably 1840?


oh and with 1840 i could use firefox and get through 3d2k1

[chew*]

where the screenshots because according to the thread you made before 1m was maxxed at 1800............and 1m is far from 3d 2001 stable.....

Not trying to give you a hard time man but if I reviewed a set of ram and showed people screenshots like that and 1m only I would not only get laughed at but I would most likely never ever ever get a sample to review again

Topping 1840 is useless.....if it can't be used, if it floats your boat, however msimax hit 950 ( 1900 ) unstable

Let me give you an example of why showing stability is important......people buy ram that they see clocks high......based on this thread....they see you posting its stable.......now somone like me would buy like 6 sets......Now if I spent all that loot and could not run 2001 at 1840 6-6-6 guess who I'm coming after.......the guy posting false info....and if he has a sponsor, I'm sending them an email that someone that represents them is misrepresenting said product......

One thing if the manufacturer sends you an uber cherry sample to review......then the problem falls on them......however when thats not the case the problem is fixed in those cases reviewers get the axe...........I know your not reviewing for somone but.......if your goals are to get to that stage....your doing it the wrong way.

[msimax]

2.0volts using d9gtr



By deeodee

[msimax]

[msimax]

1.8volt




supertalents

[msimax]

As you can see i was able to run higher freq. to post 1m runs but they were not stable with 32m runs

[cdawall]

where the screenshots because according to the thread you made before 1m was maxxed at 1800............and 1m is far from 3d 2001 stable.....

Not trying to give you a hard time man but if I reviewed a set of ram and showed people screenshots like that and 1m only I would not only get laughed at but I would most likely never ever ever get a sample to review again

Topping 1840 is useless.....if it can't be used, if it floats your boat, however msimax hit 950 ( 1900 ) unstable

Let me give you an example of why showing stability is important......people buy ram that they see clocks high......based on this thread....they see you posting its stable.......now somone like me would buy like 6 sets......Now if I spent all that loot and could not run 2001 at 1840 6-6-6 guess who I'm coming after.......the guy posting false info....and if he has a sponsor, I'm sending them an email that someone that represents them is misrepresenting said product......

One thing if the manufacturer sends you an uber cherry sample to review......then the problem falls on them......however when thats not the case the problem is fixed in those cases reviewers get the axe...........I know your not reviewing for somone but.......if your goals are to get to that stage....your doing it the wrong way.

i have managed 3D2K1 with 1840 CL 6-7-6-19 1T 2.44v i did not however get a SS with the memory tab open (i have 2 cpuz's with the cpu page )


i will rerun the benchmarks when i get more DICE which may be as soon as tomorrow after school tomorrow.


i'm sorry if i came off as a y ass i was just a little excited i managed to break 1800 with CL6 even if its not the picture of stability and required DICE.

oh and just cause i can say this people can buy as many sticks of these as they want and they wont clock like this cause the new ones use different IC's



@MSIMAX

very nice runs on your ram what other settings did it take to hit 1900 even if it was unstable.

and how come your read/write/copy is so much higher than mine just the NB clock?

[msimax]

msimax | 766.7Mhz | x4 955 3.795mhz | 5-5-5-15 1T | 1.9v | 2x1GB kingston hyperx DR3-2000 CL8 | ???? | GB GA-MA790FXT-UD5P | 790FX | link

I'll Be BAck

[Oese]

thats nice result^^

people you're absolutely right that unstable overclock is not to be compared to 32m stable. Thats why i added two stars to mark it, but maybe there should be an "unstable" or "cpuz-stable" list or not one such result as all...

to hear your opinion^^

[chew*]

Well If we are creating a database for users to find what they can expect to get for stable clocks out of a particular IC Oese........

It would make sense to create another database for just max clocks as atm I bet most people aren't reading the fine print and seeing the asterisks....

This could lead to someone purchasing ram that is indeed not capable of such speeds stable.......and being rather upset for wasting money versus if the database was "clean" it would be easy to pick from.

[Tomasis]

Well If we are creating a database for users to find what they can expect to get for stable clocks out of a particular IC Oese........

It would make sense to create another database for just max clocks as atm I bet most people aren't reading the fine print and seeing the asterisks....

This could lead to someone purchasing ram that is indeed not capable of such speeds stable.......and being rather upset for wasting money versus if the database was "clean" it would be easy to pick from.

i agree .. for both ddr2 and ddr3

[msimax]

i have managed 3D2K1 with 1840 CL 6-7-6-19 1T 2.44v i did not however get a SS with the memory tab open (i have 2 cpuz's with the cpu page )


i will rerun the benchmarks when i get more DICE which may be as soon as tomorrow after school tomorrow.


i'm sorry if i came off as a y ass i was just a little excited i managed to break 1800 with CL6 even if its not the picture of stability and required DICE.

oh and just cause i can say this people can buy as many sticks of these as they want and they wont clock like this cause the new ones use different IC's



@MSIMAX

very nice runs on your ram what other settings did it take to hit 1900 even if it was unstable.

and how come your read/write/copy is so much higher than mine just the NB clock?

i notice if your unstable or your mem throws alot of errors it will skew with the results,the test maybe fighting with it to go through this is the case with sandra too

[msimax]

msimax |888.0Mhz | x4 955 4994mhz | 6-6-6-25 1T | 1.9v | 2x1GB kingston hyperx DR3-2000 CL8 | elpida hyper | GB GA-MA790FXT-UD5P | 790FX | link

[pawl]

is anyone playing with ddr3 drive strengths and termination values?

from what i can see most bios's dont have these values exactly correct yet. so there may be some room for improvement.

the cpu termination should be 40ohms for 2 or 3 sticks of memory and the drive strengths should be 1.0 for =>1333 memclk. ( chew? )

and im not talking about the termination value in the DCT0/1 im talking about the one on the cpu itself. there are also some termination timing parameters that would play a part.

i coul be wrong and independant R&D by the manufacturers may have led to them using alternative values. but this is certainly worth exploring as we search for higer frequency ddr3 on the am3 platform.

[chew*]

I would not follow those to the T......they are more of a guideline.....not all cpu's are equal nor is every chips IMC indentical nor is the same model ram equal. Tune them to your particular system.

I refrain from posting my AOD memory settings exactly for this reason........We put the same model/speed ram, on the same board, with the same bios and the same model cpu and and loaded my bios profile..... guess what? BSOD.......

Not only that But we also bench competitively ( msimax and myself ) So having showed him the ropes I requested he not show my secrets......Not all viewers are on team XS. People are unfortunately going to have to sit down and tune for hours, or settle for a lower speed......not profit off my years of experience by copying my maxxed out bench settings....

For a prime stable speed advanced tuning is not necessary and or helpfull and I have posted up bios shots quite a few times in the motherboard threads.......this I am glad to help with......

[gOtVoltage]

I would not follow those to the T......they are more of a guideline.....not all cpu's are equal nor is every chips IMC indentical nor is the same model ram equal. Tune them to your particular system.

I refrain from posting my AOD memory settings exactly for this reason........We put the same model/speed ram, on the same board, with the same bios and the same model cpu and and loaded my bios profile..... guess what? BSOD.......

Not only that But we also bench competitively ( msimax and myself ) So having showed him the ropes I requested he not show my secrets......Not all viewers are on team XS. People are unfortunately going to have to sit down and tune for hours, or settle for a lower speed......not profit off my years of experience by copying my maxxed out bench settings....

For a prime stable speed advanced tuning is not necessary and or helpfull and I have posted up bios shots quite a few times in the motherboard threads.......this I am glad to help with......

I agree on this note as well.

Again nice work guys,
plenty of great info for all who take the time to read the * and experiment with their Bios one setting at a time

[msimax]

msimax |852.0Mhz | x4 955 3834mhz | 8-8-8-25 1T | 1.9v | 2x1GB kingston hyperx DR3-2000 CL8 | elpida hyper | GB GA-MA790FXT-UD5P | 790FX link

[cdawall]

does this count chew*?

cdawall |900Mhz | Phenom X4 955@3600mhz | 8-8-8-24 1T | 1.76v | 2x1GB Wintec AMPX 1333 CL9 | Samsung HCH9 | Asus M4A78T-E | 790GX LINK

[chew*]

Not really.........why beat around the bush man......You seem to refuse to run 32m.......that belongs in the 3dmark thread.....

I will quote titon.......I could run 1m I could run 3d but when I went to run 32m PI it would just lock my system at alot higher ram speeds that I thought were stable.....

Be it that it is it's a small competition.......but more importantly its a database for end users to find what ram they may choose to use that is Stable......I don't approve of misleading endusers.....I saw your post at OC forums and You acted like you were 100% stable.......and had people believing as much and did not say anything to the contrary.....

Let me put it easier for you..........do you think if you submitted that to HW bot under 32m it would not get rejected?

If the time it takes to run it is to long for you, hit calculate and go run and errand like I do..............

[cdawall]

Not really.........why beat around the bush man......You seem to refuse to run 32m.......that belongs in the 3dmark thread.....

I will quote titon.......I could run 1m I could run 3d but when I went to run 32m PI it would just lock my system at alot higher ram speeds that I thought were stable.....

Be it that it is it's a small competition.......but more importantly its a database for end users to find what ram they may choose to use that is Stable......I don't approve of misleading endusers.....I saw your post at OC forums and You acted like you were 100% stable.......and had people believing as much and did not say anything to the contrary.....

Let me put it easier for you..........do you think if you submitted that to HW bot under 32m it would not get rejected?

If the time it takes to run it is to long for you, hit calculate and go run and errand like I do..............

you used the same test in a couple of your runs and mentioned 3D stability in some of your posts towards my CL6 runs

your's seems to be valid on the front page still and seems to say 3Dmark not superpi
http://www.xtremesystems.org/forums/...52&postcount=9

i just wanted to see 3D stability as you said it was useless to you if you couldn't run 3D? yet i run 3D and it doesn't count...i have not had a chance to run 32m yet i am still playing around with settings these sticks are all of 1 hour old and i just got them 3D stable at 1800 CL8 1.76v i am proud of that personally and i found an alternative IMO to hyper's and these are a hell of a lot cheaper (my 2x1GB kit was $37 shipped)


as for OCF i never said it was stable and other people pointed out it was not stable hell the sticks were under DICE that will never be 100% truly stable as you can't exactly sit and throw more DICE on your sticks every 10 minutes.

if it makes you feel any better its been running 32m since i posted 3D06 i just wanted something up here and on TPU showing the thing was at least 3D

[chew*]

They can delete that for all I care man........One out of how many 32m submissions by me..........Lets count your's, 1 32M submission out of how many?

Your going to ruin your credibility man.......Personally credibility means everything to me........unfortunately it means all to others apparently......

If you want to mislead people go ahead but I won't tolerate it on this forum.....and will call you on it everytime......just like the little white lie I caught you in via a PM to me in contrast to a post you made.......how convenient you lost track of 3 weeks........

[cdawall]

They can delete that for all I care man........One out of how many 32m submissions by me..........Lets count your's, 1 32M submission out of how many?

Your going to ruin your credibility man.......Personally credibility means everything to me........unfortunately it means all to others apparently......

1 out of 5? i ran the stuff at very high volts that is known to degrade the chips i never posted them to say hey i have the stable ram clock i posted and said hey look what clock i hit. it takes 20+min to run super pi 32m i have school and a 40hr a week job i honestly don't have time to sit and let 32m run.

[chew*]

So your posting stuff in a 32m ram stability thread to say hey look what clock I hit........Seems you should be posting in bandwidth.......

Spare me the tears about degrading.....I'm using D9 to and obviously high volts yet I can still meet the requirements of the OP..........

End users come first my 2c. I'll sacrifice a few mhz no problem to show them a stable OC.......

[cdawall]

So your posting stuff in a 32m ram stability thread to say hey look what clock I hit........Seems you should be posting in bandwidth.......

Spare me the tears about degrading.....I'm using D9 to and obviously high volts yet I can still meet the requirements of the OP..........

End users come first my 2c. I'll sacrifice a few mhz no problem to show them a stable OC.......

you were right took 1.9v to hit 32m stability



cdawall |900Mhz | Phenom X4 955@3600mhz | 8-8-8-24 1T | 1.9v | 2x1GB Wintec AMPX 1333 CL9 | Samsung HCH9 | Asus M4A78T-E | 790GX LINK

after seeing that looks like 3D is a no go like you said will for sure post 32m from now on