How are these memory timings?

[orangecat]

For the past few hours I've been tweaking my memory settings to achieve maximum performance and this if where I've gotten to so far...

My current settings are 2400Mhz @ 10-12-11-12 command rate 1 with 1.725v and I doubled my memory refresh interval.

The original XMP profile was 1600Mhz @ 9-9-9-24 C2 with 1.5v...

I'm still working on it but that's what I've gotten to already.

[TheCookieGuru]

That's really good, I'm surprised your memory has not melted @ 1.725v 

[orangecat]

That's really good, I'm surprised your memory has not melted @ 1.725v

Most memory ic can easily handle 1.8 and some up to 1.9v.

[xTrekStorex]

holy.. that voltage. well thats a nice OC. when I think back to my last memory OC it was 1600 out of 1333 and I fried it after 4 days of usage

timings are still ok, though I wouldnt aim any higher if I were you

[orangecat]

holy.. that voltage. well thats a nice OC. when I think back to my last memory OC it was 1600 out of 1333 and I fried it after 4 days of usage

timings are still ok, though I wouldnt aim any higher if I were you

Yea, I'm having a really hard time hitting 2600mhz without bumping up the timings allot and I don't wanna push the voltage much higher just to be safe but so far I've gotten nearly 25% reduction in latency and almost 10gbps bandwidth increase.

[Valkyrie Lenneth]

Most memory ic can easily handle 1.8 and some up to 1.9v.

infact haswell can take up to 2.3v

[orangecat]

infact haswell can take up to 2.3v

I'm on ivy

[Valkyrie Lenneth]

For the past few hours I've been tweaking my memory settings to achieve maximum performance and this if where I've gotten to so far...

My current settings are 2400Mhz @ 10-12-11-12 command rate 1 with 1.725v and I doubled my memory refresh interval.

The original XMP profile was 1600Mhz @ 9-9-9-24 C2 with 1.5v...

I'm still working on it but that's what I've gotten to already.

double memory refresh interval is standard, depends on how hot the chips get, the hotter the less refresh interval... most sticks do 2-3x their standard interval tho when ur pushing 1.85+ they usually can only take double interval

 

cas 10 2400 isnt too good quite standard

 

aswell as that timing is not stable for sure 10-12-11-12 should be 10-12-11-22 at the least ( lower will be unstable ), it will pass memory benchmarks and such as prime95 blend and custom filling all ur ram (  u need to run it custom filling up all ur ram to 95-100% ) but the sticks will be unstable

[Valkyrie Lenneth]

I'm on ivy

ivy 1.8v safe 24/7  1.85 is pushing it

[Valkyrie Lenneth]

oh one more thing, never ever boot into windows without ensuring memtest86 is stable for atleast 10runs or so,  else u can pretty easily get OS corruption ( blew up 5-10windows partitions going crazy on my ram oc's , knewing that i would blow up some windows partitions )

 

aswell keep in mind that refresh rate cannot be seen in stability tests , u could run 50k refresh rate and still pass memtest and prime,  refresh rate is somthing u need to check overtime , keeping the pc on for more then 40hrs ect with programs on

[orangecat]

double memory refresh interval is standard, depends on how hot the chips get, the hotter the less refresh interval... most sticks do 2-3x their standard interval tho when ur pushing 1.85+ they usually can only take double interval

cas 10 2400 isnt too good quite standard

aswell as that timing is not stable for sure 10-12-11-12 should be 10-12-11-22 at the least ( lower will be unstable ), it will pass memory benchmarks and such as prime95 blend and custom filling all ur ram ( u need to run it custom filling up all ur ram to 95-100% ) but the sticks will be unstable

I've been running memtestx86 and it posts and finishes the tests just fine at 10-12-11-12, I hear online that the last value should be higher but so far I haven't had any stability problems with it at 12. I'll definitely keep your suggestion in mind.

[Valkyrie Lenneth]

I've been running memtestx86 and it posts and finishes the tests just fine at 10-12-11-12, I hear online that the last value should be higher but so far I haven't had any stability problems with it at 12. I'll definitely keep your suggestion in mind.

u will get stability problems even if it passes stress tests, guaranteed and u need atleast 10pass on memtest86 and 22hour of prime95 custom

 

crashes can occur even at 17hr of prime ect ( seen it myself before )  and on 19hr

 

btw dont forget to tighten cycle time or its pretty much a waste to tighten timings ;P

 

at 2400 cas 10 u should be around cycle time 270-280 lowest ( different per ram set and cpu )

 

at 2133 cas 9 its cycle time 230 min, 235 stable

 

personally i run at 2133 cas 9 optimized 2nd timings on haswell , its pretty much the sweet spot atm due to 100;133 fsb dram ratio ( 2400 is 100:100 )

 

2nd timings minimum clock = 4 for the chipset lvl never go lower than 4 clocks @ secondary timings, and that does not include write recovery time and recovery times, just read to read ect,   and FAWT(four act win time) = 16 minimum for chipset