16GB G.Skill RipjawsX 1600CL9 on AMD FX/990FX ( F3-12800CL9D-8GBXL x2 )

This quick test will be about 2 kits of very popular G.Skill F3-12800CL9D-8GBXL used on latest AMD FX/990FX platform.

Link to product site:
http://gskill.com/products.php?index=355
Yes these kits are designed for Intel P67/Z68.
Standard timings 9-9-9-24 2N 1.50V.

I will use G.Skill photo here as it looks quite nice, just think that you see 4 of these sticks with nice looking black PCB



Testing specification includes:
FX-8120 cpu that I'm trying to keep on stock clocks ( can add some overclocked results later but it won't change memory transfers )
Gigabyte GA-990FXA-UD5 F7a BIOS
Windows Server 2008R2 Standard SP1 installed on 2x Crucial M4 64GB RAID0 / AMD controller
WDC Black 500GB as 2nd drive ( don't think that's important )
Corsair HX850, custom water cooling based on EK waterblocks and 3x120mm heatspreader

stock / 1333 results

standard stock settings using Profile1 ( there is only 1 profile ) 1600 9-9-9-24 2N 1.5V



1333 CL7-7-7-22 1.5V

1600 CL8 / max clock@stock timings / 1866

1600 CL8-8-8-24 2N 1.5V , lowest stable timings @1600




1744 9-9-9-24 2N 1.5V , it's max that I was able to set using standard timings from XMP




1866 9-9-9-24 2N 1.5V , max clock with almost all timings from stock profile except tRFC that had to be 160 ( standard is 110 )
Anything above 1890 had problems with stability.



Now fast Q&A:
Why no tests using 8GB ? Because I was able to make the same results using 16GB.
Why no tests on looser timings ? Because I couldn't make higher clocks using other settings like 9-10-9. 10-10-10 couldn't boot at all. Maybe I will find out why and will update results later.
Why no results on overclocked cpu ? Because cpu clock won't affect memory bandwidth ( there are some changes in cache speed ). Difference is only while using higher CPU-NB that is getting less stable above 2400MHz.
Why no tests using higher voltage ? Because these IC don't really like anything above 1.6V and difference in clocks isn't worth to keep them on 1.6-1.65V.

very nice info. thanks for taking the time to test for us. I always appreciate good memory information.

we have very simular setups so i have to ask.

im somewhat a n00b at this so im having trouble making my 16gb 1600 memory run at 1600 on my amd board.

Gigabyte 990XA-UD3 AMD 990X Socket AM3+
AMD FX-4100 X4 3600MHZ 8MB 95W AM3+
Vengeance Series 16GB DDR3-1600 (PC3-12800) CL9 Dual Channel Desktop Memory Kit (Four 4GB Memory Modules)
16GB (4 x 4GB) DDR3 1600MHz Operating Speed; 9-9-9-24 Latency

what's the best way to set up this memory? im trying to get it to run at 1600 in the bios but when i change it, i get an error when restarting and it saids it cant continue so it changes back to 1333.

all i did was change the memroy speed from 1333mhz to 1600mhz. nothing else.

some help would be nice =)

thanks in advance.

It should be running fine with standard XMP settings. If it's not then try up to +0.1V CPU-NB or up to +0.1V memory voltage ( these boards have about 0.025-0.050V lower voltage than you set in BIOS ).

hmm.. ok ill try that.

Woomack said:

This quick test will be about 2 kits of very popular G.Skill F3-12800CL9D-8GBXL used on latest AMD FX/990FX platform.

Link to product site:
http://gskill.com/products.php?index=355
Yes these kits are designed for Intel P67/Z68.
Standard timings 9-9-9-24 2N 1.50V.

I will use G.Skill photo here as it looks quite nice, just think that you see 4 of these sticks with nice looking black PCB

Testing specification includes:
FX-8120 cpu that I'm trying to keep on stock clocks ( can add some overclocked results later but it won't change memory transfers )
Gigabyte GA-990FXA-UD5 F7a BIOS
Windows Server 2008R2 Standard SP1 installed on 2x Crucial M4 64GB RAID0 / AMD controller
WDC Black 500GB as 2nd drive ( don't think that's important )
Corsair HX850, custom water cooling based on EK waterblocks and 3x120mm heatspreader

Click to expand...

This RAM can do over 2.46V its a beast on stock cooling and with no mods

http://valid.canardpc.com/2854077
At 2.46V also look at the pic way better results

Results posted by me were on lower clocked CPU with totally different settings. This thread is not to beat anyone but to show other users what they can expect from this memory kit.
I have no idea what you want to prove with your 2.46V.
As I said in the other thread this memory was available on various IC so one can run on higher voltage but other won't scale above some value ( like memory that I had ).

Woomack said:

Results posted by me were on lower clocked CPU with totally different settings. This thread is not to beat anyone but to show other users what they can expect from this memory kit.
I have no idea what you want to prove with your 2.46V.
As I said in the other thread this memory was available on various IC so one can run on higher voltage but other won't scale above some value ( like memory that I had ).

Click to expand...

Your RAM still have HEAPS of OC freedom