AX3U2000GB2G9-3G - A-Data DDR3-2000G 3x2GB Gaming Kit

[Woomack]

This thread is more like info for all who wish to get this kit than full review. I was browsing internet looking for any tests about it and I couldn't find even 1 sentence. It's easy to find good tests of A-Data Xtreme kits but nothing about Gaming series. I won't hide that I had some problems to set all. Sorry for my english and I have hope it won't be big problem

my pc config:
i7 920 D0 ( water cooled )
Asus P6X58D-E bios 0303 ( like I have any big choice , there are only 2 official bioses ) ( standard stock heatspreaders )
HD5850 1GB ( water cooled )
Pioneer DVDRW
2x PQI S525 32GB SSD in RAID0 for system +74GB WD Raptor for other data
Chieftec 750W PSU ( CFT-750-14C ) old but still working good
and of course A-Data DDR3-2000G 3x2GB Gaming Series kit 1.55-1.75V

At the beginning something that many overclockers wish to see:

PSC IC XEE218A3G-A as all can see on photo

There is only 1 XMP profile - 9-9-9-24 1.65V and it's setting QPI to 1.35V that is too low at least in my case
( not really important as better is to set all manually )

 

[Woomack]

Testing 7-9-7-x 1.65V

Some may ask why I started from 7-9-7-x and not stock 9-9-9-24 ... I just had problems to set all right and just couldnt pass 1900. What more I checked some other timings but still didnt try standard 9-9-9-24

1850 7-9-7-22 1T @ 1.65V



for now only memtest as I have no idea where I left other tests ... will have to rerun something later ... ( I'm making this thread being at work so it will take some hours )

 

[Woomack]

Testing 8-9-8-x 1.65V

2000 8-9-8-24 1T @ 1.65V





memtest ...



Setting 2000+ seems easy but in real it's not ... XMP profile isn't working and most auto settings will result in crash/no boot at all ... setting all for this test took me some hours and I already thought that something is wrong with my board ...

for 2000+ I need 1.55V QPI/VTT to make it stable

more tests later

 

[Woomack]

Testing 6-8-6-x 1.65V

Not much time yesterday so something fast... 1600 6-8-6-20 1T @ 1.65V



With some more time I'm almost sure that 1600 6-7-6-18 1T @ 1.65V is possible.

I bet I could make it better but my cpu IMC is weak and it's hard to pass 4100 uncore keeping it stable with 2000+ memory. I will try to get better cpu for 2000+ tests.

 

[dejo]

I have the x2000 in a 4g kit and have about the same experience. for what you pay it seems a pretty good value. there is better memory but you will pay quite a bit more for it.

 

[Woomack]

Xtreme series made in the end of last year were on Elpida chips, now I don't know. There are only 2x2GB Xtreme kits from A-Data now ( at least can't find any available and some shops already removed 6GB 2000 X from list ).
I'm 101.3% sure that I can make more but I didn't find optimal timings yet. I saw many people complaining about problems with other 2000+ kits on P6X58D so maybe it's board issue, hard to say.

I have the x2000 in a 4g kit and have about the same experience. for what you pay it seems a pretty good value. there is better memory but you will pay quite a bit more for it.

I was thinking about OCZ 1600 6-8-6-x or some others but at the end I got info that all kits that I was checking are on almost the same PSC chips including Mushkin Redline 1600 6-8-6-x kits that cost 50% more ( in Poland ) than this A-Data 2000-G. So most will work on about 2000 cl8-10-8-x or something like 9-11-9-x ... big chance that higher but you never know what you get.
I didn't know what chips are in these A-Data 2000G and there was noone who could give me this info as there is nothing in the net or in any other source, no reviews or tests except X series.
I have hope that my tests will help some people make their choice.

46 views and 1 comment. I won't hide that I was expecting some more

 

[dejo]

many dont care what we care about. most just want to know if it will run at advertised specs, and will it just drop in and work. all in all I have been very happy with my 4gigs of it and for 120 american dollars the value is there.

 

[Woomack]

What is funny I still can't make it run on declared 2000 cl9-9-9-24 up to 1.75V, timing issue or something ( I didn't check it much ) but no problem to make 8-9-8-24 1.65V

 

[dejo]

I have some kingston that wont run a -20 timing to save its life, but will run all day long at 7-9-7-18 up to 1900+ and have had it at 8-8-8-18 at 2150 but have to really throw the volts at it for that.

 

[Woomack]

Testing 7-9-7-22 1.74V

7-9-7-22 1T 1.74V

 

[Woomack]

Stable 7-9-7-25 1T 1.65V

Some weeks more to make it really stable. Maybe someone saw my topic about 2000+ kits and P6X58D-E problems. Some timings have to be set manually on these boards exacly as memory specification and never use XMP as board is setting many things wrong. Maybe asus will give new bios one day, so far there are 2 official and both work the same. Their support is really weak.
There is some issue that have to give much higher QPI/VTT voltage than on for example Gigabyte UD4 to make it stable and the same are saying other people testing 2000+ ram with P6X58D-E and sometimes on RE3 too.

Anyway here is screenshot 2000 7-9-7-25 1T 1.65V finally stable.



Still crashing on 9-9-9-24 but I think it can be some timing synchronization problem for PSC IC. All other brands make timings with +1/2 more tRCD that is working good. A-Data should fix SPD/XMP settings because both aren't working good for these chips. G.Skill give some info how to set their kits on Asus boards. Probably noone else did the same and don't expect any help from Asus.

 

[Woomack]

~2250 7-10-7-25 1T 1.74V, max benchable for today

as in title ~2250 7-10-7-25 1T 1.74V as max clock for now and I have hope that will be better just don't know if on this cpu ...