PDA

View Full Version : MSI K8N Neo2 Platinum Chipset (Taiwan vs. Korea)



conrad.maranan
11-17-2004, 04:55 PM
On Monday, I requested to RMA my board with mWave. This morning, I headed out to their facility to make the exchange for a new board with high hopes of being able to attain more satisfactory clocks for my memory.

Blah, blah, blah. Some more blah, blah, blah. Five minutes later, the CSR returns to the front counter with four boxes. He gives me his blessing to inspect each unit, pull off the NB HSF, and make a decision as to which one I would like to keep. The first two HSFs I pulled off revealed the same exact code as the one I was returning (0440A2 TAIWAN). The third unit displayed 0439A2 TAIWAN under the HSF. At this point, I'm loosing hope because I have already subscibed to the belief that a KOREA chip will outperform a TAIWAN chip despite the fact that there is no solid proof regarding this matter. Box #4 is a keeper coded with 0442A2 KOREA on the chip. I drive back to La Habra for one hour of testing before I head off to work.

0440A2 TAIWAN
Highest Memory Frequency: 270MHz @ 2.5-3-3-10
Memory Modules Tested: 2x256MB G.Skill TCCD @ 2.8V
Stability - Over 12 hours of Prime95, over 100 loops of Memtest Test #5, and Super Pi 32M. Windows will not boot at 280MHz or above with these timings.

0442A2 KOREA
Highest Memory Frequency: 285MHz @ 2.5-3-3-10
Memory Modules Tested: 2x256MB G.Skill TCCD @ 2.8V
Stability - Over 25 loops of Memtest Test #5 and Super Pi 32M. Windows will not boot at 300MHz or above with these timings.

Keep in mind I only had approximately 90 minutes to install and conduct my testing. Because I was able to pass Super Pi 32M without incident, I started up Prime95 Torture Test/Blend/Priority10 just before I left for work.

Well, folks. I still haven't come across any technical documentation that proves the superiority of the Korean chips over the ones made in Taiwan. All we have is speculation. What I do know for sure is that I have a Korean chip manufactured 2 weeks later than my last one and it allows me to hit DDR570 as of this writing. Draw your own conclusions as I have already come to mine. ;)

xgman
11-17-2004, 07:40 PM
Hard to draw a conclusion from one sample, but they say that earlier the Korean made were better, so maybe they still are. What I wonder is whether the newer week 40 mTaiwan's are any better than the older pre 0418 Taiwan's (I had a 0408 and now a 0440 Taiwan). In any case I can't imagine any other retailer letting you pull chipset fans off to see. So I guess that for the majority of us there will be no way to tell unless a serial # pattern emerges.

I wonder why they are switching back and forth between chipset fabs?

By the way, are you testing top htt at 3x or 4x?

conrad.maranan
11-17-2004, 07:49 PM
I, too, was surprised when he just let me pop these NB HSFs off. I suppose he figured I knew what I was doing and looking for.

I'm testing at 4x.

EDIT:
My bad. I'm testing at 3x LDT and a CPU multiplier of 9 (sometimes 8).

xgman
11-17-2004, 08:11 PM
I just tested the 0440 (in windows) with clockgen a bit and with 2x512 3700ELplat I can run at 2.5-3-3-1T at 280's 3x (freezes quickly at 290) which is a bit better than the 0408 I had before. Oh well, I guess the Neo4 will be out in a few weeks and then there hopefully be a A03 chipset to wonder about.

conrad.maranan
11-17-2004, 08:16 PM
Oh well, I guess the Neo4 will be out in a few weeks and then there hopefully be a A03 chipset to wonder about.
I'm hoping the DFI solution will be out in a few weeks as well. It should be pretty sweet to see how these two boards will perform. I've come to the conclusion that the K8N Neo2 is just a very random board.