MSI Neo4 Platinum Sucks For A64 X2!

[StormPC]

After much testing (and a dead Neo4 Platinum) I am convinced that MSI has some problems with the Neo4 Platinum (MS-7125). They have yet to release a BIOS officially that supports the AMD Athlon 64 X2 dual-core CPUs. The beta BIOS's they sent me are garbage, as are all of the BIOS's on their site for this board. The one decent BIOS (1.3, the one the board came with) is not even on the MSI site. Even with the San Diego and Winchester core chips I tested on the board there were major BIOS and flash memory issues on this board.

All is not lost, however. The Asus A8N-E Deluxe is here!!!

The difference between the Asus and MSI boards is night and day. The Asus not only truly supports the X2 chips, it has done so since the last 2 BIOS releases. They also have two Beta BIOS which work extremely well. I'm currently running Beta 1006.002 and it's wonderful. It has better overclocking features that actually work. When an overclock fails the system automatically calls up the default FSB and memory timings without actually changing the BIOS settings so all you have to do is go into the BIOS and change one value until you get one that works. The MSI is a complete nightmare by comparison. Hats off to Asus for taking the time to release the product properly. Late is better sometimes I guess.

While I'm at it, I've done lots of testing with the new X2s. If anybody has any questions feel free to ask me. There's a lot to digest with these new cores. Here is just a few things you may find interesting.

1) Unlike many of the more "available" A64s the X2 really does justify the extra money for the higher-end models. You can go for the 2200MHz models and you'll probably be able to hit 2400MHz stable with very good air cooling, but anything beyond that will require water or better. The 2400MHz models overclock much better, but NO X2 will overclock anywhere near the San Diego or Venice cores, for obvious reasons.

2) When using memtest to dial in RAM speeds and checking CPU overclock limits there's something you need to realize. YOU CAN'T DO IT WITH A DUAL-CORE CHIP! The reason is that memtest only uses one core. You can get some very good stable overclocks with only one core, but when Windows loads the kernel and that second core kicks in all bets are off. What you get in memtest is meaningless on a dual-core machine.

3) These things are very fast and very powerful. Even the 4200+ is easily more powerful than just about any dualie Opteron ot Xeon at almost everything.

4) THEY REALLY REALLY NEED VOLTS!!! REALLY! For high stable overclocks they need the voltage. They also need cooling. I'm not a fan of water cooling but in the case of the overclocked dual-core system I am very tempted to "take the plunge" so to speak. The 2.2GHz 4200+ will run just fine at 2.46GHz but it needs to be fed 1.475v min to do so unless you have very good water cooling or better.

5) Did I mention these things are fast? Well they are, and very smooth also, even when running 2 instances of SETI there is no noticeable slowing no matter what I do. Multi-tasking is greatly improved, something AMD has needed for awhile.

If anybody has any questions about the 4200+, 4400+ or 4800+ chips or my testing please feel free to post them here.

Cheers!

[speedy_s]

U are saying quite the opposite to this review.
http://anandtech.com/cpuchipsets/showdoc.aspx?i=2452

They do clock quite high and don't like volts is what anand says.

[SilentDeath]

Well I would not conider 1.475v to be high. Thats only just outside spec.. asuming stock is still 1.4 for these....

My 3200+ winnie is eats up around 1.55v to get 2.4ghz, and 1.65 for 2.5ghz but thats pushing it.

My guess is the regulators on these boards are not able to supply as clean power (as they would on a single core) @ the amperage required to run dual core chips. Boards like the DFI may benifit here?

When you say memtest86, have you tryed memtest86+ (the newer one). Afaimk its still in development so they might add dual core support, or let you chosse which core (although testing each seperatly is not going to really give any better result).

When they come down in price Im sure a lot of people will switch over. I would be tempted...

[StormPC]

U are saying quite the opposite to this review.
http://anandtech.com/cpuchipsets/showdoc.aspx?i=2452

They do clock quite high and don't like volts is what anand says.

Really? I think you need to read it again. Did they not say this?

"The last 55MHz required a great deal more voltage and generated a lot more heat. For day in and day out the 240 setting (2.64GHz) would be our choice with this CPU (stock 2.2GHz) on air."

You should know that they used a hand-picked sample, and I am using a production X2, so my results are more valid in the real world. Also, I may have gotten a bad overclocker. I think not however, because the results with my 4400+ and 4800+ indicate that the yields may not be as good as AMD would have liked, thus the shortage of X2s. I believe my results are typical of what we can expect for the bottom-of-the-line X2. My 4800+ would not overclock much better than their 4200+. Definitely hand-picked for the review.

X2's are designed to use less power so compared to the single core AMD64s they don't like volts. They still require a hefty voltage increase for anything I would consider a good overclock if you are running air cooling, and you'd better be running good air cooling.

One last thing...they never said what their ambient temps were, which is very important on air. Mine are around 80 deg F. At that temp with both cores having 100% load (SETI) they run about 59 deg C. AMD's spec calls for a max die temp of 65 deg C. The voltage spec is 1.35-1.4v. I'm running 1.475 volts. Even though I'm only running 5% over on the voltage and a 12% overclock on the CPU(s) I am nearing the safe operating maximums with air cooling. Water may be necessary for really good safe overclocks.

[Smokey21]

Water may be necessary for really good safe overclocks.

But water is useless and a waste of time remember?

Dunno why your suprised, MSI couldn't make a decent mobo if they tryed.

[StormPC]

For most systems water is more trouble than it's worth. For example, I tried several high-end water blocks on my 3700+ San Diego, and unless I chilled the water it ran no higher clocks than the high-end air coolers. The San Diego and Venice cores run very cool though, and there is only one core.

The X2's are a whole other animal however. When you overclock them you are overclocking two A64s at once, and the heat output (and cooling requirements) on high overclocks is nearly doubled. Even with the extra maintenance of water it may be worth running dual-cores on water because of the extreme power of the X2's. Even on air my 4200+ at 2460MHz crunches an average SETI WU in about 30 minutes. With water it could do 28 minutes I'm guessing. May be worth the trouble to water-cool.

[Smokey21]

Whats this trouble you keep mentioning about water cooling?

I dunno what air your using it must be the worlds best air cooling.

There is no way in hell air can keep up with say a TDX, dual heatercore and 50z.

You can pump so much more volts in that air, ive learnt this from my Xp-120 to my setup.

I can use 1.8volts comfatably, and on a Clawhammer, it's perfect.

[StormPC]

If you are comfortable destroying your CPU then by all means use 1.8v. Most people are not interested in that. Even when I run phase change that's a bit extreme for a 24/7 voltage. Don't think you are not damaging the CPU just because your temps are relatively low. You put enough volts and current through a CPU and it doesn't matter if your temps are -100C, that CPU's a gonner!

Also, I didn't say air runs as cool as water. I said water doesn't necessarily buy you higher overclocks. It depends on how much your CPU likes volts and how much heat the die puts out. With dual core chips the extra trouble (yes, water is much more high maitenance and more of a pain in the arse to install) could possibly be worth it due to the fact that the faster ones are not much faster and cost nearly double.

I am considering testing a water cooling system on the 4200+ to see if it will buy me enough performance to justify the extra $$$ and trouble.

[SilentDeath]

If you are comfortable destroying your CPU then by all means use 1.8v. Most people are not interested in that. Even when I run phase change that's a bit extreme for a 24/7 voltage. Don't think you are not damaging the CPU just because your temps are relatively low. You put enough volts and current through a CPU and it doesn't matter if your temps are -100C, that CPU's a gonner!

Also, I didn't say air runs as cool as water. I said water doesn't necessarily buy you higher overclocks. It depends on how much your CPU likes volts and how much heat the die puts out. With dual core chips the extra trouble (yes, water is much more high maitenance and more of a pain in the arse to install) could possibly be worth it due to the fact that the faster ones are not much faster and cost nearly double.

I am considering testing a water cooling system on the 4200+ to see if it will buy me enough performance to justify the extra $$$ and trouble.

The heat output of all chips can be raised more than enough to bennifit from watercooling yet still last a year or two.

Heat output increases 4x for a 2x increase in voltage (from 1v to 2v for an athlonxp for example). That is assuming clockspeed is not changed - which would increase it more as below:

Assuming voltage is not changed, clockspeed will increase heat ouput linearly - i.e 40w at 1ghz, would mean 80w at 2ghz.
With that in mind you can see why watercooling can benifit any chip that is pushed.

Based on the other thread about watercooling I thihnk you just had bad experiences with water but most people dont. If done properly it does not take much more effort to install (except case moddage but you can blame the manufactueres for crappy cases for that) and there is no more maintanance than with air and can go a year without waterchange and then its only a 5 min job.

[Smokey21]

Yeah silent death.

He probably spilt water on a brand new video card or somet, cause he always says it.

I won't use 1.8 24/7, ill probly use 1.75. Clawhammer love there voltage, with temps under 45C iam fine. You can't do that on air.

[WhiteShadow]

Hmm... this isn't good news. I've got a 4400+ on order and was going to get the MSI board (because it got great reviews and was relatively cheap).

Out of interest, what happens if I get the MSI board which doesn't have an updated BIOS that supports the X2s? i.e. Will it detect the 4400+ as a single core chip and run as a single core chip? Or will it just not boot up at all?

Otherwise, what other board should I get? The ASUS or the DFI?

[StormPC]

Hmm... this isn't good news. I've got a 4400+ on order and was going to get the MSI board (because it got great reviews and was relatively cheap).

Out of interest, what happens if I get the MSI board which doesn't have an updated BIOS that supports the X2s? i.e. Will it detect the 4400+ as a single core chip and run as a single core chip? Or will it just not boot up at all?

Otherwise, what other board should I get? The ASUS or the DFI?

If you insist on the MSI K8N Neo4 I can send you a couple of BIOS that will work but they do not overclock well. I'm telling you though that the Asus A8N-E Deluxe just destroys the MSI in every way. It's much much better.

Yes it will detect the CPU as a single core so posting is no problem.

Silent and Smokey,

No boys (or girls), I've never had a bad experience with water cooling. I have many customers who use water. I just told of my experiences. In my opinion water is usually not justifiable based on the performance gain vs the added trouble and expense. Most people would be better served to spend the extra money and get the next higher (or two) CPU model(s) and using air with a bit lower overclock.

Yes you can raise the heat by raising the voltage but raising the voltage does not always allow a better overclock. The X2 may be worth it, but I need to do some testing to see.