Enhanced features separate Intel's X38 from its acclaimed P35 Express, but last-minute revisions have delayed final-revision boards for several weeks. Was it worth the wait?

http://www.tomshardware.com/2007/1 [...] index.html

I still don't think I'll be going with a newer chipset than P35 until the price of DDR3 is affordable. Or if they release vid cards that utilize more bandwidth than the standard PCI-E x16 slot on a P35 board can handle. Whichever comes first and that's still a ways away.


Good read though, wish they would've benchmarked some P35 boards on the 3d games for performance comparison though.

Good point, they should have a P35 board as a reference. These boards are insane for overlocking...only thing that is not good is price. Almost $300, that is just too much for a student. From my personal experiance no matther what kind of board its life span is about 2-3 yrs.
Taking into consideration that this board would be overlocked I'm not sure how long would it take for it to break.

It comes with water block...great but for some that don't have a watercooling ...you have a problem I suppose.

This was a X38 comparison.
They did include a P35 for reference comparison.
http://www.gigabyte-usa.com/Produc [...] uctID=2524
or am I misunderstanding the posts?

Where are the Crossfire performance figures?

I have the Asus Maximus SE and I can tell you that it can OC pretty high even if it's not the highest OCer of the bunch it's good enough for me and for the amount of features and do-dads you get it's a very good buy, it's also pretty stable, but a little picky on the memory end of things. Sorry for the poor grammar, I gotta go.

Ok... is it just me or do other people HATE this type of review? Let me explain what I mean....

1st) 1-2% does not make a board "the fastest" From my experience, this margin of difference is simply to small to be credited ANYTHING other than "equal". If your boss came up to you and told you, you get a 1% raise this year would this excite you?

2nd) After your 1% raise, if you bought a $300 board to get that 1% you'd be a fool! That said, if for features / brand you choose a board, that's very reasonable. But please, please, Tom's stop this "winner" stuff when the margins are so slim. You take your $300 board + CPU, I'll buy a $135 board + BETTER CPU, who's faster? In this light, performance is relative to price!!!

3rd) I grow oh so tired of .. Intel Quad Extreme + Asus Super Duper MB + Ultra Performance Corsair + 1200Watt PSU + 8800Ultra + Raptor raid 0 just look at this performance type of reviews!! You know what, you just dropped 3k, that system better be insane! It's much easier to pick the most expensive hardware on the market and see performance... I really want to know more about value.

I hate seeing a $300 board called a "performance winner" when a $135 board comes within a few percent. I wonder what happens when you buy a $135 board and slap in the next step CPU or Memory or GPU, oh yeah.. it will CRUSH the same price system with a $300 board.

Ahh well... I guess this is my torch to carry. I've had this same rant before, perhaps one day will hear it and listen.

Peter

I'm sorry but we can't give you the gold medal because you were only 1% faster than the 2nd place runner.

I will keep this review in mind for when the prices get more affordable for me.
It always happens that prices go down on new stuff.

I am getting more throughput/performance then the ones they tested with my GA-P35-DQ6. I think that personally if they are going to bench these things they should spend the time to tweak them first. Then you can see what really performs better IMO.

Right. So if the 6' 4" professional runner just "pounded" the high schooler by 1%.. wahoo give him the gold. That's the point... you have the 600 pound gorilla in the cage with the 100lb chimp, he better win and for that price by more than 1%!!

Everest even tells me that my P35 is better then the x38 with the new Q9650 @that.

The X38 is an enthusiast-market chipset, the last time there were any stand-out value leaders in performance boards was the P965 roundup.

My p35 out performs their x38 lol am I invisible here lol.

All the boards perform well, about this same, and no problems?.
This seems like a good thing to me!

Mines better lol. Ok I'm being silly now.

Here's the odd question: Does any of the boards (but in particular the ASUS) support ECC DDR2 memory? ASUS' website and the manuals for their respective X38 boards seem a bit torn on the issue. Usually I'd just decode some BIOS files, but I don't have tools that can handle the latest generation of AMI BIOS. FWIW, I'm a bit torn between the P5E (Maximus without the bling) and the good old P5W DH Deluxe, since ASUS claims Penryn support for it.

It isn't just you... I totally agree.

I'm often baffled by these type of reviews. At the end of the day, there is no noticeable performance difference in ANY of these boards - so why say things like "x shines", "y is a clear winner"? Honestly, who can justify a purchase based on such minimal differences? On that note - I've always had the impression that the limit you hit in your stable overclock will generally be imposed by your CPU or RAM, rather than your motherboard - so I've often wondered why this type of user (gamers/overclockers/those with bizarre heatsink fetishes) would need a $300 board vs a $100 board that in all likelihood will meet their requirements.


However, as far as features go the article covered them well. I thought the comments on the board layout, (particularly its impact on cabling and video and expansion card placement) were very practical.

I have a bone to pick with Tom's testing procedure. By limiting the test to two sticks of RAM you completely missed the fact that the current GA-X38 DQ6 is unstable with four sticks of RAM. How could you let yourself miss such an important factor. Now we have to wonder are the other boards that have been blessed by Tom also unstable? Next time you really should test the boards in a fully loaded configuration. My tests show it unstable with either Crucial or OCZ PC8500 RAM, this has been confirmed by others on the OCZ forum, and also confirmed by Gigabyte tech support. How did you miss it?

Ouch, that sux. The tester probably only HAS pairs.