New Asrock motherboard

[Nindustries]

Any updates on this?

 

[memhog]

I am intrigued that this board specs to go to 64GB, and am confused why everyone quotes it only going to 32GB?

Is there any supplier of 16GB U DIMM that would work with this board (or just a future possibility)?

My final noob question … the other ASRock mobo's have sockets that enable me to go choose the processor,
does this board come with processor installed - or do we need to purchase that on top?

 

[Nindustries]

I am intrigued that this board specs to go to 64GB, and am confused why everyone quotes it only going to 32GB?

I did notice this too, maybe because 64GB is too expensive? :D

My final noob question … the other ASRock mobo's have sockets that enable me to go choose the processor,
does this board come with processor installed - or do we need to purchase that on top?

Nope nope nope, the CPU is soldered onto the motherboard. Luckily, otherwise it would be one pricey board..

 

[memhog]

I did notice this too, maybe because 64GB is too expensive? :D


Nope nope nope, the CPU is soldered onto the motherboard. Luckily, otherwise it would be one pricey board..

In the Registered ECC world the 16GB DRAM is now < 2 x 8GB DRAM prices (not expensive in my equation).
However I haven't been able to find Unregistered 16GB Dimms. Maybe that is why folks on this thread are
quoting 32GB Max Mem when the spec quotes 64GB?!? Just asking if 64GB is possible (at any price) and with
what parts.

 

[cyberjock]

There's one company that is claiming to be manufacturing 16GB unregistered ECC DIMMs. Now I haven't seen a single one for sale yet anywhere. But allegedly, they do exist.

 

[languy]

You summed it up well jgreco.

And let me just say this since I have better things to do than argue opinions with others in IT...

1. I've put my M1015 in every slot on my X9SCM-F board, and it has worked EVERY SINGLE TIME. So I have no clue what your problems were with RAID, but I've had no such problems. And I have yet to hear anyone else on this forum complain about RAID controllers not working in any slot on a Supermicro board.
2. Do you even read these forums? Have you seen how many people can't get good CIFS speeds during a scrub?
3. I don't believe for a second that you get 1GB/sec on a scrub, on a processor that can't go beyond 8GB of RAM, which is the minimum for ZFS.
4. And if you look closely, I made some comments regarding the fact that "The whole "low power" thing is a major trip though" because so many people don't get that low power also means low performance(how much lower depends on the processor) but few people walk away talking about how fast and awesome Atoms are. If that's about to change, great. But let me tell you that NOBODY here has jumped for joy over Atom performance...yet!
5. I've run all LED since 2008. And while I'm confident that long term I might break even, I wouldn't jump to the conclusion it will save me money long term. I'm a realist, and I realize that the whole promise of LEDs in the future is compelling, but even today's LEDs might not be as good a deal as the manufacturer's want you to believe.
6. And wherever you get the idea that server grade means loud is beyond me. My server is very quiet. It used to be in my living room! And don't get me started on the "price premium for server grade" as I don't feel like it was so expensive that I'd have to argue against it...

And lastly, I said low power CPUs. I didn't specifically call out the CPU you mentioned. And here's the shocker.. I haven't dismissed it. No, I haven't. I'll accept or dismiss it when I see it in a machine running. Not from some benchmarks, which you seem to be swearing on already and getting all upset because you somehow think I've dismissed it. Yes, I dismissed it at first because of the Atom name, but notice my second post doesn't call out your as-yet-unreleased CPU.

So wake up bro. I've seen plenty of people get upset over the Intel(and AMD) "green" CPUs. And you aren't going to convince me that 30w of power from a CPU is going to translate into an enormous electricity bill, is going to heat up your room, or make your server sound like the airport. A single incandescent puts out more heat than that, and I promise you that 100 years or so will tell you people didn't even bat an eye at the heat losses.

Sorry, I've been too busy to get back to this and I wanted to see some real world usage before I commented again. But lets start with #1, shall we? You know, because if it isn't posted on this forum it doesn't really exist does it. :) The X9SCM board has had major bios issues regarding the use of it's PCI-e bus. They went through many BIOS revisions in pretty short order and putting a quad port Intel NIC and an LSI-9211 will cause the machine to "pause" under heavy load. This has been documented repeatedly. I have 4 of them right now on a rack at work. I've seen it. There were several revisions of this bios that simply wouldn't POST with both cards installed. Supermicro's response? "Replace under RMA". Didn't fix the problem. We moved on. I'm not going to debate Supermicro's support. I have first hand experience. They seem to have two responses "We can't duplicate that" or "Send it in". Fair enough. That is right up there with everybody else.

2. What is the deal with CIFS performance concerns? It is clearly an inferior protocol on the FreeNAS platform. I don't expect that to change. It hasn't in the last couple of years anyway. I don't expect any great performance during a scrub especially given the nature of the scrub process either. Since blocks are checked in the order they were written, a good amount of head thrashing can be expected; Especially on a fairly full or fragmented zpool. Unless, of course, your NAS never deletes any files and you aren't snapshotting. Given the amount of seeking being done during a scrub, how can CPU be the limiting factor? I'm not saying you're wrong, but I don't think CPU is the whole story here.

3. What you believe and what is true aren't necessarily the same thing. As I mentioned before scrubbing reads the pool in time-linear order. If I have a zpool that is relatively empty and written in a sequential fashion with no deletes, I would expect it to scrub much faster, wouldn't you? Given that, couldn't my scrub performance equal yours?

4. I get what you are saying. I read the comments from others. You very much did dismiss it in your first post and I called you on it. "Yes, I dismissed it at first because of the Atom name" was your exact quote. Technology moves on. If you have been in IT as long as you say you have, you have to know that. I've seen some preliminary reports on the performance (I know, I know, you want to see it for yourself. But I don't have the time right now to play with one). Under multi-threaded applications the board has performed very well. Actually better than I had hoped.

5. You bought LED lights for your home in 2008? That must have set you back a pretty penny. Yet you bought them. Why? Now when I mention building something that uses less power you dismiss it. I don't understand your reasoning here.

6. Where do I get the idea that server grade is loud? Simple, because the audible specifications of a server are nearly always unimportant. Do you think HP, IBM or Dell worries about how loud their systems are? No, they don't. It isn't an issue when it is racked in a data center with other machines like it. Where do you get the idea that server grade hardware isn't loud? Have you ever bought a supermicro chassis?

See, I want to work with this as a challenge. The same way a car enthusiast might, I don't know, add a second engine to a Mini Cooper. How small a system can I build with a decent level of performance? How will more cores at a lower clock work on what is supposed to be a multi-threaded OS? How can I keep such hardware cool and stable without a large enclosure or loud fans? Can I add a ZIL or an L2ARC and how will that effect performance? I was looking for discussion on these items and didn't really see it. Not complaining. It isn't anybody's job to humor me.

Finally, just so I can put this behind me. There are more than a few forums that have posts indicating that running OpenIndiana with ZFS works just fine on 1GB of RAM TOTAL. http://hardforum.com/showthread.php?t=1798774&highlight=. I'm not arguing here. I am just stating what I've seen. I guess my point is that performance is in the eye of the beholder.

 

[languy]

Okay... you're right, it isn't an "ATOM processor", it's an "Atom[sup]TM[/sup] Processor". At least according to its manufacturer.

You mentioned and then cyberjock said which is probably reasonable given that our experiences here are that lower core speeds tend to be disappointing. The Atoms and E350's and N36L's have been around for a long time and there are many people wandering around here with experience with various subsets of those products. The general opinion is that they do work, but that ZFS adds a lot of overhead.

And had I been less busy yesterday, I might well have pounded out some commentary on your long missive yesterday.



Using the ServeTheHome review, I picked the KVR16LSE11/8 which is currently $98.57, or four for $394.28 to outfit the board with 32GB.

The proper kit for an X9SCL would be Kingston KVR16E11K4/32, currently $325 to outfit the board over at Amazon.

Wait, what? You made it sound like the RAM for the Avoton was significantly less expensive. Not only is it not, it is actually a price premium. Eugh!




Well, the TDP of the E3-1230V2 is 69W. Observationally, it isn't that hard to get a base node that idles at 40 watts and runs around 80 watts peak; add hard drives to that. The Avoton looks like it'll do better. Based on Patrick's numbers, adjusting for a little more memory, I'm guessing 18 watts idle and 35 watts full tilt for the Avoton, which is remarkable.



Our ESXi nodes are designed for low noise; they're 4U boxes with 120mm fans at low rev and an oversize heatsink for the CPU. They're actually very quiet. What's your point?



Well, the problem is, that's all opinion, now, isn't it. ZFS is piggy. So you can skip ZFS and have full gigabit speeds on slower hardware, or you can have ZFS or not full gigabit speeds. But a lot of the guys here are doing this in part because they're the equivalent of hot-rodders. They want ZFS and full gigabit too. But experience says that core speed - not excessive cores - is a better predictor of the ability to get full gigabit.



And nearly none of them use ZFS, or can support two dozen drives without a second thought. The problem is that ZFS is really pretty big and fat. We overcome this by throwing massive resources at it. If you want the cheapest, smallest NAS, FreeNAS isn't suitable. It is really more of an enterprise-class product that is available to home power users, or those of us who are using it for business purposes. All those other products you're talking about, yes, any idiot can throw Busybox on a SoC (and some of the products even prove that this has been successfully done ... by idiots).



SuperMicro is as high end as you can get within this market segment. Beyond it, you generally need to buy a vendor's box. I generally prefer HP's gear, but fixed configurations, high energy consumption, and high price have all been significant factors in our return to building gear in-house.



Well, our experiences differ then. We rarely have issues that would require us to contact SuperMicro. Unlike other manufacturers, they don't think that a "server" is the guy who takes your order at a restaurant. They know what FreeBSD is too. Pretty amazing compared to other mfrs.



You did, and I even thanked you. I think for $250 this would be an attractive option for a lower performance fileserver. I'd been thinking about replacing our MicroServer N36L and this is actually just about the product I had been hoping for.



Uh, wha...? cyberjock's post doesn't quite read that way to me. He's spent more time than most of us have helping people on strange platforms, and he has a lot of experience. His message makes good points. Let's try to be fair here.

Fact is, the Avoton looks to be a welcome addition ... but it is primarily raising the bar on the low end. We've had crap and crud down there for years, E350, Atom D510, N36L, etc., and you just don't see that much discussion of those options because people don't want to spend lots of money and wind up with slow. The Avoton has the potential to be massively better. But it isn't going to be a competitor to the Xeon, which outguns it in clock speed - except that the Avoton can apparently handle 64GB. So when we see a board capable of that, it is going to be a serious competitor for certain types of workloads.

Wow! You sure put a lot of time into this. I don't know whether to feel impressed or what. Pardon my big head whilst I comment on your comments.

You kind of took my comments about ATOM a little too literally. Yes, Intel is reusing the name. After all, a trademark has been paid for, lets use it. Even if it kills sales. Early ATOM chips didn't support ECC did they? They didn't have more than 2 cores until Sept 2013. There are other bottlenecks as well. But, you know, lets split hairs...

Yeah, the whole RAM thing didn't work out the way I'd thought. Still close enough. I'm not sure I want to know what the 16GB modules cost.

The whole thermal issue needs some verification. I know the Asrock board is using a passive heat sink. I'm not sure that is sustainable in a busy system. I'd love to find out though.

Not sure how ESXi nodes factor into this. Unless you are commenting about my noise concerns. Making a silent ESXi node is pretty easy. You don't need any local storage other than a USB flash drive. Put a massive heat sink on the CPU and a slow 200mm fan and we're done; Especially in a 4U case. Or, like we do, you could go with a closed loop system liquid system. We've been bitten once or twice.) With no spinning hard disks, the boxes should be whisper quiet.

When I mentioned server class hardware, I mentioned Dell specifically. Maybe I should have said enterprise class. I don't know. We both know that those systems aren't designed to be quiet. It is simply not a design requirement.

On the whole issue of core speed vs cores, I'm not certain that is an absolute. Taken literally, that would mean a Xeon i5-2600 with 12 cores at 2.7ghz would be slower than a Xeon X5675 with 4 cores at 3.06ghz? I know it is a little facetious, but this was one of the issues I wanted to have an intelligent discussion on. How efficient is the ZFS multithreading code? I don't know.

Is anybody really putting 24 drives in a home array? At that point we are out of the price range I was looking for for a home system. I'm not sure I would buy an Avoton for a business, not yet, anyway. I mean lets be realistic here. The board is interesting, but immature. Software and drivers aren't ready yet. I'm fairly certain FreeNAS would have major issues running on this hardware at this time. I know someone tried FN 9.1 and was unable to get it to boot.

A good number of those "idiots" are making good money selling those boxes.:) They do pretty much everything they claim they will do.

We have two responses from Supermicro "Don't know" and "send it back". I don't ask them technical questions about operating systems so I can't comment on that. I know when we "discovered" a bug in the bios with regards to the pci-e x8 slots on the X9SCM, they did little to assist until a new BIOS just appeared a month later. My comment was directed at Cyberjock's remote about another vendor. With no first hand experience that he mentioned, he dismissed the brand as inferior.

When I saw the board, I saw a NAS in a box. Thrown in a U-NAS 8 bay case, I saw a potential Qnap/Synology killer. What I had hoped to spark (and apparently failed at) was a discussion about where this board would end up performance wise. I'm planning on buying one when I can confirm FreeNAS has supported drivers for the NICs and SATA controllers. To continue your car metaphor; I want to know if a better transmission (cores) makes up for a slower clock. If that is true, where is the line?


Thanks for your comments. Sorry I didn't get back to you sooner. I look forward to seeing how this all shakes out.

 

[Richman]

Okay... you're right, it isn't an "ATOM processor", it's an "Atom[sup]TM[/sup] Processor". At least according to its manufacturer.

Really? So its an ATOM™ processor. So now Intel 'Trade Marked' it and that means something?

1. I've put my M1015 in every slot on my X9SCM-F board, and it has worked EVERY SINGLE TIME. So I have no clue what your problems were with RAID, but I've had no such problems. And I have yet to hear anyone else on this forum complain about RAID controllers not working in any slot on a Supermicro board.

I think the difference may have been thta he was using something other than FreeNAS and was using HARDWARE RAID on a HARDWARE RAID controller and you were using SOFTWARE RAID and a controller flashed to IT. That may possibly be why he experienced issues and you didn't.

You kind of took my comments about ATOM a little too literally. Yes, Intel is reusing the name. After all, a trademark has been paid for, lets use it. Even if it kills sales. Early ATOM chips didn't support ECC did they? They didn't have more than 2 cores until Sept 2013. There are other bottlenecks as well. But, you know, lets split hairs...

Did they reuse the name for a totally redesigned core or is it a cousin or a previous? Does anyone know or do I have to do my own research. I have way too many things to do research on.

On the whole issue of core speed vs cores, I'm not certain that is an absolute. Taken literally, that would mean a Xeon i5-2600 with 12 cores at 2.7ghz would be slower than a Xeon X5675 with 4 cores at 3.06ghz? I know it is a little facetious, but this was one of the issues I wanted to have an intelligent discussion on. How efficient is the ZFS multithreading code? I don't know.

Where is this intellignet discussion. I for one would be interested in reading it

When I saw the board, I saw a NAS in a box. Thrown in a U-NAS 8 bay case, I saw a potential Qnap/Synology killer. What I had hoped to spark (and apparently failed at) was a discussion about where this board would end up performance wise. I'm planning on buying one when I can confirm FreeNAS has supported drivers for the NICs and SATA controllers. To continue your car metaphor; I want to know if a better transmission (cores) makes up for a slower clock. If that is true, where is the line?

$375 adds about $240 over the alternative Supermicro. That's a lot for some of us to swallow.
Edit: Oops, I forgot that the CPU come with it and that may make a difference ......... but then not over say an ASUS with a 760 or 970 chipset that supports ECC ....... for a home server that is.

 

[cyberjock]

Hey languy: I'm going to go back to what I said before...

And let me just say this since I have better things to do than argue opinions with others in IT...

I really don't want to nor need to argue with it. We have stuff that works. Regardless of your experience with X9SCM I've had zero problems. Many others have bought the board because I own it and they figure if I own it then it must work fine. They've had no problems either.

Sure, these new Atoms are more powerful, but how much, etc is something I'm not going to fight over.

Most of the rest of what I said has been discussed to death on the forums. There's a forum search feature. I give you permission to use it.

And most of the rest of that crap you tried to explain has been explained on the forums here many many times. And instead of you trying to explain it to him, he's probably explained it to others.. dozens of times.

But it's hilarious to see someone that has just hit double digit posts show up here and argue with 2 rather high post count members. You're sure off to a great start on the forums.

 

[Cupcake]

Thanks for sharing this mainboard though, it sure looks very interesting in regard of building a freenas (or anything else server-like). Will definitely keep an eye on that!

 

[jgreco]

Really? So its an ATOM™ processor. So now Intel 'Trade Marked' it and that means something?

It sure as hell does if you're correcting someone and laying on the sarcasm.

 

[jgreco]

In the Registered ECC world the 16GB DRAM is now < 2 x 8GB DRAM prices (not expensive in my equation).
However I haven't been able to find Unregistered 16GB Dimms. Maybe that is why folks on this thread are
quoting 32GB Max Mem when the spec quotes 64GB?!? Just asking if 64GB is possible (at any price) and with
what parts.

This is what generally happens with new stuff. I remember back in ~1995(?) one of my clients was desperate because they needed a FreeBSD box with 256MB RAM and ASUS had just released the P/E-P55T2P4D, but nobody had 32MB modules... we had to contact a manufacturer to build them special, which cost about 10x what it'd have cost for an equivalent amount of 16MB modules.

More recently, with the advent of Sandy Bridge, nobody had 8GB unbuffered ECC DDR3 at first, then they were ridiculously expensive for maybe a year, then the price fell into reasonable ranges.

If you want amusing, note that today, a 16GB registered is about $100, but a 32GB registered is about $500. I'm looking strictly at Supermicro tested parts here and wholesale pricing, retail's ... more. And the 32GB is only available as a DDR3-1333, HMT84GR7MMR4A-H9 ... and for a real gut-puncher, Supermicro recently bumped the specs on boards such as the X9DR7-TF+ from 32GB to 64GB modules, which allows for the possibility of a whopping 1.5TB of RAM on that board, but no candidate tested modules are listed.

http://investors.micron.com/releasedetail.cfm?releaseid=790340

would suggest that "it's coming" but I'm guessing we won't see it at a reasonable price for a little while yet.

 

[jgreco]

I really don't want to nor need to argue with it. We have stuff that works. Regardless of your experience with X9SCM I've had zero problems. Many others have bought the board because I own it and they figure if I own it then it must work fine. They've had no problems either.

Problems happen in IT. I have to wonder very seriously about the experience level of a guy who would write

I know when we "discovered" a bug in the bios with regards to the pci-e x8 slots on the X9SCM, they did little to assist until a new BIOS just appeared a month later.

This is just like, uh, duh. What did he expect? For them to wave a magic wand? So a problem was discovered. Oftentimes problems like this must be resolved by having it analyzed, debugged, reported back to the BIOS development group, having them figure a fix, pump out a fixed version, do QA testing, and then release the fixed BIOS. A month is extremely reasonable for a previously unknown problem, especially if it involves specific hardware. Sometimes you get lucky and discover a problem someone else already found, so yeah it's totally awesome if you immediately get a "here try this new beta BIOS" response from tech support, but that only happens because someone else had already identified the problem and is now using you as a guinea pig beta tester.

Lots of us who do this professionally will suggest avoiding being an early adopter if reliability is a key concern. New hardware tends to have more issues. New firmware also tends to have more issues. But you can also enjoy the benefits of that new hardware. Tradeoff.

 

[languy]

Problems happen in IT. I have to wonder very seriously about the experience level of a guy who would write



This is just like, uh, duh. What did he expect? For them to wave a magic wand? So a problem was discovered. Oftentimes problems like this must be resolved by having it analyzed, debugged, reported back to the BIOS development group, having them figure a fix, pump out a fixed version, do QA testing, and then release the fixed BIOS. A month is extremely reasonable for a previously unknown problem, especially if it involves specific hardware. Sometimes you get lucky and discover a problem someone else already found, so yeah it's totally awesome if you immediately get a "here try this new beta BIOS" response from tech support, but that only happens because someone else had already identified the problem and is now using you as a guinea pig beta tester.

Lots of us who do this professionally will suggest avoiding being an early adopter if reliability is a key concern. New hardware tends to have more issues. New firmware also tends to have more issues. But you can also enjoy the benefits of that new hardware. Tradeoff.

What I expected, and I might add, paid for, is an accurate, professional response from Supermicro. Also who said it was previously unknown the them? What they know and what I know aren't the same thing is it? What is more likely true here is that the developers where keenly aware of the issue, but had not communicated same to support. The support team has a knowledge base and an RMA form. They use the KB first and the RMA second. I am completely fine with "We have no reports of this". That isn't the answer we got.

I hardly think a board that had been available for over 6 months at the time as a new product. I also think the fact that we bought several with an older BIOS that did work would give us reason to believe the product was stable.

 

[languy]

Hey languy: I'm going to go back to what I said before...



I really don't want to nor need to argue with it. We have stuff that works. Regardless of your experience with X9SCM I've had zero problems. Many others have bought the board because I own it and they figure if I own it then it must work fine. They've had no problems either.

Sure, these new Atoms are more powerful, but how much, etc is something I'm not going to fight over.

Most of the rest of what I said has been discussed to death on the forums. There's a forum search feature. I give you permission to use it.

And most of the rest of that crap you tried to explain has been explained on the forums here many many times. And instead of you trying to explain it to him, he's probably explained it to others.. dozens of times.

But it's hilarious to see someone that has just hit double digit posts show up here and argue with 2 rather high post count members. You're sure off to a great start on the forums.

Hey Cyberjock, then don't argue (especially at 3AM). Because, you know, if you do it, then it is perfect. See I've noticed that when someone calls you on your posts you resort to condescension.

But you are right, the post has exceeded it's life span. You have your snide comments. Point made. Me and my crap will just wander off. It is interesting, though, isn't it, that you have equated volume of posts with knowledge and intelligence.

 

[languy]

Problems happen in IT. I have to wonder very seriously about the experience level of a guy who would write

Wow. Seriously? Here it comes again. We wanted, expected, and paid for boards that POSTED. It is not a reliability issue. The boards were completely reliable. They did nothing very very well. Your passive aggressive insult doesn't suit you. Again, it wasn't the technical issue that was the problem, it was the handling of it. But instead of acknowledging my experience and indicated that it wasn't your experience. You cast aspersions at my technical experience. Because your experience an mine differ, I'm a n00b.

Look, you clearly have good experiences with Supermicro. We haven't. We haven't refused to buy their products when they fit the need. We just don't preclude other brands/options when searching for solutions.

It would appear that I have touched a nerve here. To have two high post count users reply so many times, I must have done something.

 

[languy]

Well, people, it has been interesting hasn't it? Two senior members with high post counts have spent quite a bit of time telling me I'm either unintelligent or a noob. Because volume = quality, hey Cyberjock? When two users with nick names like Resident Grinch and "Forum Watch Dog" speak, it is from on high.

Yeah, off to a great start. Or perhaps a great finish.

 

[jgreco]

Wow. Seriously? Here it comes again. We wanted, expected, and paid for boards that POSTED.

They failed to POST once you removed the problematic cards? My oh my, I must be seriously caffeine-deficient this fine morning, I completely failed to read those words above. I still don't see them, in fact.

It is not a reliability issue. The boards were completely reliable. They did nothing very very well. Your passive aggressive insult doesn't suit you. Again, it wasn't the technical issue that was the problem, it was the handling of it.

User has problem. User reports problem. Manufacturer collects reports. Manfacturer presumably identifies a problem, sends it off to BIOS group. BIOS group corrects problem. BIOS group QA's fix and releases BIOS.

I'm not sure which part of that handling was so evil.

But instead of acknowledging my experience and indicated that it wasn't your experience. You cast aspersions at my technical experience. Because your experience an mine differ, I'm a n00b.

Anyone who's been in this business for more than a day has run into problems. I take issue with your citing a resolved issue that simply didn't happen as fast as you'd like as being some sort of failure.

Look, you clearly have good experiences with Supermicro. We haven't. We haven't refused to buy their products when they fit the need. We just don't preclude other brands/options when searching for solutions.

The problem is that in this segment, there aren't many other realistic options. HP, Dell, IBM, etc., all make great gear, but it is typically built for a certain application (like *cough* Windows), which tends to mean that if you order a box with lots of storage, the options available to you are usually power-hungry compute boxes with a built-in RAID controller, etc., and a premium price too. Tyan and ASUS have historically been weak in the server market, though both have had various offerings. Both of them, and really other manufacturers such as Gigabyte, etc., are primarily focused on desktops and possibly workstations. The needs of those of us who run servers, whether they're for virtualization or NAS purposes, are somewhat different, and hard to address.

Supermicro has been known to change their products (thinking specifically the X9SCL+-F) to meet market demands (ESXi 4 didn't support the 82579 out of the box). That's remarkably rare.

But really, in the end, the key bit is that there isn't any real competition. Show me another manufacturer that offers more than half a dozen server boards in a given class.

It would appear that I have touched a nerve here. To have two high post count users reply so many times, I must have done something.

Some of us are here mainly because we find the crowd here to be generally pleasant. Or because we have other historical ties to the community. Or both.

Well, people, it has been interesting hasn't it? Two senior members with high post counts have spent quite a bit of time telling me I'm either unintelligent or a noob. Because volume = quality, hey Cyberjock? When two users with nick names like Resident Grinch and "Forum Watch Dog" speak, it is from on high.

Yeah, off to a great start. Or perhaps a great finish.

I vote ... finished.