I'm anxiously waiting for the X10SAT as well...I've emailed acmemicro and superbiiz and both tell me that it should be released this month. Hopefully, since they removed "Coming Soon", the release is imminent!
I'm anxiously waiting for the X10SAT as well...I've emailed acmemicro and superbiiz and both tell me that it should be released this month. Hopefully, since they removed "Coming Soon", the release is imminent!
Sorry for the double post, but this looks like the manual: http://www.supermicro.com/manuals/motherboard/C226/MNL-1544.pdf
Interesting that the published user manual for X10SAT states that only unbuffered, non-ECC memory is supported and says nothing about ECC (contrary to the product page). The photo of PCB even shows "UNB NON-ECC DDR3 DIMM REQUIRED" written above memory slots. I've emailed Supermicro regarding this subject and got the prompt reply:
So it appears that ECC RAM is indeed supported with Xeon CPUs on X10SAT, absence of which in the manual made me considering buying X10SAE instead.
So it appears that ECC RAM is indeed supported with Xeon CPUs on X10SAT, absence of which in the manual made me considering buying X10SAE instead.
Last edited:
It is my understanding that in configuration with Xeon CPU both memory types can be used (unbuffered ECC/non-ECC).
A wild X10SAT has appeared at Acmemicro for 256 U$D (Just 5 over the C7Z87-OCE). Only Newegg is missing from the party.
Too bad is that while Supermicro said that the X10SAT is C2 only, the C7Z87-OCE was already available with C1 Chipsets, so if the packaging isn't marked in some way, you will not be able to recognize a C2. A reason to avoid it, to be honest.
Also, chances are that I will annoy Supermicro support until I get what I want: A X10SAT with a hybrid BIOS capable of letting you adjust any non-Chipset limited option available in the C7Z87-OCE, so I can do undervolt tweaks in the BIOS.
Too bad is that while Supermicro said that the X10SAT is C2 only, the C7Z87-OCE was already available with C1 Chipsets, so if the packaging isn't marked in some way, you will not be able to recognize a C2. A reason to avoid it, to be honest.
Also, chances are that I will annoy Supermicro support until I get what I want: A X10SAT with a hybrid BIOS capable of letting you adjust any non-Chipset limited option available in the C7Z87-OCE, so I can do undervolt tweaks in the BIOS.
You can add me to list of people that would like those features if it will help your chance out any. 
It would be nice to drop down the vcore down a bit for some power savings if they do include it.
It would be nice to drop down the vcore down a bit for some power savings if they do include it.I've been emailing Acmemicro for weeks to see when they would stock it. They say they don't have any in stock yet, but should get back to me later today with a definitive date. Fingers crossed!
*BUMP*
I'm getting a bit anxious. Newegg still doesn't have the X10SAT. Not only I have to wait to actually purchase it, but also wait an extra 2-3 weeks for the international shipping required by the middleman. I could purchase it from somewhere else through, Acmemicro isn't that more expensive, but I need to make sure that they're willing to ship internationally.
I also have on my pending list mailing Supermicro to check if they're willing to enable all the overclocking related options in the X10SAT. Also, I have posted in the Xen Mailing List but didn't receive any responses regarding my possible usage issues.
Of note, both C7Z87-OCE and X10SAT specifically ask for a EPS12V 8 Pin connector according to the Manual, while my current Power Supply has the old 4 Pin ATX12V one. As far that I know, there could be some Motherboards that will not even POST if the VRM are hardwired in some way where you requiere either the EPS12V or the ATX12V 4 + 4 without exception (They don't work with only the ATX12V 4 Pin), but it should work with this adapter.
I'm getting a bit anxious. Newegg still doesn't have the X10SAT. Not only I have to wait to actually purchase it, but also wait an extra 2-3 weeks for the international shipping required by the middleman. I could purchase it from somewhere else through, Acmemicro isn't that more expensive, but I need to make sure that they're willing to ship internationally.
I also have on my pending list mailing Supermicro to check if they're willing to enable all the overclocking related options in the X10SAT. Also, I have posted in the Xen Mailing List but didn't receive any responses regarding my possible usage issues.
Of note, both C7Z87-OCE and X10SAT specifically ask for a EPS12V 8 Pin connector according to the Manual, while my current Power Supply has the old 4 Pin ATX12V one. As far that I know, there could be some Motherboards that will not even POST if the VRM are hardwired in some way where you requiere either the EPS12V or the ATX12V 4 + 4 without exception (They don't work with only the ATX12V 4 Pin), but it should work with this adapter.
Someone on Hardforum tipped that this German store says availability will be September 30th. http://www.sona.de/.1004278378-Supermicro-Mainboard-X10SAT-Retail
I haven't heard back from Acmemicro whether they will also get it on the 30th, I'll try sending them another email...
I haven't heard back from Acmemicro whether they will also get it on the 30th, I'll try sending them another email...
In Holland we have some store that will ship them.... BUT can someone tell me why to pay 100 euro's more for a board that is almost the same as the Asus p9d ws? Except the c2 stepping and the extra useless pcie slots there's only the overpriced thunderbolt interface that makes a difference.
Is there any reason to go for the x10sat instead of the Asus in my scenario; esxi with some HBA cards?
On Newegg the ASUS P9D WS cost 240 U$D, X10SAT should be around 250-255 U$D based on Acmemicro price, so it shouldn't be a 100 Euros difference.
ASUS main issue is the fact that they never seem to care about fixing BIOS bugs or add features on request that should be expected, plus their support seems to love to tell you that they "only support Windows OS" if you have any issue with Linux thanks to ASUS stupid BIOS issues. Examples: PCIe ASPM, Cool'n Quiet, and worst of all, VT-d ACPI DMAR tables/AMD-Vi ACPI IVRS tables, which tons of broken Motherboards that you can find by googling around. I'm avoiding them for that reason, got bored of that type of mediocrity where the only thing you need to fix a problem is a simple Firmware upgrade with data that conforms what the standards or specifications says instead of those Windows-based workarounds because the lazy devs expect a Windows Driver to take over and make it work even if missing what they didn't implemented.
ASUS main issue is the fact that they never seem to care about fixing BIOS bugs or add features on request that should be expected, plus their support seems to love to tell you that they "only support Windows OS" if you have any issue with Linux thanks to ASUS stupid BIOS issues. Examples: PCIe ASPM, Cool'n Quiet, and worst of all, VT-d ACPI DMAR tables/AMD-Vi ACPI IVRS tables, which tons of broken Motherboards that you can find by googling around. I'm avoiding them for that reason, got bored of that type of mediocrity where the only thing you need to fix a problem is a simple Firmware upgrade with data that conforms what the standards or specifications says instead of those Windows-based workarounds because the lazy devs expect a Windows Driver to take over and make it work even if missing what they didn't implemented.
Last edited:
vailr
Diamond Member
- Oct 9, 1999
- 5,365
- 54
- 91
What about a Gigabyte board? If it were me buying a board for your special need support, I would only consider a board with UEFI bios, which Supermicro doesn't seem capable of yet.
A fairly active bios modding community exists for Gigabyte boards:
http://forums.tweaktown.com/gigabyte/48085-gigabyte-modified-bios-56.html
So: even if officially released bioses weren't providing proper support, there's a good possibility of some kind of workaround.
There's ongoing activities to allow native booting an unaltered OSX install disc for Gigabyte Z77-UD5H boards, for example:
http://www.insanelymac.com/forum/topic/291655-ozmosis/
A fairly active bios modding community exists for Gigabyte boards:
http://forums.tweaktown.com/gigabyte/48085-gigabyte-modified-bios-56.html
So: even if officially released bioses weren't providing proper support, there's a good possibility of some kind of workaround.
There's ongoing activities to allow native booting an unaltered OSX install disc for Gigabyte Z77-UD5H boards, for example:
http://www.insanelymac.com/forum/topic/291655-ozmosis/
If I were going to go for a non-Supermicro option, it would be a cheap-ass AsRock, because they have been quite popular recently for both budget builds and their excellent VT-d support.
X10SAT seems to be UEFI based according to manual mentions about that:
Regarding BIOS modding, I recall that during 2006 or 2007 I sended PMs to some popular BIOS modders asking if they could check if it was possible to add any sort of overclocking capabilities to the ASUS A8N-VM CSM, which had a OEM-like BIOS with absolutely not even basic Multiplier or Voltage control. Never got a reply. They seemed to be more interesed in tweaking the already-overtweaked DFI Lanparty BIOSes and the like.
So if I can't get it running out of the box, and considering the actual limits of the BIOS modding tools that the only easy stuff is making visible options that are there but hidden, I don't have any type of hopes that the BIOS modding community would be capable of more complex fixes. Reason why I was looking for Software alternatives for Haswell undervolting.
I would love to give Coreboot a try, but only if they distributed ready-to-be-flashed binaries and just happen to have one for my Motherboard of choice, as no non-programmer would possibily be able to port it to their Motherboard.
One of the things that I like about Supermicro is their reputation due to their Hardware being intended mostly for 24/7 Servers, so I am expecting a robust piece of Hardware out of it with no possibility of dubious choice of components, and more testing to fix bugs.
X10SAT seems to be UEFI based according to manual mentions about that:
...it just simply doesn't have the fancy GUI. Nor I know if actually having it would make any real difference.
Regarding BIOS modding, I recall that during 2006 or 2007 I sended PMs to some popular BIOS modders asking if they could check if it was possible to add any sort of overclocking capabilities to the ASUS A8N-VM CSM, which had a OEM-like BIOS with absolutely not even basic Multiplier or Voltage control. Never got a reply. They seemed to be more interesed in tweaking the already-overtweaked DFI Lanparty BIOSes and the like.
So if I can't get it running out of the box, and considering the actual limits of the BIOS modding tools that the only easy stuff is making visible options that are there but hidden, I don't have any type of hopes that the BIOS modding community would be capable of more complex fixes. Reason why I was looking for Software alternatives for Haswell undervolting.
I would love to give Coreboot a try, but only if they distributed ready-to-be-flashed binaries and just happen to have one for my Motherboard of choice, as no non-programmer would possibily be able to port it to their Motherboard.
One of the things that I like about Supermicro is their reputation due to their Hardware being intended mostly for 24/7 Servers, so I am expecting a robust piece of Hardware out of it with no possibility of dubious choice of components, and more testing to fix bugs.
Last edited:
vailr
Diamond Member
- Oct 9, 1999
- 5,365
- 54
- 91
That tweaktown.com forum linked earlier is exclusively for Gigabyte boards. A poster there named "Stasio" has some very good contacts within the bios writers group, at Gigabyte headquarters in Taiwan. I would have no hesitation buying a (same chipset & similar feature set) Gigabyte motherboard instead of a Supermicro. Of course, Gigabyte is more well known for their consumer-grade and gaming boards. But: it may be worthwhile investigating their server boards and comparing features with your "under consideration" Supermicro X10SAT board. I believe that you'd get a better response now, vs. what you didn't get 6 or 7 years ago, about an older Asus motherboard.
They have just three LGA 1150 Server Motherboards and they're either C222 or C224, no C226 from them. Only considerable option would be the GA-Q87M-D2H, which doesn't have anything I find interesing. Actually, I find it rather stupid that instead of using Haswell ability to do 8x/8x PCIe 3.0, they decided to go with all 16 lanes to the same PCIe and leave the other as a mere 4x 2.0 coming from Chipset. And that PCI... Yuck! I'm avoiding Supermicro X10SAE for having two of them.
One thing I don't like about X10SAT is that according to the block diagram published in the manual it dedicates half of available chipset PCIe bandwidth to Thunderbolt controller. As a result Supermicro had to hide three PCIe x1 slots, IEEE 1394 and external USB 3.0 controller behind some kind of a switch (PLX8606) with single PCIe x1 uplink. X10SAE, by contrast, has full bandwidth available to PCIe x1 slots and USB 3.0 controller. This part of X10SAT design looks as a limitation to me since I don't think I'm ever going to use Thunderbolt with this motherboard, but may likely be using one of PCIe x1 slots and USB 3.0 simultaneously.
If X10SAT was available right now in local retail stores, I'd get that one to make sure I have C2 chipset revision. But if I'm going to wait mid-October then C2 revision X10SAE might surface as well. Not sure which one should I buy then.
X10SAE looks too plain comparing to X10SAT and I don't like front facing SATA ports. Other than that, I don't have any complaints about it. Having 2 PCI slots in place of third PCIe x16 slot is OK for me. I'm going to use neither of them anyway.
If X10SAT was available right now in local retail stores, I'd get that one to make sure I have C2 chipset revision. But if I'm going to wait mid-October then C2 revision X10SAE might surface as well. Not sure which one should I buy then.
X10SAE looks too plain comparing to X10SAT and I don't like front facing SATA ports. Other than that, I don't have any complaints about it. Having 2 PCI slots in place of third PCIe x16 slot is OK for me. I'm going to use neither of them anyway.
I'm aware of how Thunderbolt eats PCIe lanes. I don't need it nor think I will ever use it, and worse if I consider that I will pay an hefty price premium to have it. But I simply don't like the X10SAE legacy IO, and the uncertainity of what Chipset Stepping you will get. X10SAT guarantees me a C2 Chipset, which most others options do not. Besides, there isn't a lot of variety of C226 based Motherboards, and from all those I looked, the X10SAT is the one that fits me the more. Q87 variety is even worse than C226, and Q87 is one step behind C226 feature wise.
X10SAT is far from perfect, there do are many things that I don't like from it. Besides Thunderbolt (Through that is by design - I suppose any other Motherboard with Thunderbolt will have similar bandwidth issues), I don't like the PLX chip because usually is a bit harder to do passthrough of devices behind it. I don't like neither to have extra Controllers like the ASMedia Sata Controller that can their own quircks when compared to the Chipset IO, and also the fact that it has only 6 slots when it could have 7 (Which X10SAE has).
If you picked the X10SAT, removed Thunderbolt, the ASMedia SATA Controller, the 4 Port USB3 Controller, and the PLX chip, you would have 6 PCIe lanes free (From the 8 max, 2 are used for the NICs). Intel Chipsets flexible IO can allow you to repurpose 2 PCIe lanes to 2 USB3 Ports (As X10SAT has 4 from the Chipset), so you could have 6 USB3 using only the Chipset, which I think should be enough. With the free 4 PCIe lanes, you can feed 4 PCIe 1x Slots, for a total of 7 slots. I would love such arrangement, its minimalistic in components (As you rely 100% on Chipset, eliminating a lot of possible quirks or compatibility issues), simplifies designs, yet should have enough IO.
X10SAT doesn't have IPMI or the dedicated NIC for it. It seems to be reserved to Server Motherboards, X10SAT is a Workstation one.
X10SAT is far from perfect, there do are many things that I don't like from it. Besides Thunderbolt (Through that is by design - I suppose any other Motherboard with Thunderbolt will have similar bandwidth issues), I don't like the PLX chip because usually is a bit harder to do passthrough of devices behind it. I don't like neither to have extra Controllers like the ASMedia Sata Controller that can their own quircks when compared to the Chipset IO, and also the fact that it has only 6 slots when it could have 7 (Which X10SAE has).
If you picked the X10SAT, removed Thunderbolt, the ASMedia SATA Controller, the 4 Port USB3 Controller, and the PLX chip, you would have 6 PCIe lanes free (From the 8 max, 2 are used for the NICs). Intel Chipsets flexible IO can allow you to repurpose 2 PCIe lanes to 2 USB3 Ports (As X10SAT has 4 from the Chipset), so you could have 6 USB3 using only the Chipset, which I think should be enough. With the free 4 PCIe lanes, you can feed 4 PCIe 1x Slots, for a total of 7 slots. I would love such arrangement, its minimalistic in components (As you rely 100% on Chipset, eliminating a lot of possible quirks or compatibility issues), simplifies designs, yet should have enough IO.
X10SAT doesn't have IPMI or the dedicated NIC for it. It seems to be reserved to Server Motherboards, X10SAT is a Workstation one.
Last edited:
I just finished a home server using the ASROCK H61MV-ITX with a Xeon E3 1230 v2. It has the full hardware accelerated virtualization enabled. The price was right too at only $60. I know you're looking at a Haswell build, but I like the lower power of IVB, and I was able to pick one up for $200 at a microcenter. Other components include the Seasonic 360w gold PSU, 3x3tb WD Reds, 16gb non ECC ram and a reprovisioned laptop drive for now for the OS's, all in a Lian-Li PC-Q25 (highly recommend this case, btw).
I'm a total virtualization NewB, but I have Ubuntu Server 12.04 running with KVM with the hardware acceleration enabled. So far, the server is running a Samba & NFS file server, Plex media server, and 2 minecraft servers for the kids. Everything works great and no Bios updates were required.
The only limitation I've run into is with the non ECC Ram. I was considering experimenting with a free-BSD Nas install on a VM, but from what I read, having ECC is a must for how that file system works.
I considered holding off for the new ASROCK board with ECC, but when it didn't release in August, I decided I couldn't wait any longer, plus I saved some $$ with this build.
I'm a total virtualization NewB, but I have Ubuntu Server 12.04 running with KVM with the hardware acceleration enabled. So far, the server is running a Samba & NFS file server, Plex media server, and 2 minecraft servers for the kids. Everything works great and no Bios updates were required.
The only limitation I've run into is with the non ECC Ram. I was considering experimenting with a free-BSD Nas install on a VM, but from what I read, having ECC is a must for how that file system works.
I considered holding off for the new ASROCK board with ECC, but when it didn't release in August, I decided I couldn't wait any longer, plus I saved some $$ with this build.
Last edited:
I completely sympathise with you - I would like Vt-d and unlocked multipliers as I use VMs a lot. Meanwhile I have absolutely no interest in vPro. Its irritating that Intel has chosen to segment the CPUs this way.
The X10SAT manual explicitly lists support for AMT -- doesn't AMT support remote KVM?
http://software.intel.com/sites/man...ult.htm?turl=WordDocuments/kvmandintelamt.htm
What are we referring to in this thread when we say IPMI, and how does that differ from the KVM features of AMT?
No idea on the first.
IPMI is what this Motherboard got, a dedicated NIC that is intended to be used exclusively for remote management purposes. They seem to all have an ASpeed controller (That one is an AST2400) that does roles of both IGP with its own 2D GPU and VRAM (Reason why you don't need a discrete Video Card even if you use a Xeon without IGP. But on that Chipset, you can't use the Haswell GPU anyways), and also of a NIC itself. It can route video output even from boot time, so you can actually enter and configure the BIOS remotely.
Notice that it seems to be a common sight on C222 and C224 Motherboards that can't use Haswell own GPU, but I think that I didn't saw C226 with IPMI. I don't know if they can route Haswell GPU video output to the IPMI NIC, or that is the reason why they don't do it.
What do you think about using one on the cheaper mATX Supermicro C222 or C224 boards and putting a low power 7750 with passive heatsink? The 7750 use only the PCIe slot for power and have very low idle power usage. This would use up a slot on the board, but would also allow you to use a cheaper Xeon with no IGP. Do you think IPMI would still work that way as well?
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 20K
-
-
-
Facebook
Twitter