ASRock Rack Launches Xeon D Motherboards

Intel’s Xeon D SiP (System-in-package) has turned out to be one of the exciting launches this year in the server CPU space. We have already analyzed Xeon D in detail in our review of the Supermicro SuperServer 5028D-TN4T. Almost all currently available Xeon D systems / motherboards are from Supermicro, but we now have another set of options from ASRock Rack.

The Xeon D family currently consists of two members:

• Xeon D-1520 : 4C/8T Broadwell-DE x86 cores @ 2.2 GHz, 6 MB of L2 cache, 45W TDP
• Xeon D-1540 : 8C/16T Broadwell-DE x86 cores @ 2.0 GHz, 12 MB of L2 cache, 45W TDP

ASRock Rack’s Xeon D lineup consists of one board using the Xeon D-1520 and six boards using the Xeon D-1540. Customers have the option of going with either the mini-ITX (mITX) form factor or the micro-ATX (uATX) form factor. The mITX boards are all compatible with 1U rackmount chassis.

Gallery: ASRock Rack Xeon D Lineup

In addition to the motherboard size, the differentiation aspects come in the form of support for different varieties of LAN ports, PCIe slot configurations, additional storage ports using the LSI 3008 HBA and different USB 3.0 port configurations. Unlike the mITX boards, all the uATX boards come with a COM port in the rear I/O.The following tables summarize the features of the various products in the ASRock Rack Xeon D lineup.

mITX Boards

uATX Boards

These boards are ideal for network and warm storage devices as well as micro-servers. Given the low power nature of the Xeon D platform, some of them can also be useful in home lab settings for experimenting with virtualization or even act as boards for high-end development machines.

Intel Launches ‘Greenlow’: C236 Chipset and Skylake E3-1200 v5 Xeons

One of the recent notable aspects about the Xeon E3 family was the lack of a significant presence with Intel’s Broadwell architecture. A handful of them were released under the E3 v4 nomenclature, with half of them only for embedded solutions – we reviewed the socketed models at the time, but noted through IDF that the main direction these models were going seemed to be towards Intel’s Valley Vista PCIe co-processors, where three of them would be fixed to a PCIe card. For most single socket workstation users, this means a jump from Haswell-based E3 v3 to Skylake-based E3 v5. 

In a somewhat muted launch today, without any PR announcement or anything in the way of details direct, Intel quietly launched details regarding their new workstation single-processor platform based on Skylake. Through Intel’s processor information repository, Intel ARK, eleven socketed models were released in the E3-1200 v5 Xeon family. As with previous Xeons, the goal here is to supply server level features that wouldn’t normally be on consumer grade hardware. All these parts support DDR4 un-buffered modules in both non-ECC and ECC configurations up to 16 GB per module, giving 64 GB maximum. Xeon platforms historically run at base JEDEC memory configurations, so expect to see DDR4-2133 at 15-15-15 subtimings being the norm. Similar to the Skylake consumer parts, DDR3L up to 1600 MT/s is also supported.

These are all quad core parts, ranging from 25W low power models at 2.1 GHz base clock (with hyperthreading) up to up to 80W for the fastest 3.7 GHz model. Prices will range from $193 for the E3-1220 v5 up to $612 for the E3-1280 v5. As shown in the table, only a select few will have integrated graphics, and those that do will be under the HD P530 designation, which if we refer to the standard Skylake launch corresponds to a GT2 configuration with 24 execution units. At this point in time, there are no dual core parts.

ServeTheHome’s Table of Xeon Processors is a great representation of the offerings

Several parts clearly fly out as interesting – the E3-1275 v5 comes in at just over half the price of the E3-1280 v5 for only a 100 MHz deficit on the base clock. For $9 less, the E3-1270 v5 does the same as the E3-1275 v5 but without the integrated graphics. The one below that, the E3-1260L v5, is one of the low power parts at 45W but still comes close in single threaded frequency to those above it. One of Intel’s big comparison points to these parts will be with Xeon D-1540, also known as Broadwell-DE. This is a 45W part that offers eight cores at 2.0 GHz base/2.6 GHz turbo, with similar DDR4 support but 12MB of L3 cache ‘at $581’.

Users interested in the perennial alternative to the Core processors, the E3-123x range, might come up empty handed this generation. Not necessarily due to the processor itself, but multiple sources in discussion or published at ComputerBase are concluding that Intel is further locking the Xeon processors down for Skylake. This means that in order to use a Xeon processor, you need a server based chipset, such as the C230 series, rather than the 100-series found in consumer platforms. Several companies have confirmed that a ‘workaround’ in this case is most likely not possible.

This new segregation tactic shows that Intel is perhaps trying to adjust any overlap these two markets may have had for non-business users, putting more power into the hands of the OEMs that develop pre-build systems. While E3-1200 based server motherboards have not been difficult to source in the past, they tend to be workstation/server focused and not have bells and whistles that accompany consumer boards. This also removes any thought of potential overclocking that may/may not be overlooked in some Xeon parts in the past. That being said, we have seen ASUS/ASRock release workstation type motherboards in the past that still have consumer trimmings.

Addition 10/21: The quote direct from Intel is that ‘we do not validate or support Xeon E3 SKUs with client chipsets’. This has always been the case in the past so nothing on the official line has changed, and at this time Intel is not making any additional comment to the adjustment in compatibility.

The new Greenlow platform (E3-1200 v5 Xeon plus C236 motherboards) will look like the consumer counterparts with sixteen PCIe lanes for co-processors, dual DDR4/DDR3L memory support, DMI Gen 3.0 to the PCH, a mix of PCIe 3.0 lanes from the chipset, support for up to eight SATA 3.0 ports, and a myriad of possibilities through 26 lanes of HSIO.

At this point in time the launch seems to be biased towards workstations rather than anything 1P server related, but information is currently patchy. At this point in time we are looking to get sampled with some of these processors, along with appropriate motherboards as well. Some motherboard announcements have been made from GIGABYTE Server and ASRock Rack, and I suspect workstation OEMs will come forth in due course.

Source: ComputerBase, ServeTheHome, Intel ARK

QNAP’s TVS-871T – A 10G NAS with Thunderbolt 2

QNAP operates in SMB / SME NAS space and differentiates itself by providing consumers with plenty of choices when it comes to the hardware platform. Over the last year or so, they have been introducing a range of NAS models with 10GbE capabilities. Video production houses are one of the main targets of these 10G NAS units. In order to better serve that particular market, QNAP has introduced the TVS-871T – a 8-bay NAS with dual 10GbE ports as well as two Thunderbolt 2 ports.

The TVS-871T is a tweaked version of one of their premium 10G NAS units – the 8-bay Core i5/i7-based TVS-871. The TVS-x71 series has two free expansion slots, and the TVS-871T has been created by just adding a Thunderbolt 2 PCIe card to one of the free expansion slots. The other slot is taken up by the 10G card.

Obviously, the more challenging aspect is the support for Thunderbolt 2 in the QTS operating system itself. On the QTS side, the Thunderbolt connection is made to appear as an ad-hoc network link. This enables the TVS-871T’s triple play feature – a direct-attached storage (DAS) unit using the Thunderbolt 2 links, a high-speed NAS unit using the two 10GbE ports and an iSCSI SAN over the same network ports. In addition to the TVS-871T’s i5- and i7-variants, QNAP also has expansion enclosures designed to work with them – the TX-800P and TX-500P. These expansion enclosures also come with dual Thunderbolt 2 ports for easy daisy-chaining.

Given the variety of connections (multiple network ports, including 10GbE ones, as well as the two Thunderbolt 2 ports), there are many possible usage scenarios. QNAP talks about four main ones, which are outlined in the gallery below.

Gallery: QNAP TVS-871T Thunderbolt 2 NAS – Use-Case Scenarios

The hardware specifications of the various models being announced today are summarized in the table below.

QNAP’s TVS-871T has actually been doing the rounds in various trade shows over the last couple of months. Today, we are having the official launch in the North American market along with pricing details. The TVS-871T-i7-16G-US has a manufacturer suggested retail price of $3,199 while the TVS-871T-i5-16G-US comes in at $2,799. The two Thunderbolt 2 ports seem to be carrying a hefty premium over the vanilla versions (around $1000). However, the prices are reasonable when one looks at the cost of Thunderbolt SAN setups that can be replaced with the TVS-871T. Looking forward, the arrival of Thunderbolt 3 with the USB Type-C connector is sure to make such NAS / DAS / SAN triple-play units popular.