Thread: Anandtech News

A large number of high-end enthusiasts and prosumers are expecting a lot from the next enthusiast level upgrade. While Haswell on the mainstream desktop has been here a year, Haswell-E is still a number of months away. Computex seems to have shown a few X99 chipset models, the chipset designed for Haswell-E, and it looks set to be a significant upgrade in terms of functionality. Most of the major motherboard manufacturers had something on display for X99 – some were out in the open whereas others were behind closed doors and we can’t talk about them. EVGA was happy enough for us to take images of their X99 demo sample, which is almost final in terms of functionality but still needs some work on the aesthetic look.

For X79, EVGA came out with the X79 Dark, the X79 SLI, the X79 FTW and the X79 Classified. This model definitely looks to be in line with the Dark, as that was the only model with a full complement of DRAM slots, but the color styling is more the regular EVGA style.

Due to the size of EVGA, the functionality on their motherboards is not in the same area as an ASUS, GIGABYTE or ASRock, but they do attempt to be versatile, especially with the PCIe layout. There the GPU layout should be suitable for two 3-slot cards, three 3-slot cards or four dual slots cards. Users can adjust the air gap in a dual card setup to fit their needs.

The board does come with dual NIC (Intel I218-V I believe), 10 SATA ports, PCIe switches and extra power to the VGA comes via a 6-pin PCIe connector. EVGA also put their 24-pin ATX power connector at right angles, and is the only company to go down this route for X99 so far. The heatsinks in these images are just for function, but will be replaced with the EVGA branded ones come release. Jacob from EVGA expects the motherboard to be available for X99 launch, which is always a good thing.
Gallery: Computex 2014: EVGA Shows X99 Prototype

More...

Over the last year or so, Zotac seems to have reduced its motherboard output in focus of the mini-PCs. It explains why we have no Zotac Z87 or Z97 motherboard on show! However the mini-PC business seems to be doing well for Zotac, especially in the digital signage sector so I hear, and we have reported on some interesting new models in our Pipeline feed over the year. There were two particularly interesting items on display at the Zotac booth this year at Computex.
ZBOX Sphere
Despite a sphere not being box shaped, the ZBOX line recently got a new member in the Sphere. The model on display at the booth was based on the i5-4200U mobile processor with support for two SO-DIMM modules of 8 GB each.

The Sphere is designed to light up around the rim, and I believe Zotac is planning to develop different lighting strategies. The design opens itself up to some visual modifications down the road – I instantly thought of a Homer Simpson model.

The rear is where all the connectors are, including a 19V DC-In, four USB 3.0, two USB 2.0, Ethernet, HDMI, DisplayPort, an SD card reader and front panel audio connections. The system also includes built in WiFi, but it is a single stream 802.11ac solution. It does allow dual band usage, but I think dual stream is now becoming the standard, so I hope this is upgraded before the Sphere comes to market in volume. Storage comes from a single mSATA port slot in the device.

ZBOX CA320 nano
Perhaps a little exciting was the fanless AMD mini-PC also on display. The system uses an 8W quad core Jaguar/Temash APU (the A6-1450, 1 GHz stock, 1.4 GHz turbo) which is cooled by the bulk device itself.

Being a Temash APU, there is only a single channel of memory which means a single SO-DIMM can give a maximum of 8 GB. Storage is via a 2.5” SATA 3 Gbps SSD, with HDMI and DisplayPort video outputs.

The integrated card reader also has 802.11ac WiFi included (no word on the chipset for that), but the chassis has a large number of USB ports to additional functionality.

Gallery: Computex 2014: Zotac’s ZBOX Sphere and AMD Fanless Mini-PC





More...

I remember sitting in a briefing with Samsung last year when the company first started talking about translating its success in the Android smartphone space into the tablet market. Samsung has done well in the Android tablet space but it’s safe to say that the company is better known for its phones. This year Samsung hopes to change all of that and is putting its most valued mobile sub-brand (or letter) to work in tablets. Later this month (globally) and starting next month in the US, Samsung will begin selling its Galaxy Tab S line of premium tablets. Read on for our take and hands on experience.

More...

When a power supply becomes sufficiently efficient, it loses less energy from the input as heat. Given the right conditions, a very efficient and low rated power supply can be built without a fan, giving a quieter experience. The limit is always a balancing act between efficiency and seeing how highly rated the PSU can be – there is a certain segment of the market that will always prefer the quiet systems where possible. Up until this point, the peak rating I have used for a fanless PSU was 500W, from the Rosewill Silentnight range that were 80 PLUS Platinum rated. Even though Titanium rating has become a thing which might help push the limits even further in the interim, more fanless Platinum models are coming to market. SilverStone showcased a 520W Nightjar model and Enermax showed off a prototype 550W digital fanless version.
SilverStone 520W Nightjar
Despite the word ‘night’, this all-white PSU will come in 400W and 520W variants and will be fully modular. It seems like it will have at least four PCIe 6-pin connectors, and I assume some or all of which will also be 6+2.

Due to the fanless design, it must be oriented in a specific way to maximise the airflow. All the power is on a single 12V rail, with the unit rated up to 516W and 43A, which should satisfy all single GPU devices out in the market at stock speeds.

I might have to take issue with the ‘zero dB’ claim, given that electrical noise is around 15 dB. In the past I have also seen silent power supplies with significant bulk to help radiate temperature away, whereas this unit seems fairly empty.

From previous experience 500W fanless Platinum power supplies are not entirely cheap, and the Nightjar has recently gone on sale for $160.
Enermax 550W Digifanless
Digital power supplies have had a rough entry into the market. They add a good amount of cost to a unit, but few users interact with their power supply on such a scale to monitor and adjust the unit for best efficiency and performance. With all that being said, I have not yet come across a fanless digital power supply, and Enermax believes there is a market for one.

Unfortunately they did not have an individual unit on show that was not in use, but the design is fully modular and the system is rated at 550W. I did ask about the rating, and it turns out that 550W is a peak value, whereas the average value is more around 500W. Perhaps I am naïve in power supply marketing, but I had assumed that the wattage listed was the average due to issues in the previous decade.

At this point I believe the Digifanless is still in the prototype stage and will be coming to various markets later in the year.
Gallery: Computex 2014: Fanless PSUs from Silverstone and Enermax

More...

One of the big launches this year will be the Haswell-E platform. It is pretty much common knowledge in hardware discussions that this means Haswell-E, X99 and DDR4 will effectively launch to the consumer on the same day. One element of that link that we have the fewest leaks and information from is the DRAM side. I explicitly asked the DRAM manufacturers I have most contact with if they would be showing any DDR4 – either modules, specifications or in action. Many of them obliged – it helps that a fair number also make SSDs and so Kristian was able to get a few snapshots. But talking to GeIL at Computex also revealed something a little more interesting: CAS Latencies.

Memory, as I attempt to convey in reviews, is more than just the frequency stated. There are several sets of subtimings associated with the memory, often divided up into primary subtimings, secondary and tertiary. The primary ones are the most important, and arguably one of the most important of those is called the CAS Latency, or CL for short. Memory is usually quoted as a combination of frequency and CL, and in the past it has been easy to compare kits by comparing the results of Frequency divided by CL, such that the kit with the higher result is often the better performing.
Moving from DDR3 to DDR4 means a move from 280-pin to 288-pin connections. The layout of the modules is slightly different, with the pins being different lengths (notice in the image above how the pins in the middle are longer than those at the edge) in order to help installing memory. Voltage moves down from 1.5 V to 1.2 V, and processors are expected to support DDR3-2133 by default.
Most manufacturers on the Computex show floor were coy with what sub-timings they will aim for their modules, whereas GeIL had a handy list:

So when GeIL offers below 2133 MHz, I am a little taken aback, because their answer is ‘because some people will use it’. I hope that the reason they will use it is because the processor will not support above 1600 MHz. But this list gives that clear indication of CAS Latency between 1600 Mhz and 2400 MHz, with indications that 2666 MHz and 3200 MHz might be more common than we think.
CAS Latency for DDR4 does seem a little down compared to DDR3, though this might be due to the initial batches of ICs coming through. I remember 2133 C11 being the first 2133 MHz DRAM off of the shelf, but now we can buy 2133 C8 very easily.

More...

With the new overclockable Pentium processor soon coming to market, named as Pentium-AE and listed under the title ‘Pentium G3258’, it offers an opportunity for some low cost overclocking and perhaps a nice gaming system or two. In light of the low cost nature, it would perhaps make sense if motherboard manufacturers brought out new models designed for this new CPU to help hit the price conscious but also offer some overclocking features. ASRock believe their new models, the ATX-sided Z97 Anniversary and micro-ATX sized Z97M Anniversary, might be the answer.

These motherboards are particularly stripped down, as seen by the rear panel in the picture above, with only one video output and no additional controllers for USB 3.0 or SATA 6 Gbps. There is no M.2 or SATA Express, and given the potential for the lack of GPU scaling due to CPU power of the dual core Pentium, there is only a single PCIe 3.0 x16 slot on both versions.

Both new models use low end audio codecs: an ALC887 on the full-sized version and an ALC662 on the micro-ATX. This latter codec is more often found in notebooks, for example. The ATX model also offers an Intel NIC, compared to the Realtek on the Z97M.
The Z97M model seems more of a low cost play, whereas with the Z97 Anniversary there is better provision for power delivery. Both models are still focused on that sub $120 market (perhaps even lower), but ASRock did give data showing that the Pentium-AE processor can hit 4.5 GHz in the Z97 Anniversary.

Unfortunately I was too rushed to get a hands-on with both models, also because my attention was caught with the X99 products on show. But if we get them in to review we will let you know how they perform.
I would expect both models to be available at retail by the time that the Pentium G3258 CPUs come to market.

More...

While NASs are more of Ganesh' expertise, I dropped by the biggest NAS manufacturers at this year's Computex as Ganesh was unable to attend the show. First up is ASUSTOR who introduced a new 7 Series NAS for the high-end market.
The 7 Series NAS is mostly aimed towards the enterprise and comes in two form factors: rackmount and desktop tower. The rackmount model is available in two sizes with the smaller one being a 9-bay and the bigger being a 12-bay solution. Both version come with either an Intel Core i3-4330 or Xeon E3 processor depending on the customer's choice. The default configuration ships with 4GB of DDR3 but that is expandable to up to 32GB. Connectivity wise there are four Gigabit Ethernet, four USB 3.0 and two eSATA ports for additional storage and two HDMI ports along with a VGA port for display connectivity.

The desktop tower version drops one bay at the low-end and two at the high-end and thus comes in 8-bay and 10-bay configurations. The processor is the same i3-4330 but there is no option for a Xeon CPU and the RAM changes from UDIMMs to SO-DIMMs with the default capacity being 2GB (up to 16GB). Gigabit Ethernet ports drop to two as well but there are still three USB 3.0, two USB 2.0 and two eSATA ports. For video output there are two HDMI ports and one audio output.

ASUSTOR also had their other NAS lineup at their booth. The 6 Series is available from two to eight bay configurations and ships with a 2.13GHz Atom CPU along with 1GB of DDR3.

Moving to the mainstream models, the 3 Series comes in two and four bay options and ships with a 1.6GHz Atom CPU and 1GB of DDR3. Similar to the higher-end models, the 3 Series has an HDMI port, making it viable as a media player as well with 1080p playback support.

And finally the 2 Series. This is the low-end series with no video playback support, so it's suited for people who just want a regular NAS with no added features. Two and four bay models are available and the CPU is a 1.2GHz Atom with 512MB of DDR3.

ASUSTOR was also showing the new ADM software. This is the HTPC view with the desktop mode being a bit more iOS-like. The new 2.2 version brings Mail Server and Photo Gallery apps for instance and there are over a hundred apps in ASUSTOR's App Central, which are designed for NAS use.

More...

Continuing with NAS coverage from Computes, QNAP had a few new products to show, including the new HS-251 silent NAS. However, I spent most of my time looking at their software and especially virtualization because that was a feature QNAP highlighted heavily, mostly because they now support virtualization even in lower-end models.

The virtualization window is very simple to use and the software supports imported VMs from most virtualization software. Exporting is also possible in case you want or need to move the VM to another machine. There is also a snapshot feature that allows you to take a snapshot of the system state and then return to that state later in case a user manages to break something for instance.

Permissions can be set for every user and VM individually to ensure that the users only have access to what they need.

Overall the software is very app-centric and more apps can be downloaded from QNAP's App Center. Cloud backups are of course supported as well with several service options.
And now, the hardware. Another thing QNAP emphasized was their use of Intel Celeron CPUs instead of Atom that most of their competitors use. I think the specific SKU in this case is a Silvermont based Celeron J1750, which should provide far better performance than the old Bonell based NAS-specific Atoms. QNAP said that this allows VMs to be run even in lower-end models, which is something that Atom doesn't support (at least not that well). Obviously you won't be able to run several VMs at once like in the pictures above but one VM should still be enough for home or even small business use. Moreover, Celeron CPUs are powerful enough to do transcoding on the fly, which can be beneficial in case you happen to use an Apple TV or other format picky media player.

The HS-251 is successor to the popular HS-210 silent NAS. It features two 3.5" hard drive bays and is passively cooled, meaning that the only noise will be coming from the hard drives (or no noise at all if you are crazy enough to go with SSDs). It features the same Celeron CPU as the TS-x51 series along with 1080p video and 7.1 audio playback.

For connectivity there are two Gigabit Ethernet, two USB 3.0, HDMI and optical audio ports. I think the HS-251 should be a great fit for people who are looking for a NAS that functions as a media player because unlike with traditional NASs, there is no extra noise from the fans that could negatively impact the experience.

More...

In a move that will mark a mild shake up in how AMD operates, AMD has announced that they will be undergoing a company reorganization next month. Come July 1^st, AMD will be consolidating their various business groups into just two groups, and overseeing those groups will be Dr. Lisa Su, who will become the company’s new Chief Operating Officer (COO).
AMD is citing the reorganization as the latest step in their efforts to transform the company, a process that started in earnest over two years ago in 2012. Since then the company has been making changes to move away from its traditional cost-heavy PC CPU and GPU roots and towards a structure that is focused on mobile (x86 and ARM), semi-custom silicon, and other market areas with lower margins but also lower costs that are more sustainable for a company of AMD’s size and capabilities. AMD is nearing the end of that transformation – after years of losses they’re now approaching profitability at their desired margins – with AMD realigning their business groups ahead of some of their final steps, including becoming a fully ambidextrous company through designs such as the K12 CPU.

As part of that general transformation AMD’s business groups have already begun to overlap some, so now AMD is taking the next step by making it official and consolidating the relevant groups. AMD’s client, consumer graphics, and professional graphics groups will now be combined under a single group, the Computing and Graphics Business Group. By bringing together those three groups like this, this change effectively consolidates all of AMD’s core technology teams in to the same group, CPU and GPU alike. In this case in particular the lines between CPU and GPU have already been blurring for some time, with the bulk of AMD’s “CPU” business having shifted to APUs (CPUs with integrated graphics), so in a sense this is the formalization of the fact that AMD cannot build complete CPUs without technology from their graphics group.
Meanwhile AMD’s second group will be the Enterprise, Embedded and Semi-Custom Business Group. This group consolidates the server, embedded, and semi-custom groups under one roof. At this point it’s not clear quite where the line is drawn on the server CPU side – is an Opteron sale a Computing group sale or an Enterprise group sale? – but otherwise this marks the combining of AMD’s “fringe” groups such as SeaMicro and the semi-custom groups, which in contrast to the core technology focused Computing group are focused on building designs and applications around AMD’s core technologies.
Both of these groups will also see their relevant sales appendages also integrated into the larger group. AMD currently has a separate sales group, which will no longer be the case after the reorganization.

Heading up these groups both directly and indirectly will be Dr. Lisa Su, who is getting a promotion from Senior VP and GM of Global Business Units to the C-level position of Chief Operating Officer (COO). AMD has not had a COO for a few years now, so this marks the return of that position to AMD’s executive organization and arguably makes Lisa AMD's second-in-command. Meanwhile AMD’s Chief Sales Officer, John Byrne, will also be getting a promotion of his own, which will see him move up to SVP and GM of the Computing group.
In regards to AMD’s new structure, Lisa will be taking direct control of the Enterprise group on an interim basis. Meanwhile Lisa will have indirect oversight of the Computing group, with John serving as GM of that group and reporting to Lisa. Lisa in turn will now report directly to CEO Rory Reed.
Ultimately the consolidation of AMD’s businesses is not unexpected, especially on the core technology side where APUs and AMD’s HSA initiative has greatly worn away the distinctions between CPUs and GPUs. Meanwhile the shifts in leadership bring with it new mangers and new reporting structures, so although things will be changing at AMD it doesn’t sound like AMD’s development processes will be affected on the whole – though management shifts often come with smaller internal changes.
But perhaps the single most visible change from this may end up being how AMD reports their financials. Currently AMD separates their CPU and GPU businesses as the Computing Solutions and Graphics Solutions respectively, with Graphics also including semi-custom business and game console royalties. If AMD changes their financial reporting to match their new businesses then we’d be able to more easily see how AMD’s semi-custom and console businesses stack up, but AMD’s CPU and GPU businesses would be indistinguishable. AMD hasn’t commented on the matter in their press release, so we’ll have to see what they do for their Q3’14 results later this year (where the combined groups will have been in effect for a whole quarter).

More...

Earlier this year, we had our first look at the GIGABYTE BRIX Pro, along with a note about further benchmarks to follow. It was out first look at the Intel Core i7-4770R, a Crystal Well part with 128 MB of eDRAM built into the package. The part was aimed squarely at making integrated graphics competetive with low-end discrete GPUs. As a direct result of this, PC manufacturers could make compact gaming units. The BRIX Pro was even distributed at the Steam Developers Conference as a Steam Machine. In our second look at the BRIX Pro, we changed the memory and storage subsystem to better reflect the expected use case. Read on for our refreshed performance evaluation results, gaming benchmarks, performance as a HTPC and the final word on the thermal throttling issue that we encountered during our initial evaluation.

More...