Author: Carl Poirier
Editor: Howard Ha
Publish Date: Thursday, August 4th, 2011
Originally Published on Neoseeker (http://www.neoseeker.com)
Article Link: http://www.neoseeker.com/Articles/Hardware/s/asus_e35m1_i/
Copyright Neo Era Media, Inc. - please do not redistribute or use for commercial purposes.
Just because the Llano launched a a little while ago doesn't mean its little brother, Bobcat, is not good anymore. Comparing the TDP of each APU shows that Bobcat still has nearly less than four times the power footprint. Whereas it would be pretty difficult to run a 65W Llano APU without any fan attached to its heatsink, the 18W demanded by the older Bobcat chips is pretty reasonable.
Where does the Bobcat's lower TDP come in really handy? Consider there are many motherboard vendors shipping all-in-one board solutions with an embedded CPU (Bobcat chips included) and only a passive heatsink to cool it. Such all-in-one solutions have proven deal for HTPCs or any other application where the lower the noise, the better. Paired with an SSD and a fanless power supply, and you could build yourself a near silent system where there is absolutely no way visually to tell if the system is turned on except from checking if there's anything being displayed on the montior it's hooked up to.
ASUS' name often pops up when looking for innovative solutions, and we have one such product today; the fanless mini-ITX board based on AMD's Brazos platform which includes many special features typically found on the higher-end boards, such as the Turbo Key II and MemOK! switches, and of coure all usual features expected for a board aimed at the HTPC market.
|CPU||AMD CPU on Board; Fusion APU E-350 Dual-Core Processors|
|Chipset||AMD FCH A50 (Hudson M1)|
2 x DIMM, Max. 8 GB, DDR3 1066/800 Non-ECC,Un-buffered Memory
Single Channel memory architecture
* Please refer to www.asus.com or user manual for Memory QVL.
|Expansion Slots||1 x PCIe 2.0 x16 @X4, with latch|
Integrated AMD® Radeon HD 6310
Supports DVI with Max. resolution: 1920X1080 @60 HZ
At first glance, the E35M1-I Deluxe seems to pack a lot; the PCB is crowded with components of every kind. The large fanless heatsink is pretty impressive with its three heatpipes, which extend awkwardly outside the case via the I/O panel. The expansion slot comes in the form of a full-length PCI-E wired at x4. The memory slots also accomodate desktop-sized DIMMs.
The cooling system is solidly attached with six metal screws. The back of the E35M1-I Deluxe reveals that the PCB is brown, since it's not hidden under a plethora of components.
The E35M1-I Deluxe leaves a strong impression with an internal USB 3.0 header and a Turbo Key II instant overclocking switch. The USB 3.0 controller is slightly hidden under the heatsink and can still be seen if you look hard enough. There is also a 4-pin CPU power connector and a CMOS clear jumper.
Even more impressive is the inclusion of the Energy Processing Unit (EPU) which yielded great results on full-sized desktop motherboards, such as Neoseeker's AM3 test platform, the M4A89GTD PRO/USB3. The motherboard battery is hung vertically, closely to the heatsink.
ASUS does not stop there, as the MemOK! button is present. A mini-PCIe slot is also included, with a WLAN card already pre-installed into it. Its antennae run up to the I/O panel, as will be shown later. The front panel and chassis intrusion header are also located in this area, but unfortunately the former is not the same shape as the one present on most ASUS motherboards compatible with the Q-Connector.
Further to the left are two USB 2.0 headers, five SATA 6Gbps ports courtesy of the Fusion Controller Hub, the front panel audio header and the S/PDIF output. The small green LED indicates whether or not power is supplied to the board.
One will find six USB ports at the back of the board. ASUS has packed a second USB 3.0 controller to speed up two of them. The video connectivity is handled by a DVI and HDMI port; that's right, there is no VGA. This isn't really a terrible decision, as pretty much all display adapters now should support its digital counterparts instead. In fact, many manufacturers including AMD plan to drop support for analog video interfaces by 2013. There is a PS/2 port for older mice and keyboards, though. The analog audio connections have also suffered, as the board only supplies three. This is just enough for a 5.1 Surround system, but that doesn't leave any room for a microphone unless that goes through the front panel I/O. The onboard Realtek ALC892 can still offer 7.1 Surround sound via optical S/PDIF though. Lastly, the networking options include Gigabit Ethernet, a built-in Bluetooth adapter, and the standard connectors for both Wifi antennae. One can also admire the useless heatpipe ends going through the I/O shield.
The packaging includes two SATA cables, the I/O shield, a user guide and driver DVD. There are also some oddly shaped Wifi antennae.
These antennae connect to the back of the board through an 80cm long wire. The rod in the middle tilts upwards so that the circle part works as the base, which has some small magnets for attaching onto a steel surface. Of course, should these prove to be overkill for one's system, they can always be swapped for the more common black foldable antennae.
Neoseeker has already tested offerings from Gigabyte and ECS in the Brazos department, and our contender from ASUS today will be pitted against both of these. One important thing to note here is that Neoseeker's standard test platform has been moved to a 2x4GB memory configuration compared to the previous reviews. The latency of this kit is back at CAS 9 instead of the CAS 8 used in the previous set. This does increase latency and decrease bandwidth a tiny bit according to SiSoft Sandra, but there weren't any real impact on other benchmarks outside of the PCMark series which tends to favor to more RAM.
In our previous Brazos board review, Neoseeker was not able to see the true potential of Bobcat, as the only overclocking feature we got to play around with was a "Turbo Mode"; this simply proved to be a predefined overclock. Though this feature delivered good results for Bobcat, there was definitely room left for improvement. The good news is that ASUS E35M1-I Deluxe has all settings you'll need to bring the best out of the Bobcat chip, so hopefully these will show the true capabilities of board's embedded processor.
The first overclocking feature to try out was the Turbo Key II switch found directly on the motherboard, but this only pumped the reference clock up a mere 5MHz. That's rather disappointing.
Firing up the usual stress test with the APU at stock settings without any fan attached to the stock heatsink saw the die temperature rise over the maximum 90oC recommended by AMD, so for overclocking purposes a Tornado 92mm fan undervolted to 5V was placed over the board. It will be interesting to see if the die will continue to reach such high temperatures during normal use.
Whatever hope we had in going the full manual route for overclocking was dashed when all that could be achieved was a disappointing 18% overclock on the reference clock, way below the 33% boost seen with ECS' Turbo Mode, and that's with increased voltage settings on the ASUS board. What a shame!
Here's what we played around with:
For the final test, the OCCT power supply benchmark was run so that both the x86 cores and the IGP were under full load in order to both stress and maximize heat output.
Here are the final reference clocks achieved by each board:
This program includes benchmarks for most hardware. The CPU arithmetic and multi-core efficiency benchmark will be run as well as memory bandwidth and latency.
The E35M1-I Deluxe lands in the middle of this crowd, however there are some odd results that would show up run after run, like the memory latency of the Gigabyte and the memory bandwidth of the ECS board. One thing that's certain is the E35M1-I Deluxe's 18% overclock was not nearly enough to reach the HDC-I's performance, even with the faster memory multiplier.
HandBrake is an application that converts sound and video files to other formats. It makes use of all available threads so it can exploit the processor to its full potential. The input video is a 1:48 mp4 file coming from the Lord of the Rings, in 1080P. The file has a size of 96MB and it's being converted to the mkv format.
At stock, the ASUS board finishes last. This time around however, the overclocked settings help it surpass the HDC-I.
POV-Ray, for Persistence of Vision Raytracer, is a 3D rendering software that has impressive photo-realistic capabilities.
In POV-Ray, it also ends up at the very bottom, and once overclocked it was still behind the HDC-I.
7-Zip is a compression program, much like WinRAR. It features a built-in test, which gives a score for compression and decompression.
7-Zip shows the same rankings as POV-Ray; today's subject ends up last and the overclocks are definitely not comparable.
Cinebench 11.5 is another rendering program supporting an insane amount of threads. The image is processed by chunks, each running on a particular thread. It also includes an OpenGL test that puts the graphics card through its paces.
Cinebench puts all boards on the same level when it comes to the CPU test. The ASUS board is at the bottom once again in the OpenGL test.
3DMark Vantage is the stunning sequel to 3DMark 06, which will also be run. Futuremark's benchmarking programs have always been at the center of every bragging match; the best way to show that one has got the greatest gaming rig is to have the highest 3DMark score. These benchmarks put systems through a series of strenuous tests, and provides the user with a score to brag about!
The 3DMark benchmarks follow the same trend as before. PCMark however sees the HDC-I place last.
PCMark is similar to the 3DMark suite, except that it includes many other tests like hard drive speed, memory and processor power, so it is considered a system benchmark and not just a gaming benchmark.
This time, it's the HDC-I that gets the crown.
HDTune is a benchmarking program for hard drives. Their speed also depends on the chipset so this is why the read speed test is ran.
Today's subject positions itself in the average. The HDC-I yields weird scores, though.
Street Fighter IV from Capcom has a free benchmarking tool that will be used to test the HD 4200 capabilities. World in Conflict is developed by Massive Entertainment and it also has a benchmarking tool, accessible in the game settings. Both games will be tested at 1280x1024. For World in Conflict, the graphics details will be set to "Very Low", and for Street Fighter IV they will be at the lowest settings.
Street Fighter IV delivers the same scoreboard. The World In Conflict test unfortunately didn't prove precise enough to show a difference between the boards at stock.
The popular, free media streaming website now hosts 720p as well as 1080p content, which happens to be quite processor intensive. Fortunately, Flash offers GPU acceleration since version 10.1 to offload work onto the GPU, so that's why the CPU usage will be measured to gauge the hardware acceleration and processor performance.
The ASUS offering is in the middle of the pack here. Overclocked, it finally manages to beat the HDC-I.
With the Unified Video Decoder 3 built into the HD 6310, ATI drivers can of course accelerate the playback of Blu-Ray content. Armed with an ASUS Blu-Ray reader and the bundled Cyberlink PowerDVD 8 Blu-Ray Edition, the E35M1-I was put to the test with Incendies.
It did so without any hiccup, and without any fan at all. The normal air convection alone allowed the system to stay below the 75oC mark, which is well below the 90oC maximum temperature ceiling recommended by AMD. This shows that the cooling system developed by ASUS is quite adequate. The average CPU load was also reasonable at around 33.5%, so the platform is definitely well adapted for this type of use.
To measure power usage, a Kill A Watt P4400 power meter was used. The following numbers represent the power drain for the entire benchmarking system, not just the video cards themselves. For the 'idle' readings, the power drain from the desktop with no applications running was measured; for the 'load' situations, the average power consumption was taken while running the OCCT power supply test, stressing both the video card and processor, for a couple of minutes.
The results of the Energy Processing Unit were quite a sight. Under full load, the board consumes a whole 5W less! This is approximately a 10% reduction, which really is a lot. It did not do any better than the competition at stock settings, however.
As with most of its offerings these days, ASUS has not skimped on the feature set of the ASUS E35M1-I Deluxe. For the connectivity, the E35M1-I Deluxe has an integrated Bluetooth module and a a half-height wireless LAN card pre-installed into the mini-PCIe slot , leading to two exterior antennas. They are not the usual small black antennas found on the vast majority of wireless adapters and routers though; they have long cables and are terminated in small sticks having a round base to help these antennae stay straight up. There are also two external USB 3.0 ports and an internal header for fast transfers with compatible drives.
For stability and overclocking, the E35M1-I Deluxe features the MemOK! button that ensures the RAM is configured at the right settings for a headache-free boot. There is also the Turbo Key II switch, but as great as it sounds, it was pretty useless compared to what we've already seen on previously reviewed Brazos boards. As it turns out, ASUS' Turbo Key and OC Tuner both pumped the reference clock a mere 5MHz, whereas the Turbo Mode on the ECS HDC-I increased clocks by a hefty 33MHz, six times the improvement in overclocking. What's worse is that going the manual route wouldn't even allow one to get close to the 33% overclock demonstrated by the ECS solution. When a manual overclock cannot even reach a pre-defined setting of a competing board, you know there's a problem. There's some consolation in how the E35M1-I Deluxe uses UEFI instead of a standard BIOS, making for a much more user-friendly experience when manually overclocking.
Performance wise, the E35M1-I Deluxe performed just a tiny bit below the other Brazos boards and in real world tests this difference will not be really apparent; in most cases only synthetic benchmarks could detect it. However, the thing that makes the E35M1-I Deluxe stand out the most is its fanless design, allowing for a completely silent setup. Be careful how far you wish to go in going silent though, because while the E35M1-I Deluxe can cope with normal usage, its die components can get way too hot when running a stress test program without any fan. Any airflow, be it very small, did keep it cool enough.
At $175, the E35M1-I Deluxe is not very gentle on the wallet, however it has loads of features that make it worth it. It's just too bad its overclocking results were so poor, as this would only have made the all-in-one solution even better.
Please do not redistribute or use this article in whole, or in part, for commercial purposes.