Viewing page 4 of 5 pages. Previous 1 2 3 4 5 Next
The Intel NUC D54250WYK reaches all new lows in idle, and video playback. The
AC idle power is 3W lower than any previous tested system, and H.264 playback
power draw is 4W lower. The numbers don't seem like much until you consider
that they are single digits: 6W, 8W, and 9W respectively for the three states.
They represent power reductions from previous bests of 33%, 25% and 31% respectively.
Power draw during lower demand tasks is the lowest ever.
Energy efficiency drops a bit with more intensive applications but still remains
among the best ever. The vast power difference between minimum and maximum load
is impressive, 6W to 40W, reflecting both the energy savings functions as well
as the performance enhancement functions in the Intel processor.
Power draw during high demand tasks is similar to the first NUC.
This Intel NUC, along with its other recent NUC brethren (Logic Supply ML300),
outperforms all previously reviewed mini-computers by substantial margins in
every category. While a bit shy of the performance of a Sandybridge i3-2100,
it trades blows equally with an AMD FM2 A8-5600K. It clearly offers the kind
of performance expected of a modern desktop system.
The NUC is nearly on par with the Core i3-2100 desktop on the above benchmarks,
and some 10% above the AMD FM2 A8-5600K.
The NUC remains highly competitive against desktop systems with NOD32,
TMPGEnc and Handbrake benchmarks.
Intel HD Graphics 5000
As the graphics engine is essentially the same as the HD Graphics 5000 in desktop
chips, we refer you to the GPU benchmarks in our Intel
Core i7-4770K Haswell Processor review. Its 3D performance is definitely
improved over the HD Graphics 4000 in the original NUC, perhaps by as much as
25% depending on the game or application. This GPU is good enough for online
gaming, but it can't replace a discrete GPU for most modern gaming titles except
at lower resolution and detail settings. Video playback is perfectly smooth
on 1080p content including H.264 MKV/MOV and Flash (YouTube HD).
Storage Subsystem Performance
To test storage subsystems CrystalDiskMark was employed on the 1000 MB setting
with 0x00 fill test data, and a Kingston HyperX 3K 120GB solid state drive (compressible
data produces the best possible speeds out SandForce drives). The drive was
connected using an Icy Dock external dock which supports eSATA and USB 3.0 (limited
to 3 Gbps and 5 Gbps respectively). This drive was also used to test the internal
SATA data and power ports.
Kingston 120 mSATA drive on left. USB 3.0 on right.
The performance of the Kingston 120 mSATA drive (left screen) in the NUC D54250WYK
was very speedy. USB 3.0 performance (right screen) was surprisingly modest,
notable enough that several other USB 3.0 SSDs and drives were tried, all with
similar results. In contrast, on recent motherboards, sequential reads/writes
with 1M and 512K data approach 200 MB/s, nearly double that of the NUC.
The SATA port had to be tested with the cover off, as there is no way to shoehorn
a 2.5" drive into the NUC.
SATA test setup
As expected, the Kingston SATA SSD performed similar to the Kingston mSATA
120 SSD, both approaching half gig per sec speeds with 1MB reads and writes.
Neither the SATA or the mSATA port is likely to be a bottleneck for any type
of data transfer in or out of the system. The reality is that there are only
two methods of transferring data in/out: LAN and USB 3.0. Gigabit ethernet is
limited to 125 MB/s theoretical max, and the NUC's USB 3.0 peformance is slightly
lower than that.
For your interest, we did run some quick tests with the gigabit ethernet port
and found performance to be fairly typical: Bursts close to saturating the LAN
capacity, typically maintaining 70~90 MB/s average in either direction.
|Help support this site, buy from one of our affiliate retailers!|