Viewing page 4 of 5 pages.
Previous 1 2 3 4 5 Next
TEST METHODOLOGY
The focus here is primarily on the cooling ability of the Euler case, and
secondarily, on the power brick which came with the case. SPCR's usual suite
of thermal and power tools were called into play. The system's thermal and power
characteristics are recorded in idle, CPU stress and CPU+GPU stress conditions.
The extended stress conditions, using Prime95 and Furmark, are totally artificial
and over the top. Virtually no useful application actually can make components
run as hot, but it gives us a good reference by which to compare the performance
of cooling solutions.
Measurement and Analysis Tools
Before getting to the performance of the case, my impressions of this system,
using either the Intel 525 mSATA SSD or the Kingston HyperX 3K SSD: It is quick,
much quicker than you might expect of a "lowly" dual-core Pentium.
It is still a 2nd gen Core Sandy Bridge chip and retains much of the performance
of the pricier i3/i5 chips. For those who are curious, it is quicker than the
Intel NUC. There is nothing I cannot do with this system that I can do with
a Z77-chip i5-2500K ATX desktop. Except serious 3D gaming, of course, as this
system is limited to the Intel HD graphics... which does a perfectly good job
with everything else.
TEST RESULTS
The system was kept in each state for a minimum of 30 minutes. It is when the
temperatures don't change for 5 minutes that stability is deemed to be reached.
For the stress loads, stability was reached after nearly an hour. The entire
case heats up, so it takes quite a while to reach stability. For Prime95+Furmark,
the AC power began at around 50W, and it increased to 55W over time due to decreased
efficiency as power components (especially the VRM) became hotter. This is normal.
|
Akasa Euler System Measurements
|
|
State
|
Idle
|
Prime95
|
Prime95 + Furmark
|
|
CPU
|
29°C
|
54°C
|
79°C
|
|
PCH
|
52°C
|
63°C
|
91°C
|
|
VRM
|
40°C
|
54°C
|
82°C
|
|
DIMM
|
40°C
|
53°C
|
81°C
|
|
SSD
|
31°C
|
34°C
|
38°C
|
|
Case
|
30°C
|
39°C
|
62°C
|
|
AC Power
|
16.5W
|
41W
|
55W
|
|
Noise
|
none
|
none
|
none
|
|
Ambient temperature: 22°C.
CPU - highest core temperature of processor
PCH - temperature of Platform Controller Hub (motherboard chip)
VRM - temperature of voltage regulation module
DIMM - temperature of RAM memory
SSD - temperature of Solid State Drive
Case - highest temperature on external casing
|
The temperature readings come mostly from AIDA64 Extreme Edition, which matches
SpeedFan readings almost exactly. What's clear is that in any normal usage,
the Euler keeps this CPU and motherboard perfectly cool enough for stable, long
term use. Yes, at full tilt, with both CPU and graphics cores pumped to the
max, everything gets pretty hot, yet the CPU never throttled, and not once did
any kind of overheating misbehavior occur during testing. It's a short, positive
outcome after the lengthy leadup: The Euler case works as intended, cooling
the CPU and motherboard passively without making any noise. That's right, even
with my ear almost directly atop the case or the power brick, there was no noise
to be heard. OK, it's possible there was some hint from the power brick, but
it's irrelevant because the level is so low as to be immeasurable and inaudible.
The 62°C measured at the hottest spot of the external casing is too hot
to touch. In normal use, the case got warm but never too hot to touch. Changing
the 2.5" SSD for the mSATA SSD had virtually no impact whatsoever on temperatures.
Power might have been up by 1W with the SSD in some instances, but it's too
small a difference to consider significant.
The 55W TDP of the Pentium G2120 turned out to be a false alarm. Since the
AC power draw never exceeds 55W, there's no way it could possibly pull 55W by
itself, at least not with our torture test apps. The latest version of AIDA64
Extreme Edition has an extremely interesting and useful addition in
its sensor monitoring, which I believe probably works only with newer motherboards
(and components): It tells us the power pulled by the CPU. Its accuracy is open
to question, but the data is worth looking at. See the screen capture below.
AIDA64 Extreme Edition Sensor page for Intel DQ77KB w/ Pentium G2120
in Akasa Euler case at full system load. |
The Power Values heading indicates the power drawn by the CPU. I believe CPU
Package indicates the total power drawn by the CPU. Does 23W sound too low for
a minimalist system that draws 55W AC total? Perhaps. It has to depend on motherboard
sensors, and they can certainly vary even from sample to board sample. On the
other hand, AIDA64 indicates 61W CPU Package for a Core i5-2500K in a Z77 board
while running just Prime95, which seems pretty close to what I'd expect.
The Great Wall 120W power adapter was swapped out for comparison with a recent
Dell 19VDC 80W power adapter that has high "V" efficiency rating.
The difference in AC power demand in the various states between these power
adaptes was less than a watt, so I can safely say the included power brick is
probably >85% efficient, perhaps 87% at the full 55W power draw, about as
good as power bricks get these days. The total system idle power of 16.5W is
very modest, not as low as an Intel NUC, but then it's also a more capable system.
| Help support this site, buy from one of our affiliate retailers! |
|
