Intel Haswell Low Power CPU Review: Core i3-4130T, i5-4570S and i7-4790S Tested
by Ian Cutress on December 11, 2014 10:00 AM ESTProfessional Performance: Windows
Agisoft Photoscan – 2D to 3D Image Manipulation: link
Agisoft Photoscan creates 3D models from 2D images, a process which is very computationally expensive. The algorithm is split into four distinct phases, and different phases of the model reconstruction require either fast memory, fast IPC, more cores, or even OpenCL compute devices to hand. Agisoft supplied us with a special version of the software to script the process, where we take 50 images of a stately home and convert it into a medium quality model. This benchmark typically takes around 15-20 minutes on a high end PC on the CPU alone, with GPUs reducing the time.
Cinebench R15
Professional Performance: Linux
Built around several freely available benchmarks for Linux, Linux-Bench is a project spearheaded by Patrick at ServeTheHome to streamline about a dozen of these tests in a single neat package run via a set of three commands using an Ubuntu 14.04 LiveCD. These tests include fluid dynamics used by NASA, ray-tracing, molecular modeling, and a scalable data structure server for web deployments. We run Linux-Bench and have chosen to report a select few of the tests that rely on CPU and DRAM speed.
C-Ray: link
C-Ray is a simple ray-tracing program that focuses almost exclusively on processor performance rather than DRAM access. The test in Linux-Bench renders a heavy complex scene offering a large scalable scenario.
NAMD, Scalable Molecular Dynamics: link
Developed by the Theoretical and Computational Biophysics Group at the University of Illinois at Urbana-Champaign, NAMD is a set of parallel molecular dynamics codes for extreme parallelization up to and beyond 200,000 cores. The reference paper detailing NAMD has over 4000 citations, and our testing runs a small simulation where the calculation steps per unit time is the output vector.
NPB, Fluid Dynamics: link
Aside from LINPACK, there are many other ways to benchmark supercomputers in terms of how effective they are for various types of mathematical processes. The NAS Parallel Benchmarks (NPB) are a set of small programs originally designed for NASA to test their supercomputers in terms of fluid dynamics simulations, useful for airflow reactions and design.
Redis: link
Many of the online applications rely on key-value caches and data structure servers to operate. Redis is an open-source, scalable web technology with a b developer base, but also relies heavily on memory bandwidth as well as CPU performance.
Facebook
Twitter
RSS
76 Comments
View All Comments
MrSpadge - Thursday, December 11, 2014 - link
> lower power cpus, saves us time from undervolting and just draws less power from day one.That's not the same. Low power CPUs limit your clock sped when you need it most, i.e. under full load. Whereas undervolting delivers full performance, or in TDP limited cases even higher performance. It does cut into the OC headroom, though.
xeizo - Thursday, December 11, 2014 - link
I expressed it badly, sorry. The point is as these cpus are TDP-limited you know how much they will draw the most, and can design cooling/case etc. based on that, and in "99%" of home user cases you don't need the extra performance of 4xfull load+OC without restraints - the cpu is fast during "normal" use. Even gaming. It's not the best buy for a pure rendering box or similar though ...xeizo - Thursday, December 11, 2014 - link
Btw, I noticed at a customer they had replaced all the workstations with mini-ITX HP:s, running 4570S and having SSD:s. They felt very snappy.azazel1024 - Thursday, December 11, 2014 - link
It would be nice to see what the idle power consumption is with like setups on some of the processors. In addition it is a shame that you have dV for power consumption for the Pentium models...but there are NO performance benchmarks for them. Both of the Haswell pentiums have roughly the same power consumption of the i3-4130t, but no idea what they actually can put up in terms of performance. Also rather suprised that the older Ivy Bridge i3 has such low power consumption, but seems to average slightly better than the i3-4130t.Too bad no test of the 4MB cache and higher clocked Hasy i3-ts. That would have been interesting, as that is almost exactly what I am thinking of replacing my G1610 server with. Though realistically by the time I get around to it, I'll probably be looking at a Broadwell based server.
piasabird - Thursday, December 11, 2014 - link
A lot of the 35 watt i3 CPU's are not available for sale anywhere. Especially the Haswell with the 4 meg cache and the 4600 graphics. One question I would pose is cant you just buy a regular i3 and just underclock it to use less power?MrSpadge - Thursday, December 11, 2014 - link
Yes. XBitLabs tested exactly that and the result is the same as the S/T models. The automatic voltage-frequency scaling of Turbo does all the magic behind the scenes (i.e. lowers the voltage when you lower the clock).sireangelus - Thursday, December 11, 2014 - link
Could we have mixed amd/intel linux test? it's very interesting to see since the compilers optimize more fairly.XZerg - Thursday, December 11, 2014 - link
Ian, man - seriously you need to stop these retarded Delta charts! They are useless! To make these even worse is that you are testing low wattage cpus. the people interested in such cpus are usually interested in idle and load numbers separately. what you have has no reference point to workout either the idle or load numbers.quit with these useless charts. there are many who have agreed to this.
sweetie peach - Thursday, December 11, 2014 - link
Agreed. Also the idle and load voltages would have been helpful.BrokenCrayons - Thursday, December 11, 2014 - link
I like the review, but I agree with this. I think readers would understand that their individual results are going to be different because they won't be running identical configurations, but we can do the simple subtraction ourselves if we want to see the difference in load versus idle. I'd much prefer getting the total system wattage at idle and load from a review. It seems more useful to me to see those raw numbers actually posted.