DDR4 Haswell-E Scaling Review: 2133 to 3200 with G.Skill, Corsair, ADATA and Crucial
by Ian Cutress on February 5, 2015 10:10 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.
Photoscan, on paper, would offer more possibilities for faster memory to make a difference. However it would seem that the most memory dependent stage (stage 3) is actually a small part of the overall calculation and was absorbed by the natural variation in the larger stages, giving at most a 1.1% difference between times.
Cinebench R15
Cinebench is historically CPU dependent, giving a 2% difference from JEDEC to peak results.
3D Particle Movement
3DPM is a self-penned benchmark, taking basic 3D movement algorithms used in Brownian Motion simulations and testing them for speed. High floating point performance, MHz and IPC wins in the single thread version, whereas the multithread version has to handle the threads and loves more cores.
3DPM is also relatively memory agnostic for DDR4 on Haswell-E, showing that DDR4-2133 is good enough.
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.
Natural variation gives a 4% difference, although the faster and more dense memory gave slower times.
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.
NAMD showed little difference between our memory kits, peaking at 0.7% above JEDEC.
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.
Despite the 4x8 GB results going south of the border, the faster memory does give a slight difference in NPB, peaking at 4.3% increased performance for the 3000+ memory kits.
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.
When tackling a high number of users, Redis performs up to 17% better using 2800+ memory, indicating our best benchmark result.
120 Comments
View All Comments
ddriver - Thursday, February 5, 2015 - link
Upgrading became a non-issue around sandy bridge. My system is 3+ years old, and still within 10% of the corresponding tier of CPU today. Might as well be my last x86/Windows system before I switch to an ARM cluster under Linux...mdav9609 - Sunday, February 14, 2016 - link
Awesome! I've got an intel server board running two quad core Xeon E5620's (or something don't remember the exact numbers right now, socket 775) and their performance is almost as good as an i7 2600k, at least according to Passmark. I'm running them with an EVGA GTX 580. Got no problem running Fallout 4 and The Witcher 3 on it in 1080p. It's not my primary machine but I got one of these systems from work for free and put the second Xeon in it. Got it off eBay for like 15 bucks. Put in a few 15K SAS drives in RAID 0 and it is pretty cool. I like maxing out older systems just for the hell of it.pandemonium - Friday, February 6, 2015 - link
I thought it was pretty clear in this, and many, many, previous test comparisons of speed and DDR versions, that it makes very little difference. I'm on 8GB DDR2 and it's still going strong for everything I use it for. If it works...FlushedBubblyJock - Sunday, February 15, 2015 - link
Bingo - poor guy had to go through all that just so kingpins can win prizes flying around the world on enduser dimes.To the sane electorate, memory means number of GB.
I have to add I know plenty who, so long as the number is higher, they really and truly believe there is a performance increase. Sometimes they get confused, mixing generations of cpu's and memory, then their big brag on their junk doesn't work so well, but they still believe it.
So the memory marketing works, because there are an awful lot of people out there who fit the above description.
phoenix_rizzen - Thursday, February 5, 2015 - link
You're not alone.I have an HTPC running in the bedroom at home with an Athlon64 and 1.5 GB of DDR1. Plenty of horsepower for Windows XP, Google Chrome, and Plex web client, as it's connected to a 27" CRT TV.
One of my desktops at work is a tri-core Athlon-II system with 4 GB of DDR2 (AM2 motherboard)
My other work desktop is slightly more advanced, running a tri-core Athlon-II system with 8 GB of DDR3 (AM3 motherboard).
And the home server is just slightly more advanced still, running a quad-core Phenom-II system with 8 GB of DDR3.
nwrigley - Thursday, February 5, 2015 - link
I'm still running a quad Q6600 @ 3ghz with 8GB of DDR2. I've upgraded to an SSD and newer graphics card over its life. While money is certainly a limiting factor, in some ways there hasn't been a compelling reason to upgrade to a new machine.I work in video production and use high-end Macs at work. I often don't feel a difference between work and at home, with the exception of when the Mac doesn't have an SSD installed - then my system feels much faster (my boss isn't the type to upgrade an existing system, he'll just order a new one - very frustrating when a $200 SSD upgrade would make a huge difference).
I'm surprised just how well this processor has stood the test of time, but we haven't seen the type of performance jump that happened after the Pentium 4 era. The big performance jump we did see was with SSDs, so that's where I put my money (along with a bigger/better monitor.) My computer has also been a quiet and reliable workhorse - you never know what problems may come with a new system.
Murloc - Thursday, February 5, 2015 - link
I wouldn't feel compelled to change such a system either except for the sata/USB speeds, IF your use case can obtain advantages from faster speeds in that sector of course.nwrigley - Thursday, February 5, 2015 - link
Yep, you're absolutely right. A current motherboard would make both my SSD and GPU run faster with increased SATA and PCI Express speeds. USB 3.0 would be nice, but I don't have a current need for it.Guspaz - Thursday, February 5, 2015 - link
PCIe speeds in a Core 2 era system would still outstrip SATA on a modern system, though. Slap in an SSD using an x4 interface, for example, and you're talking 1GB/s of full duplex bandwidth even with PCIe 1.0, while modern SATA is still only doing around 600MB/s.Do you have any free PCIe slots that are more than 1x? Those could directly power a PCIe SSD, or you could stick in a SATA3 controller and use a SATA3 SSD. Ditto for USB3, if you did need it.
nwrigley - Thursday, February 5, 2015 - link
That's a good thought, unfortunately I only have PCIe 1x slots. Doesn't look like that would prGA-P35-DS3Rovide any benefit.