Workstation Performance

For our performance analysis, we will split this into two parts. Firstly we will add in the data for the system as it was sent, in a 1x8GB DRAM configuration. After this is a discussion with 2x4GB results, showing the importance of maintaining a dual configuration setup. For comparison points, we are picking up Dustin’s array of workstation review results, although a couple of the newer benchmarks have fewer data points.

Lenovo ThinkStation P300 (Xeon E3-1276 v3 + Quadro K4000)
vs
DigitalStorm Slade Pro (Xeon E5-2687W v2 + Quadro K4000)
BENCHMARK DigitalStorm Slade Pro ThinkStation P300
PCMark 8 (Home, OpenCL) 4879 3834
PCMark 8 (Creative, OpenCL) 4094 3160
PCMark 8 (Work, OpenCL) 4591 4505
Cinebench R15 (OpenGL) 102.85 118.6
Cinebench R15 (Single-Threaded) 123 158
Cinebench R15 (Multi-Threaded) 1218 769
x264 5.0 (Pass 1) 95.53 69.16
x264 5.0 (Pass 2) 25.43 16.58

Point Calculations – 3D Movement Algorithm Test: link

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.

3D Particle Movement: Single Threaded

3D Particle Movement: MultiThreaded

Compression – WinRAR 5.0.1: link

Our WinRAR test from 2013 is updated to the latest version of WinRAR at the start of 2014. We compress a set of 2867 files across 320 folders totaling 1.52 GB in size – 95% of these files are small typical website files, and the rest (90% of the size) are small 30 second 720p videos.

WinRAR 5.01, 2867 files, 1.52 GB

Image Manipulation – FastStone Image Viewer 4.9: link

Similarly to WinRAR, the FastStone test us updated for 2014 to the latest version. FastStone is the program I use to perform quick or bulk actions on images, such as resizing, adjusting for color and cropping. In our test we take a series of 170 images in various sizes and formats and convert them all into 640x480 .gif files, maintaining the aspect ratio. FastStone does not use multithreading for this test, and thus single threaded performance is often the winner.

FastStone Image Viewer 4.9

Video Conversion – Handbrake v0.9.9: link

Handbrake is a media conversion tool that was initially designed to help DVD ISOs and Video CDs into more common video formats. The principle today is still the same, primarily as an output for H.264 + AAC/MP3 audio within an MKV container. In our test we use the same videos as in the Xilisoft test, and results are given in frames per second.

HandBrake v0.9.9 LQ FilmHandBrake v0.9.9 2x4K

Rendering – PovRay 3.7: link

The Persistence of Vision RayTracer, or PovRay, is a freeware package for as the name suggests, ray tracing. It is a pure renderer, rather than modeling software, but the latest beta version contains a handy benchmark for stressing all processing threads on a platform. We have been using this test in motherboard reviews to test memory stability at various CPU speeds to good effect – if it passes the test, the IMC in the CPU is stable for a given CPU speed. As a CPU test, it runs for approximately 2-3 minutes on high end platforms.

POV-Ray 3.7 Beta RC4

Lenovo ThinkStation P300 BIOS and Software System Benchmarks
POST A COMMENT

57 Comments

View All Comments

  • akula2 - Wednesday, November 5, 2014 - link

    I build my own hi-end (X79/Z97) and ultra (Xeon) workstations because of hardware choice, saving on prices and cutting off vendor costs. Each custom-build workstation costs a lot, hence a lot of saving considering the number of builds. Reply
  • Dr.Neale - Wednesday, November 5, 2014 - link

    The Intel Z97 chip set does NOT support ECC memory. Only Haswell motherboards with an Intel C220 series chipset support ECC memory, and then only with a Xeon CPU.

    For example, ASUS P9D WS (C226) motherboard supports ECC memory (with a Xeon CPU); whereas, ASUS Z97 WS (Z97) does NOT (even with a Xeon CPU). I'm not sure whether or not it will accept ECC UDIMMS, but I am sure the ECC function is disabled, using Z97 chipset.
    Reply
  • otherwise - Tuesday, November 11, 2014 - link

    When I saw that the ASUS Z97 WS supported ECC I also found that incredibly odd, and wrote an e-mail to their support address. They got back to me a couple days later and assured me it did support ECC with a Xeon. I still don't really trust them, and if I was to go that route I would just get a C226, but I'd love to see someone test their claim. Reply
  • Scalarscience - Wednesday, November 5, 2014 - link

    Of course it's possible to beat prebuilt computers in terms of component choice, price or both. But why waste your breath? Do you think people out there are going to read this article, then read the comments and go "OH WOW I HAD NO IDEA??!!". Rather, this is like an echo chamber...the IT & purchasing people this article is aimed at will probably never even read the comments as they're parsing multiple reviews to make a purchasing (NOT building) decision.

    Ie, systems like this are built for entry level systems for office situations where there's either no IT or it's not worth the time to roll a custom solution. And it's been that way since...oh about 1988...

    I see the same thing in Apple Mac reviews, and it just boggles the mind. Fwiw I've run custom built XEON hardware since the PPro era (and before that I had a serverworks dual P133, which was very oddball). I used to use SGI & HP workstations at 'work' and Xeons at home, then I used Xeons at work running NT, then Linux, and now Linux & OSX on Macs...so yea obviously this review system wouldn't be my first choice for Autodesk Smoke, After Effects or Maya. But on a time crunch if I needed an intern or two to pick up some slack a short lease might just help finish a project....
    Reply
  • Dr.Neale - Wednesday, November 5, 2014 - link

    All the more reason why Kenobi should spend at least $20 more for an 8 GB stick of ECC memory, or $110 more for 2 x 8 GB dual-channel ECC memory, instead of saddling the end-user (buyer or lessor) with a crippled single-channel non-ECC memory system. Come on, it's less than the cost of upgrading the GPU from a K4000 to a K4200, and it would enhance the spec sheet considerably.

    Other things being equal, isn't 16 GB ECC dual-channel memory MUCH more appealing than 8 GB non-ECC single-channel memory???

    And it would generate MUCH more favorable reviews!!!
    Reply
  • Dr.Neale - Wednesday, November 5, 2014 - link

    Damn auto-correct changed "Lenovo" to "Kenobi" !?!?! Reply
  • atl - Friday, November 7, 2014 - link

    This configuration is obviously unbalanced and have some funny moments with pricing.
    1 Unbalanced.
    You do not put by default on generic $2000 price range workstation $800 USD video cards.
    Yes, with some GPU heavy workloads may require it, but in higher class workstation. Or, if is in this price range - in very rare cases.
    Far better balanced will be if for same price get K2200 video card and with spared resources get
    32GB RAM, good SSD + 2/4TB high class HDD - WD black or same range from other suppliers.
    Also CPU is not well chosen - 1276 is 50% more expensive, but only 8.1% faster than 1230.

    2. Funny thing about pricing.
    I don't say over expensive, just can't find proper word for it.
    Given CPU, RAM, HDD and VGA are commodity and can be obtained with same service/warranty quality from local supplier (sometimes with even better terms) and cost of it is around $1200, this left us with $1000 price tag for PSU/CASE/MB combo.
    Note that Lenovo itself obtains this components on very lower than Newegg!!!

    Sorry, but i can't imagine how semi commodity PSU for price tag under $100, virtually featureless motherboards and case without any special cooling, 4 PCI slots and even no frontal hot swap disk cases can cost $1000. Ye, i know that MB is "industrial strength, highly redundant, certified, sunroof, ABS, 4x4, etc", but this class MB costs under $300 from other suppliers.
    Examples can continue but will stop here...
    Reply
  • Dr.Neale - Friday, November 7, 2014 - link

    Unbalanced? Agree 100% !

    Best Bang-for-the-Buck Xeon (Haswell e3-1200 v3 series) = 1230? Disagree!!! Based on SuperBiiz (Price+Shipping):

    Retail Box, CPU with no IGP:
    1230 (3.3 / 3.7 GHz) $250
    1231 (3.4 / 3.8 GHz) $250
    1240 (3.4 / 3.8 GHz) $273
    1241 (3.5 / 3.9 GHz) $272
    1270 (3.5 / 3.9 GHz) $334
    1271 (3.6 / 4.0 GHz) $335

    Retail Box, CPU with IGP:
    1245 (3.4 / 3.8 GHz) $285
    1246 (3.5 / 3.9 GHz) $285
    1275 (3.5 / 3.9 GHz) $347
    1276 (3.6 / 4.0 GHz) $347

    Assume a rock-bottom system cost of $1800 + CPU, then:

    CPU Cost GHz
    1230 $2050 (+0%) 3.3 (+0%)
    1231 $2050 (+0%) 3.4 (+3%)
    1240 $2073 (+1%) 3.4 (+3%)
    1241 $2072 (+1%) 3.5 (+6%)
    1245 $2085 (+2%) 3.4 (+3%)
    1246 $2085 (+2%) 3.5 (+6%)
    1270 $2134 (+4%) 3.5 (+6%)
    1271 $2135 (+4%) 3.6 (+9%)
    1275 $2147 (+5%) 3.5 (+6%)
    1276 $2147 (+5%) 3.6 (+9%)

    The IGP only adds about $12 - $13 or 0.6% to the system cost, so might as well get it.

    The most expensive CPU, the 1276, raises the system cost by $97 or 4% over the 1230 / 1231, but raises the GHz by 9% / 6% over the 1230 / 1231.

    So, the 1276 (or 1271) offer the best bang-for-the-buck, with (without) an IGP.

    This is because although the CPU price increases by $97 or 35% (not 50%) in going from the 1230 (1231) to the 1276, the SYTEM price increases by 4%, for a 9% (6%) increase in CPU speed, which dominates the system performance.

    P.S. The ASUS P9D WS lists on SuperBiiz for only $233 (price + shipping).
    Reply
  • Dr.Neale - Friday, November 7, 2014 - link

    Oops, 1271 is 4% increase in SYSTEM price, 1276 is 5%.

    The conclusion is unchanged.
    Reply
  • Dr.Neale - Friday, November 7, 2014 - link

    Should read CPU price increases by 39% (not 50%) in going from 1230 (1231) to 1276, the SYSTEM price increases by 5%, for a 9% (6%) increase in CPU speed, which dominates the system performance. Reply

Log in

Don't have an account? Sign up now