A comparison of quad-core server CPUs

A comparison of quad-core server CPUs

Summary: For anyone looking to buy a workstation or server CPU, quad-core CPUs have become mainstream. Therefore it's important to know what you're getting for the money so I've compiled a chart with general purpose computing performance using the SPEC CPU database with the highest scores as of December 28, 2007.

SHARE:

For anyone looking to buy a workstation or server CPU, quad-core CPUs have become mainstream. Therefore it's important to know what you're getting for the money so I've compiled a chart with general purpose computing performance using the SPEC CPU database with the highest scores as of December 28, 2007. I included single and dual processor solutions to help you decide whether you want to go single CPU socket or dual socket motherboard. You and also read more about energy efficiency on server processors here.

Note: This information is also available as a PDF from the TechRepublic Downloads Library.

All Intel dual processor models starting with the 54xx are the latest "Harpertown" 45nm CPUs launched in November 2007. All Intel dual processor models starting with 53xx are the 65nm "Clovertown" quad-cores Intel launched in late 2006 and mid 2007. In the single processor space, only the QX9650 "Yorkfield" processor uses Intel's latest 45nm process and everything else uses the 65nm process. The Q6600 and X3220 are essentially identical processors marketed towards desktop and entry level server markets respectively. Since one of the key differentiators on a workstation/server system is the inclusion of error correction memory, one can use any of the desktop CPUs in an ECC capable single processor motherboard.

The two AMD processors are Opteron quad-core CPUs based on 65nm "Barcelona". The 2.0 GHz Opteron 2350 is delayed due to the TLB bug and the 2.5 GHz Opteron 2360SE won't come out until the B3 stepping is out which fixes the TLB bug and brings higher clock speeds. There are reports that B3 stepping may be delayed until Q2 of 2008 (tranlated link here) though AMD's last analyst meeting presentation has a rough timeline of Q1 or Q2.

Note: SPEC CPU is broken down by performance on general purpose integer and scientific memory-bandwidth/floating-point intensive workloads. The general purpose workloads are summarized by a geometric mean score called SPECint and the scientific workloads are summarized by a geometric mean score called SPECfp. The results are further broken down by single-threaded results and multi-threaded results labeled as "rate2006". Note that a geometric mean is sort of like an average but it punishes the extremes more with a lower score than the average if a particular chip performs very poorly on some workloads. Ideally, one would simply benchmark their own specific application but that's not always possible so these published numbers from SPEC are very valuable data points.

SPECint includes workloads like Perl, compression, compilers, video compression, and other general purpose workloads. SPECfp includes workloads like bwaves, gamess, gromacs, povray, and a dozen other memory bandwidth and floating point intensive benchmarks. So while it's important to have a general ideal of how a chip performs in general, discriminating buyers will look inside the detailed disclosure (which I link to) and look at the application that is most similar to their own. So while a chip from AMD might have a lower overall score on SPECfp_rate2006, there are individual workloads within SPECfp that overwhelmingly favor AMD's memory bandwidth advantage. The inverse of this situation where an Intel CPU has a lower overall SPECfp score than an AMD CPU but still win some of the specific workloads can also be true. So in a nutshell, the chip you select should be based on your application requirements.

CPU Model CPU Clock FSB SPECint 2006 SPECint rate2006 SPECfp 2006 SPECfp rate2006
Mainstream dual processor server quad-core CPUs
Intel X5482 3.2 1600 26.1 147 22.2 85.2**
Intel E5472 3.0 1600 26.7 143 23.7 88.1
Intel X5460 3.16 1333 27.7 138 23.9 79.2
Intel X5450 3.0 1333 26.5 134 23.2 77.3
Intel X5365 3.0 1333 24.5 117 21.4 67.7
Intel E5410 2.33 1333 21.6 115 19.9 69.4
Intel E5405 2.0 1333 19.2 104 18.2 64.7
Intel E5335 2.0 1333 18.1 92.2 16.9 58.4
AMD 2350 2.0 NA   88.8 *   77.9 *
AMD 2360SE 2.5 NA   102 *   86.3 *
Entry level single processor workstation/server quad-core CPUs
Intel QX9650 3.0 1333 25.5 76.7 22.3 52.0
Intel QX6850 3.0 1333 23.6 69.1 21.2 49.4
Intel X3220 2.4 1066 15.9 59.0 15.3 42.5
Intel Q6600 2.4 1066 18.5   16.0  
* These results were invalidated last month because of lack of availability. Furthermore, the TLB bug patch performance penalty has not been factored in to these results. Assuming AMD fixes the bug in Stepping B3 and solves the manufacturing challenges in mid 2008 to deliver 2.5 GHz parts, scores similar to these invalidated numbers can be resubmitted. So while these numbers are officially invalidated, they were invalidated for lack of availability and not for inaccuracy so I left these numbers in for comparison purposes.

** Results for the X5482 3.2 GHz systems seem odd since they're worse than the E5472 3 GHz results. Intel gave an unofficial estimate at IDF2007 of 89.8 for SPECfp_rate2006 so we might see this number get updated as time goes by. Note that the SPEC CPU base scores for the X5482 were higher than the E5472 so that seems to fall more in line with expectation.

These results indicate a significant improvement with Intel's latest 45nm technology in multi-threaded applications. Comparing 3 GHz Harpertown with 3 GHz Clovertown, improvements for single-threaded applications were noticeable in the 8% range and that is mostly attributable to architectural enhancements in the chip's execution engine. At 3.0 GHz for multi-threaded applications, we saw a ~14% improvement on both SPECint and SPECfp using the same motherboard chipset and the additional gains are mostly due to the 50% larger CPU cache. But once the new 5400 series "Seaburg" chipset got involved with a 50% larger snoop filter and 20% faster memory bus, the 3.0 GHz scores jumped 22.2% for SPECint and 30.1% for SPECfp.

Considering the fact that the energy efficient 45nm Intel E5410 2.33 GHz chip costs around $300 whereas the 65nm Intel E5345 2.33 GHz chip costs around $600, buyers who are looking for Intel based solutions should immediately switch to 45nm technology. The Intel E5410 even manages to beat the $1200 Intel X5365 3.0 GHz processor on SPECfp_rate2006 and comes awfully close on SPECint_rate2006. So for the general purpose server market, the new E5410 on average seems to be the performance/dollar leader.

HPC (High Performance Computing) customers who have memory bandwidth intensive workloads on the other hand have been purchasing loads of inexpensive AMD Barcelona processors despite the TLB bug. Those memory-bandwidth hungry customers are using custom Linux kernels that work around the TLB bug with minimal impact on performance so they don't care about the bug or the lower overall SPECfp scores.

Topics: Processors, Hardware, Intel, Servers

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

52 comments
Log in or register to join the discussion
  • Very timely post!

    George, hats off to you as usual. There has been a lot of talk on the DV editing forums on the advantages of 45 nm and I would say most are skeptical right now. My only problem is I can never make heads or tails of the techno babble on the SPEC website so I appreciate your efforts.

    It would be nice to see benchmarks for the 5430 and the 5440 because these seem to be the favorites among the DIYers.

    Also, I wonder which "Seaburg" boards are being used on these benchmarks. Supermicro has a really nice one, the X7DWA-N, that is targeted for the workstation oriented consumer but right now the board is really hard to get for a decent price ($700 if it's in stock).

    I may as well also mention that Supermicro recently announced one of the nicest chassis I have ever seen, the SC743TQ-865B. Can't wait to get my grubby hands on one.

    Here are the links:

    http://supermicro.com/products/motherboard/Xeon1333/5400/X7DWA-N.cfm
    http://supermicro.com/products/chassis/4U/743/SC743TQ-865B.cfm
    Cornhead
    • For video editing, 45nm is a no brainer because of SSE4

      For video editing, 45nm is a no brainer because of SSE4. We're talking about a doubling of video encoding performance at the same clock speeds! Nearly all the commercial video editing and encoding packages on the market are adding SSE4 support because the performance boost is so stunning. Even for SSE3, the 45nm "Penryn" based technology adds about 15% in performance because of the added super shuffler. There shouldn't even be any debate in the video editing world, it's a done deal.
      georgeou
    • If you click on the links I provide, it tells you the exact model

      If you click on the links I provide on the scores, it tells you the exact model of the motherboard used.

      Seaburg is for dual-processing solutions and using the 5410 with a Seaburg chipset will deliver very good performance.
      georgeou
      • Thanks

        I noticed that after my post, so now I feel a little dumb :)
        Cornhead
    • Also remember the San Clemente DDR2 based solutions

      Also remember the 5100 series San Clemente chipset based solutions. These motherboards use the lower power and slightly less expensive registered DDR2 DIMMs though you're limited to 6 dual-rank DDR2-667 DIMMs. Tyan however pushes that chipset to allow 8 DDR2-800 DIMMs. I can see doing a cheap 16 GB system with eight 2GB registered DDR2 DIMMs using a cheaper San Clemente chipset motherboard and two E5410s. Performance is only a hair less than comparably clocked FSB systems using Fully Buffered DDR2 DIMMs.
      georgeou
      • Those San Clemente motherboards can be had in the low $300s

        http://www.monstronix.com/s5375ag2nr-p-455691.html
        TYAN COMPUTER S5375AG2NR Tempest i5100X (S5375) $333.84

        Add two E5410s for $600 and some cheap 2GB registered DDR2-667 DIMMs for $80/DIMM and you got yourself a system that rivals a $15K Mac Pro which used the 150W TDP Clovertown 3.0 CPUs! Forget the massive server chassis, any old ATX case and an "80 Plus" 400W PSU will do.
        georgeou
        • This definitely is one route to go

          However, if you run a high end graphics card like the nVidia Quadro FX4600 then a 400W PSU won't be up to the task. Thus a massive server chassis that comes equipped with a 865W PSU sounds very nice. Also, the supermicro chassis comes with an integrated hotswappable backplane. A definite plus when swapping out hdd with lots of video capture.
          Cornhead
          • Great point if you're going to use that powerful a graphics card

            Great point if you're going to use that powerful a graphics card. If that kind of video is a requirement, then I definitely agree with your choice.
            georgeou
      • Interesting strategy but...

        since the Seaburg boards are so new they are hard to find in stock and are really expensive. So it makes a lot of sense to swap out a 5100 series chip or build a 5400 series machine using a San Clemente board.

        However, some of the posters have claimed to not get the 5400 to run in non-Seaburg boards. I don't know for sure if these posters were using San Clemente boards, but has Intel stated whether the 5400 series is backward compatible with older chipsets like San Clemente?
        Cornhead
        • San Clemente was launched at same time as Seaburg

          San Clemente was launched at same time as Seaburg, without the fanfare that is. Seaburg is the ultra high-end that Intel pushes and it offers better *official* DDR2-800 FB memory support along with a bigger snoop filter on the chipset. The Tyan and Supermicro motherboards will however fully support registered DDR2-800. The Tyans are unique in that they go beyond Intel's official spec of 6 dual-rank DIMMs and let you use 8 dual-rank DIMMs. Intel doesn't validate or support that kind of memory configuration but Tyan does.
          georgeou
    • Build a Mac Pro equivalent workstation for 1/3 the cost

      http://blogs.zdnet.com/Ou/?p=942
      See how the DV folks respond to this one!
      georgeou
      • what dioes that have to do this?

        It is a comparison of quad core cpu chips.

        Does Mac have a quad core cpu chip?

        The go away.
        jrbeaman
        • The Mac Pro is powered by dual quad core Xeons

          nt
          Cornhead
  • RE: A comparison of quad-core server CPUs

    I am a firm supporter of competition an so I cannot wait for AMD to sort out their quad CPU problems.
    danielarbib
  • RE: A comparison of quad-core server CPUs

    Well shouldn't it be:
    A comparison of quad-core Server x86 CPU's ?
    Otherwise there is a few missing here..

    // Jesper
    JesperFrimann
    • Fair enough - nt

      nt
      georgeou
  • Very telling!

    To me the most telling thing is that you don't include energy consumption in your table, or at least (or even better) a performance value per energy unit. I'd think that in the world of servers, where energy efficiency can make a huge difference (e.g. in a large server farm), this kind of feature would be helpful to, say, IT management. Maybe next year?
    RichVan
    • Did you miss the link in the very first paragraph?

      http://blogs.zdnet.com/Ou/?p=927
      "You and also read more about energy efficiency on server processors here."

      I gave one of the most in-depth coverage of SPECpower in the media last month. This month I wrote something in-depth about CPU performance and linked to last month's energy consumption article. I didn't have physical space to include performance/watt in this table.
      georgeou
  • RE: A comparison of quad-core server CPUs

    They could care less? Would have thought you meant couldn't.
    Reged
  • RE: A comparison of quad-core server CPUs

    "...with minimal impact on performance so they could care less about the bug or the lower overall SPECfp scores."


    I think you mean "....couldn't care less...."
    hkb25