AMD courts servers with new FirePro cards

Earlier this month, AMD added a bunch of new FirePro products to its lineup of workstation graphics cards. All of them employ 28-nm GPUs based on the Graphics Core Next architecture. Now, the very same chips have made their way into FirePro cards destined for servers. Introducing the FirePro S7000 (left) and S9000 (right):

Why might one need a substantial amount of graphics horsepower in a server? Virtualization, for one. The new FirePro cards support RemoteFX GPU virtualization, which allows multiple clients to share the resources of a single graphics chip. VMware or Citrix software can also be used to virtualize the FirePros, but they’re limited to pass-through modes that allow only one client per GPU. More clients can be supported by adding FirePro cards, of course.

When it briefed us on the new FirePro cards, AMD noted that GPU virtualization is currently handled by the hypervisor software. AMD’s current GPUs don’t support hardware-based resource sharing. However, it sounds like that might change in a future generation.

Crunching compute workloads is also part of the job description for the new FirePro cards. For highly parallelized workloads, GPUs can deliver substantially better performance than the fastest CPUs. AMD mentions workstation graphics applications, too, but I suspect FirePro W-series cards are more appropriate in that arena. The workstation-oriented models have a lot more output options than the server cards, which are equipped with only a single DisplayPort out each.

  GPU Core clock ALUs Memory bus Memory TDP Price
FirePro S7000 Pitcairn 700MHz 1280 256-bit 4GB GDDR5 150W $1,249
FirePro S9000 Tahiti 925MHz 1792 384-bit 6GB GDDR5 225W $2,499

At the high end, the FirePro S9000 employs the same Tahiti GPU that underpins the Radeon HD 7900 series. The chip runs at 925MHz, and it’s backed by 6GB of GDDR5 RAM. Despite the card’s 225W thermal envelope, the dual-slot cooler has no fans. At least 20 CFM of chassis airflow needs to be directed toward the passive heatsink to prevent the GPU from overheating.

The FirePro S7000 is a single-slot solution with a less exotic Pitcairn GPU pulled from the Radeon HD 7800 series. Pitcairn has fewer ALUs than Tahiti, and in the S7000, it’s clocked at only 700MHz. As evidenced by its tighter 150W TDP, the S7000’s cooling requirements are more relaxed. The heatsink is smaller, and it requires just 10 CFM of airflow.

On the memory front, the S7000 gets by with 4GB of GDDR5. The card doesn’t support ECC, a feature that’s limited to the S9000. AMD says enabling error correction cuts memory performance by about 6%.

Like their desktop counterparts, both FirePros employ ZeroCore power to reduce power consumption at idle.

  Peak single precision Peak double precision Memory bandwidth
FirePro S7000 2.4 TFLOPS 152 GFLOPS 154GB/s
FirePro S9000 3.2 TFLOPS 806 GFLOPS 176GB/s

Obviously, the S9000 is the faster of the two. But its performance specifications don’t match the 2X difference in price. The S9000’s peak single-precision throughput is only 33% higher than the S7000, and its memory bandwidth is just 14% higher. That said, the S9000 has more than five times the double-precision throughput of its little brother. You can thank the Tahiti GPU, which has much better double-precision performance than Pitcairn.

Both FirePro cards are expected to start shipping at the end of September. Servers typically have longer upgrade cycles than the average enthusiast’s desktop, so AMD has pledged to support these FirePro cards until at least 2015.

Comments closed
    • Goty
    • 7 years ago

    [quote<]Obviously, the S9000 is the faster of the two. But its performance specifications don't match the 2X difference in price.[/quote<] [quote<]That said, the S9000 has more than five times the double-precision throughput of its little brother. You can thank the Tahiti GPU, which has much better double-precision performance than Pitcairn.[/quote<] Wait, what? If you have a workload that needs a little DP muscle, then that would [i<]more[/i<] than justify the price increase.

    • kalelovil
    • 7 years ago

    Are you sure you have the bandwidth figure for the S9000 correct? Other sites are reporting a figure of 264GB/s.
    176GB/s is the memory bandwidth of the 256bit W8000, perhaps you are confusing it with that?

      • Dissonance
      • 7 years ago

      Looks like there’s an inconsistency in the materials AMD sent us. One slide says 264GB/s, while another says 176GB/s. We’ve asked the company to clarify.

        • brucethemoose
        • 7 years ago

        [url<]http://www.amd.com/us/products/workstation/graphics/firepro-remote-graphics/S9000/Pages/S9000.aspx#3[/url<] [quote<] Memory Size/Type: 6GB GDDR5 Interface: 384-bit Bandwidth: 264 GB/s [/quote<] Their site also says 10CFM

          • Arag0n
          • 7 years ago

          That makes things more confusing if anything… since it points the DP and SP specs as techreport does…. something is so weird with those cards specs…

      • Arag0n
      • 7 years ago

      Looking to the linked website at the start of the article, the S9000 specs seem to mimic the FirePro W8000 specs, 3.2TFLOPS SP, 0.81TFLOPS DP but has 189W TDP and 256bit memory interface.

      In the other hand, the S9000 in the article points to have same characteristics, cost 1000$ more than the W8000 and have less outputs and higher TDP…. I can understand the higher price if it was a tweaked W9000 down clocked to use passive cooling but it doesn´t seems to be the case….

      So I’m gonna think that the correct specs given that the S7000 does match the W7000 ones should be 3.99TFLOPS SP, 1TFLOPS DB and 264GB/s Memory bandwidth. 64%, 566% and 71% improvement respectively.

    • DeadOfKnight
    • 7 years ago

    Geoff? Working on the weekend?

      • derFunkenstein
      • 7 years ago

      We all have to do it sometime.

    • Arag0n
    • 7 years ago

    Peak values do not properly reflect the real workload that those cards can achieve. Peak values can be only 33% difference but nominal values may perfectly double. That´s why techreport benchmarks exist right? To show the real differences 🙂

    Sure, you can´t spend 4000$ to test the GPGPU capabilities of both cards but you may ask AMD to send some samples to you and compare them to some NVIDIA Quadro. It would be interesant to have some comparison of the current GPGPU market for once.

    If you think those cards can be reflected by consumer grade ones, then use those ones and try to do a full GPGPU review, it would be pretty interesting, but I do acknowledge that tests are not as common or standard as CPU benchs.

    • willyolio
    • 7 years ago

    you labelled both as s7000…

    interesting how they have such a huge difference. what happened to having a series of models to cover the spectrum?

      • Srsly_Bro
      • 7 years ago

      I noticed that, too. Nice catch.

      Edit:
      I refreshed the page after posting and it was fixed. Nice response time!

Pin It on Pinterest

Share This