AMD EPYC 3251 Benchmarks and Review the Challenger We Need

22

AMD EPYC 3251 Power Consumption

Please see our caveats to our test configuration. These numbers include the AMD FirePro W2100 GPU. They do not include a BMC. They also utilize a different boot OS SSD. We still wanted to show power consumption figures for the system to give some ballpark of where these CPUs would fall in terms of power consumption.

  • Idle: 32.4W
  • 70% Load: 52.8W
  • 100% Load: 69.3W
  • Peak: 77.8W

These results were taken using our calibrated Extech TrueRMS Power Analyzer 380803 in a 71F ambient environment with 41% RH on 120V power.

This power consumption, taking into account the major platform caveats, is somewhere between the Intel Xeon D-1541 and the D-2141I. When you look at the performance of the AMD EPYC 3251 it is closer to the D-2141I. We think if we had a production platform with both Intel Xeon D-2141I and AMD EPYC 3251 with the same feature set, the AMD EPYC 3251 would use less power while performing about the same. We hope to validate that claim in the future.

AMD EPYC 3251 Market Positioning

Thes chips are not released in a vacuum instead, they have competition on both the Intel and AMD sides. When you purchase a server and select a CPU, it is important to see the value of a platform versus its competitors.

AMD EPYC 3251 v. Intel Xeon

Is the AMD EPYC 3251 a competitor to Intel Xeon D-1500 and Intel Xeon D-2100? Absolutely in terms of power consumption and performance.

To the embedded systems vendors out there considering AMD EPYC, there is more to the story. This is the first competitive part AMD has had in a market that Intel dominates today. Embedded lifecycles are such that the market is determined by far more than raw performance figures. Intel has proven that it can handle longer product cycles, and has issued programs for large numbers of components when they find issues in the field. These steps go a long way in the embedded market for companies looking to partner on a product that will be in the market for the better part of a decade.

On the flip side, we need a strong AMD in the market. Our 40GbE Arista switches are powered by AMD embedded CPUs, so we did not have to worry about the Intel Atom C2000 series AVR54 bug with them. Recently, STH was hit by the Intel Atom C2000 AVR54 Bug on one of the site’s main firewalls. In the future, if we can have the same firewall appliance with both AMD and Intel silicon, in about the same performance, power, and form factor envelopes, we would deploy it in a heartbeat. The AVR54 issue extends to more than just that series, including the Another Atom Bomb Intel Atom E3800 Bay Trail VLI89 Bug we covered. These embedded platforms are now designed as SoCs so supplier diversity is a key concern.

AMD EPYC 3000 8 Core Die
AMD EPYC 3000 8 Core Die

The bottom line is that increased supplier diversity for the underlying x86 SoC infrastructure would be excellent for the ecosystem. Currently, many vendors look toward Arm solutions for a second-source embedded firewall or other appliance. In the future, the AMD EPYC 3251 or other EPYC 3000 series parts are what we want to see and better yet, what we want to deploy. The same code can run on existing Intel and AMD platforms so the software code base can remain constant.

Of course, there are a few caveats here. Intel has invested in QuickAssist technology. STH is the only major review site to have worked with QAT in Intel QuickAssist at 40GbE Speeds: IPsec VPN Testing and Intel QuickAssist Technology and OpenSSL Benchmarks and Setup Tips. While QAT is far from receiving universal support, it is a multi-generational Intel accelerator technology that is popular in the embedded systems market. AMD does not have a compatible accelerator platform at this time but Intel offers QAT on PCIe cards or via chips if that is absolutely required.

AMD EPYC 3251 v. AMD EPYC

Here is where we need more data. Where does the AMD EPYC 3251 fall with respect to the rest of the embedded AMD EPYC 3000 series lineup? We have some idea that the lack of memory bandwidth is being offset by the single NUMA node so we are seeing similar AMD EPYC 7251 to EPYC 3251 numbers.

The best we can tell our readers here is that we are working on it. We still have not seen the dual die parts in action. We also have not seen a huge uptick in platforms in the market. As a first generation embedded part this makes sense. Embedded device lifecycles are much longer than consumer parts and clearly slower than our desire to try out the new technology.

Final Words

We do not have final pricing on the AMD EPYC 3251, but we know that the list price on the AMD EPYC 3301, the 12 core dual die SKU above this one, is $450. That puts one squarely in the 8-core Intel Xeon D range. AMD has a compelling product. We asked AMD, and will update this article when we get a number.

Update: AMD gave us a $315 price tag on the EPYC 3251, well below what performance competitive chips are going for on the Intel Xeon D side.

We know that many STH readers are in the embedded appliance market. AMD has the potential with the EPYC 3000 series. These markets are also extremely slow to transition which is why it is still hard to find the chips in the public market. At Computex 2018 we saw the first 3rd party system we highlighted in Piecing Together the iEi Puzzle AMD EPYC 3000 Spotted in the Wild. AMD’s APU offerings have been popular in the embedded space for years, so the AMD V1000 launched at the same time we saw several times at the show and there are publicly available platforms for the V1000.

The next step is twofold. We need embedded appliance vendors and motherboard manufacturers to make systems based on the AMD EPYC 3000 series.

22 COMMENTS

  1. Do ya’ll have the 16 core one?

    I’m going to ask our PM team if their fixin’ to make up our appliance with embedded 3000 EPYC.

  2. I’ll read the rest later, but you’re right on the need for an Intel alternative. These side channel attacks and the embedded clock bugs show a need in the market for a second x86 supplier.

    Can someone make a damn mini ITX motherboard with these, 4 DIMM slits and some NVMe headers?

  3. Are the single die parts way cheaper? The higher core counts looked like huge cost savings but I’m not seeing any on AMD’s site. Only 4 SKUs there https://www.amd.com/en/products/specifications/embedded/8161

    Model: AMD EPYC™ Embedded 3251
    Product Type: SOC
    Family: AMD EPYC™ Embedded Processors
    Line: EPYC Embedded SOC
    OPN: PE3251BGR88AF
    TDP: 55W
    CPU Type: Zen
    CPU Base Freq.: 2.5GHz
    CPU Max Freq.: 3.1GHz
    # of CPU Cores: 8
    # of Threads: 16
    GPU Support: No
    Security Processor: Yes
    Total L2 Cache: 4MB
    Total L3 Cache: 16MB
    System Memory Type: DDR4
    DDR4 Rate (Max): 2666 MHz
    Memory Controller: Dual Channel
    ECC: Yes
    USB 2.0: 0
    USB 3.0: 4
    USB 3.1 Gen1: 0
    USB 3.1 Gen2: 0
    SATA: 8
    Low Speed Interfaces: EMMC, eSPI, GPIO, I2C, LPC, SMBus, SPI, UART
    # of PCI controllers: 8
    Gen3: 32
    Gen2/3: 0
    Gen2: 0
    Ambient Temp Range: 0-105°C
    Enhanced Temp Support: Extended Temp (0-105°C)
    Infrastructure: SP4
    Last Time Buy: 2028
    Recommended for new designs: Yes

  4. Unless I missed something, but who makes the NIC? The lspci output showed the NIC as an AMD network card, but who did AMD license the NIC from? What network driver is it the chip using? That could provide some clues.

  5. I’m just surprised AMD building their own NIC IP. I suspect they licensed it as many others do. E.g. Asmedia makes the chipsets.

  6. Hi Misha, I understand it is an SoC. Typically what you do as part of SoC designs is license IP (in chip design IP means a blob of VHDL / Verilog code), so let’s say I design an ARM SoC, I get an ARM core from ARM, I may get a video core IP from Imigation, a flash controller IP from someone else. Yes, you can design pieces in-house too. Something like networking is so complicated and I don’t recall AMD having any experience in there, I doubt they started from scratch and just licensed IP from a Marvell or some other company. I’m curious what IP it is unless it is truly their own design.

  7. Finally, Patrick, thanks.

    I would assume that’s $315 USD in bulk lots of 1000 for the chip? Wholesaler price?
    So could we assume at least $550 / $600 US motherboards using this?

    I’d still strongly consider it, it certainly seems to perform. I’d prefer perhaps something 15w lighter and $100 lighter, but we’ll have to wait for more boards.

    As I posted on the forums, incredibly hugely disappointing that AsRock, SuperMicro, Tyan all either told me “nope, no interest in this!” or flat out didn’t respond!

    These would make a monster FreeNAS machine, especially the 3201 model with 8c/8t at only 30Watts!

    Patrick, do you know if this problem impacts the Epyc 7xxx or 3xxx series?
    https://bugzilla.kernel.org/show_bug.cgi?id=196683

    https://forums.freenas.org/index.php?threads/ryzen-stability-on-11-0-u4.59017/

  8. $315 is the tray pricing. Anyone buying these embedded parts is likely buying quantities for motherboard runs.

    I believe that kernel bug is for Ryzen users and has not been confirmed on EPYC.

  9. @kpin:

    I hadn’t thought of that, it seems unlikely but I wouldn’t put it past them to try that.
    I did contact SuperMicro, Tyan and AsRock about this CPU multiple times since Feb and have either had no response or an outright “no” from them, which was disapointing.

    However, SuperMicro already do some AMD Epyc stuff, so they aren’t ‘frightened’ of Intel, to my knowledge.

  10. We need a competitive board to the likes of SMCI’s x11sdv-*c-tlnf8 boards. (I hope I got the model name right from memory.)

    Would be great if we could get the same treats the Intel line currently has.

  11. Anyone know of a board coming out with the dual 10 Gb option? the supermicro boards which just came out are all 1 Gb.

  12. Any updates. This platform looks very cool, and I would like to use it for routers and small appliances, but I can’t find any consumer motherboard or system that actually exposes the built in 10Gbps NICs as SFP+. I did found some boards from Supermicro and ASRock Rack, but they all use Intel NICs for their 10Gbps and 1Gbps ports. Weird. Would be great to see native 4x SFP+ and 2x SFP (via external controller), and this SoC can do it quite well and cost competitively, yet I can’t find products doing that. Weird.

  13. Some information about idle power use to someone like me who uses the platform in residential setting where the system is running 24/7, workload is bursty and idle power use is very important.

    With the following configuration I’m getting 22.3W at idle.

    M11SDV-8C-LN4F with EPYC 3251.
    Stock fan replaced with Revoltec RL036 (mostly due to noise reduction, but likely it helps power use too).
    64GB DDR4-2400 RAM (2x32GB HMA84GR7MFR4N-UH)
    WD SN750 1TB NVME SSD
    Only single gigabit network cable is connected. BMC is left unconnected.
    PSU is 120W 12V power brick (Chieftec CDP-120ITX).
    VGA is disabled via jumper (saves 1W)
    “powertop –auto-tune” has been run (saves 0.2W if VGA is enabled, 1.1W if VGA is disabled)

    The measurement has been made on console login screen of stock Debian 11 that had powertop installed. The power meter is not calibrated.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.