ASRock Rack X570D4U-2L2T Review an AMD Ryzen Server Motherboard

20

ASRock Rack X570D4U-2L2T Benchmarks

We wanted to give some sense of available relative performance compared to more traditional server solutions on the market. With that said, the performance of our Ryzen 3600 and 3950X chips have not changed since we last benchmarked them, and the CPUs are not the focus of this review, so these results are unchanged from our previous review. If you read the previous review, and already have a sense of how Ryzen 3000 series performs in these platforms, then this is a section to skip. If you want to know more about AMD Ryzen v. Intel Xeon (lower-end Xeon) then here are the numbers for you.

Python Linux 4.4.2 Kernel Compile Benchmark

This is one of the most requested benchmarks for STH over the past few years. The task was simple, we have a standard configuration file, the Linux 4.4.2 kernel from kernel.org, and make the standard auto-generated configuration utilizing every thread in the system. We are expressing results in terms of compiles per hour to make the results easier to read:

AMD Ryzen 9 3950X Linux Kernel Compile Benchmark
AMD Ryzen 9 3950X Linux Kernel Compile Benchmark

There are many folks out there who offer that the Xeon E3-12xx series is still very fast. Likewise, older Xeon E5 V1-4 systems are extremely competitive. We wanted to show a fairly wide range of CPU options cutting across families here. As you can see, the Ryzen 9 3950X performance is significantly better than the older generation systems.

c-ray 1.1 Performance

We have been using c-ray for our performance testing for years now. It is a ray tracing benchmark that is extremely popular to show differences in processors under multi-threaded workloads. We are going to use our 8K results which work well at this end of the performance spectrum.

AMD Ryzen 9 3950X C Ray 8K Benchmark
AMD Ryzen 9 3950X C Ray 8K Benchmark

Here we can see c-ray 8K results that are solid. The performance of AMD Zen, Zen+, and Zen 2 chips tend to be great on this type of benchmark. If you wanted to build out a render farm and still have manageable nodes, then this may make a lot of sense as a solution.

7-zip Compression Performance

7-zip is a widely used compression/ decompression program that works cross-platform. We started using the program during our early days with Windows testing. It is now part of Linux-Bench.

AMD Ryzen 9 3950X 7zip Compression Benchmark
AMD Ryzen 9 3950X 7zip Compression Benchmark

7zip shows something intriguing. Here, the performance of the Ryzen 9 3950X is well above that of the much more costly Xeon D-2183IT. The Xeon has RDIMM support so it can hit more appropriate memory capacities for a sixteen core CPU, but in terms of raw CPU performance, the Ryzen chips are much faster than Xeon D 16-core offerings.

OpenSSL Performance

OpenSSL is widely used to secure communications between servers. This is an important protocol in many server stacks. We first look at our sign tests:

AMD Ryzen 9 3950X OpenSSL Sign Benchmark
AMD Ryzen 9 3950X OpenSSL Sign Benchmark

Here are the verify results:

AMD Ryzen 9 3950X OpenSSL Verify Benchmark
AMD Ryzen 9 3950X OpenSSL Verify Benchmark

Here we just wanted to highlight how much faster the Ryzen 9 3950X is versus the Xeon E-2200 line. Simply having more cores and cache helps a lot. Performance per core on the Xeon E-2288G is very close, but we have twice the cores with the Ryzen 9 3950X.

Another key one here is that the Ryzen 9 3950X is about twice as fast as the Xeon E5-2670 V1’s. The Xeon E5 series has more PCIe and memory capacity, but if one is space/ compute-bound, the Ryzen 9 3950X in the ASRock Rack X570D4U-2L2T is similar to four E5-2670 V1’s.

Chess Benchmarking

Chess is an interesting use case since it has almost unlimited complexity. Over the years, we have received a number of requests to bring back chess benchmarking. We have been profiling systems and now use the results in our mainstream reviews:

AMD Ryzen 9 3950X Chess Benchmark
AMD Ryzen 9 3950X Chess Benchmark

Here we focused on more server CPUs. This is where we wanted to highlight something that is extraordinarily important. Other than the other Ryzen CPUs and Xeon E-2288G on this list, the remainder of this list has ECC RDIMM support. Platforms such as the EPYC 7282 also have significantly more PCIe Gen4 connectivity and memory bandwidth.

Memory capacity is a big deal. Generally, with 16-core processors, we would expect to see a minimum of 128GB-256GB (8GB-16GB per core) paired with them for general virtualization workloads. Most of the single socket EPYC nodes we deploy are 256GB of memory at 16 cores and go up from there. That is a weakness of using a consumer platform even if performance is great.

Performance Summary

Overall, performance is good if all you are looking to obtain is a lower core count and memory footprint server.

We may have readers who want to know the performance of the chip on the X570D4U-2L2T versus traditional workstation motherboards. The numbers are very close. As a control, we tested this motherboard against an ASUS PRIME X570-P that we use for storage testing. Looking across two dozen benchmarks, the variance was +/- 1.5% which is within a margin of error. We felt this was close enough that it is less of a factor than the raw price/ performance this combination offers.

ECC Memory

While the X570D4U-2L2T and its predecessors the X470D4U series supports ECC memory, there is a bit of a gotcha. As readers noted in the original X470D4U reviews, while ECC memory is supported and performing error correction, the reporting of that error correction was not functioning. In other words, even if you were experiencing continuous memory errors, no log of those errors was being recorded in the IPMI event log where one might expect them to show up. A user over on the Level One Techs forums had a conversation thread with someone from ASRock Rack, who reported that while the AM4 platform had ECC support, it did not have error reporting support.

For the X570D4U-2L2T, we verified that operating systems (Windows and Linux) were reporting that multi-bit ECC support was enabled, but we were unable to test if the error reporting situation had changed. My personal assumption is that it is likely still not working, as the problem seemed to stem from the CPU memory controller not providing the ECC error reports; the chipset likely has very little to do with the process.

Power Consumption

We wanted to compare the X570D4U-2L2T power consumption to its closest last-gen part, the X470D4U2-2T, which we tested with a Ryzen 3600 CPU.

  • Idle Power: 29W
  • STH 100% Load: 117W
  • Maximum observed power: 128W

These results were observed on 120V power using a basic Kill-A-Watt meter. The system is powered by a consumer-grade 80Plus Bronze power supply similar to what is found in many low-power short-depth server chassis.

In general, power consumption was modestly higher on the X570 based board, though not to a degree worth worrying about.

Next, we are going to discuss some of our closing thoughts around the solution.

20 COMMENTS

  1. “One note, the shared interface for the AST2500 is on the 10 GbE NIC, not the 1 GbE NIC.”

    A word of warning that if you accidentally enable IPMI access to the 10 GbE NIC (eg. selecting link aggregation in the IPMI settings), it will wreak havoc with performance on that interface.

    I spent weeks trying pfSense tuning settings to figure out why performance was nowhere near 10 Gbps. It was only by accident that I noticed the IPMI setting and turned it off.

  2. Any comment on the lack of ECC report support?
    Will it still be able to correct smaller bit errors, don’t even know if ECC is capable of that?
    In my +25 year in the IT business I cannot remember I have ever ran into a computer that had corrupted modules in a way where ECC would have seemed to make any difference. As fare I can remember it has always been a matter of boot errors or the computer couldn’t see the module at all. Not saying it couldn’t be the case, but I can’t remember I’ve ran into the situation myself.

  3. Anyone has distributor in EU that could offer combo with chassis for this baby? If they can do QA before shiiping, I’m all in.

    Tnx

  4. “One note, the shared interface for the AST2500 is on the 10 GbE NIC, not the 1 GbE NIC.”

    Any particular reason why they did so ? 1 GbE was not enough for IPMI ?

  5. domih,
    I assume it was so that, under the assumption you only had a single available switch port to plug the server in with, that you could have both 10 GbE and IPMI access.

  6. @Datasaver/@Will Taillac:
    Both my Xeon D-15xx Boards draw more Power than servethehome has measured. I always wondered what I am doing wrong in configuring the systems, as power consumption is important for me.

    My X10SDV-TP8F idles at 26W with everything stripped except for 2*32GB RAM, a single mSATA-SSD and a small FAN at low revs. I use a 300W Gold PSU. “Power off/BMC only” is 7.5W in this configuration. Servethehome has measured 22.9W/4.9W for the X10SDV-4C+-TP4F in a test configuration with double the RAM and many SSDs. If I move closer to the test configuration used for the review of the board, I am at >30W/10W.
    I wonder where the difference comes from, especially since even if I assume that servethehome is measuring power with a simple configuration, at least my BMC only values are to high. One can argue that my board has an additional quad port i350 NIC, but even the X10SDV-7TP8F was measured at only 5.0W and it has my configuration +12 Cores +SAS-HBA

  7. @cspguy: RDIMMs are not supported on Ryzen or Threadripper, you have to go to EPYC to get RDIMM support. There’s nothing ASRock Rack can do about this, it’s a limitation from AMD.

  8. This board will do 3200M/T memory, as long as its only two sticks, according to Asrock. 4 sticks is 2933 if single rank, 2667 if dual rank (if i read their response correctly). They also confirmed that Ryzen 5000 support is coming sometime December or January.

    How would this board do in a workstation? I’ve been trying to make a compact desktop for work and gaming with 10gbe, but available components are sparse. Next best solution is some finagling with a pcie extender to make room for the chunky gpu.

  9. Hello,

    Does the motherboard includes an integrated, non-removable, backplate on the socket as the X570-D4I-2T ?

    Thanks !

  10. @Elsa,
    Yes. But it’s a standard AM4 backplate design, nothing Intel flavored like on the X570D4I-2T. The stock Ryzen 3600 cooler mounts just fine (wraith spire), as does the Wraith Prism.

    @Patrick,
    As a workstation board, this will do fine, though perhaps not better than any other X570 board. The BMC doesn’t do much for a workstation. On the other hand, it may actually be the least expensive X570 board with onboard 10 GbE networking, but you can always just get a ‘normal’ X570 board and an add-on PCIe 10 GbE NIC.

  11. What i really hate, also on the X399D8A, is that anoying slow media speed via BMC. Why, Asrock??? It’s almost impossible to use that for a reasonable remote installation e.g. of anything larger than a MSDOS image. Installing Windows or also TrueNAS/FreeBSD is a pain.
    Why the heck? My old X10SL7-F and X11SSH-CTF (both having the older AST2400) are also not fast, but much more bearable.
    And also, routing the BMC via 10GB at the X11SSH-CTF does not reduce 10G speed in my case.

    So, allthough i really would like to use the Ryzen Zen2 or 3 as a multipurpose home server, the Asrock Rack AMe/TR4 boards try to mimic server boards and present lots of server features, but are lacking in the end a good implementation.
    – only partial ECC features (even my B550M Tuf Gaming reports ECC to the OS)
    – almost impossible BMC media useage
    – slow memory speeds with 4 sticks (and we all know: memory speed is important for Ryzen)
    – it seems performance problems when routing BMC over the regular NICs
    – only limited attention for Bios updates.

    So, all in all, the plattform looks attractive at the first glance, but seems to lack enthusiasm in the details then leaving it with some quirks.

  12. @ZFSfan
    My understanding of the “ECC reporting” problem is that the OS will be aware of an ECC event, but the report does not travel ‘upstream’ to the BMC/IPMI environment, where it normally would on a normal ‘server-class’ platform. Since the BMC is where many sysadmins make their first stop when investigating a potential hardware issue, lack of memory reporting in the BMC log is a big deal.

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.