Patrick-STH

Labor is included and discussed both for my time as well as remote hands we pay for. It is very low at this point.

Those are included in our colo costs with a second data center where we have a smaller DR cluster.

Discussed later in the video. But this is possible which adds another layer of complexity if one tries to track pulls.

HPE and some other vendors do this, not just Dell.

Hi tzs - Jim who wrote that Ars article (and confirmed our original finding) actually commented on the link above, noted how he did not test these in a Synology type solution. His is an important data point, but it, like ours, is only one.

After we published our numbers, we found that it was not just ZFS, but other NAS vendors as well. People were sending us their experiences on non-ZFS systems. For example, Synology users were having issues and Synology took them off their compatible list. QNAP for its part is working on ZFS as we discuss in there as well.

Hi dang. I know you did not have to do that but I just wanted to say from the STH team we appreciate it. Have a great day.

Ha! Sometimes it surprises people that we know each other and hang out a bit when we are in the same town.

On an Altra review. Great question. I have been bringing it up for some time and live 15 minutes from their headquarters. The invitation is open on our end.

GCC is the open-source compiler that is used all over. Not every AMD EPYC system people are using AOCC2.0 on. Likewise, people do not only compile code on ICC that is used on Intel Xeons. Arm has focused efforts on getting optimizations in GCC because it is so popular.

Generally, that is why we prefer to publish "compiler optimized" as best-case performance as well as "GCC" as more of the least common denominator. Both sets of data points are important.

Official SPECint published numbers will not use GCC because the organizations that submit them always want to see the best performance. Ampere used a scaling factor off of published numbers.

If you want to see the impact, we have some numbers from my ThunderX2 review: https://www.servethehome.com/cavium-thunderx2-review-benchma...

You can see the impact clearly there even though that was from a few years ago. Cray has a better performing compiler for ThunderX2 but we did not get to use it due to licensing restrictions.

I hope that helps. The bigger need is for more data since this is one view of performance. There are other needs as well such as FP performance.

HPE will probably tell you that you need to use their drives. We have been putting in non-HPE hard drives so far without issue. More on that in the full test.

Street pricing in the US from lower margin HPE resellers for the performance E-2224 diskless model (P16006) has been discounted to $578-$660 depending on reseller/ day.

Still a lot more than some of the previous versions.

Just added a quick note. Everyone in the industry that contacted me today about this seems to be calling it WAFL or something that sounds like "Waffle" for the extra bonus PCIe lanes.

Here is the easy example:

A dual 100GbE NIC requires a PCIe Gen4 x16 slot to run both ports at full speed. Mellanox ConnectX-6 200GbE cards require either 2x Gen4 or 1x Gen5 slot for full dual port bandwidth. Most of the larger NVMEoF nodes we are seeing are using more than one 100GbE NIC.

Enterprise SSDs will rapidly fill Gen4 x4 lanes later this year as we start to see them launch alongside more mainstream platforms. Four of those will fill a x16 slot.

GPUs and PCIe switches are also great examples of uses.

You are right that for the high-volume server segment, 160 PCIe lanes are overkill, as is even 128 today frankly. For some context, there is a huge portion of servers sold with less than 1 DIMM per RAM channel installed.

Likely is there because it requires that AMD's partners release servers in this configuration, and those servers are unreleased. In theory, something during validation could prevent it or partners could decide not to release this configuration (which are unlikely but still possible.)

It is a little hard for us to do so since the M.2 devices came in 16GB and 32GB capacities. The U.2 Optane 900p comes in at 280GB only as of now not the multiples of 400GB that many other drives came in. We actually had, for example, an 800GB S3700 tested but left that result out as it was essentially the same as the 400GB drive.

On the full drive part, a several hundred GB SSD will almost always nearly "empty" in practice for this application. That is why you see 8GB and 16GB RAM based solutions (e.g. the venerable ZeusRAM.)

Hi tw04 - just as a heads up, none of the SSDs were furnished by Intel. We have a budget and bought all of the drives we used on this test. Some we had to buy second hand, some were scavenged from servers we have had to decommission.

Intel does provide us with hardware, as does AMD, Intel, Toshiba, Samsung, WDC, Cavium and etc. In fact, we have a full list of all relationships here: https://www.servethehome.com/about/editorial-copyright-polic...

Unlike virtually every other major review site (and a large number of bloggers), as of today we still do not have direct ad sales to any of these companies.

We also buy an absolute ton of gear for the DemoEval service we run which is why we have access to many different bits. For example, we have purchased a dozen Optane 900p's already since Intel would not furnish us with a U.2 drive.

Intel does not want to have the 900p eat into the P4800X which is a higher margin product. As a result, this is the kind of article (Optane 900p in servers) that Intel specifically would not want us to do. Since we buy hardware, we can.

Awesome! Great to hear. You got me on the typo. Should have been 0.24A per PSU (we meter by outlet on multi-outlet systems). Still 1A on 120V in a 1U is very easy to achieve.

Hi zaroth - We do not use Kill-o-Watt meters since we use calibrated Extech TrueRMS units for our "desktop" testing.

We also have a few racks in Sunnyvale, California that we use APC metered by outlet PDUs for and that we actually calibrated by testing with the Extech units and found two of over a dozen that gave us consistent readings across PDUs and ports.

This unit is in a SC515 chassis with redundant PSUs. Even with 4x 7200 rpm Seagate 2.5" drives, a 64GB mSATA drive and a Samsung SM951 m.2 SSD it is still pulling 0.24A on 208V (so 0.48A on 208V). If you were to gain a bit of efficiency running a single PSU, you could easily fit these in 1A on 120V power envelope.

What we have seen several folks do with similar D-1540 machines is actually use 1U 1A/ 120V hosting (which is very inexpensive) to deliver cheap local PoPs for their applications. We have been strongly considering decommissioning our Las Vegas 1/2 cab DR site and moving to this sort of distributed model as there are a lot of benefits with this.

We published more complete benchmark results ( http://www.servethehome.com/intel-xeon-d-1587-benchmarks-16-... )of the D-1587 yesterday, expect more in terms of reviewing the overall platform next week. The Xeon D is very platform dependent for power and the particular unit we have does have the additional LSI/ Avago/ Broadcom SAS 2116 (16 port SAS2) HBA onboard.