Recent content by UhClem

Won't sustained writing of data to a pool be severely constrained by the poor write performance (35K * 4KB = 140 MB/s maximum) of the M10 as the ZIL device?

and The specs for this (M10 16GB) are Seq-W 150 MB/s max & 4KB_IOPS-W 35k max. Does that warrant the above quotes?

Is this a fact, or hearsay? If "fact", evidence please ... {assuming RDIMM 2133-2666?) -- thanks

You probably do NOT have the Memory Cooling Kit option, which HP says is "required" for >32GB configurations. [I don't know whether you get a hard-fail @ POST, or just a warning during.]

This drive (MG07 Series) is NOT an Enterprise grade drive. It is Nearline with a spec'd "550 total TB Transferred per Year Workload Rating" Despite the Marketing-Slime literature liberally using the "Enterprise" (and "Cloud") buzzword(s) in many places. [Link] ====== [Hey, Toshiba,] "Don't piss...

Be careful with generalizations. Ceacent has TWO SlimSAS (SFF-8654) 8i to PCIe-slot riser cards, the CNS41CX16I & the CNS41CX16W. The 16I has a standard 6-pin power connector, supplying only 12V, which it draws upon via a buck converter (GS7317) to also get 3.3V (max 5A max 3A) for the slot...

You need a reality-check ... :(

Is this the drive referenced in this thread [Link]? [If so, you should have postponed the FS post till you verified Proof of Life.]

Seller clearly states (3 times) that drive is a "Dell 0RTRG0", but used an image of the HP OEM version. If it is a Dell, note that there is a Firmware Update available. Pretty decent price, in current market (rising NAND/NVMe $) [but not compelling enough, to overcome my "I don't really need...

For that same ~$30, you can get E5-2690v4. E5-2667v4: E5-2690v4 (6 more cores! ; and still a good tight speed range): And, for @itronin, E5-2680v4: Yes, the 2680 is only ~$15 now, but, the 2690, with a 10+% boost across the full 14-core range is worth it. ??

Might just be a "Sleeping with the enemy" thing. [I.e., it works (?), but MSFT won't say it in public.] Interesting ... and an excellent example of the dynamic nature of the market; even with just a few small/US sellers. Could be very different in a week or so. When I was looking, Apr-May'23...

(related to pricing/market-dynamics) One significant difference between the LGA2066 and LGA3647, is that 2066 is more for workstation and 3647 for server; resulting in a slow trickle of W-2xxx re-entering vs. large waves of 1st/2nd Gen Scalables re-entering the marketplace. Combine that with the...

Feeling stupid ... but very relieved. (It's alive!) (My first build with a server motherboard) I mistook the length of time at 6D for a "freeze", and the following [90, 89, 88, ... one-second-intervals] countdown as an impending implicit reset. (Even after finding out that 6D was Northbridge...

Agreed; and easy enough to do. I used a blend/combination of the docs for SuperMicro X11SPL-F, Noctua NH-D9, and WC621D8A (putting least trust/priority on the last**). However, I read the "PHM Heatsink" Section 8.6 (Pgs 80-82; attached below[1st & 2nd Gen same]) from Intel's...

(Since I haven't gotten to the stage where I could actually interact with the board/BIOS, my answer is "No, not explicitly.") But, I did do a full "Clear CMOS" procedure[Link] prior to each retry; so maybe I did. Definitely will do. "Looks can be deceiving." I.e., not all/many failures in...