I am an Unraid user, ~ 11 years. I have my original server as my main device.
I am building a second server for backup and my goal is low cost NAS. I may do more with this box later but the immediate goal is low cost and low power. Looking at the killer nas 4.0 thread i have picked out the following and would appreciate a few of you hardware experts checking this out. What am I missing?
MBD: Supermicro X9SCL … ebay
CPU: Xeon E3-1260L … ebay
RAM: Patriot Viper 3 16GB (2 x 8GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800) Desktop Memory Model PV316G160C9K
The ram is new from newegg because it isnt any more money than ebay used items
I will use the on-board SATA ports for up to 6 drives and if this needs to grow beyond that i will add a PCIE SAS card later. I need to buy a case and PS but no worries about those items.
What about the ram? there is some question about latency timings for the 1600 - this ram is 9-9-9-9-24 iirc while the supermicro tested is all 11’s - i cant remember if this would be an issue.
Any thoughts?
Thanks very much. If this config looks good I will order it in the next couple days and post up the build here in this thread. My main concern is purchasing a MBD / CPU / RAM combo that is incompatible - i believe this pair is compatible.
Thanks for pointing that out. I did some more reading here on serverbuilds.net and decided on the following build, a NAS 5.0 I believe. This should be an improvement on what i have so this hardware will become my primary server and i will start using my existing hardware as my backup server instead.
New build:
MBD: Supermicro X10SLM-F used $70
CPU: Xeon E3-1230L v3 used $ 60 (25 W TDP but ~ 8000 bench lol)
Memory: MEM-DR380L-HL01-EU16 16 GB $80 (used supermicro sticks found on amazon)
PS: Seasonic 650 $65
Case: Coolermaster N400 $60
I have various parts lying around to complete the build. I have a SAS2008 to transfer from my existing hardware to use for the rest of my disks.
Just posting this in case anyone is interested. I could have saved about $50 or so getting a supermicro X9 and a 12x0 v2 but decided to go with the newer hardware for that extra cost. My current server with the C7P67 supermicro board has been very reliable and seems to have plenty of life left after more than 10 yr continuous service, and I expect this newer one will return similar service for me.
Regarding cases, it is a real PITA to find a decent case (i.e, more than 2 or 3 bays for 3.5 in disks) for under $100. You have like 2 to choose from. There are a bunch of interesting ones from Be Quiet, silverstone, etc that have tons of space but you have to buy drive caddies separately for around $20 each. What a load of crap. So it didnt pay for me to look too hard into cases. If you want to spend $150 or so, there are a number of good choices but I will rarely have a look at this thing and really what i needed was tool-less drive installation to make it easy to handle drives when needed.
This is an advantage of the N400 over the Rosewill 4500 that i have. Handling drives in the 4500 is not possible without at least partially removing the drive cage holding the drive you are working on.
and yesterday’s price on the 4500 was about $180, by the way. Its a good case to have but i dont really want two of them. The N400 will be good for my backup server which probably wont ever contain the 10 disks it could hold. that might or might not work so well during a parity check. If I ever do this I will post a comment about it.
Thanks for the explanation. I appreciate you responding and providing the link. Honestly, I’ve never paid much attention to scores until now that I’m researching a build.
Agree, thanks for posting this. I noted this, though;
But for PerformanceTest V10 we did really major changes to the CPU test algorithms. These changes included
Using new CPU instructions (e.g. AVX512) only available in modern CPUs.
Use a more up to date compiler (Visual Studio 2019 instead of 2013) which also brings some code optimization.
Have better support for out of order execution, which is a feature of newer CPUs.
Updated the 3rd party libraries we use for some of the tests (including more modern versions of GZip, Crypto++ and Bullet Physics.
Fixed up a bunch of bugs that hurt performance (like some variable alignment issues and compiler optimization flags).
Completely rewrote some of the tests. e.g. removed TwoFish encryption and replaced it with the more common Elliptic curve encryption.
Improving the algorithms to push more data through the CPU also results in more load on the cache and memory subsystem. So older CPUs, those with inadequate cache or memory bandwidth are expected not to perform so well with PT10
based on these changes, for the cpu’s i have been considering which are more than 5 years old, i am not sure the values posted in the cpu list arent internally consistent.
but, idk, i already have my slow E3-1230L v3 (which apparently, is getting slower :), so job done.
Definitely helpful to know abut the update, Lee, thanks.