Supermicro SuperServer 5018A-FTN4 Rackmount Server
-
- SPCR Reviewer
- Posts: 1115
- Joined: Fri Mar 04, 2005 9:07 pm
- Location: Vancouver
-
- *Lifetime Patron*
- Posts: 2000
- Joined: Tue May 15, 2007 1:39 am
- Location: Finland
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
I was definitely interested in rackmounts a good while back, primarily due to their small form factor and standardised mounting. However, the fans were terrible at best, and unfortunately "almost tolerable" seems to be the case today as well.
Good thing that today we've got smart NAS boxes that can take on light server duties. Just miss the mounting options.
Good thing that today we've got smart NAS boxes that can take on light server duties. Just miss the mounting options.
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
why you dont try to work with media server program like plex media server with 1 or 2 bluray playback ?
i think of this product for home server to serve 1-2 tv and ipad
i think of this product for home server to serve 1-2 tv and ipad
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
Is this a question for the reviewer? Perhaps you didn't notice -- bluray playback from this machine is not possible (unless you're talking about streaming it from this server through the network, which is routine & not worth separate testing). The integrated video is straight VGA unsuitable for anything other than ordinary monitoring. Yes, you could add a video card, but that defeats the low power server purpose. You don't really need anything near this level of performance to run a Plex media server, which can be done even by a NAS box.hhnn wrote:why you dont try to work with media server program like plex media server with 1 or 2 bluray playback ?
i think of this product for home server to serve 1-2 tv and ipad
-
- Posts: 434
- Joined: Tue Oct 04, 2005 6:10 am
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
Why can't internal PSUs be as good as or better than external ones?as the 200W 80 Plus Gold unit inside is only a small step behind a quality external power source.
-
- *Lifetime Patron*
- Posts: 2269
- Joined: Sun May 21, 2006 9:09 am
- Location: Northern California.
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
they CAN be, they just hardly ever are. nearly every OEM cuts corners on their included PSU's.Olaf van der Spek wrote:Why can't internal PSUs be as good as or better than external ones?as the 200W 80 Plus Gold unit inside is only a small step behind a quality external power source.
-
- Posts: 434
- Joined: Tue Oct 04, 2005 6:10 am
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
80+ Gold isn't cutting corners IMO.
In what way is this one behind a quality external power source.?
In what way is this one behind a quality external power source.?
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
Larry's statement should not be taken out of context -- or maybe he should have made it more explicit that the reference is to EFFICIENCY only. There's a simple reason why the pico + adapter ends up more efficient at the <20W loads. The 200W stock PSU is well below its best efficiency range; maximum efficiency with most PSUs is reached at 50~70% of rated load. The <20W load is below the optimal range for the pico+Seasonic 80W 12VDC adapter, but not quite as much.Olaf van der Spek wrote:Why can't internal PSUs be as good as or better than external ones?as the 200W 80 Plus Gold unit inside is only a small step behind a quality external power source.
A class V AC/DC adapter is rated for maximum 87% efficiency (presumably at 50~70% of max load) while a Gold 80+ is rated for 90% at 50% load & 87% at full load. The picoPSU also has some losses through it -- the 12V line is passed straight through, but the 5V and 3.3V lines are converted from the 12V line, so there's always a loss. Perhaps as little as 5%, but this really means that even at its ideal, peak efficiency, the pico+adapter will always be lower than the 87% of the adapter alone.
If the Supermicro system was loaded up with its maximum hardware complement, and power load reached... say 60-70W or higher, the stock Gold 80+ PSU would likely draw LESS power (a watt or 2 or 3?) than the pico.
NOTE -- I will edit Larry's statement to clarify.
-
- Posts: 198
- Joined: Thu May 19, 2005 12:48 pm
- Location: Shi-Khan: Vulcan or MosEisley Tattonnie
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
Mike C:
Thankf for the clear and simple explanation on how the 80+ ratings are figured.
Thankf for the clear and simple explanation on how the 80+ ratings are figured.
-
- Posts: 434
- Joined: Tue Oct 04, 2005 6:10 am
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
I understand, so why does nobody put such an adapter into a ATX PSU form factor?MikeC wrote: A class V AC/DC adapter is rated for maximum 87% efficiency (presumably at 50~70% of max load)
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
What's in the adapter is a AC-to-DC converter -- same as in an ATX PSU, except the latter has multiple output DC voltages. The pico adds the other DC lines -- 5V, 3.3V, 5Vsb. Technically, Gold & Platinum 80+ PSUs already outperform a pico + adapter in efficiency at their respective rated outputs.Olaf van der Spek wrote:I understand, so why does nobody put such an adapter into a ATX PSU form factor?MikeC wrote: A class V AC/DC adapter is rated for maximum 87% efficiency (presumably at 50~70% of max load)
The reason why a full fledged ATX PSU with high efficiency at 10~60W will not be produced is simple: There is no money in it, and very little energy savings. 98% of people will not pay more -- or even the same price -- for a lower power PSU even if it is higher efficiency. Between a 100W and a 300W ATX PSU Platinum, the cost of manufacturing is the same, the parts difference virtually identical -- maybe even more for the 100W model. And at 10W load, what will be difference be? Probably 1W.
Marketing guys would scream that they can't sell less power in a product that's supposed to deliver power.
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
Having just built a Windows 2012 R2 Essentials server using the Supermicro A1SAi-2750F board, 8GB RAM, a Fractal Design Node 304, a RAID controller (PERC 6i), and a lot of drives (currently two Samsung SSDs + four 3.5" 2TB "green" drives, but I will be switching to eight 2TB 2.5" drives when finances permit), may I suggest you review just the Supermicro board without all the associated rack-mount hardware? It's simply amazing what can be done with the Avoton microarchitecture, server-class components, and the mini-ITX form factor for less than 70W.
-D
-D
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
We said as much, and I think the review appraises the board well enough. From the last page...derekva wrote:Having just built a Windows 2012 R2 Essentials server using the Supermicro A1SAi-2750F board, 8GB RAM, a Fractal Design Node 304, a RAID controller (PERC 6i), and a lot of drives (currently two Samsung SSDs + four 3.5" 2TB "green" drives, but I will be switching to eight 2TB 2.5" drives when finances permit), may I suggest you review just the Supermicro board without all the associated rack-mount hardware? It's simply amazing what can be done with the Avoton microarchitecture, server-class components, and the mini-ITX form factor for less than 70W.
-D
Perhaps you can share with us photos of your setup and your assessment of its performance... and acoustic qualities, too.The real star of the show is the A1SRi-2758F mini-ITX motherboard and its embedded Atom C2758 SoC. The C2758 represents a giant leap forward for Intel, a snappy 8-core chip that truly addresses the gap behind the previous generation of Atom and the current crop of ULV Ivy Bridge based chips like the Core i3-3217U. The difference wasn't just in the numbers either — we could actually feel the difference. It's the first embedded processor we've used that didn't exhibit noticeable lag or unresponsiveness compared to a "proper" socketed desktop solution.
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
I'm still polishing the server, but when it is cleaned up a bit (wiring is a mess and I'm using a crappy Antec Flex PSU while I modify the wiring on the Seasonic SS-300ET), I'll post some photos / write-up.MikeC wrote:We said as much, and I think the review appraises the board well enough. From the last page...derekva wrote:Having just built a Windows 2012 R2 Essentials server using the Supermicro A1SAi-2750F board, 8GB RAM, a Fractal Design Node 304, a RAID controller (PERC 6i), and a lot of drives (currently two Samsung SSDs + four 3.5" 2TB "green" drives, but I will be switching to eight 2TB 2.5" drives when finances permit), may I suggest you review just the Supermicro board without all the associated rack-mount hardware? It's simply amazing what can be done with the Avoton microarchitecture, server-class components, and the mini-ITX form factor for less than 70W.
-DPerhaps you can share with us photos of your setup and your assessment of its performance... and acoustic qualities, too.The real star of the show is the A1SRi-2758F mini-ITX motherboard and its embedded Atom C2758 SoC. The C2758 represents a giant leap forward for Intel, a snappy 8-core chip that truly addresses the gap behind the previous generation of Atom and the current crop of ULV Ivy Bridge based chips like the Core i3-3217U. The difference wasn't just in the numbers either — we could actually feel the difference. It's the first embedded processor we've used that didn't exhibit noticeable lag or unresponsiveness compared to a "proper" socketed desktop solution.
Thanks,
-D
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
What happened to the audio file for this article? It's no longer there on the last page...
I don't like the idea of idle load (20w) being well below this PSU's efficiency sweet-spot...
If there's an internal PSU that fits into this chassis, is gold+ or higher, & designed for ~100w, that'd be ideal, ditto for an external PSU.*
Unless this CPU/MB actually needs the 200w headroom, for the heaviest compute loads they're capable of spiking to?
Might pop over to JonnyGuru & BenchmarkReviews, IIRC they focused really heavily on PSU's of all flavours.
Thank-you.
*it doesn't have to fit inside the chassis ofc
MikeC wrote:Larry's statement should not be taken out of context -- or maybe he should have made it more explicit that the reference is to EFFICIENCY only. There's a simple reason why the pico + adapter ends up more efficient at the <20W loads. The 200W stock PSU is well below its best efficiency range; maximum efficiency with most PSUs is reached at 50~70% of rated load. The <20W load is below the optimal range for the pico+Seasonic 80W 12VDC adapter, but not quite as much. A class V AC/DC adapter is rated for maximum 87% efficiency (presumably at 50~70% of max load) while a Gold 80+ is rated for 90% at 50% load & 87% at full load. The picoPSU also has some losses through it -- the 12V line is passed straight through, but the 5V and 3.3V lines are converted from the 12V line, so there's always a loss. Perhaps as little as 5%, but this really means that even at its ideal, peak efficiency, the pico+adapter will always be lower than the 87% of the adapter alone. If the Supermicro system was loaded up with its maximum hardware complement, and power load reached... say 60-70W or higher, the stock Gold 80+ PSU would likely draw LESS power (a watt or 2 or 3?) than the pico.
Is there an after-market pico+adpater I can look at, in order to replace the one that comes stock with this unit?MikeC wrote:What's in the adapter is a AC-to-DC converter -- same as in an ATX PSU, except the latter has multiple output DC voltages. The pico adds the other DC lines -- 5V, 3.3V, 5Vsb. Technically, Gold & Platinum 80+ PSUs already outperform a pico + adapter in efficiency at their respective rated outputs. The reason why a full fledged ATX PSU with high efficiency at 10~60W will not be produced is simple: There is no money in it, and very little energy savings. 98% of people will not pay more -- or even the same price -- for a lower power PSU even if it is higher efficiency. Between a 100W and a 300W ATX PSU Platinum, the cost of manufacturing is the same, the parts difference virtually identical -- maybe even more for the 100W model. And at 10W load, what will be difference be? Probably 1W. Marketing guys would scream that they can't sell less power in a product that's supposed to deliver power.
I don't like the idea of idle load (20w) being well below this PSU's efficiency sweet-spot...
If there's an internal PSU that fits into this chassis, is gold+ or higher, & designed for ~100w, that'd be ideal, ditto for an external PSU.*
Unless this CPU/MB actually needs the 200w headroom, for the heaviest compute loads they're capable of spiking to?
Might pop over to JonnyGuru & BenchmarkReviews, IIRC they focused really heavily on PSU's of all flavours.
Are you still going to do this?derekva wrote:I'm still polishing the server, but when it is cleaned up a bit (wiring is a mess and I'm using a crappy Antec Flex PSU while I modify the wiring on the Seasonic SS-300ET), I'll post some photos / write-up.
Thank-you.
*it doesn't have to fit inside the chassis ofc
Last edited by NT on Thu Jun 11, 2015 2:58 am, edited 1 time in total.
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
*BUMP* @MikeC & @derekva
I don't suppose you can spare a minute to address my post now?
viewtopic.php?p=593894#p593894
Thank-you.
I don't suppose you can spare a minute to address my post now?
viewtopic.php?p=593894#p593894
Thank-you.
-
- Posts: 5275
- Joined: Wed Jun 13, 2007 10:12 am
- Location: ITALY
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
With reference to our kind host, I think that an e-mail will be more effective than a post.
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
Okay I sent a PM addressed to both of them & linked to my post...quest_for_silence wrote:With reference to our kind host, I think that an e-mail will be more effective than a post.
viewtopic.php?p=593894#p593894
Weirdly it's sitting in the "outbox" & not the "sent" folder (alongside old PM's I've sent), but it said it was sent successfully *shrugs*.
I tried a second time & it did the same thing, hopefully I don't annoy them with all the spam!
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
It's not clear what you're asking. I've really said all I want to say on the subject -- it simply isn't worth the effort, time or money to try and improve efficiency at 20W. It just won't save enough energy. 2-3W maybe? Just turn off one of your lights a half hour earlier before bed and that will save you as much. As for "best combos" of pico + adapter, do it if you want quieter.... but for highest efficiency, you'll have to hunt on your own.
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
Yikes relax, no need to be so rude/snarky...MikeC wrote:It's not clear what you're asking. I've really said all I want to say on the subject -- it simply isn't worth the effort, time or money to try and improve efficiency at 20W. It just won't save enough energy. 2-3W maybe? Just turn off one of your lights a half hour earlier before bed and that will save you as much. As for "best combos" of pico + adapter, do it if you want quieter.... but for highest efficiency, you'll have to hunt on your own.
No worries, I'll work it all out, don't need your thoughts/help.
Cheers.
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
Really? Rude/snarky? This was not intended, all I said was I've addressed this... tho I admit I was feeling some annoyance at the multiple identical msgs in PM, email & posts from you and perhaps this is reflected in some way in my last post. I find your response over the top.NT wrote:Yikes relax, no need to be so rude/snarky...
No worries, I'll work it all out, don't need your thoughts/help.
Cheers.
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
There's the written word for you, yours seemed pretty over the top to me, & mine in response over the top to you.
Yeah multiple PM's, sorry about that, some slight confusion with using it initially...
As to multiple bumps there hasn't been any, there's been just one, & RE emails I haven't sent a single one to you.
Yeah multiple PM's, sorry about that, some slight confusion with using it initially...
As to multiple bumps there hasn't been any, there's been just one, & RE emails I haven't sent a single one to you.
Re: Supermicro SuperServer 5018A-FTN4 Rackmount Server
Sorry. I've not been on SPCR in a while.NT wrote:*BUMP* @MikeC & @derekva
I don't suppose you can spare a minute to address my post now?
viewtopic.php?p=593894#p593894
Thank-you.
I am still planning on swapping out the power supply and cleaning up the wiring, but before I do that, I need to budget for a set of (8) 2.5" 2TB drives for the RAID arrays.
-D