Supermicro X10SRH-CF
- Thread starter jcl333
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Hows the noise?I've had both Noctua and SM narrow ILM coolers, and have been happy with both.
But I can't ever really fault the SM one and it's way cheaper, so that's all I use now
I found a 2680v4 for under $300 so we'll see how it goes soon
If its going in a desktop case, you could be bold and go liquid! An All-In-One with the Asetek style mounting (Corsair, Antec, NZXT, Thermaltake, and Zalman all have compatible models) will work using a Narrow ILM mounting ring kit ($8.00) from the Asetek online store.
Haven't touched an AIO watercooling in more than a decade, are they really more performant than a good Noctua?
Probably not, or at least not enough to matter. The only reason I tried it originally was because I was building a project machine into an original 2006 vintage Mac Pro case I'd been given and told to do "something interesting with it". I couldn't use air coolers on the dual processor board and keep the original fan shroud, but an AIO fit underneath just fine and the Radiators + 25MM thick fans fit well in the space that the Apple fans formerly occupied. Once I lowered the thresholds on the fan curve to accommodate the slower fans, it stayed pretty silent under normal operation.
That's why it was a recommendation to be bold, rather than it being any better
Yes the last time I used it was a H50 on a 1090T because I couldn't fit a big coolerProbably not, or at least not enough to matter. The only reason I tried it originally was because I was building a project machine into an original 2006 vintage Mac Pro case I'd been given and told to do "something interesting with it". I couldn't use air coolers on the dual processor board and keep the original fan shroud, but an AIO fit underneath just fine and the Radiators + 25MM thick fans fit well in the space that the Apple fans formerly occupied. Once I lowered the thresholds on the fan curve to accommodate the slower fans, it stayed pretty silent under normal operation.
That's why it was a recommendation to be bold, rather than it being any better
Ended up taking a U12DX i4... like new from Amazon Warehouse. It's still a bold move?
The threshold on the noctua fans is a non issue. A simple one-line command with ipmitool or similar will adjust the threshold for whatever port you use the fan on. You can do it booting from a Linux live USB if necessary. The change is persistent across reboots. Not trying to guide you one way or the other, just communicating that rotational speed in regards to the alarm threshold should not necessarily be a factor.
Good luck
^ This I'm using the NH-U12DX i4 noctua on my X10SRH-CLN4F (which is physically the same board) doesnt block any standard profile ram.
You do however have to remove the cooler to remove/replace/add to the closest slots (i've been able to angle it in too with the fans removed).
This is the code I use from a linux box with ipmitool installed to resolve the threshold issue:
You do however have to remove the cooler to remove/replace/add to the closest slots (i've been able to angle it in too with the fans removed).
This is the code I use from a linux box with ipmitool installed to resolve the threshold issue:
Code:
ipmitool -H 192.168.1.250 -U ADMIN -P ADMIN sensor thresh FAN1 lower 100 100 100There is also a Windows version that works similar to IPMI tools (it is called ipmiutil but uses similar commands).
ipmiutil - IPMI Management Utilities
ipmiutil - IPMI Management Utilities
I need a new board for a 15 bay system that isn't taxed too hard and spends most of its time at idle. I was thinking E3 (Xeon D is still more than I want to spend right now), but this board looks interesting.
Anyone have any idea how much more power an E5 system of this generation would draw at idle over an otherwise similar spec'd E3 system? By the time of Haswell, I think the C6xx and C2xx chipsets were only ~3w apart, and I'd think an LSI 3008 is going to draw the same regardless?
But I don't know if DDR4 RDIMM vs DDR3 UDIMM is a factor or how well Haswell E5s idled (vs E3) . I've got an E5 v3 in my desktop, but it doesn't idle much.
Anyone have any idea how much more power an E5 system of this generation would draw at idle over an otherwise similar spec'd E3 system? By the time of Haswell, I think the C6xx and C2xx chipsets were only ~3w apart, and I'd think an LSI 3008 is going to draw the same regardless?
But I don't know if DDR4 RDIMM vs DDR3 UDIMM is a factor or how well Haswell E5s idled (vs E3) . I've got an E5 v3 in my desktop, but it doesn't idle much
.
DDR4 ram uses less voltage so respectively should use less power.
That said it all depends on your core/thread/frequency needs, these chips can get as low as 50 TDP which is comparable or lower than some of the E3's.
The e5-2608L v3 for example is a 6c/12t @ 2.0ghz and is only 52w TDP: Intel® Xeon® Processor E5-2608L v3 (15M Cache, 2.00 GHz) Product Specifications
The mobo isn't likely to be a significant draw.
I have two of these systems with a 2690 v3 and a 2660 v3 cpu they pull an average of about 155w each (320w minus the switch from the wall).
That said it all depends on your core/thread/frequency needs, these chips can get as low as 50 TDP which is comparable or lower than some of the E3's.
The e5-2608L v3 for example is a 6c/12t @ 2.0ghz and is only 52w TDP: Intel® Xeon® Processor E5-2608L v3 (15M Cache, 2.00 GHz) Product Specifications
The mobo isn't likely to be a significant draw.
I have two of these systems with a 2690 v3 and a 2660 v3 cpu they pull an average of about 155w each (320w minus the switch from the wall).
I have two version of the SM board, one version is without the SAS controller, the other one is with SAS controller.
Both board with 128gb ram ( 8 x16 gb ) , E5-4667v3 ( 16cores) , 1 SSD.
Running Windows 10 for testing purpose
test board on the workbench without chassis fans , or hard disks
Without SAS , Windows 10 idle at 35-36w
With SAS controller, board idle at 46w
Both board with 128gb ram ( 8 x16 gb ) , E5-4667v3 ( 16cores) , 1 SSD.
Running Windows 10 for testing purpose
test board on the workbench without chassis fans , or hard disks
Without SAS , Windows 10 idle at 35-36w
With SAS controller, board idle at 46w
My guess is about 18w-20w for E3-1240 v3 or E3-1270 v3, 4 x8 gb 32gb ram, 1 SSD
No case fans , no hard disks.