Intel Xeon E5-2670 Deal and Price Tracking

handruin

Member
May 24, 2015
51
15
8
Make sure to plug in both 8-pin EPS power connectors. If it still beeps, try to remove CPU2 and only use CPU1 then boot. If it still beeps, looks like you've got a dud motherboard.

Thanks for the feedback I definitely had both 8-pin power cables connected. I tried removing the motherboard from the case to check for any kind of short and there was none. With the board outside the case I tried powering it on and still no luck.

I decided to take out the entire motherboard and connect it to the PSU and case cabling from my other case leaving the motherboard hanging outside temporarily. I was able to power it on and get it to post as well as get into the BIOS to confirm the CPUs and memory are being seen just fine. I'm guessing it's not the removable module part of the PSU but perhaps the circuitry or cabling inside the other case that may be causing the issue?
 

Stereodude

Active Member
Feb 21, 2016
412
67
28
USA
Idle power?
It draws more power in Windows. Idle is around 125w and under load is 380w. I added a graphics card and plugged in an Ethernet cable, but I rechecked with the Linux boot USB and it's still ~85W / ~283W.

I finally got the fans to speed up in Windows with the Acoustic fan profile. They don't speed up until the CPUs hit ~82C.

Can you modify / change fan behavior with IPMI tools to get the fans to ramp at a lower temperature or behave more like a desktop board with a gradual ramp as the CPU heats up instead of letting it get to a certain high temp and only increasing fan speed to hold it at that high temp?
 

jwegman

Active Member
Mar 6, 2016
144
65
28
47
I finally got the fans to speed up in Windows with the Acoustic fan profile. They don't speed up until the CPUs hit ~82C.

Can you modify / change fan behavior with IPMI tools to get the fans to ramp at a lower temperature or behave more like a desktop board with a gradual ramp as the CPU heats up instead of letting it get to a certain high temp and only increasing fan speed to hold it at that high temp?
You can give this a read, particularly the comment by Doc_SilverCreek:
Control Fan Speed on S5520SC ??

His suggestion is to modify a copy of the SDR (sensor data records) to customize the fan profile.
 

Boddy

Active Member
Oct 25, 2014
780
145
43
For buyers interested in Natex goods outside US and Canada, you can email them ('contact us").

I managed to get them to put a sign inside the box with an 'Electrostatic statement' in case box gets inspected by freight company or customs.
They preferred to send to my Shipito address in USA than international freight (unless you have a FedEx account).

I picked up 3 x motherboards, I asked whether they could separate motherboards in separate boxes or with a cardboard layer, in between each M/B.

They replied: 'We will pack them in a double wall box 20 x 20 x 20 . Each board will be wrapped very well and I can put a leaflet in the package with the Electrostatic statement.'

My concern is with the weight of 3 motherboards with heat sinks, whether the weight would damage parts of motherboard in transit? Or am I being over concerned?
 

legopc

Active Member
Nov 2, 2014
220
37
28
27
The Netherlands
The two 120mm fans are not the same part number. The top one that blows on the CPU is a Nidec BETAV V34809-35INT10F. The bottom one is a Delta AFB1212SHE. Both are 38mm thick.

The Nidec looks to run at 3750RPM at full power and the Delta at 5200RPM from my oscilloscope readings.

Sitting in the BIOS with the fans set to Performance, 0 PWM offset, and 0-300m altitude setting:
Nidec: ~46% PWM duty / 1785RPM
Delta: ~25% PWM duty / 1380RPM

Sitting in the BIOS with the fans set to Acoustic, 0 PWM offset, and 0-300m altitude setting:
Nidec: ~28.5% PWM duty / 1120RPM
Delta: ~22.5% PWM duty / 1260RPM

I'm pretty sure the upper Nidec is the one with the annoying whine. No idea if it's normal, or if I have a bad fan.

Since the two fans are not identical, I'm not entirely sure what I should do. I guess I want to see what the PWM duty cycles look like with the CPUs heavily loaded.
I have the same fans in my intel chassis and the nidec does make an annoying sound and I am definitely going to order some nice delta fans from ebay. Going to put 2 fat radiators in my system and I hope those delta's will provide low temps :D
 

rubylaser

Active Member
Jan 4, 2013
842
230
43
Michigan, USA
Anyone with the Intel S2600CP how the heck do you get the Remote Control working? I installed the Intel AXXRMM4 Remote Management and the AXXRMM4Lite as well. It is the RMM is recognized in the IPMI, but when I click on Console Redirection, it just downloads a file with this path (http://192.168.172.7/page/jnlp_launcher.html), and does nothing. I have tried this in Chrome and Safari on OS X 10.11 and Windows 10 in Chrome and Edge. Any help is appreciated!



Also, is there anyway to remove the fan warning for the rear fan without re-running the FSDR thing again? I thought the rear PWM fan header was called rear fan, so I had it monitor it, but it's apparently called FAN7.
 

jwegman

Active Member
Mar 6, 2016
144
65
28
47
Anyone with the Intel S2600CP how the heck do you get the Remote Control working? I installed the Intel AXXRMM4 Remote Management and the AXXRMM4Lite as well. It is the RMM is recognized in the IPMI, but when I click on Console Redirection, it just downloads a file with this path (http://192.168.172.7/page/jnlp_launcher.html), and does nothing. I have tried this in Chrome and Safari on OS X 10.11 and Windows 10 in Chrome and Edge. Any help is appreciated!
It's a java application, so you need to make sure that your browser has a java plugin that can execute it.
 
  • Like
Reactions: rubylaser

poto

Active Member
May 18, 2013
222
86
28
Anyone with the Intel S2600CP how the heck do you get the Remote Control working? I installed the Intel AXXRMM4 Remote Management and the AXXRMM4Lite as well. It is the RMM is recognized in the IPMI, but when I click on Console Redirection, it just downloads a file with this path (http://192.168.172.7/page/jnlp_launcher.html), and does nothing. I have tried this in Chrome and Safari on OS X 10.11 and Windows 10 in Chrome and Edge. Any help is appreciated!



Also, is there anyway to remove the fan warning for the rear fan without re-running the FSDR thing again? I thought the rear PWM fan header was called rear fan, so I had it monitor it, but it's apparently called FAN7.
I'm having better luck with firefox for running ipmi remote sessions. I think Chrome is quietly blocking the console popup windows.
 

rubylaser

Active Member
Jan 4, 2013
842
230
43
Michigan, USA
It's a java application, so you need to make sure that your browser has a java plugin that can execute it.
I knew that, but apparently I didn't have Java installed on this Windows workstation. And, Java on OS X is a pain, so no surprise. I just installed Java on the PC and it's working like a champ. Thanks for fixing my stupidity :)

 
  • Like
Reactions: Did

Luke Pulaski

New Member
Mar 16, 2016
4
0
1
33
Not sure if anyone brought up this before, but just want to let everyone who is interested in converting a Dell T5600 to a dual E5-2670 workstation know that the 635W power supply comes with T5600 single cpu workstation is not enough to power dual E5-2670. You will have to have a 825w power supply. Some 2x quad core T5600 also equipped with the 635w PSU so be careful. The 825W power supply alone is about $300.00 on eBay. T5600 won't take a generic PSU due to its special design.
 

Fritz

Well-Known Member
Apr 6, 2015
2,725
807
113
68
I picked up a Lenovo Thinkstation D30 1120w WS PSU to power mine. Works great and was only $43 delivered.
 
  • Like
Reactions: rubylaser

Stereodude

Active Member
Feb 21, 2016
412
67
28
USA
If anyone wants to help tackle the .sdr modifications to change the way the fans work on the S2600CP2J I'd welcome some input.

I think I have figured out that there are multiple domains in the system (3 in the Intel Chassis I'm using 0, 1, & PS) and each domain has a particular Stepwise Curve associated with the domain that ultimately controls the fans. The problem is for the chassis I have "P4000M_BASE" there are multiple record entries for each of the domains, and I can't determine which of the entries is used. For example, there are 6 record entries for Domain 0 and 1. Each one has a unique Record ID, but I can't see how to determine which of the 6 it uses (or maybe all 6 are used).

Edit: I think they're the multiple entries it chooses from depending on the fan profile you pick (Acoustic / Performance) and the altitude you pick. The base PWM values I see at idle make sense. For Domain 0 the least aggressive Stepwise Curve ID would put the fan in the in the upper 20's at my room's ambient temp in my house, and the most aggressive stepwise curve ID would put the fan at the low 60's at my room's ambient temp. That's pretty much the PWM duty that I measure with my oscope for the two extremes I can pick in the BIOS (acoustic under 300m vs. performance max altitude) for Domain 0.

Frankly what I really want to know is how to set the temperature threshold it's trying to hold. I'm less concerned with the idle fan speed. It seems like it tries to hold a -20 P1 / P2 thermal margin (but admittedly I'm not really sure that's what drives the fan RPMs). I'd like to increase the thermal margin target to -30 or maybe even -35.
 
Last edited:

nthu9280

Well-Known Member
Feb 3, 2016
1,627
468
83
San Antonio, TX
If anyone wants to help tackle the .sdr modifications to change the way the fans work on the S2600CP2J I'd welcome some input.

I think I have figured out that there are multiple domains in the system (3 in the Intel Chassis I'm using 0, 1, & PS) and each domain has a particular Stepwise Curve associated with the domain that ultimately controls the fans. The problem is for the chassis I have "P4000M_BASE" there are multiple record entries for each of the domains, and I can't determine which of the entries is used. For example, there are 5 record entries for Domain 0. Each one has a unique Record ID, but I can't see how to determine which of the 5 it uses (or maybe all 5 are used).
I'm in a slightly different situation but still needing to modify the sdr. FRUSDR doesn't recognize my P4216xxmhgr chasis and marked it as other. I'll give it a try early next week as I'll be out of pocket starting tomorrow.
 

OliG

New Member
Feb 21, 2016
7
4
3
39
Cant find much info on theses sticks; is this ECC memory ? I guess it is, but just want to be sure...
 

Stereodude

Active Member
Feb 21, 2016
412
67
28
USA
I'm in a slightly different situation but still needing to modify the sdr. FRUSDR doesn't recognize my P4216xxmhgr chasis and marked it as other. I'll give it a try early next week as I'll be out of pocket starting tomorrow.
I'm pretty sure I've figured out how to modify the SDR to do what I wanted. I lowered the clamp temperature by 10C on several of the thermal sensors and the system ran about 10C cooler under 100% load. I then lowered it another 5C for a total of 15C reduction and now it runs 15C cooler under 100% load. The Domain 0 (CPU) fans ramps up to about 41.6% duty from an idle of about 22.8%. I will probably leave the .sdr alone for now until I see how warm the system gets once it's in my basement doing what I built it for instead of just prime95 stress tests.
 

jwegman

Active Member
Mar 6, 2016
144
65
28
47
I'm pretty sure I've figured out how to modify the SDR to do what I wanted. I lowered the clamp temperature by 10C on several of the thermal sensors and the system ran about 10C cooler under 100% load. I then lowered it another 5C for a total of 15C reduction and now it runs 15C cooler under 100% load. The Domain 0 (CPU) fans ramps up to about 41.6% duty from an idle of about 22.8%. I will probably leave the .sdr alone for now until I see how warm the system gets once it's in my basement doing what I built it for instead of just prime95 stress tests.
Can I ask why you wanted to modify this? Were you concerned by the core temps at load? There shouldn't be *any* problem whatsoever provided that the the cooling system can maintain a negative thermal margin, whcih from what I've seen, the default heatsink and cooling duct does *very* well. At full load, I've never seen my thermal margin decrease lower than -18% as the cooling fan(s) will increase RPM to maintain that margin. What I understand you are doing is increasing the fan RPMs earlier to maintain a larger margin at the expense of more noise, which seems counter intuitive to your desire for less noise (or perhaps it was just the type of noise that you were bother by)... If you just wanted the fans to spin faster for a given pwm signal, you could consider increasing the FAN PWM OFFSET in the BIOS settings?
 

Stereodude

Active Member
Feb 21, 2016
412
67
28
USA
Can I ask why you wanted to modify this? Were you concerned by the core temps at load? There shouldn't be *any* problem whatsoever provided that the the cooling system can maintain a negative thermal margin, whcih from what I've seen, the default heatsink and cooling duct does *very* well. At full load, I've never seen my thermal margin decrease lower than -18% as the cooling fan(s) will increase RPM to maintain that margin. What I understand you are doing is increasing the fan RPMs earlier to maintain a larger margin at the expense of more noise, which seems counter intuitive to your desire for less noise (or perhaps it was just the type of noise that you were bother by)... If you just wanted the fans to spin faster for a given pwm signal, you could consider increasing the FAN PWM OFFSET in the BIOS settings?
First, heat is the enemy of electronics. With few exceptions the hotter you let something run the shorter it's lifespan will be. Second, the temperature of the CPU seems to be a factor in the turbo settings. At 80C the cores run slower than at 65C. I picked up about an extra 200MHz in clock speed by running it 15C cooler. At ~65C running the exact same Prime95 test it would run about 3.0gHz with all cores loaded. Occasionally one would drop to 2.9gHz. At ~80C it was 2.8gHz with occasional drops to 2.7gHz. I can't think of any reason to let it bake at higher temperatures. It's not like it goes from nearly silent to screaming. I'm not sure it actually gets all that much louder. It's just the pitch of the Nidec BETAV fan that changes. That Nidec BETAV fan generates is a very unpleasant slight grinding whine.

Yes, I know I changed the noise character, probably made it a little bit louder. At this point I've pretty much decided to stick this system in my basement where the noise won't really matter as long as it doesn't go full power on all the fans. I still may change out the fans, or at least the upper CPU fan, just to see what happens since I'm pretty sure I can control the idle RPM and other fan characteristics by modifying the .sdr file.