Overclock.net › Forums › Case Mods & Cases › Builds & Case Mods › Build Logs › [Build Log] Tron Server - HAF XB Dual Xeon Sandy Bridge-EP LGA 2011 ATX
New Posts  All Forums:Forum Nav:

[Build Log] Tron Server - HAF XB Dual Xeon Sandy Bridge-EP LGA 2011 ATX - Page 7

post #61 of 83
Thread Starter 
TOP FAN EXPERIMENTATION


Here are some initial measurements regarding the impact a top fan has in this build. Initially, I tried using a stock CoolerMaster 200mm fan, but it is 30mm high and didn't clear the top of both Artic Freezer i30 heat pipes. I decided to complete the build without a top fan, then measure the temperatures of both Xeon CPUs. Thereafter, I mounted a thinner (20mm) BitFenix Spectre 200mm case fan on top, and I was happy to see that it does clear both heat pipes:


The Xeon E5-2670 processor has an 100C Tj maximum operating temperature limit when pegged at 115W. The Artic Silver 5 thermal compound takes a while to cure, so I'm expecting the following initial temperature readings to drop a bit over the next few days during burn-in.

Idle Performance - No Top Fan

  • CPU1 Average Core Temperature 32C
  • CPU2 Average Core Temperature 34C

Idle Performance - BitFenix Top Fan

  • CPU1 Average Core Temperature 30C
  • CPU2 Average Core Temperature 32C

Prospects for amazing cooling performance look promising. Average core temperatures are simply fantastic at idle with the top fan in place, 68C below Tj Max. Using the BitFenix case fan dropped idle temperatures 2C or about 6%. The fan is extremely quiet mainly because it spins at 700 RPM. I will take this modest temperature improvement at idle. All of the other case fans are Noctua PWM hardware, controlled by the ASUS motherboard's BIOS. All the fans are spinning between 600 and 800 RPM at idle.

Amazingly, the KingWin power supply fan is not spinning at idle, so there is zero PSU noise. It will be interesting to see at what load factor the PSU decides to spin up its fan. Thanks to the CoolerMaster case design and some wonderful Noctua PWM fan hardware, the end result is a server platform that is silent at idle-- at least until the K5000 is installed later today.

redface.gif

I'm looking forward to future measurements at higher/hotter loads...

Note: Only two of the four BitFenix fan mountings align with the case top holes, but using two Nexus silicon fan mounts has nevertheless kept the fan very stable:


Edited by SonicAgamemnon - 4/20/13 at 5:54am
post #62 of 83
Thread Starter 
MORE CABLE MANAGEMENT HELL, QUADRO K5000 INSTALLED


The KingWin 1000W power supply replaced the older Enermax PSU, which means there was a second round of cable management fisticuffs. The good news is the new unit is smaller and the flat modular cables are much easier to manage. Overall, clearances and open space improved, which should help air flow and provide slightly better thermal performance.









The final piece to the puzzle is at hand-- time to kick the passive Radeon adapter to the curb and seat the Quadro K5000. Despite the shorter PCB, the board itself is still rather long, and the 6-pin power connector is located behind the board, not on top, so clearances are rather tight up front next to the 140mm Noctua intake case fan. Fortunately, clearances between the CPU coolers and other slot hardware is good, with plenty of room for good air flow on both sides.




post #63 of 83
Thread Starter 
GLAD TO SEE BORING BIOS


After the motherboard RMA issue, I was never happier to see these rather old school ASUS server BIOS screens. Server BIOS software is boring compared to exotic gaming platforms. Xeon processors are locked. Server ECC registered memory is made for the long haul, so there's no juicing that up either. Basically, there is really nothing to overclock and that makes for boring BIOS. Nevertheless, there was much relief and satisfaction when these crude screens reported both processors and all the memory correctly:




post #64 of 83
Thread Starter 
TRON SERVER: THE STRONG, SILENT TYPE


Initial thermal testing looks extremely promising. CPU and GPU temperatures at idle and load look wonderful. I am truly amazed how well the CoolerMaster HAF XB case, Noctua fans, KingWin PSU and Artic Freezer heat pipes are performing together, operating in silence at idle, and remaining so even while Prime95 is running a torture test. The Quadro K5000 is also silent at idle, and also while working with Adobe CS6 applications. The only perceptible fan noise finally appears from the Quadro card when running GPU 3D stress tests.

That's it-- really superb cooling performance!

The numbers are more than hoped for. The first CPU idles with an average core temperature of 29.75C. As expected, given the push-pull Artic Freezer heat pipe configuration, the second CPU cores idle slightly hotter on average at 32C. The Quadro K5000 idles at a very cool 26C. During Prime95 stress testing the first CPU average core temperature stabilizes at 55.5C, and the second CPU core temperature average is 57.625C. Across all load testing, the highest reported single-core temperature in the first CPU was 59C and the highest single-core reading from the second CPU was 65C-- 35 degrees below the 100C Tj Max limit.

These are really good results!

Besides the case/fan cooling setup, other factors behind these thermal results include the Xeon E5 series improvements in power efficiency (9W idle, 115W max.), a 20 percent power reduction in the K5000 Kepler architecture compared with the prior Fermi generation, and a very efficient, amazingly silent KingWin PSU.


Xeon E5-2670 and Quadro K5000


Idle Temperatures


Prime95 Stress Test Temperatures


Windows 7 System Information


NVIDIA Control Panel


Tron Server Desktop

Edited by SonicAgamemnon - 4/20/13 at 5:57am
post #65 of 83
I'm still amazed how far we've come since the IBM PC of 1981.....

The engineer in me wants you to fabricate a wind tunnel to link both coolers. biggrin.gif

Nice build, I would love a dual Xeon BOINC rig, but that will have to wait till next year. rolleyes.gif

dunx
Wee-PC
(17 items)
 
Wee black box
(13 items)
 
Boinc Box
(10 items)
 
CPUMotherboardGraphicsGraphics
i7 - 960 Asus P6T7 WS Revolution with MIPS water blocks. XFX HD 7950 + waterblock Powercolour R9 280X 
RAMHard DriveHard DriveOptical Drive
Crucial Ballistics Tracer 6 x 4Gb 2Tb WD caviar green Crucial M4 64 Gb SSD  Sony NEC DVD  
CoolingOSMonitorPower
Apogee Drive + Allphacool 180mm rad W7 Dell 29" 21:9 IPS monitor Coolermaster 1250W PSU 
CaseAudio
Antec P280 Scythe Kama Bay Speaker 
CPUMotherboardRAMHard Drive
i7-4770K Asus Impact VI HyperX beast 2400MHz 2x 8Gb WD 1Tb caviar black x3 
Hard DriveCoolingOSMonitor
Intel X-25 SSD 80 Gb Corsair H80 W7 Panasonic 32" HDTV 
KeyboardPowerCaseOther
Keysonic wireless Silverstone 450 W SFX PSU Fractal Design Array R2 Griffin Powermate 
CPUMotherboardGraphicsGraphics
i7 870  Asus Maximus III Formula GTX 480  HD 5870 
RAMHard DriveCoolingMonitor
OCZ 4x 2Gb  WD 2TB Green Corsair H70 Dell 17"  
PowerCase
Coolermaster Silent Pro 750W NZXT Panzerbox 
  hide details  
Reply
Wee-PC
(17 items)
 
Wee black box
(13 items)
 
Boinc Box
(10 items)
 
CPUMotherboardGraphicsGraphics
i7 - 960 Asus P6T7 WS Revolution with MIPS water blocks. XFX HD 7950 + waterblock Powercolour R9 280X 
RAMHard DriveHard DriveOptical Drive
Crucial Ballistics Tracer 6 x 4Gb 2Tb WD caviar green Crucial M4 64 Gb SSD  Sony NEC DVD  
CoolingOSMonitorPower
Apogee Drive + Allphacool 180mm rad W7 Dell 29" 21:9 IPS monitor Coolermaster 1250W PSU 
CaseAudio
Antec P280 Scythe Kama Bay Speaker 
CPUMotherboardRAMHard Drive
i7-4770K Asus Impact VI HyperX beast 2400MHz 2x 8Gb WD 1Tb caviar black x3 
Hard DriveCoolingOSMonitor
Intel X-25 SSD 80 Gb Corsair H80 W7 Panasonic 32" HDTV 
KeyboardPowerCaseOther
Keysonic wireless Silverstone 450 W SFX PSU Fractal Design Array R2 Griffin Powermate 
CPUMotherboardGraphicsGraphics
i7 870  Asus Maximus III Formula GTX 480  HD 5870 
RAMHard DriveCoolingMonitor
OCZ 4x 2Gb  WD 2TB Green Corsair H70 Dell 17"  
PowerCase
Coolermaster Silent Pro 750W NZXT Panzerbox 
  hide details  
Reply
post #66 of 83
Thread Starter 
Quote:
Originally Posted by 2002dunx View Post

The engineer in me wants you to fabricate a wind tunnel to link both coolers...

I hear you . . . and I have suspected the second CPU would suffer all along based on this heat pipe configuration. I was even prepared to retool for liquid cooling (gasp), but the temperature delta is only a few degrees, so it could have been much worse. The way I look at it, the top 200m fan dropped temperatures a few degrees, so that sort of makes up for the "gap", right?

redface.gif

(PS: It looks like the delta grows gradually as loading increases, up to a 4C average core temperature delta...)

Regardless, in hindsight, I'm glad to say that all that worrying was nothing more than a tempest in a teapot...
Edited by SonicAgamemnon - 2/10/13 at 4:23pm
post #67 of 83
Thread Starter 
TRON BONIC CRUNCH


Who needs Prime95 to stress test a system? It appears good old BONIC provides a harder work out. Do you want to evaluate your CPU cooling design? Well, it certainly seems like searching for extra-terrestrial signals in the universal background ether for never-ending hour upon hour is a good approach, no?




Although the Xeon E5-2670 base clock rate is 2.6GHz, note how under heavy load CoreTemp reports TurboBoost is in play bumping it up to a sustained 3GHz. After running BONIC for a few hours with all 32 threads at 100% utilization the average core temperature in CPU 1 was 59.875C. It was 4C higher in CPU2 at 63.875C. The second/rear CPU doesn't intake as much cool air compared to the first/front CPU, and that accounts for this 4C delta at maximum CPU loading. After hours of madly searching for E.T. throughout the cosmos, the average core in the hotter CPU 2 was still 36C below the threshold.

I am satisfied with the CPU cooling situation. It is time to stress the Quadrophonic Kepler 5000...

The BONIC log spew at start-up:
Code:
2/10/2013 8:01:58 PM |  | No config file found - using defaults
2/10/2013 8:01:58 PM |  | Starting BOINC client version 7.0.28 for windows_x86_64
2/10/2013 8:01:58 PM |  | log flags: file_xfer, sched_ops, task
2/10/2013 8:01:58 PM |  | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
2/10/2013 8:01:58 PM |  | Data directory: C:\ProgramData\BOINC
2/10/2013 8:01:58 PM |  | Running under account SonicAgamemnon
2/10/2013 8:01:58 PM |  | Processor: 32 GenuineIntel        Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz [Family 6 Model 45 Stepping 7]
2/10/2013 8:01:58 PM |  | Processor: 256.00 KB cache
2/10/2013 8:01:58 PM |  | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 syscall nx lm vmx smx tm2 dca popcnt aes pbe
2/10/2013 8:01:58 PM |  | OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00)
2/10/2013 8:01:58 PM |  | Memory: 63.96 GB physical, 127.92 GB virtual
2/10/2013 8:01:58 PM |  | Disk: 463.79 GB total, 332.37 GB free
2/10/2013 8:01:58 PM |  | Local time is UTC -8 hours
2/10/2013 8:01:58 PM |  | NVIDIA GPU 0: Quadro K5000 (driver version 310.90, CUDA version 5.0, compute capability 3.0, 4096MB, 8384363MB available, 2167 GFLOPS peak)
2/10/2013 8:01:58 PM |  | OpenCL: NVIDIA GPU 0: Quadro K5000 (driver version 310.90, device version OpenCL 1.1 CUDA, 4096MB, 8384363MB available)
2/10/2013 8:01:58 PM | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 6907411; resource share 10
2/10/2013 8:01:58 PM | SETI@home | General prefs: from SETI@home (last modified 12-Jan-2013 05:18:39)
2/10/2013 8:01:58 PM | SETI@home | Computer location: home
2/10/2013 8:01:58 PM | SETI@home | General prefs: no separate prefs for home; using your defaults
2/10/2013 8:01:58 PM |  | Reading preferences override file
2/10/2013 8:01:58 PM |  | Preferences:
2/10/2013 8:01:58 PM |  | max memory usage when active: 52395.99MB
2/10/2013 8:01:58 PM |  | max memory usage when idle: 58945.49MB
2/10/2013 8:01:58 PM |  | max disk usage: 40.00GB
2/10/2013 8:01:58 PM |  | (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
2/10/2013 8:01:58 PM |  | Not using a proxy
2/10/2013 8:01:59 PM | SETI@home | Restarting task ap_03ja13ad_B0_P0_00065_20130210_02363.wu_1 using astropulse_v6 version 601 in slot 21
2/10/2013 8:01:59 PM | SETI@home | Restarting task 27dc12ac.25059.39081.8.10.250.vlar_0 using setiathome_enhanced version 603 in slot 5
2/10/2013 8:01:59 PM | SETI@home | Restarting task 27dc12ac.25059.39081.8.10.157.vlar_0 using setiathome_enhanced version 603 in slot 20
2/10/2013 8:01:59 PM | SETI@home | Restarting task 27dc12ac.25059.39081.8.10.220.vlar_0 using setiathome_enhanced version 603 in slot 1
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.113_1 using setiathome_enhanced version 603 in slot 3
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.96_0 using setiathome_enhanced version 603 in slot 6
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.154_1 using setiathome_enhanced version 603 in slot 0
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.104_1 using setiathome_enhanced version 603 in slot 9
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.111_1 using setiathome_enhanced version 603 in slot 11
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.135_0 using setiathome_enhanced version 603 in slot 12
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.114_1 using setiathome_enhanced version 603 in slot 2
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.110_1 using setiathome_enhanced version 603 in slot 7
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.155_0 using setiathome_enhanced version 603 in slot 4
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.130_1 using setiathome_enhanced version 603 in slot 8
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.143_1 using setiathome_enhanced version 603 in slot 10
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.21800.15385.8.10.78_0 using setiathome_enhanced version 603 in slot 15
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ac.25900.2930.3.10.113_1 using setiathome_enhanced version 603 in slot 14
2/10/2013 8:01:59 PM | SETI@home | Restarting task ap_30dc12ac_B6_P1_00182_20130210_19626.wu_1 using astropulse_v6 version 601 in slot 13
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ac.25900.2930.3.10.116_1 using setiathome_enhanced version 603 in slot 30
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.26492.26428.8.10.26_1 using setiathome_enhanced version 603 in slot 16
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.26492.26428.8.10.70_1 using setiathome_enhanced version 603 in slot 29
2/10/2013 8:01:59 PM | SETI@home | Restarting task 28dc12ac.25466.7190.7.10.229_1 using setiathome_enhanced version 603 in slot 27
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.26492.26428.8.10.20_1 using setiathome_enhanced version 603 in slot 17
2/10/2013 8:01:59 PM | SETI@home | Restarting task 28dc12ac.25466.7599.7.10.109_1 using setiathome_enhanced version 603 in slot 18
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.26492.26428.8.10.61_1 using setiathome_enhanced version 603 in slot 26
2/10/2013 8:01:59 PM | SETI@home | Restarting task 29dc12ab.26492.26428.8.10.67_1 using setiathome_enhanced version 603 in slot 31
2/10/2013 8:01:59 PM | SETI@home | Restarting task 27dc12ac.28861.41535.9.10.117.vlar_0 using setiathome_enhanced version 603 in slot 23
2/10/2013 8:01:59 PM | SETI@home | Restarting task 28dc12ac.25466.9644.7.10.128_1 using setiathome_enhanced version 603 in slot 25
2/10/2013 8:01:59 PM | SETI@home | Restarting task 28dc12ac.25466.8826.7.10.100_0 using setiathome_enhanced version 603 in slot 22
2/10/2013 8:01:59 PM | SETI@home | Restarting task 17dc12ac.14581.15993.14.10.141_0 using setiathome_enhanced version 603 in slot 28
2/10/2013 8:01:59 PM | SETI@home | Restarting task 28dc12ac.25466.9644.7.10.174_1 using setiathome_enhanced version 603 in slot 19
2/10/2013 8:01:59 PM | SETI@home | Restarting task 28dc12ac.25466.9644.7.10.180_0 using setiathome_enhanced version 603 in slot 24
2/10/2013 8:01:59 PM | SETI@home | Sending scheduler request: To fetch work.
2/10/2013 8:01:59 PM | SETI@home | Requesting new tasks for CPU and NVIDIA
2/10/2013 8:02:01 PM | SETI@home | Scheduler request completed: got 0 new tasks

Edited by SonicAgamemnon - 4/20/13 at 5:58am
post #68 of 83
BUT IT'S TRON. TRON IS ALL ABOUT THAT SUBTLE BLUE LIGHTING. frown.gif
Feed The Dada
(12 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7-2600k Asrock Z77e-itx MSI 7970 Samsung 30nm 8GB 
Hard DriveCoolingCoolingOS
Samsung 840 250GB SSD H100 w/ 2x SP120 Antec Kuhler 920 w/ 1x SP120 Linux Ubuntu 10.10 
PowerCaseAudioAudio
Seasonic X650 Bitfenix Prodigy E07K HFI-580 
  hide details  
Reply
Feed The Dada
(12 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7-2600k Asrock Z77e-itx MSI 7970 Samsung 30nm 8GB 
Hard DriveCoolingCoolingOS
Samsung 840 250GB SSD H100 w/ 2x SP120 Antec Kuhler 920 w/ 1x SP120 Linux Ubuntu 10.10 
PowerCaseAudioAudio
Seasonic X650 Bitfenix Prodigy E07K HFI-580 
  hide details  
Reply
post #69 of 83
Thread Starter 
Quote:
Originally Posted by dmanstasiu View Post

BUT IT'S TRON. TRON IS ALL ABOUT THAT SUBTLE BLUE LIGHTING. frown.gif

I freely admit you've made an excellent point . . . there is deep irony at play here because I'm fresh out of blue neon, yet the server's namesake is Tron . . . and I don't know what to do about this dilemma.

blushsmiley.gif
post #70 of 83
Someone mail this guy some blue el cord pronto.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Build Logs
Overclock.net › Forums › Case Mods & Cases › Builds & Case Mods › Build Logs › [Build Log] Tron Server - HAF XB Dual Xeon Sandy Bridge-EP LGA 2011 ATX