- 2014_12_11 - Added LGA-2011-3 Haswell-E test to the review
- 2014_11_29 - Added results of Bitspower Summit EF water block to the test
- 2014_09_13 - Added hydraulic testing results to the test.
- 2014_08_27 - Added results of EK-Supremacy EVO water block to the test
- 2014_08_05 - Added CPU block orientation diagram and some more notes.
- 2014_08_03 - Added hydraulic restriction testing diagram, added notes and legend to chapter 2 and Q&A.
- 2014_08_02 - Initial publish.
Hello dear forums members!
It's summer time and it's time for another water block comparison review (see my previous work on XtremeSystems Forums).
In this review, there will be no rewards given, it is purely a performance revealing test to see which water block provides the highest performance on the two of the most popular CPU platforms today:
- DT: Intel LGA-1150: 4th Generation Haswell CPU
- HEDT: Intel LGA-2011: 3rd Generation Ivy Bridge-E CPU
- HEDT #2: Intel LGA-2011-3: 4th Generation Haswell-E CPU
In the course of this review, which will get updated from time to time with new water blocks added to the test bench, the following characteristics will be measured:
- Thermal performance on a LGA-1150 Haswell CPU (brand new, un-modified)
- Thermal performance on a LGA-2011 Ivy Bridge-E CPU (the same as before)
- Thermal performance on a LGA-2011 Ivy Bridge-E CPU (brand new, un-modified)
- Hydraulic restriction curve
The following water blocks will be tested:
- Alphacool NexXxos XP3 Light
- Aquacomputer Cuplex Kryos PRO
- Bitspower Summit EF - added 2014_11_29
- EK-Supremacy EVO - added 2014_08_27
- EK-Supreme LTX CSQ
- Koolance CPU-370
- Koolance CPU-380I
- Swiftech Apogee HD
- Swiftech Apogee XL
- Phobya UC-2 LT
- XSPC Raystorm
- Watercool Heatkiller 3.0
Due to their age the following previously tested water blocks are excluded from this test:
- EK-Supreme HF
- DT 5Noz
- Phobya UC-1 LT
1.1. Thermal testing methodology:
- System is assembled and bled of air. Fresh distilled water is used for every water block.
- Every water block is tested in their optimal mounting position as per experiences from my previous review.
- CPU and water block base are cleaned using Isopropyl Alcohol before each fresh application of Arctic Silver Ceramique 2 TIM.
- Prime 95 v28.5 x64 (custom settings) is used to stress (heat up) the overclocked processor while the RealTemp software logs the core temperatures every second.
- The recorded duration of this stress test is 2000 seconds.
- K-type thermoprobes measure water temperature on the CPU inlet.
- Voltcraft K202 thermocouple thermometer with RS232-USB capability is used as a data logger and logs the water temperature every second. An old laptop is used as a data logger.
- After testing, both logs are time-syncronized. A lot of table formatting is done and the final result is merged in Microsoft Excel.
- The initial 'warm-up' stage (test standard) is discarded for each water block.
- Each test is repeated 5 times, the best- and the worst result are discarded.
Common hardware - computer:
- Intel Core i7 4770K (4c 8t)
- Intel Core i7 4930K (4c 8t)
- Intel Core i7 5820K (6c 12t)
- Gigabyte GA-Z87X-SOC Force
- Gigabyte GA-X79A-UD3
- ASUS X99-DELUXE
- 4x4GB Crucial Value DDR4-2133 (Micron D9)
- 2x 4GB Samsung DDR3-1600L C11 (rev.D HYK0)
- 120GB SSD Kingston Hyper-X 3K
Common hardware - liquid cooling loop:
- Magicool 480mm (4x120mm) radiator
- 4x 120mm Scythe Ultra Kaze 3000rpm
- EK-DDC 3.2 PWM X-RES 140
- EK-PSC Fitting 10/13mm G1/4
- Tygon Norprene 10/13mm tubing
- Distilled water (too expensive to use premixed coolants in this review)
1.2. Hydraulic restriction testing methodology:
- System is assembled and bled of air. Fresh distilled water is used for every water block
- A short (test standard) water loop consisting of pump/reservoir combo, a ball valve (for regulating pump flow rate) and a flow meter is assembled.
- Dwyer type wet-wet differential pressure manometer is used to measure pressure drop over water block at set flow rates. Flow rates are set by adjusting ball valve.
- Aquacomputer Aquaero 4.0 LT is used to measure flow rates of Digmessa FHKU flow meter with 5.6mm nozzle.
- The system is powered by auxillary ATX PSU, jump started.
- After testing, pressure drop values are used with Microsoft Excel to plot the charts.
2. Water block comparison table
Below you will find some basic information about construction and compatibility of each tested water block out-of-the-box. The minimum G1/4 determinates the largest fittings that can be used with certain water block.
GREEN ... Optimal
YELLOW ... Mediocre
RED ... Bad
N* = On demand, Swiftech supplies AMD socket mouting plate free of charge (shipping fee still applies).
N** = Is available for separate purchase.
Edited by Moonmanovich - 12/11/14 at 1:30am