![]() At the stock voltage, which drops to around 1.09V on our chip under load, Wraith Spire comfortably kept our CPU cool at full fan speed (around 2700 RPM). We did a quick batch of testing with AMD's included Wraith Spire RGB cooler and found it to be a decent unit. Assuming that the mid-to-high-60s temperature reading for Ryis accurate, based on AMD's sensor design, I would have no worries about operating under these temperature levels for extended usage periods. Variations this small are too close to declare anything other than tied performance between the pair. When overclocked using a load voltage of 1.395-1.417V, our Rysample operated similarly to the 1700X. However, limiting the comparison to other Ryzen 7 chips, the 1700 still shows excellent thermal performance. There have been suggestions that AMD's temperature-reading sensor is inaccurate, which has an influence on our analysis. Thanks to its low power draw and superb heatspreader design, which uses solder and a liquid-metal TIM, Ryshows excellent temperature performance at stock. ![]() However, the reported temperature reading with the Ryoverclocked brings into question the operation of AMD's temperature sensor and/or offset value for the Ry(which should be operating around 20☌ lower than reported, similar to 1700X and 1800X levels, if it has no offset reporting coded in). The Rydoes not have a temperature offset so its reported values are the actual junction temperature values, according to AMD. AMD is advising that 20☌ can be subtracted from the reported Ryzen 7 1800X and 1700X temperature values to understand the actual junction temperature. ![]() Update : AMD has announced that Ryzen 7 1800X and 1700X have a +20☌ temperature offset on the reading currently displayed by software, over the actual junction temperature. Performance differences below around 3☌ are small enough to say that cooling results were similar. Ambient temperature was maintained at around 20☌.ĭue to the use of slightly different CPU coolers, our temperature measurements should only be used as a guideline. Each CPU cooler's fans were running at full speed. We use different coolers for speed purposes. Temperature recordings were taken using the Noctua NH-D14 for the LGA 1151 test system, Cryorig R1 Ultimate for the LGA 2011-3 system, and Noctua NH-D15 for the AM4 system. Overclocked using the same voltage as the Ryzen 7 1700X, Rydraws practically the same power. That's an important metric for many prosumers who pay for electricity or users who want to squeeze a high-performance CPU into a SFF chassis with limited thermal budget. Put simply, the Ryprocessor's performance-per-Watt is absolutely superb. The Core i7-6800K manages 8.4 points-per-Watt, the 6950X is at 11.0 points-per-Watt, and Ryzen 7 1800X sits at 10.4 points-per-Watt. Compared to the other Ryzen 7 SKUs, the 1700's 30W reduction in TDP translates closely into the power draw decrease observed.Ī quick Cinebench-point-per-Watt calculation scores the Ry(stock) at 11.7 points-per-Watt, with the Core i7-7700K scoring 9.7 points-per-Watt at its stock (forced-turbo) frequency. All of the Intel HEDT chips that we tested drew more power under load than the Ryzen 7 1700, despite being outperformed by AMD's part in many scenarios. Power consumption from the stock-clocked Ryis superb thanks to its voltage being reined in to hit the 65W TDP target. We use a Platinum-rated Seasonic 760W PSU and install a GTX 1070 video card that uses very little power. ![]() Power draw readings are accurate to around +/- 5W due to fluctuations in the value even at sustained load. The same test parameters were used for temperature readings. The power consumption of our entire test system (at the wall) is shown in the chart. Cinebench has a short run time on high-performance CPUs which influences the validity of the temperature reading, so we run AIDA64 stress test to validate data. For CPU load results we read the power draw while running the Cinebench multi-threaded test as we have found it to push power draw and temperature levels beyond those of AIDA 64 and close to Prime 95 levels. We leave the system to idle on the Windows 10 desktop for 5 minutes before taking a power draw reading. ![]()
0 Comments
Leave a Reply. |