theotorm Posted February 18, 2016 Author Share Posted February 18, 2016 You don't need shims and this has nothing to do with the installation. You CPU temps are warmer in the first 10 minutes than they are at 4 hours. This is most likely because the initial low fan speeds cannot quite contain the instant 100% load. As I mentioned earlier, you can rule this out of your testing by setting a constant fan speed prior to the start of the test. It will also provide more consistent results. While average CPU temps may be somewhat consistent between tests on different days, individual peak core values may not unless you have a way to maintain consistent room temperature, case temperature, and fan speed. I don't think you need to take it that far. I do not understand why 'with the old h110i gt the temperature rose to more and consequently the fans ... in this rather well in game gets to 29 ° and the fans to 1000. I played 2 hours assassin's creed syndicate and the most 'high temperature' was 67 ..... with that of the first (2015) did not exceed 61 .... then there is a problem? I do not know '' what to think ... I can keep it so 'but then what do I do in the summer? Link to comment Share on other sites More sharing options...
c-attack Posted February 18, 2016 Share Posted February 18, 2016 You are using a custom fan curve. Are they both identical? Do both systems have the same GPU? Is it the same physical case for both systems? All of these things factor into your cooling. This is why people use a test bench without some of these variable or the exact same machine to test. Even with identical BIOS settings, there is no guarantee both chips will be exactly the same in peak core temperatures. If you are really going to proceed down this path, the only logical thing to do is switch the coolers. If the core temperatures values also switch along with it, then there is something to talk about. If machine #2 stays the warmer one, you will know it is tied to the individual machine and it's other components. Link to comment Share on other sites More sharing options...
theotorm Posted February 19, 2016 Author Share Posted February 19, 2016 You are using a custom fan curve. Are they both identical? Do both systems have the same GPU? Is it the same physical case for both systems? All of these things factor into your cooling. This is why people use a test bench without some of these variable or the exact same machine to test. Even with identical BIOS settings, there is no guarantee both chips will be exactly the same in peak core temperatures. If you are really going to proceed down this path, the only logical thing to do is switch the coolers. If the core temperatures values also switch along with it, then there is something to talk about. If machine #2 stays the warmer one, you will know it is tied to the individual machine and it's other components. I understand and thanks of 'aid ... But my question is always this ... because the old h110i gt the temperature of the water rose even to 32 ° while this remains low? This is my dilemma. If no salt of the 'water temperature does not go up even the fans .... thing that the old one does. * Surely there is something wrong ... 'cause okay two different houses but the cpu are equal. I lighthouse 'I pay another h110i and see if anything changes. Meanwhile, I did like you said 2 test with fans with custom bending and fans to the max. Curve Custom Aida 18 minutes: 20 ° 500 rpm 25 ° 1013 rpm 26 ° 1371 rpm 27 1841 rpm 29 2126 rpm http://i68.tinypic.com/2woirme.jpg http://i66.tinypic.com/2dvlv9s.jpg Fan MAX Speed Aida 13 minutes: http://i64.tinypic.com/2vjofpc.jpg http://i67.tinypic.com/34got1w.jpg It is not 'cause the old h110i gt'd left the pre-applied paste ??? Yet here I am 'using the MX-4 they say is the best ... Anyway I want to take another H110i GT 'cause the old one (2015) does not have more'. I was told in this discussion than the average temperature of 'I7 4790K is above 10-15 ° with respect to the tests that are found on the net ... this' cause? Sure I do not put shims on the backplate to make more 'pressure on the cpu? 'cause a lot of talk about it on the net ... I wait opinion Thanks so much Link to comment Share on other sites More sharing options...
c-attack Posted February 19, 2016 Share Posted February 19, 2016 I am not sure all of your message translated properly, but I will try to answer the best I can. These graphs are for the second H110/4790K? They appear to be several degrees cooler on the peak temperatures compared to the earlier results. You did not go above 80C on any core, if that was your target. Changing to a premium thermal interface material (MX-4) might give you a 2-3C reduction at most. However, that is yet another difference between the two units and could be a factor. However, skill in applying paste almost always trumps brand. Are you saying the older unit with the pre-applied thermal paste is better? I don't know where the 10-15C numbers comes from. 10-15C over water temp at 100% CPU load? Ridiculous. At idle 5-9C is common, but again it depends upon voltage setting and power saving features like C-states. Perhaps that information is at stock Intel settings, but that seems like too low for any 4790k. Keep in mind most professional reviews are done on open test benches in a stable controlled room. It's hard to replicate that in real life and I am not sure you should try. Don't add shims. There are a couple of specific motherboards where this is a problem. Your's isn't one of them or you would have high idle and extreme load temperatures during AIDA. All you can do is make it worse. I think you should give more credence to your average core temperature while under AIDA, rather than the peak values. The peak values will be different almost every time you run the test. Even your various software programs are reporting different peak values for each core. The only time the peak core values really matter is when your cross whatever threshold your chip has to trigger CPU throttling. You aren't there. You averages appear to be in the 65-70C range, which is good, particularly considering this is at 100% for hours. The only valid thing you can do is to switch the H110iGT's from Computer A to Computer B and see if the temperatures are the same, better, or worse. Don't change LINK settings. Don't change fans. This is a bit of pain and wouldn't normally advise anyone to do this, but you still seem to be concerned and you are not going to be able to file a claim for replacement because two coolers have a few degrees difference between them. You would need to demonstrate a link to that specific cooler and warmer temperatures on whichever machine it is used on. Link to comment Share on other sites More sharing options...
theotorm Posted February 20, 2016 Author Share Posted February 20, 2016 I am not sure all of your message translated properly, but I will try to answer the best I can. These graphs are for the second H110/4790K? They appear to be several degrees cooler on the peak temperatures compared to the earlier results. You did not go above 80C on any core, if that was your target. Changing to a premium thermal interface material (MX-4) might give you a 2-3C reduction at most. However, that is yet another difference between the two units and could be a factor. However, skill in applying paste almost always trumps brand. Are you saying the older unit with the pre-applied thermal paste is better? I don't know where the 10-15C numbers comes from. 10-15C over water temp at 100% CPU load? Ridiculous. At idle 5-9C is common, but again it depends upon voltage setting and power saving features like C-states. Perhaps that information is at stock Intel settings, but that seems like too low for any 4790k. Keep in mind most professional reviews are done on open test benches in a stable controlled room. It's hard to replicate that in real life and I am not sure you should try. Don't add shims. There are a couple of specific motherboards where this is a problem. Your's isn't one of them or you would have high idle and extreme load temperatures during AIDA. All you can do is make it worse. I think you should give more credence to your average core temperature while under AIDA, rather than the peak values. The peak values will be different almost every time you run the test. Even your various software programs are reporting different peak values for each core. The only time the peak core values really matter is when your cross whatever threshold your chip has to trigger CPU throttling. You aren't there. You averages appear to be in the 65-70C range, which is good, particularly considering this is at 100% for hours. The only valid thing you can do is to switch the H110iGT's from Computer A to Computer B and see if the temperatures are the same, better, or worse. Don't change LINK settings. Don't change fans. This is a bit of pain and wouldn't normally advise anyone to do this, but you still seem to be concerned and you are not going to be able to file a claim for replacement because two coolers have a few degrees difference between them. You would need to demonstrate a link to that specific cooler and warmer temperatures on whichever machine it is used on. Hello, is not the ultimate test is always with the same H110i gt (2016) What the 2015 do not have more 'time. I sewmpre compared with screenshots I had kept of 'AIO 2015. My motherboard is a Maximus VII Hero and in this post do see that with the thickness you can 'also lower than 15 degrees with the application of ..but washers / shims do not know if I trust .. You can give him a look and tell me what do you think? http://forums.evga.com/H110i-GT-Unboxing-performance-on-4790k-UPDATE-BETTER-TEMPS-using-washers-on-backplate-m2291170.aspx Now I lighthouse 'I pay another h110i Gt and will compare the temperatures. Sure I do not put the washers? Let me know what you think and thank you so much for your patience in answering. Francis Link to comment Share on other sites More sharing options...
c-attack Posted February 20, 2016 Share Posted February 20, 2016 Some of the Asus boards do have trouble with the back plate. However, when you have a back plate and contact issue, you will get increased load and increased idle temperatures. If you have this issue, you can't have 20-30C idle temperatures. You would idle in the 35-50C range and your stress test runs would be likely be 90C+. That said, you can experiment with it if you like. It will obviously be either better, worse, or stay the same and then you can make a decision. Did I misunderstand? Are there two current existing computers (A and B), each with a 4790K and H110iGT? Or is all of this a comparison to an older computer with similar hardware that you no longer have? Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.