-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
4960x/RIV-BE temp monitoring & rec. limit?
- Thread starter funsoul
- Start date
Hardware monitor works good for me but other soft is also fine. In theory you should keep it below 66*C but in real anything below 90*C for tests is good.
On air/water max safe voltage seem 1.35-1.4V. Above that most IB-E are not clocking good or are overheating. On cold whatever chip likes. Some will go high at 1.5V and won't like high voltages while other will scale good above 1.7V.
On air/water max safe voltage seem 1.35-1.4V. Above that most IB-E are not clocking good or are overheating. On cold whatever chip likes. Some will go high at 1.5V and won't like high voltages while other will scale good above 1.7V.
- Thread Starter
- #3
Thanks Woomack
From the screenshot above...do you really think hwmonitor is the most accurate? There's pretty huge differences (delta=~20C) in load temps depending on app. Hwmonitor ~62, Core Temp and Real Temp ~56, Speedfan ~48 and ai suite (and oc panel) 41
From the screenshot above...do you really think hwmonitor is the most accurate? There's pretty huge differences (delta=~20C) in load temps depending on app. Hwmonitor ~62, Core Temp and Real Temp ~56, Speedfan ~48 and ai suite (and oc panel) 41
- Joined
- Dec 4, 2010
A+ When I run P95 I always used realtemp and CPU-Z for monitoring and have been good for me.
- Joined
- Apr 12, 2014
F the H2O, freeze dat beach 
AI Suite usually shows wrong temps. Tbh recently I'm using almost only AIDA64 but on 4930K I had no problems with hw monitor. I saw that realtemp is recommended by some overclockers but as you see, hw monitor shows about the same values. I'm not even touching speedfan.
Easier as in, there isn't a bunch of irrelevant crap on the screen, not installation. Hwmonitor and programs of the like are great for doing everything, but if its just temps one wants, RT or Coretemp for me. I have them both installed. When I just want temps, I use RT, when I want a screen full of data, I use HwMonitor.
- Joined
- Apr 12, 2014
- Thread Starter
- #13
Thanks folks
Yeah Witchy...may have to install a temp probe to get an accurate result. Lemme find the directions (I've got a couple extra probes laying around somewhere). Since I'm only running watery to learn the mobo/cpu, figure out which ram to use (likely psc per shammy) and get the 780's overclocked (how the heck do I connect matrix cards, 780's and others, to the oc panel?)...might not need to care SOO much about temps for now.
Didn't have much time to play yesterday but did try the 190 bclk profile. It worked perfectly (highest bclk I've ever managed) so the imc seems decent. 195 failed to post, 192.5 works so final, max bclk is somewhere between 192-194.9 which seems pretty solid.
Figure I'll start testing some ram over the next couple evenings.
Yeah Witchy...may have to install a temp probe to get an accurate result. Lemme find the directions (I've got a couple extra probes laying around somewhere). Since I'm only running watery to learn the mobo/cpu, figure out which ram to use (likely psc per shammy) and get the 780's overclocked (how the heck do I connect matrix cards, 780's and others, to the oc panel?)...might not need to care SOO much about temps for now.
Didn't have much time to play yesterday but did try the 190 bclk profile. It worked perfectly (highest bclk I've ever managed) so the imc seems decent. 195 failed to post, 192.5 works so final, max bclk is somewhere between 192-194.9 which seems pretty solid.
Figure I'll start testing some ram over the next couple evenings.
High max bclk doesn't mean you actually can use it for better results or higher oc. Optimal for most stuff is 125 strap. It will take some time to figure out best strap/memory ratio as not all will work good.
Even on cold reading is not accurate as it's pot temp, not cpu. Putting it into waterblock or between cpu and block will give you IHS temps, not core. ASUS soft is reading "package" temp that's why it's bad and you see there is like 10*C difference between AI Suite and Real Temp or HW monitor.
I'm still wondering how you haven't figured out what soft use for temp reading when you were benching so much stuff. Also start benching this X79 platform and stop asking stupid questions. Answers will come faster than you think
Even on cold reading is not accurate as it's pot temp, not cpu. Putting it into waterblock or between cpu and block will give you IHS temps, not core. ASUS soft is reading "package" temp that's why it's bad and you see there is like 10*C difference between AI Suite and Real Temp or HW monitor.
I'm still wondering how you haven't figured out what soft use for temp reading when you were benching so much stuff. Also start benching this X79 platform and stop asking stupid questions. Answers will come faster than you think
Last edited:
- Thread Starter
- #18
Thanks everyone!
Yeah...like I said...only had about an hour of playtime so was really just tooling around a little. Since I'm on water atm, figured it would be prudent to figure out which app gave the most accurate reading so I don't kill the chip before I get to take it cold. Yeah...the probe won't give me anything I don't already have...gotcha on the not worth it for water, only for cold thing but since the cascade has a built in temp guage (temp at the head) will just stick with real temp.
Agreed...super high bclk almost definitely will not help with oc nor results but a) wanted to give a couple of the profiles a try, b) was pretty cool to see such a high bclk since I've never gotten more than around 127 before and c)I always like to find max so I know the boundaries for oc testing.
Omg...cannot wait to get home and settled tonight then spend a couple more hours playing with this setup! Psyched to see what kind of aida results can be had with the tridentx and pi's in quad channel.
Yeah...like I said...only had about an hour of playtime so was really just tooling around a little. Since I'm on water atm, figured it would be prudent to figure out which app gave the most accurate reading so I don't kill the chip before I get to take it cold. Yeah...the probe won't give me anything I don't already have...gotcha on the not worth it for water, only for cold thing but since the cascade has a built in temp guage (temp at the head) will just stick with real temp.
Agreed...super high bclk almost definitely will not help with oc nor results but a) wanted to give a couple of the profiles a try, b) was pretty cool to see such a high bclk since I've never gotten more than around 127 before and c)I always like to find max so I know the boundaries for oc testing.
Omg...cannot wait to get home and settled tonight then spend a couple more hours playing with this setup! Psyched to see what kind of aida results can be had with the tridentx and pi's in quad channel.
Probably better than DDR4
here is one link with results on X79 and samsung kit:
http://www.overclockers.com/forums/...jawsZ-4x4GB-DDR3-2133-CL9-F3-17000CL9Q-16GBZH
A Core i7-4960X has 6 cores so if you use RealTemp, you need to run RealTempGT.exe which supports the 6 core CPUs.
Make sure you are using the most recent version of CPUID HWMonitor. The Package temperature represents the highest temperature anywhere on the CPU. When this is being reported correctly, the package temperature should always be equal to or higher than the highest core temperature. Your screenshot shows the Package temperature being reported significantly less than the peak core temperature so something is not right. The latest version of CPUID HWMonitor should work correctly. It does not have this problem on my 4 core CPUs.
AI Suite used to report some sort of CPU socket temperature. Whatever it reports is useless information. The peak core temperature determines when an Intel CPU will start to throttle and slow down so you need to monitor that with your fav core temperature monitoring software.
When a CPU is loaded, there shouldn't be any significant difference between HWMonitor, RealTemp GT or Core Temp. When lightly loaded, individual core temps can change so rapidly that some monitoring apps will report slightly different temps. This happens because they are reading the temperature sensors at slightly different times.
Make sure you are using the most recent version of CPUID HWMonitor. The Package temperature represents the highest temperature anywhere on the CPU. When this is being reported correctly, the package temperature should always be equal to or higher than the highest core temperature. Your screenshot shows the Package temperature being reported significantly less than the peak core temperature so something is not right. The latest version of CPUID HWMonitor should work correctly. It does not have this problem on my 4 core CPUs.
AI Suite used to report some sort of CPU socket temperature. Whatever it reports is useless information. The peak core temperature determines when an Intel CPU will start to throttle and slow down so you need to monitor that with your fav core temperature monitoring software.
When a CPU is loaded, there shouldn't be any significant difference between HWMonitor, RealTemp GT or Core Temp. When lightly loaded, individual core temps can change so rapidly that some monitoring apps will report slightly different temps. This happens because they are reading the temperature sensors at slightly different times.
