Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Intel x5679 or x5675 v. x5677
- Thread starter usna92
- Start date
- Sort by reaction score
Lou,
I saw your post, but I guess I was asking did you do anything other than tell MFSC to set the min default of the the bottom 3 fan controls to +100rpm? There are a lot of custom options when you open this windows and the last thing I want to do is not cool them enough.
I saw your post, but I guess I was asking did you do anything other than tell MFSC to set the min default of the the bottom 3 fan controls to +100rpm? There are a lot of custom options when you open this windows and the last thing I want to do is not cool them enough.
^^^^I have a couple of apps that monitor Temps. I observed those, fooled around a bit, and chose the ones that IMO, allowed the temp points to run at a satisfactory temperature. As you can see I placed most of the Emphasis on the two Boost Fans. The X5677 were actually my second CPU upgrade. I had W5590s first, and they ran hotter than the X5677s, so I needed a more aggressive fan setting for them.
Lou
Lou
Apart from fixed RPM, you can have something like this.
I don't have the dual processor model to test. However, from flowrider's post, we know that 1600RPM is good for max CPU loading. And that's base on 30C ambient, a reasonably warm ambient as reference. So, we don't need anything significantly higher than that.
Since the fan min at 800 and max at 5200. So, if we want the fan spin up 150RPM per 1C. 5200max, 800min, means there are 4400RPM to cover 30C.
That means if we set the fan base on CPU diode, min 70C, max 99C. We can expect the fan spin up 150RPM every 1C when above 70C, and eventually reach the stabilised RPM (e.g. ~1600RPM at about 76C).
And since we don't know how's your ambient temperature compare to flowrider's. So, I suggest give another 5C buffer on the "start to spin up" temperature, which makes it 65C.
If the calculation is right, most likely the fan will spin up to something between 1600-2000, and the CPU temperature max at between 70-75.
For the intake and exhaust. We can consider something like this.
Since we know the boost fan will keep the CPU cool. So, both intake and exhaust only need to assist the boost fan when the CPU is really warm. e.g. approaching 70C (assume CPU max at around 70-75C from the above fan setting).
So the above setting simply make intake / exhaust fan spin up ~100RPM per 1C (when CPU reach 70C). And that should eventually stabilise somewhere <1000RPM and the CPU <75C.
For your reference, my own fan profile is like this. The lower threshold is 8C higher than what I suggested to you. And that's for my W3690, which has a Tcase max about 10C lower than your X5680. So, even though I can't test those value. I believe my suggestion already put you on the safe side.
For my fan setting and my usage. Even though I keep the CPU >90% usage most of the time. It can clearly keep the CPU at 78C or below, with the boost fan speed usually stabilised at ~1300, and all other fans a bit below 1000.
Last edited:
Ok Looking at your fan controls, it looks like you have several different fan tools. I have iStat Menus and Macs Fan Control. I have modified the Macs Fan Control initially to look similar to the settings you recommended. Which was the 3rd program that looks like a preference panel that allows you to set minimums as well as the rise over time?
Yes, I have 2 fan controlling softwares. I use the one that can set min and lower threshold at the same time (link at post #16) to control my CPU's fan (which will also spin up the fans if the north bridge getting warm). There is no option for MacsFanControl to control one fan base on 2 sensors, that's why I modify this open source software to fit my personal preference.
However, I am too lazy to modify it to control the PSU / PCIe fan as well. And I only need simply control for those fans. Therefore, I install MacsFanControl to control the PSU / PCIe fan. (Make them cooler under light loading, but quieter under high loading. Basically make them stay at more or less the same temperature under any situation)
iStat is only for easy monitoring. I didn't use that for fan control.
I use Macs Fan Control. I have the fans set to stay at minimum speed just 1-2C higher than stock then ramp up much more quickly. The machine is generally much quieter than before and never gets as hot as it used to when being pushed hard. I make the exhaust fans spin up before the intake fans, since the intakes are 2 feet closer to my ears and sound dissipates at an exponential rate.
Well the new processors are in. It seems like on average with the new fan control information set, the processors stay about 1-2 hotter than the old ones did, which is awesome consider I went from 2.66 to 3.33. Sometimes they dip below where the old ones were, at idle, etc. So overall I would say the transplant to the x5680s was a success. Thanks to all with the help for making this possible.
I just upgraded to 12-core (x5690) in my 2009. I'm really unhappy with the fan noise compared to my 6-core 2010 (x3690).
Of course more cores should be louder, but what I did not expect was that single-core applications also create significantly more noise in the 12-core 2009 than the 6-core 2010. Basically, single-core pegged use in my 6-core 2010 is nearly silent, while single-core pegged use in a 12-core is massively loud.
h9826790, I tried your MacsFanControl recommendations and while it did help, it mostly delayed the onset of the loud fans by a minute or so. Eventually they still end up very loud from single-core use.
I actually took the time to repeatedly pull the CPUs, apply thermal paste in different patterns, and pull them off to check coverage until I was satisfied that I had a pattern with no unpasted areas. But this did not help.
Of course more cores should be louder, but what I did not expect was that single-core applications also create significantly more noise in the 12-core 2009 than the 6-core 2010. Basically, single-core pegged use in my 6-core 2010 is nearly silent, while single-core pegged use in a 12-core is massively loud.
h9826790, I tried your MacsFanControl recommendations and while it did help, it mostly delayed the onset of the loud fans by a minute or so. Eventually they still end up very loud from single-core use.
I actually took the time to repeatedly pull the CPUs, apply thermal paste in different patterns, and pull them off to check coverage until I was satisfied that I had a pattern with no unpasted areas. But this did not help.
Last edited:
What's your actual temperature (under stress)?
What's the stabilised RPM?
My suggested fan profile is conservative (because I can't test it myself), if you are OK to let the 5690 run warmer, I think it's possible to further reduce the noise. Or if the noise is coming from a particular fan (e.g. booster), we may reduce that fan RPM a bit and use more intake / exhaust.
In my experience, my 3690 has T-case max 68C, and real world experience shows that any T-diode temperature below 80C is 100% stable. Your 5690's T-case max is 78C. If I were you, I will let it run up to ~85C to reduce fan noise. Of course, this is very personal.
Yeah, when I upgraded, I don't recall anyone mentioning the added noise that two 130W would cause. That's why I always try to bring it up when people are considering.
I have 2 X5690 in my mac pro. I originally had 2 X5650's.
The X5690 can get quite hot if you are putting full load on them. You can use several apps to control the fan curve. In Mac OS I use iStat Pro. I have different fan profiles for different scenarios. Even for general use like browsing, I use slightly higher fan speeds. If I put max load on the CPU's, like converting videos with handbrake, I have a high or medium setting. I like to keep temps below 80°C on the cpu's.
I like macs fan control, but my fans tend to ramp up and down from time to time when using it. For Bootcamp though, its my number 1 choice
The X5690 can get quite hot if you are putting full load on them. You can use several apps to control the fan curve. In Mac OS I use iStat Pro. I have different fan profiles for different scenarios. Even for general use like browsing, I use slightly higher fan speeds. If I put max load on the CPU's, like converting videos with handbrake, I have a high or medium setting. I like to keep temps below 80°C on the cpu's.
I like macs fan control, but my fans tend to ramp up and down from time to time when using it. For Bootcamp though, its my number 1 choice
I must not have stressed my x5680s yet. With the modified fan profile described above, I really haven't broken any significant temperature thresholds, nor rev'd the fans to a speed I found difficult to tolerate. It might be due to the fact that I am not a rendering machine, or that when I rip DVDs, the limiter is really the ancient 2010 era Superdrive still in my machine. Even while running benchmarks, the fans did not seem to spike and temperatures remained normal. It could be possible that the thermal sensors on the side of the processors are not hitting the heat sinks properly resulting in some strange fan behavior?
If you want to run stress test. Download and open Luxmark 3 -> mode -> Stress test (Open CPUs), is one of the easiest way to put the CPUs in stress. This is a free software.
Run that for 15min, then all temperature / fan speed should be stabilised.
If your ambient temperature is low, that will help the CPU to cool down a lot.
I appreciate the response. I am SUPER busy right now, but yes I will post some numbers later and probably ask for advice in the next couple of days.
I ran the stress test. With an Ambient of about 30C I ran the Luxmark 3 Stress Test with Open CPUs. With the fan settings described above, both my x5680s didn't get above 80C. CPUA was at 79-80 and CPUB ran a solid 65C. The noise at load was still well below my DROBO ambient noise and slightly above regular room background noise. I didn't feel like it was running excessively loud while maintaining temperatures. I would be interested to see if they behave similarly loaded for more than the 2 minutes during the benchmark, but I think I am convinced the fans are set near enough to where they need to be with the processor change.
Yes, I agree this test shows that the setting is a good balance between temperature and noise.
So some time has gone by. I've had inconsistent results and I'm not sure what to make of it.
Initially with fairly normal loads I was getting up to 97C on the CPUs and extremely high fan speeds - like jet engines. I used MacsFanControl to make the fans spool up earlier, which helped delay the onset of the extremely high fan speeds. But the delay was only a minute or so, so I set it all back to auto.
After about week, mysteriously the temps are fine and the fans have calmed down. MacsFanControl still set to auto. I am fine with the noise load now. It is almost indistinguishable from my w3690 2010--maybe just slightly louder.
I have no explanation for this other than it seems to have needed some sort of settling in. I have never experienced this before.
Last edited:
Wow. I haven't really experienced that with the x5680s so I am not sure what to make of the inconsistent behavior. Did you take a look at what was running at the time that might have been spiking your CPU loads, maybe artificially raising your temperatures?
It wasn't a rare scenario. Literally any/all demanding applications that would peg at least one core did it, which is almost everything.
I've got dual CPU tray by reasonable price and now thinking about which CPU install.
i thought about pair of X5690, because right now i have single W3690
From factory cMP5.1 have one 130W TDP heatsink for single CPU (W3680) and two 95W TDP heatsinks for dual CPU configuration (X5675).
Single CPU heatsink about 1,5X times bigger than dual CPU cooler.
Will be overheating and high level of noise as result if i install pair 130W TDP processors into 95W TDP heatsink system?
Should i install X5675 instead X5690?
i thought about pair of X5690, because right now i have single W3690
From factory cMP5.1 have one 130W TDP heatsink for single CPU (W3680) and two 95W TDP heatsinks for dual CPU configuration (X5675).
Single CPU heatsink about 1,5X times bigger than dual CPU cooler.
Will be overheating and high level of noise as result if i install pair 130W TDP processors into 95W TDP heatsink system?
Should i install X5675 instead X5690?
Mine is a 2010 5.1 single cpu, with a X5679.
I tried 2 memory sets, first with 16gb in total, and now 32gb in total.
Both memory sets claim to be 1333, but the system report 1066...
Both of your machines are 5.1? both with X5679?
If you already tried the PRAM reset, then try to install only 6 sticks (if all 8 slots are used now). If the memory rank is not right, using all 8 slots can force the RAM to run at low clock speed.
[doublepost=1488810650][/doublepost]
There was some discussion about that, you may try the search function.
I only have a single CPU 4,1, but from memory, overheating is not an issue at all, however, the user may require to create their own fan profile (e.g. via MacsFanControl) because the SMC never decided to handle 2x 130W TDP CPU. Heatsinks are not the problem, they are good enough, but just need the correct fan speed.
And the extra fan speed may cause annoying noise for some users. At the end, noise is very personal.