Overclock.net banner

8101 - 8116 of 8116 Posts

·
Registered
Joined
·
799 Posts
Quote:
Originally Posted by superstition222 View Post

The weird thing is that the only thing that really aggravates my tinnitus, besides listening to music with headphones, is my computer. This is despite it being in a closet with a feather blanket blocking most of the door, a 480mm rad with four quiet 950 RPM sleeve fans, a 360mm with vardars on the lowest speed my fan controller can manage (and two in push so they're inside the case), a hacked GPU that has a 140mm fan strapped to it, and so on. The overall noise level with the doors mostly closed, across the room, is very low. But, I get terrible persistent tinnitus, including a midrange sound that makes it sound as if something is running when it's not, after using the computer. It lasts for days. I tried adding the extra radiator to turn the vardars way down but that didn't fix the problem. It did lower the overall noise level, along with the GPU mod.

My hypothesis is that something is creating high-frequency sound and that that's what is bothering my hearing the most. However, it may be that the noise from the fans and the pump is steady that is the culprit. I am getting to the point of considering switching to a nearly fanless build with Intel's overpriced Broadwell C, after putting so much into my watercooling. The Broadwell beats the FX in most everything but with a wattage of around 60 when the CPU is maxed (not including the GPU). I loathe giving Intel a cent because of their variety of anti-consumer tactics*, but the 8370E at stock is rather slow and the VRMs need more airflow than those for Z97.

* a short list:

1) No proper thermal transfer because of cruddy polymer TIM, a decision that incentivizes people to destroy their chips with delidding.
rolleyes.gif

2) Asking an extra $100 for hyperthreading.
rolleyes.gif

3) Selling the expensive "E" models without the current cores (e.g. Broadwell E should never be released).
rolleyes.gif

4) Not putting EDRAM with Skylake so the "revolutionary" 6700K gets beaten by a lower power lower clocked Broadwell in most games.
rolleyes.gif

5) Providing just a tiny 16 PCI-E lanes with Z97. Board makers were asking hundreds of dollars for these boards.
rolleyes.gif


I suppose I might just drop the 8370E to stock and see if I can undervolt a bit and wait for Summit Ridge. I have a quiet air cooler with wide fins I can pair it with. Boring, low power, and slow. But, at least it won't cost money. I'd invest in a passive watercooling setup but the only stuff on the market is aluminum. And, even then, it could be the pump making the noise that's bothering me for all I know. It's probably the vardars and the PSU, though.

I'd like to find a way to run the Fury Nano semi-passively (very low fan speed) but I think I'll wait for 14/16nm to upgrade my GPU since 4 GB of VRAM is on the small side for the future.

People mentioned having the computer in a different room but the problem is my cat. I don't want to damage her hearing and she goes to every room. If something high-frequency causes me this much trouble in such a short amount of time I don't want to create a cat with tinnitus. And the only other room is my spouse's and it would be inconvenient due to the timing of naps and sleep. I'm more of the night owl.
Could be your PSU that's causing your tinnitus.. I had a problem with high frequency noise from one of my cheapo 750 watt PSUs that would stay with me for days..solved it by blowing it up running Prime and got a new one with no noise....
 

·
Registered
Joined
·
660 Posts
Quote:
Originally Posted by Mega Man View Post

Quote:
Originally Posted by reaver83 View Post

Thats funny, my brother got one, he just misplaced it. Thats fine, got ahold of him today and he made a ISO of it and saved it to his Storage Drive. The disc was to install AMD FX's Version of Chill Control VI v6.1.0, and the updated silabs USB Bootloader. I got Antec Chill Control working just fine on this computer for now, it just reads that the Firmware for the USB bootloader is newer than the Software can read, which makes updating it impossible. No worries to me, since it is running all functions of the cooler.
Yea , again I call bs. There is no software like this for the oem aire cooler unless you are talking about amd overdrive that I have ever seen or heard of. Now if you mean some crappy aio software I can see that being true
Sorry if I wasn't clear, This was for the Liquid cooler that came with the box FX-8150 and Liquid Cooler Combo back in 2012. it has software that comes with Disc with a bootloader program to control the fans for the cooler, since they plug into it, not the Mobo. Almost all Closed loop coolers do, unless they plug directly to the Mobo (Antec Khuler 620 does for instance) the rest of the LIQUID CLOSED LOOP COOLERS come with programing to control fan speeds. This is Antec's version of Chill Control.

 

·
Registered
Joined
·
660 Posts

·
Missing friends
Joined
·
15,996 Posts
Ah I see, sorry
 

·
Registered
Joined
·
3,326 Posts
Quote:
Originally Posted by superstition222 View Post

The advantage is that it's easier to manage the heat load of a higher clock. For people with moderate cooling and moderately capable boards it may help them with MMOs and other single-threaded apps. But, for multithreading it tanks the performance.

The voltage required isn't dramatically lower than with the whole chip enabled in my experience but the heat output is significantly lower. Since my board (UD3P 2.0) has a throttling issue and forces one to use BCLK to get past 4.4, though, YMMV.
The difference in his post is that he somehow is getting "better performance" disabling three of his six cores. That makes no sense whatsoever. He should just buy a single core Sempron and take it to the logical limit!
wink.gif
 

·
Missing friends
Joined
·
15,996 Posts
it isnt worth trying anymore chris

he wants us to prove a negative, when the burden of proof is on him, he just doesnt like it because he knows he cant
 

·
Registered
Joined
·
2,725 Posts
Quote:
Originally Posted by Mega Man View Post

it isnt worth trying anymore chris

he wants us to prove a negative, when the burden of proof is on him, he just doesnt like it because he knows he cant
Are you talking about me? If so, if you want to be demanding then you can also produce your evidence. The burden of proof is on the person making strident demands and acting like he knows everything.

In the club thread you accused people of posting "misinformation" and included The Stilt in your list of people whose posts you had issues with. But, instead of proving your case against his post you just offered anecdotal evidence like what some overclocking guides say. That's not hard evidence.

If you're going to entertain yourself by calling people out then you can do more than just call them out and argue with their posts. You can produce hard evidence.
 

·
Registered
Joined
·
154 Posts
Quote:
Originally Posted by superstition222 View Post

The advantage is that it's easier to manage the heat load of a higher clock. For people with moderate cooling and moderately capable boards it may help them with MMOs and other single-threaded apps. But, for multithreading it tanks the performance.

The voltage required isn't dramatically lower than with the whole chip enabled in my experience but the heat output is significantly lower. Since my board (UD3P 2.0) has a throttling issue and forces one to use BCLK to get past 4.4, though, YMMV.
Quote:
Originally Posted by Mega Man View Post

Yea , again I call bs. There is no software like this for the oem aire cooler unless you are talking about amd overdrive that I have ever seen or heard of. Now if you mean some crappy aio software I can see that being true
Again people have claimed this for a while, but never can show proof, and the ones that do, show the opposite is true
I would like to present you guys this excel sheet, i made a long time ago. Put A LOT of hours into adjusting testing and noting things down.
To trim down the time needed to make this chart, the heat tests where mostly only 10 ITB and 30 min prime, especially in the 'inbetween' numbers.

AMDfx8150_v2.xlsx 25k .xlsx file


i DID notice a rather LARGE difference in voltage needed to run stable.
Testing was done on Sabertooth 990 FX rev. 1 with an AMD FX-8150

i will highlight here:

3600 mhz, 8 core -> 1,156 volts (so 4 modules)
3600 mhz, 6 core -> 1,125 volts (3 modules)
3600 mhz, 4 core -> 1,075 volts (2 modules)
3600 mhz, 2 core -> 1,075 volts (1 Module)
--> do notice low difference between 2 VS 1 module.

4500 mhz, 8 core -> 1,44375 volts
4560 mhz, 8 core -> 1,38125 volts
4500 mhz, 4 core -> 1,325 volts
4500 mhz, 2 core -> 1,29375 volts
--> difference of a whopping 0,15v

For temps comparison we must do with these numbers.
4300 mhz, 8 core --> 68 °C using 1,368v
4300 mhz, 6 core --> 62 °C using 1,3125v
4200 mhz, 4 core --> 47,5 °C using 1,212875v
4200 mhz, 2 core --> 40 °C using 1,2v
--> difference of about 25 °C and 0,16v.

These temps where PRE-WC and with STOCK HS.
To bad i lost some 'in-between' notes and my laziness made me not write down every measurement at the time...

I didn't tested and compared speeds higher than 4600 with disabling modules,
because i found that running only 1 module was just ******ed for my normal day to day use of this PC (-and I was going to WC).

Also this: i did do a quick test last summer with a watt-meter and the difference between 8-6-4-2 was just stupid low. In best a few watts per disabled module. I was very disappointing. If you REALLY don't believe this. Go buy a meter, just a few bucks, or i could retest this and attach some pics. Would rather like if you guys pick the first option
biggrin.gif
 

Attachments

·
Registered
Joined
·
2,725 Posts
I only tested at high clocks (4.7 to 5.3) which could explain why voltage difference was smaller but heat output was significant.

However, as I mentioned, my board has odd throttling issues which mucks things up in general.

Additionally, the 8320E and 8370E are lower leakage parts than your chip I'll bet.

It is nice to see evidence, though, that shows a heat reduction and a voltage reduction.
smile.gif
 

·
Registered
Joined
·
153 Posts
Hey guys, I personally own a 4790k, but I'm overclocking an FX 6200 for a friend of mine. I've read some guides and gotten the gist of it, just wondering if 4.8Ghz/1.525v looks okay. From what I had read, max safe voltage is 1.55v and max safe temp is 65c, and it's currently topping out at 64c running Prime95 27.7, and 62c doing handbrake encodes.

Checked his VRM and capacitor temps on his board with an infrared thermometer, everything was at or under 65c as well. Did I dun good?
 

·
Riding on the edge
Joined
·
3,266 Posts
Just to verify, the 64c you mentioned is the Core temp correct?
Hows the Socket temp holding out?

It's never hurts to add active cooling for the VRMs and CPU Socket if you haven't already. Thinking heat soak?

Overall I'd say yes, a 1GHz OC @ 1.525 is fine but does depend on how those temps hold out.
 

·
Politically incorrect
Joined
·
9,464 Posts
Quote:
Originally Posted by Brohem0th View Post

Hey guys, I personally own a 4790k, but I'm overclocking an FX 6200 for a friend of mine. I've read some guides and gotten the gist of it, just wondering if 4.8Ghz/1.525v looks okay. From what I had read, max safe voltage is 1.55v and max safe temp is 65c, and it's currently topping out at 64c running Prime95 27.7, and 62c doing handbrake encodes.

Checked his VRM and capacitor temps on his board with an infrared thermometer, everything was at or under 65c as well. Did I dun good?
Maximum core and socket temperatures are 70C, not 65. You've done a good job.
 

·
Registered
Joined
·
153 Posts
Socket temp was also below 70c. Infrared thermometers are pretty awesome tools, super duper recommend picking one up.

This was with a Hyper 212 on his chip in a Crosshair V Formula motherboard.
 

·
Registered
Joined
·
1 Posts
I have two boards running secondary systems with FX Cpus - i have an 8350 and a 6300 on a ASUS M5a97 Evo and a M5a99FX Pro - they will both run at 4,7GHz on Air with voltage around 1.404V - i have them running at 4.56Ghz for everyday stability and run Phenom MSR tweaker which still kinda works to undevolt the lower power states - i run 2000Mhz and 3600Mhz and 4560Mhz power states - just the 3 - with 1.05V 1.22V and 1.404V respectively - the boards are rock solid - have loads of features, reboot after a bad overclock - bios flashback etc - and overall they make for a very reliable PC that arent far off 4770K performance when overclocked. I have a X370 Ryzen now and so many good features from the FX boards are not as good and the RAM timings have got silly complicated - I like the FX and have run them for years overclocked - both the boards and the CPUS have held up - i even have a phenom X6 1090T that still works but has 6 pins missing - trick to bodge the socket with that many pins, 1 pin missing can work with a bodge easy enough though.
 

·
Registered
Joined
·
3,944 Posts
cool
 
8101 - 8116 of 8116 Posts
Top