Soldato
- Joined
- 29 Aug 2010
- Posts
- 8,215
- Location
- Cornwall
I tried it at rf level 10 and still only got 40% CPU usage while encoding a DVD.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
**LETS SEE YOUR PILEDRIVER OVERCLOCKS - LET ME START WITH 5GHz+!!**
- Thread starter Gibbo
- Start date
I tried it at rf level 10 and still only got 40% CPU usage while encoding a DVD.
screen shot
You need a higher bit rate source such as a BD rip
You need a higher bit rate source such as a BD rip
Last edited:
Soldato
- Joined
- 29 Aug 2010
- Posts
- 8,215
- Location
- Cornwall
Fair enough, a point worth noting if Gibbo does test using Handbrake.
Yes, Silver Arrow SB-E. Not quite stable in Prime but ran day to day tasks faultlessly. Will get it settled in Prime, but not sure what's causing the instability, got a bit of voltage headroom. But Prime Stable is irrelevant. If it can handle an intensive high quality bluray encode and run games without a hiccup then I'll consider it stable. A high bitrate 2 pass encode of Troy will be my acid test.
Handbrake will use as many resources as can be thrown at it. Run it on a Xeon/Opteron multi CPU setup and the speed would be ridiculous.
Last night's blu-ray rips ranged between 58+61 fps. That was with the CPU running at around 4.4Ghz. Onc eI can stabilise a higher 24/7 overclock that should improve even further.
I am using win7 SP1, do I need to download anything specific for Piledriver? I may just go to 12.11 for catalyst, will that do?
Phanteks go on tonight for phase 2 of overclocking this CPU.
4.6GHz at 1.38V is the highest prime stable so far within temps, 55C core. Extreme LLC
4.82GHz at 1.435V loads and runs fine without BSOD but fails prime and x264 runs too hot. Otherwise I have run plenty of apps at 4.82.
Phanteks go on tonight for phase 2 of overclocking this CPU.
4.6GHz at 1.38V is the highest prime stable so far within temps, 55C core. Extreme LLC
4.82GHz at 1.435V loads and runs fine without BSOD but fails prime and x264 runs too hot. Otherwise I have run plenty of apps at 4.82.
Associate
- Joined
- 26 Jan 2009
- Posts
- 875
nkata, how are you getting to those speeds, via Multi, FSB or a mix of both?
4.6GHz = 230Mhz x 20
4.8GHz = 240MHz x 20
I am just adjusting 'FSB' and voltages. Monitoring NB, HT and mem clocks. At 4.6 mem = 1840Mhz and HT, NB is at 2530MHz from memory, I am not in front of the PC at the moment
4.6ghz is the best I can get Prime95 stable under water with ~1.45V (not going any higher than that), even at that clock/voltage and under water temps seems to be a real problem and I've even noticed throttling occur, whether it's the VRM's overheating (CH Formula-Z) or the CPU temp causing it I'm not sure.
If you want a CPU to overclock and that remains Prime stable and without temp issues I think Intel is far less hassle, although saying that I've no experience with Ivy Bridge and it's dodgy heatspreader/paste issue that could be just as bad.
If you want a CPU to overclock and that remains Prime stable and without temp issues I think Intel is far less hassle, although saying that I've no experience with Ivy Bridge and it's dodgy heatspreader/paste issue that could be just as bad.
Last edited:
A lot of people are struggling getting ivy past about 4.5.
This is worth reading
FX-8350 @ 5Ghz / 1.52v. he also completed a CINEBECH R11.5 run @ 5.245Ghz / 1.56v
I think this chip is worth getting
FX-8350 @ 5Ghz / 1.52v. he also completed a CINEBECH R11.5 run @ 5.245Ghz / 1.56v
I think this chip is worth getting
The chips themselves are easy to cool IMO. My 8320 needds 1.47v for 4.6ghz stable and I'm keeping core temps well under 60C with my limited HTPC case loop.
The VRMs on the boards definitely need more cooling though, probably because these chips consume more power than their Intel cousins. I've added some quiet, small and slow spinning fans to the top of the VRM heatsinks on my Sabertooth and it has made a big difference to their temps (and the cpu socket temps nearby) which in turn has meant stability with lower volts than before.
Is that a Sabertooth rev-1 Its the same board i have, so you would recommend a small fan on the VRM heat-sink. which one did you use?
Which sensor are you reading? are you using Prime95 to test for stability?
From what I've seen the 'CPU temp' sensor is just a runaway train as soon as I try proper stability testing with Prime. I was going to ignore the 'CPU temp' and just follow the 'core temp' but then I saw some throttling occurring.
What waterblock do you have out of interest? I'm left wondering if the EK Supremacy is just pants for AMD because it doesn't seem to be much of an improvement on air.
I'm going to have to try sticking a fan over the socket/VRM's but if that lowers the CPU temps as well then something is very flawed with AMD's temp monitoring.
Last edited:
Mine's an R2 but the heatsink design is the same AFAIK. I took it off and changed the pads and NB TIM, but that didn't help until I popped a couple of 60mm fans on top of it. When it gets to high 60s stability dies off for me so it's no problem now.
CPU temp is the socket temp. It's directly linked ot the VRMs on my board, in terms of heat, so cooling the VRMs has helped loads with socket temps that runaway during stress testing. Socket temps over 65C cause throttling AFAIK. Core temps (from HWmonitor or Coretemp) need to stay under 60C if possible, low 60s is probably OK.
@4.6ghz with 1.47v my core and socket temps are about 51C after an hour of BF3 with my quietest fan profile. Flat out fans I can keep it ~46C but that's too noisey for my living room!
Waterblock is a Dtek Fuzion V2 with quad nozzle. I messed up there, really wanted the larger nozzle in this loop and CBA to drain it and change it now!
The CPU Temps come from a sensor under the CPU Socket, its not actually for the CPU, its for the Motherboard to monitor socket temperatures, the Motherboard software and its BIOS simply call it CPU temp.
Its not actually the CPU's temperature, its the temperature of the socket.
Core temps come from sensors on top of the CPU's DIE, those are the CPU temperatures.
Edit- ok thanks wolvers.
Its not actually the CPU's temperature, its the temperature of the socket.
Core temps come from sensors on top of the CPU's DIE, those are the CPU temperatures.
Edit- ok thanks wolvers.
Previously the VRMs were running into the 70s and the heat was soaking into the CPU socket.
Board heatsink removed;
Ready for refitting;
Homemade block mounting plate and added heatsink fans;
You can see why I need the VRM fans, with the location of the radiator. I think I could have done with this on the X58 board TBH.
Board heatsink removed;
Ready for refitting;
Homemade block mounting plate and added heatsink fans;
You can see why I need the VRM fans, with the location of the radiator. I think I could have done with this on the X58 board TBH.
Caporegime
- Joined
- 18 Sep 2009
- Posts
- 30,276
- Location
- Dormanstown.
TLDR version.
CPU temp = Socket temperature, but needs keeping ~ 65c or you'll see throttling because of it.
Core temp = An offset temperature, but needs to be kept ~62c (Which I still don't think it's the same calibration for everyone, I've seen an FX6 report 6c idle)
Think I could get my socket temp/CPU temp on my CH IV to about 71/72c without any issue, seems a bit meh if the improved AM3+ socket can't handle higher?
But then the Core temp maximum of 71c AMD gave on the 1055T 95W's you'd never see because it's actual temperature would be a lot hotter if you actually managed to get it to 71c lol.
CPU temp = Socket temperature, but needs keeping ~ 65c or you'll see throttling because of it.
Core temp = An offset temperature, but needs to be kept ~62c (Which I still don't think it's the same calibration for everyone, I've seen an FX6 report 6c idle)
Think I could get my socket temp/CPU temp on my CH IV to about 71/72c without any issue, seems a bit meh if the improved AM3+ socket can't handle higher?
But then the Core temp maximum of 71c AMD gave on the 1055T 95W's you'd never see because it's actual temperature would be a lot hotter if you actually managed to get it to 71c lol.
Last edited:
How did you fasten those fans to the heat sink? I have a bit more room in my case (HAF 912+) but as you can see my CPU cooler is somewhat in the way, i could turn it facing the normal way but that will increase temps as the fan will be right behind optical drives and not picking up the airflow from the 200mm intake fan.
The alternative is to get a H2O cooler, but i want to see how it does on the Hyper 412S before i decide. (another exhaust fan at the top out of shot)
The alternative is to get a H2O cooler, but i want to see how it does on the Hyper 412S before i decide. (another exhaust fan at the top out of shot)
Last edited: