-
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.
AMD® Phenom™ II Overclocking Thread
- Thread starter Big.Wayne
- Start date
gurusan
I just been delving back through the thread looking for some info and I came across one of your posts I missed (#443) . . . I know it's not funny but I LMAO when I read the problems you were having, especially the bit about clocking the chip so hard your PC "Shut Down" and wouldn't boot at all until you removed the CPU!!
Loadline Calibration (LLC)
Deffo need to get time to experiment with this feature as I'm getting more vCore ripple than I like . . . . on the later LGA775 boards Loadline Calibration only had three options [Auto] [Enabled] [Disabled] so it's interesting too see all the different % of LLC. I see a few of you guys have played around with this? . . . I just read eddyc's post (#624) where he suggests 16%/19% as a good starting point? . . . . also is this a universal feature amongst all the AM2+/AM3 boards and if so have most people left theirs set to [auto]?
I'm dealing with an assault of vDrop, vDroop and McRipple!
OCCT Website here
I just been delving back through the thread looking for some info and I came across one of your posts I missed (#443) . . . I know it's not funny but I LMAO when I read the problems you were having, especially the bit about clocking the chip so hard your PC "Shut Down" and wouldn't boot at all until you removed the CPU!!
Loadline Calibration (LLC)
Deffo need to get time to experiment with this feature as I'm getting more vCore ripple than I like . . . . on the later LGA775 boards Loadline Calibration only had three options [Auto] [Enabled] [Disabled] so it's interesting too see all the different % of LLC. I see a few of you guys have played around with this? . . . I just read eddyc's post (#624) where he suggests 16%/19% as a good starting point? . . . . also is this a universal feature amongst all the AM2+/AM3 boards and if so have most people left theirs set to [auto]?
I'm dealing with an assault of vDrop, vDroop and McRipple!
OCCT Website here
Last edited:
Most people do not get above 3.8 on any stepping for 24/7 use because getting it stable requires a very large voltage bump which most people consider either dangerous or not worth it for the small gain in performance.
the-tazs what kind of cooling are you using? I personally would feel quite uncomfortable using 1.57 vcore 24/7. Its all down to preference though. Some like to live on the edge!
thats my situation aswell i keep getting tempted try higher voltage to get what i believe to do 4.1 maybe but i draw the line at 1.5v at moment
at 1.5 i can do 3.9 solid 24/7 but..prefer 3.7 @ 1.40 . mine boots and does most things fine at 4.1 with 1.5v so im sure if i went that high i could get it at least over 4 stable but why risk it just yet
as for nb overclocking i notice it is quicker but the temps do shoot up a lot quicker and stay higher. upto 2.4 on mine and not much change anything over and we start to see quite a big leap .
gareth ive noticed the voltages thing aswell weird that
as for the llc not messed with that any good articles to read about it anyone?
Load line calibration
This is the best review I have found of my mobo.
On the page I have linked to there is some specific testing of the load line calibration feature (which I have never seen anywhere else) I imagine that this is fairly generic to asus mobos (or perhaps all am3 mobos?).
They conclude that 19% is the best setting to eliminate vdroop. I figured that I would use 16% because it causes the vcore to very slightly increase under load. I reckon that on the bleeding edge of stability a slight increase in vcore as the load increases might just prevent a crash. I have no proof of that though.
This is the best review I have found of my mobo.
On the page I have linked to there is some specific testing of the load line calibration feature (which I have never seen anywhere else) I imagine that this is fairly generic to asus mobos (or perhaps all am3 mobos?).
They conclude that 19% is the best setting to eliminate vdroop. I figured that I would use 16% because it causes the vcore to very slightly increase under load. I reckon that on the bleeding edge of stability a slight increase in vcore as the load increases might just prevent a crash. I have no proof of that though.
really??
i've upped my cpu nb to 2.7ghz 1.35v from 2000mhz 1.1v and the temp as only gone up about 1 or 2c under load.. im only using air cooling with a Thermalright Ultra-120 + 2 fans as push/pull
my max tcase temp under load is 46c
Last edited:
Soldato
[SIZE="5]I just been delving back through the thread looking for some info and I came across one of your posts I missed ([URL="http://forums.overclockers.co.uk/showpost.php?p=16093888&postcount=443"]#443[/URL]) . . . I know it's not funny but I LMAO when I read the problems you were having, especially the bit about clocking the chip so hard your PC "Shut Down" and wouldn't boot at all until you removed the CPU!!
lol yeah I was getting frustrated. Just the usual getting used to a new system though...the board is very good but I still stick by my original statement of it not having enough fine tuning options.
Soldato
Not really. It just sorts the sheep from the goats, or the wheat from the chaff
.If running 3DMark06 over Prime is enough to generate errors or BSOD your PC it is not stable. BitTech use looped Prime95 Blend + 3DMark06 left 24 hours (or maybe it's just overnight - I'm not sure) as their standard stock stability test for motherboard reviews. The majority of current AM3 and Intel based boards pass (i.e. both are still running at the end of the test), although I believe they didn't have a single Phenom I based rig pass the testing. Athlon X2s and Phenom IIs have usually been fine though.
Last edited:
Soldato
You're all just hiding the fact that you're running rigs that will fall over once you do anything intensive when Prime's running
I've done 2 hours Prime Blend with a 3DMark06 loop on this rig straight after an 8 hour Prime blend. No errors. No BSODs. No resets.
Prime is only sensitive to errors if you're generating them because your rig is unstable.
I've done 2 hours Prime Blend with a 3DMark06 loop on this rig straight after an 8 hour Prime blend. No errors. No BSODs. No resets.
Prime is only sensitive to errors if you're generating them because your rig is unstable.
Soldato
You can fire up a CPU intensive program that is resilient to errors and run it all you want, but that's not going to make your rig crash. Whereas programs like IBT and Prime95 are just the opposite and will crash on any errors.
In other words... you are saying that running an "error resistant program" alongside an "error sensitive program" will somehow crash your rig faster than just dedicating 100% of the CPU to finding errors... it just doesn't make sense.
In other words... you are saying that running an "error resistant program" alongside an "error sensitive program" will somehow crash your rig faster than just dedicating 100% of the CPU to finding errors... it just doesn't make sense.
Surely the purpose of stability testing is to predict crashes in the normal use of the computer. Different levels of testing will predict more or less likely crashes.
It is down to the personal preference of the user and what they will be doing with their PC.
Clearly an 8 hour prime blend is not enough for someone who will be doing weather modelling which will cause their cpu and memory to run at 100% for a week straight. But it is enough for someone who will spend a lot of time playing games which only ever hit 100% cpu occasionally.
Stability testing is not mandatory. If you do not care if you get a BSOD every now and then or if you would rather test the stability by just using your computer for a month, why bother?
It is down to the personal preference of the user and what they will be doing with their PC.
Clearly an 8 hour prime blend is not enough for someone who will be doing weather modelling which will cause their cpu and memory to run at 100% for a week straight. But it is enough for someone who will spend a lot of time playing games which only ever hit 100% cpu occasionally.
Stability testing is not mandatory. If you do not care if you get a BSOD every now and then or if you would rather test the stability by just using your computer for a month, why bother?
Last edited:
u wouldn't say that if your o/s corrupts and needs reinstalling, just because it wasn't stable
Last edited:
Soldato
Games hitting 100% CPU use on a quad with high GPU load is starting to happen occasionally. Ghostbusters does it on occasion. GTA IV is pretty close. More and more games are actually using quad cores these days. 3DMark + Prime is a pretty good simulation of that.
With stability testing I want results quickly. If CPU and RAM are not stable, adding 3DMark06 into the mix reveals instability a lot quicker than Prime95 alone. I tend to use IntelBurn, S&M and 3DMark06 + Prime95 Blend for quick testing. Only when I think my PC is really stable will I usually go to blending for 8 hours. Prime95 is simply not very good for a quick stability test.
I'm very intolerant of instability - hence rigorous testing is required to ensure it. I've been getting occasional BSODs on boot with my X2 unlocked to X4. It never causes any problems once Windows is actualy running or during gaming / stability testing but it's still bugging me enough to consider turning off ACC the majority of the time and just have it on when I'm doing something that really benefits from quad.
Edit: Bit of history. Back in the day I was running a socket 7 rig that became increasingly more unstable, eventually to the point that I couldn't run games for more than about 10 minutes without a crash. 3DMark (can't remember which version) + Prime turned out to be an easy way to reproduce the crashing. I've used them together for stability testing ever since.
With stability testing I want results quickly. If CPU and RAM are not stable, adding 3DMark06 into the mix reveals instability a lot quicker than Prime95 alone. I tend to use IntelBurn, S&M and 3DMark06 + Prime95 Blend for quick testing. Only when I think my PC is really stable will I usually go to blending for 8 hours. Prime95 is simply not very good for a quick stability test.
I'm very intolerant of instability - hence rigorous testing is required to ensure it. I've been getting occasional BSODs on boot with my X2 unlocked to X4. It never causes any problems once Windows is actualy running or during gaming / stability testing but it's still bugging me enough to consider turning off ACC the majority of the time and just have it on when I'm doing something that really benefits from quad.
Edit: Bit of history. Back in the day I was running a socket 7 rig that became increasingly more unstable, eventually to the point that I couldn't run games for more than about 10 minutes without a crash. 3DMark (can't remember which version) + Prime turned out to be an easy way to reproduce the crashing. I've used them together for stability testing ever since.
Last edited:
gareth170 - I personally run prime blend for 12+ hours and consider that stable. I have yet to have a crash or a corrupted OS so that seems to be good enough for my purposes. If I start to get crashes then I will have to re-evaluate this.
All I was saying is that people's uses of computers and their level of tolerance for instability is so diverse that it is impossible to recommend a generic stability test suitable for all circumstances.
I have seen people bumbling along quite happily with laptops so slow, full of junk and unstable that to me they would be completely unusable but they seemed happy enough so what can you do.
EDIT: Granted this does not generally apply to overclockers.
All I was saying is that people's uses of computers and their level of tolerance for instability is so diverse that it is impossible to recommend a generic stability test suitable for all circumstances.
I have seen people bumbling along quite happily with laptops so slow, full of junk and unstable that to me they would be completely unusable but they seemed happy enough so what can you do
. EDIT: Granted this does not generally apply to overclockers.
No it does not.
When windows shows 100% CPU load, it does not actually mean that the CPU is loaded 100% constantly, games are not optimised enough and have many wasted cycles.
I've never had an OS corrupt having run AMD in the past and now C2D despite some OCs not being entirely stable.