Overclock.net › Forums › AMD › AMD CPUs › Phenom II steppings and max OC 550-940
New Posts  All Forums:Forum Nav:

Phenom II steppings and max OC 550-940 - Page 210

post #2091 of 2829
Interesting thoughts about system stability here.

For what it's worth, here are a couple of thoughts.

First, in my opinion, the only time a computer is ever perfectly stable is if it is turned off. There's zero chance of it ever crashing or behaving in an unstable manner when it's not on.

What this also means is that any computer that is operational has some chance of exhibiting in an unstable fashion. Stress testing and then tweaking the settings on any machine, whether it is running underclocked, at stock speed, or overclocked, never really can definitively "guarantee" stability.

Case in point: I will bet the vast majority of OCNers have never run the S&M stress test. I'm a self-admitted ex-stability nazi, and therefore I've tested a lot of different stress testing programs for PCs. S&M is probably the toughest system test I've ever run, and it's probably the only stress test that I know of that can literally break a weak or malfunctioning component (I lost a weak set of Corsair RAM to this test a couple of years ago, as an example).

I run my machines through S&M when I've finalized an overclock. If it passes the battery of tests (I don't run the HDD test, though) in "Long" and "max stress" settings, then I can reasonably say this overclock is "stable."

However, there will still be times when some program or condition will induce an unstable behavior from the machine. An excellent example is my sig rig, S&M-tested and passed at its current setting, freezing only after closing Firefox. Something in the newest flavor of Firefox, of all things, causes the freeze (I tested this six or seven times, and without fail, upon closing Firefox, the machine will freeze).

The point is, virtually anything can cause even a thoroughly-tested machine to exhibit instability.

I agree with the point that testing is a subjective condition. One person's "stable" (even "S&M-stable") is not necessarily another person's. There are just an incredible amount of variables and factors involved for stability to ever be an absolute.

(Except, that is, when the machine is turned off. )
    
CPUMotherboardGraphicsRAM
Phenom II X3 720 BE Gigabyte GA-MA790FXT-UD5P Sapphire Radeon HD 3870 2 x 2GB Crucial Ballistix Tracers PC3 10600 
Hard DriveCoolingOSPower
Almost 8TB of storage AMD Stock Cooler Windows Home Server CoolerMaster SilentPro M 700W 
Case
Lian Li PC-V2000B (with thlnk3r mods) 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Phenom II X3 720 BE Gigabyte GA-MA790FXT-UD5P Sapphire Radeon HD 3870 2 x 2GB Crucial Ballistix Tracers PC3 10600 
Hard DriveCoolingOSPower
Almost 8TB of storage AMD Stock Cooler Windows Home Server CoolerMaster SilentPro M 700W 
Case
Lian Li PC-V2000B (with thlnk3r mods) 
  hide details  
Reply
post #2092 of 2829
Thread Starter 
mr. biggums: It's AMD OverDrive. Which can be downloaded from http://game.amd.com
Sulaco
(20 items)
 
  
CPUMotherboardGraphicsRAM
Intel 3770K ASUS Sabertooth Z77 ASUS X290X Corsair Vengeance 2x4GB 1866MHz 9-10-9-28 
Hard DriveHard DriveCoolingCooling
1x SAMSUNG 830 128GB, 2x Samsung 1TB 32MB Samsun 840 evo Aqua computer Kyros XT Feeser Exchanger 240 
CoolingCoolingCoolingOS
Swiftech MCR320-QP Swiftech MCP 350 Swiftech MCP 350 W8.1 64 
MonitorMonitorKeyboardPower
Iiyama ProLite XB2779QS Black 27" Iiyama E480S 20" Roccat Ryos MK pro CoolerMaster 850W ESBA 
CaseMouseMouse PadAudio
NZXT Switch 810 gun metal Roccat Kone[+] CorePad XXXL Logitech G35 
  hide details  
Reply
Sulaco
(20 items)
 
  
CPUMotherboardGraphicsRAM
Intel 3770K ASUS Sabertooth Z77 ASUS X290X Corsair Vengeance 2x4GB 1866MHz 9-10-9-28 
Hard DriveHard DriveCoolingCooling
1x SAMSUNG 830 128GB, 2x Samsung 1TB 32MB Samsun 840 evo Aqua computer Kyros XT Feeser Exchanger 240 
CoolingCoolingCoolingOS
Swiftech MCR320-QP Swiftech MCP 350 Swiftech MCP 350 W8.1 64 
MonitorMonitorKeyboardPower
Iiyama ProLite XB2779QS Black 27" Iiyama E480S 20" Roccat Ryos MK pro CoolerMaster 850W ESBA 
CaseMouseMouse PadAudio
NZXT Switch 810 gun metal Roccat Kone[+] CorePad XXXL Logitech G35 
  hide details  
Reply
post #2093 of 2829
[QUOTE=XxBeNigNxX;6278715]Seriously,

Originally Posted by XxBeNigNxX
jimibgood,

I personally run 1 hour OCCT (either regular or Linpack) and have NEVER had an unstable OC unless I personally caused it and/or a Driver caused the issue.

You can run 12+ hours of stressing and still have issue's, and by running Long stress tests you also risk damaging Your mobo and/or other hardware.

Don't believe me....Head over to Extremesystems and check out there threads.

Yes there are some that run long stress tests, but the majority do not (AMD side).

Stability is in the eye of the owner/operator.

But NO where in this thread is it required to run over 1 hour OCCT....So do not criticize any one here or on this Forum.


Have a nice day

PS...

AIpha nice job

Funny, you started it.....
Phenom II 955
(13 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 965 Lapped OC'd 4.14GHz (Stable 24/7) Gigabyte GA-MA790XT-UD4P 2 x 1G 4890 Liquid cooled crossfire Corsair XMS DDR3 1600(8G 9/9/9/23/1T) 
Hard DriveOptical DriveOSMonitor
2 x 300G Velocity Raptor(Raid0) WD 500G Barracuda Asus Windows7 64 bit Hanspree 22in 
KeyboardPowerCaseMouse
Logitech Ultra X3 850W Zalman GS1000 Logitech 
  hide details  
Reply
Phenom II 955
(13 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 965 Lapped OC'd 4.14GHz (Stable 24/7) Gigabyte GA-MA790XT-UD4P 2 x 1G 4890 Liquid cooled crossfire Corsair XMS DDR3 1600(8G 9/9/9/23/1T) 
Hard DriveOptical DriveOSMonitor
2 x 300G Velocity Raptor(Raid0) WD 500G Barracuda Asus Windows7 64 bit Hanspree 22in 
KeyboardPowerCaseMouse
Logitech Ultra X3 850W Zalman GS1000 Logitech 
  hide details  
Reply
post #2094 of 2829
Quote:
Originally Posted by txtmstrjoe View Post
Interesting thoughts about system stability here.

For what it's worth, here are a couple of thoughts.

First, in my opinion, the only time a computer is ever perfectly stable is if it is turned off. There's zero chance of it ever crashing or behaving in an unstable manner when it's not on.

What this also means is that any computer that is operational has some chance of exhibiting in an unstable fashion. Stress testing and then tweaking the settings on any machine, whether it is running underclocked, at stock speed, or overclocked, never really can definitively "guarantee" stability.

Case in point: I will bet the vast majority of OCNers have never run the S&M stress test. I'm a self-admitted ex-stability nazi, and therefore I've tested a lot of different stress testing programs for PCs. S&M is probably the toughest system test I've ever run, and it's probably the only stress test that I know of that can literally break a weak or malfunctioning component (I lost a weak set of Corsair RAM to this test a couple of years ago, as an example).

I run my machines through S&M when I've finalized an overclock. If it passes the battery of tests (I don't run the HDD test, though) in "Long" and "max stress" settings, then I can reasonably say this overclock is "stable."

However, there will still be times when some program or condition will induce an unstable behavior from the machine. An excellent example is my sig rig, S&M-tested and passed at its current setting, freezing only after closing Firefox. Something in the newest flavor of Firefox, of all things, causes the freeze (I tested this six or seven times, and without fail, upon closing Firefox, the machine will freeze).

The point is, virtually anything can cause even a thoroughly-tested machine to exhibit instability.

I agree with the point that testing is a subjective condition. One person's "stable" (even "S&M-stable") is not necessarily another person's. There are just an incredible amount of variables and factors involved for stability to ever be an absolute.

(Except, that is, when the machine is turned off. )

Very well put. I read and interesting thought. Why overclock if you were not reasonable certain the system is stable? The person re-iterated why do a 1 hour stress test and assume your system is stable? 1 HOUR STRESS TESTS DO NOT DEEM A SYSTEM STABLE. 8 HOUR STRESS TESTS DO NOT DEEM A SYSTEM STABLE. 24 HOUR STRESS TESTS DO NOT DEEM A SYSTEM 100% STABLE.


Reasonably, 8 hr you can come up with a conclusion it most probably is but a 24 hour makes it clear it is or as close as can be. There is no professional in this industry that would conclude any system is stable at a 1 hr stress.

OF COURSE THE 2 INDIVIDUALS WANT TO DISAGREE.

2 reasons to overclock..... Bragging rights which is cool. Highest possible overclock or getting better performance out of system for gaming and other applications. If you chose the later, why rely on a 1 hour stress. makes no sense. Most novices are happy with a 1 hour.
Phenom II 955
(13 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 965 Lapped OC'd 4.14GHz (Stable 24/7) Gigabyte GA-MA790XT-UD4P 2 x 1G 4890 Liquid cooled crossfire Corsair XMS DDR3 1600(8G 9/9/9/23/1T) 
Hard DriveOptical DriveOSMonitor
2 x 300G Velocity Raptor(Raid0) WD 500G Barracuda Asus Windows7 64 bit Hanspree 22in 
KeyboardPowerCaseMouse
Logitech Ultra X3 850W Zalman GS1000 Logitech 
  hide details  
Reply
Phenom II 955
(13 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 965 Lapped OC'd 4.14GHz (Stable 24/7) Gigabyte GA-MA790XT-UD4P 2 x 1G 4890 Liquid cooled crossfire Corsair XMS DDR3 1600(8G 9/9/9/23/1T) 
Hard DriveOptical DriveOSMonitor
2 x 300G Velocity Raptor(Raid0) WD 500G Barracuda Asus Windows7 64 bit Hanspree 22in 
KeyboardPowerCaseMouse
Logitech Ultra X3 850W Zalman GS1000 Logitech 
  hide details  
Reply
post #2095 of 2829
Quote:
Originally Posted by jimibgood View Post
Very well put. I read and interesting thought. Why overclock if you were not reasonable certain the system is stable? The person re-iterated why do a 1 hour stress test and assume your system is stable? 1 HOUR STRESS TESTS DO NOT DEEM A SYSTEM STABLE. 8 HOUR STRESS TESTS DO NOT DEEM A SYSTEM STABLE. 24 HOUR STRESS TESTS DO NOT DEEM A SYSTEM 100% STABLE.


Reasonably, 8 hr you can come up with a conclusion it most probably is but a 24 hour makes it clear it is or as close as can be. There is no professional in this industry that would conclude any system is stable at a 1 hr stress.

OF COURSE THE 2 INDIVIDUALS WANT TO DISAGREE.

2 reasons to overclock..... Bragging rights which is cool. Highest possible overclock or getting better performance out of system for gaming and other applications. If you chose the later, why rely on a 1 hour stress. makes no sense. Most novices are happy with a 1 hour.
sigh i never said my system was completely stable i said in my eyes its stable for what i use it for it works my games don't crash its good. say if i was doing it for client at a profit for me, i would run a 24 hour stress and go out of my way to make sure its perfect for the client because if it is unstable it will come back and bite me. me on the other hand it blue screens or locks up i know how to fix it. hence why 1 hour is fine in my eyes for ME

so let me just get this clear i will not disagree a large amount of time of stress testing is needed, I don't go out saying to people new to overclocking you only need 1 hour stress tests either. I'm just saying for my own use i cannot be bothered running 8-24H stress tests, if i was doing it for someone else i would because it will be at my cost if it fails.
post #2096 of 2829
Quote:
Originally Posted by mr. biggums View Post
sigh i never said my system was completely stable i said in my eyes its stable for what i use it for it works my games don't crash its good. say if i was doing it for client at a profit for me, i would run a 24 hour stress and go out of my way to make sure its perfect for the client because if it is unstable it will come back and bite me. me on the other hand it blue screens or locks up i know how to fix it. hence why 1 hour is fine in my eyes for ME

so let me just get this clear i will not disagree a large amount of time of stress testing is needed, I don't go out saying to people new to overclocking you only need 1 hour stress tests either. I'm just saying for my own use i cannot be bothered running 8-24H stress tests, if i was doing it for someone else i would because it will be at my cost if it fails.


Runnin backwards I see!!!! LOL just kidding.. I knew I would get it out of you! If you fail on a
8-24 hr test it may be the difference of 100-200Mhz which is not much performance difference thus making the overclock experience complete. Some guys say in these threads I have 3.9GHz stable and but don't say it is on a 1 hour stress making the OP believe he can get 3.9 or in your case 3.75GHz. They never say 1 hr they say stable. The objective is 2 fold like I said... max overclock and 24/7 stable.

It's kinda like you paint your house. You do not prime it before you paint cause it looks OK. But if someone paid you to do a perfect job... You would prime it before painting. Makes the job complete and much nicer. The paint job will look nicer and last longer.

Exactly at your cost if it fails. Do a extreme overclock to see what max OC is and capable of benchmark and get a 24/7 stable OC. Stress test overnight while sleeping! Run it at your 24/7 OC setting for everyday usage. Then you will never have to worry! This is my point.
Edited by jimibgood - 5/22/09 at 12:51pm
Phenom II 955
(13 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 965 Lapped OC'd 4.14GHz (Stable 24/7) Gigabyte GA-MA790XT-UD4P 2 x 1G 4890 Liquid cooled crossfire Corsair XMS DDR3 1600(8G 9/9/9/23/1T) 
Hard DriveOptical DriveOSMonitor
2 x 300G Velocity Raptor(Raid0) WD 500G Barracuda Asus Windows7 64 bit Hanspree 22in 
KeyboardPowerCaseMouse
Logitech Ultra X3 850W Zalman GS1000 Logitech 
  hide details  
Reply
Phenom II 955
(13 items)
 
  
CPUMotherboardGraphicsRAM
Phenom II 965 Lapped OC'd 4.14GHz (Stable 24/7) Gigabyte GA-MA790XT-UD4P 2 x 1G 4890 Liquid cooled crossfire Corsair XMS DDR3 1600(8G 9/9/9/23/1T) 
Hard DriveOptical DriveOSMonitor
2 x 300G Velocity Raptor(Raid0) WD 500G Barracuda Asus Windows7 64 bit Hanspree 22in 
KeyboardPowerCaseMouse
Logitech Ultra X3 850W Zalman GS1000 Logitech 
  hide details  
Reply
post #2097 of 2829
Quote:
Originally Posted by jimibgood View Post
Runnin backwards!!!! LOL just kidding.. I knew I would get it out of you!
i never meant it to be for everyone though...(even though i properly worded it that way) what i was trying to say for my own use 1 hour is fine because the only thing that comes close to pushing my cpu too the max is gta4 and XxBeNigNxX most likely agrees with me in that aspect. if i was overclocking a server chip for 24/7 use my answer would obviously be different because it will be under constant stress and i would need to factor that in.

beyond that i really don't have the time to run long stress tests having work/school and partially being an insomniac i rather enjoy my computer then wait on it

Quote:
mr. biggums: It's AMD OverDrive. Which can be downloaded from http://game.amd.com
i should of clued in lol amd overdrive hates me, attempted playing with it before on my 4400x2 and ended up down clocking my cpu by 1ghz lol so i really don't want to touch it.
Edited by mr. biggums - 5/22/09 at 12:43pm
post #2098 of 2829
Alright guys let's not start a huge discussion/flame war about stress testing. Everyone has their own process. Let's respect that

Good luck
Workstation
(13 items)
 
  
CPUMotherboardGraphicsRAM
Athlon II X4 620 Asus M4A785-M Asus 5770 G.Skill 8GB DDR2-800 
Hard DriveOSMonitorPower
WD3200AAJS Windows 7 Pro Acer AL2216W, Samsung P2770 Corsair 550VX 
Case
Tagan Black Pearl 
  hide details  
Reply
Workstation
(13 items)
 
  
CPUMotherboardGraphicsRAM
Athlon II X4 620 Asus M4A785-M Asus 5770 G.Skill 8GB DDR2-800 
Hard DriveOSMonitorPower
WD3200AAJS Windows 7 Pro Acer AL2216W, Samsung P2770 Corsair 550VX 
Case
Tagan Black Pearl 
  hide details  
Reply
post #2099 of 2829
I'm going to need my entry changed soon I've been playing with it some more running 8 hour stress tests while at work to appease jimibgood so far i dropped the speed down by 50mhz but doing this allowed me to drop the voltage down too 1.45 from 1.5, tonight I'm going to try for 3.7 at 1.425v

cpuz




post #2100 of 2829
Thread Starter 
Hmmm, looking at ur temps, it seems that after 8hrs stress, the average core temp has dropped by 2-3C
Sulaco
(20 items)
 
  
CPUMotherboardGraphicsRAM
Intel 3770K ASUS Sabertooth Z77 ASUS X290X Corsair Vengeance 2x4GB 1866MHz 9-10-9-28 
Hard DriveHard DriveCoolingCooling
1x SAMSUNG 830 128GB, 2x Samsung 1TB 32MB Samsun 840 evo Aqua computer Kyros XT Feeser Exchanger 240 
CoolingCoolingCoolingOS
Swiftech MCR320-QP Swiftech MCP 350 Swiftech MCP 350 W8.1 64 
MonitorMonitorKeyboardPower
Iiyama ProLite XB2779QS Black 27" Iiyama E480S 20" Roccat Ryos MK pro CoolerMaster 850W ESBA 
CaseMouseMouse PadAudio
NZXT Switch 810 gun metal Roccat Kone[+] CorePad XXXL Logitech G35 
  hide details  
Reply
Sulaco
(20 items)
 
  
CPUMotherboardGraphicsRAM
Intel 3770K ASUS Sabertooth Z77 ASUS X290X Corsair Vengeance 2x4GB 1866MHz 9-10-9-28 
Hard DriveHard DriveCoolingCooling
1x SAMSUNG 830 128GB, 2x Samsung 1TB 32MB Samsun 840 evo Aqua computer Kyros XT Feeser Exchanger 240 
CoolingCoolingCoolingOS
Swiftech MCR320-QP Swiftech MCP 350 Swiftech MCP 350 W8.1 64 
MonitorMonitorKeyboardPower
Iiyama ProLite XB2779QS Black 27" Iiyama E480S 20" Roccat Ryos MK pro CoolerMaster 850W ESBA 
CaseMouseMouse PadAudio
NZXT Switch 810 gun metal Roccat Kone[+] CorePad XXXL Logitech G35 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD CPUs
Overclock.net › Forums › AMD › AMD CPUs › Phenom II steppings and max OC 550-940