Release CPU stress test


Updating the aggregated info (all at max OPP unless mentioned otherwise):


- original (CC) Pandora: 778@OPP3 (onpon4), 917 (notaz), 928 (TrashyMG), < 936 (Linux-SWAT), 946 (Googer), 954 (Fahrstuhl), 954 (PokeParadox), 967 (Takedown), 968 (FaeMinx), 969 (ZXDunny), 973 (Budweiser), ~978 (BaDToaD), 1006 (Ian J)


- Rebirth (GC) Pandora: 789@OPP3 (B-ZaR), 862@OPP3 (stardroid), 947 (B-ZaR), 955 (Gausen), 956 (apersonthinks), 985 or 989 (TrashyMG), 999 (_wb_), ~1000 (lemondragon), 1017 (Mosschops)


- 1GHz Pandora: 1157 (Lafazar), 1175, 1178 (foft), 1214 (kamui_kun), 1249 (notaz)
 
My rebirth got up to 1020@OPP5, with no errors reported, however it then froze up :)


Can I do anything in order to avoid the crashes and find the highest clock speed?


Steve
 
My rebirth got up to 1020@OPP5, with no errors reported, however it then froze up :)


Can I do anything in order to avoid the crashes and find the highest clock speed?


Steve

Use the new version of the tool, you can set an upper limit with that one. Try an upper limit of 1015 or something like that. It can take a long time before it finds an error.
 
My CC has been at 979mhz for a hour or so now, got 2 more hours at work if it stays at 979 till I go home ill probably just assume thats my max, Pretty happy


edit:977 :(
 
Last edited by a moderator:
What does the message "something changed cpu config, aborting" actually mean? It always happens when the clock rate is at 915 MHz @ OPP5. It then resets back to 900 MHz and goes up again to 915 MHz, where the message appears again.
 
My CC has been at 979mhz for a hour or so now, got 2 more hours at work if it stays at 979 till I go home ill probably just assume thats my max, Pretty happy


edit:977 :(

My suggestion? Leave it overnight. :)

What does the message "something changed cpu config, aborting" actually mean? It always happens when the clock rate is at 915 MHz @ OPP5. It then resets back to 900 MHz and goes up again to 915 MHz, where the message appears again.

Are you using the latest version of the stress test?
 
Use the new version of the tool, you can set an upper limit with that one. Try an upper limit of 1015 or something like that. It can take a long time before it finds an error.

Been running as advised for sometime now (probably getting towards 90 minutes), and it has been sitting on 1015 without any problems so far (I got it to test from 985 - 1015, so didn't take long to get to 1015). From other peoples reports it sounds like if I leave it long enough it'll probably go wrong, but it seems somewhat stable at least. Is the suggestion that if I can run this test at 1015 for 2 hours, I should also be able to run most anything at 1015 for around 2 hours (obviously not an exact science)? If this is the case, I am certainly quite interested in the new PSX emulator with the 2X scaler in, as maybe 1015 will be enough for that!


Steve
 
Yes, I think that if the stress test does not find any errors in 2 hours, it should be stable enough to play games at that clock speed. Not only does it put less load on the cpu, also not every error will be dramatic (e.g. maybe it's just some pixels that have the wrong color in one frame - you will probably not even notice). Of course you should save frequently to be sure.


It would be great if the new 2X scaling can be done on the 600 MHz Pandoras - but remember that the 1GHz models running at 1GHz are quite a bit faster than the 600MHz models running at 1GHz, because it's not only the cpu that has been upgraded.
 
Nope, it keeps trying to find errors. So it's up to you to decide what you consider to be "safe enough" - 1 hour, 10 hours, 1 month?


At some point there will be an error at any clock speed, if only because of cosmic radiation or weird quantum effects or whatever. No CPU or RAM is ever 100% error-free, because if they handle billions of bits per second, then there is some tiny but nonzero probability that one of those bits gets corrupted by an incoming cosmic ray or some other kind of ambient radiation. According to Wikipedia, studies by IBM in the 1990s suggest that computers typically experience about one cosmic-ray-induced error per 256 megabytes of RAM per month. This is probably not smaller now, because if the technology becomes more reliable, the smaller scale makes the impact of cosmic rays worse.
 
OK I was brave enough to raise my overclock limit so could test OPP5 properly


my result was:


1009MHz @OPP5


so you can update my result
 
Not sure if you are interested, but after almost 5 hours with no errors (still running, but I'm going to stop the test and start on OPP 4) I'm satisfied that my highest stable clock speed on OPP 3 is 821Mhz...


Which is not too shabby if you ask me. :)


Here's hoping the double scaling won't require overclocking past 800Mhz on the CC units....? EDIT: I'm dreaming aren't I? :unsure:


I would prefer to stay at OPP 3.
 
Last edited by a moderator:
Like the new version - can not set the max to stop it crashing:). Since it was crashing on the old version at 1246MHz I've started it at 1245MHz and... so far so good!


So that gives:


i) 1GHz no 1: 1166MHz (overnight)


ii) 1GHz no 2: 1245MHz (15 mins in... will edit to update in an hour)


Shame the keyboard doesn't work properly on (ii)!
 
What does the message "something changed cpu config, aborting" actually mean? It always happens when the clock rate is at 915 MHz @ OPP5. It then resets back to 900 MHz and goes up again to 915 MHz, where the message appears again.

Are you using the latest version of the stress test?

Okay the new version doesn't have that message anymore. Currently I'm at 1010 MHz @ OPP5, no errors so far (Rebirth Pandora).


€dit: Okay, errors at 1015 MHz :p
 
Last edited by a moderator:
Back
Top